Obesity-Induced miR-455 Upregulation Promotes Adaptive Pancreatic β-Cell Proliferation Through the CPEB1/CDKN1B Pathway

Pancreatic β-cell function and mass are markedly adaptive to compensate for the changes in insulin requirements observed during obesity (1). Clinical results have shown that the β-cell mass in the pancreata of obese individuals without diabetes or with prediabetes is larger than that in lean individuals with normoglycemia (2). Evidence for this compensatory process has been consistently provided by obese rodent models and human pancreas necropsies (3,4). In obesity, β-cell mass increases by 30–40%, whereas insulin secretory output is augmented by 100% (5). Postmortem histology further reveals a 20–65% decrease in β-cell mass in islets from obese individuals with type 2 diabetes (T2D) compared with BMI-matched subjects without diabetes (6,7). This adaptive capacity of human islets to obesity has been confirmed in experimental murine models (8,9). Strategies aimed at improving β-cell function and mass plasticity could be of major interest for designing innovative therapeutics to prevent β-cell decline and restore β-cell functional adaptive ability in diabetes. However, the physiological mechanisms that promote adaptive pancreatic β-cell expansion are still not completely understood.

The adaptive capacity of β-cell mass and function depends on the activity of transcriptional and translational regulators, and miRNAs are extremely important in accomplishing this task (10,11). In pancreatic β-cells, miRNAs regulate insulin production by directly or indirectly affecting the expression of key transcription factors and contribute to fine-tuning of hormone release by modulating the levels of important components of the β-cell secretory machinery (12–14). We and other researchers have reported changes in the expression of islet miRNAs in animal models of diabetes, with detrimental effects on the secretory activity and survival of β-cells (15,16). Moreover, obese mice lacking certain miRNAs failed to compensate for insulin resistance and developed a severe diabetic phenotype (3,12,17). These observations prompted us to investigate whether changes in miRNA expression contribute to compensatory β-cell mass expansion during obesity.

In this study, we found that miR-455 was upregulated during obesity to inhibit the expression of cytoplasmic polyadenylation element binding protein 1 (CPEB1) in pancreatic β-cells. Both in vivo and in vitro experimental results revealed that overexpression of (oe)-miR-455 promoted adaptive pancreatic β-cell proliferation and that knockdown of miR-455 prevented compensatory β-cell expansion. Moreover, we observed that the miR-455/CPEB1 pathway inhibits translation of cyclin-dependent kinase inhibitor 1b (Cdkn1b) mRNA and ultimately leads to lower levels of CDKN1B in β-cells, in turn significantly promoting β-cell adaptive proliferation. These observations show that miR-455 plays an integral role in the β-cell compensatory mechanism during obesity.

Eight-week-old C57BL/6J mice, ob/ob mice (4–12 weeks), and db/db mice (4–12 weeks) were obtained from GemPharmatech Co., Ltd. (Nanjing, China). All animals were on the C57BL/6 background unless otherwise stated. The db/db mice were on the BKS background. The care of all animals was within institutional animal care committee guidelines, and all procedures were approved by the animal ethics committee of China Pharmaceutical University (permit no. 2162326) and in accordance with international laws and policies (European Union Council Directive 86/609/EEC, 1987). Unless otherwise stated, C57BL/6J mice were fed a normal chow diet (10% calories from fat, D12450J; Research Diets) and provided water ad libitum. Diet-induced obesity was obtained by feeding a high-fat diet (HFD) (60% calories from fat, D12494; Research Diets) for at least 15 weeks. Male mice were used for all the indicated studies.

Six human islets were provided by Tianjin First Central Hospital. All human subjects provided informed consent, and all procedures using human islets were approved by the Research Ethics Committee of the Tianjin First Central Hospital (18). High-purity islets (>80%) were collected and cultured in CMRL-1066 medium (Corning, Manassas, VA) supplemented with 10% human serum albumin (Baxter, Vienna, Austria), 100 units/mL penicillin, and 100 μg/mL streptomycin at 37°C in 5% CO₂. The islets from six human donors were mixed randomly, then mixed islets were used to perform each type of experiment and three technical replicates.

Mouse islets were isolated via collagenase digestion and enriched using a Histopaque (Sigma-Aldrich, St. Louis, MO) density gradient, as described previously (15,19). To assess the effects of pathophysiological concentrations of palmitate, glucose, and proinflammatory cytokines, islets were incubated in modified medium with 0.5% (w/v) BSA and various concentrations of glucose (low glucose 2.5 mmol/L, high glucose 25 mmol/L), palmitate (0.5 mmol/L), interleukin-1β (5 ng/mL), and tumor necrosis factor-α (30 ng/mL).

The mouse pancreatic β-cell line MIN6 (passages 15–20) was cultured in DMEM (Gibco, Burlington, MA) containing 15% FBS (Gibco), 100 IU/mL penicillin, 100 μg/mL streptomycin, and 50 μmol/L β-mercaptoethanol (Sigma-Aldrich) at 37°C in a humidified atmosphere containing 5% CO₂.

Human EndoC-βH1 cells were obtained from Univercell Biosolutions (Toulouse, France). Human EndoC-βH1 cells were cultured in extracellular matrix/fibronectin-coated plates in low-glucose DMEM with supplements as previously described (20,21).

For the glucose-stimulated insulin secretion assay, MIN6 cells or human islets were preincubated overnight in Krebs-Ringer bicarbonate HEPES balanced buffer containing 0.2% BSA supplemented with 2.5 mmol/L glucose and incubated for 2 h with 2.5 or 16.7 mmol/L glucose. Immediately after incubation, an aliquot of the medium was removed for insulin analysis, and the cells were incubated in acid-ethanol for insulin content determination using a mouse insulin ELISA kit (ExCell Bio, Shanghai, China) according to the manufacturer’s instructions.

Total RNA was isolated from islets of 8-week-old control mice and 8-week-old ob/ob mice using an RNeasy Plus Universal Mini Kit (QIAGEN), following the protocol for total RNA isolation. RNA integrity and concentration were assessed using an RNA Nano 6000 Assay Kit for the Bioanalyzer 2100 system (Agilent Technologies, Santa Clara, CA), according to the manufacturer’s instructions. To extract small RNA (18–30 nucleotides [nt]), total RNA was separated in 15% agarose gels. After ethanol precipitation and centrifugal enrichment of small RNA samples, a small RNA library was prepared using a Small RNA Sample Preparation Kit (RS-200-0048; Illumina), according to the manufacturer’s instructions. Briefly, a 3′ adaptor or 5′ adaptor was attached directionally. Each library was indexed with the Illumina adaptor (6-base barcode). The small RNA library was size fractionated in a 6% Tris base, boric acid, EDTA-urea polyacrylamide gel, and the 140- to 160-base pair (bp) fraction was excised from the gel. The library RNA concentration was measured using a Qubit RNA Assay Kit in Qubit 2.0 for preliminary quantification and then diluted to 1 ng/μL. Insert size was assessed using a 2100 Bioanalyzer System (Agilent Technologies), and after the insert size was confirmed to be consistent with expectations, the qualified insert size was accurately quantified using a TaqMan fluorescence probe and an Applied Biosystems StepOnePlus Real-Time PCR System (valid library concentration >2 nmol/L). Then, the qualified libraries were sequenced on an Illumina HiSeq 2500 platform. We calculated the fragments per kilobase per million mapped reads for mature miRNA from the mapped reads using a custom Python script. Differentially expressed miRNAs were analyzed using DESeq version 1.26.0 on the basis of the negative binomial distribution test (22), and the P value was corrected for false discovery rate analysis for multiple testing (23). Differentially expressed genes were considered significant with an adjusted P ≤ 0.05 and an absolute value of log₂ (fold change) >1. The raw data are presented in Supplementary Table 1 and have been deposited in the National Center for Biotechnology Information Sequence Read Archive database (PRJNA731006).

Serum and clinicopathological data were collected from Zhongda Hospital, which is affiliated with Southeast University. All human subjects provided informed consent, and all human studies were conducted according to the principles of the Declaration of Helsinki and approved by the ethics committees of Zhongda Hospital, Southeast University.

The probe, an ∼25-bp fragment, including the binding size, was biotin end labeled according to the instructions of the Pierce Biotin 3′ End DNA Labeling Kit (Thermo Fisher Scientific, Waltham, MA) and then annealed to double-stranded probe DNA. Egr2-DNA complexes were generated according to the instructions of a LightShift Chemiluminescent EMSA Kit (Thermo Fisher Scientific). Probe sequences are listed in Supplementary Table 2.

MIN6 cells and human islets were fixed with 37% formaldehyde for 10 min, followed by 30 rounds of sonication (each for 3 s) to fragment the chromatin. Chromatin was incubated with an anti–early growth response 1 (EGR2) antibody at 4°C overnight and then immunoprecipitated with Proteinase K (EMD Millipore, Billerica, MA). Purified DNA was amplified via quantitative RT-PCR (qRT-PCR) using PCR and primer pairs that spanned the predicted Egr2 binding sites. Primer sequences are listed in Supplementary Table 2.

The coding sequences for Egr2 (NM_001373983.1), Cpeb1 (NM_001252525.1), and Cdkn1b (NM_009875.4) were amplified via PCR from full-length cDNA of MIN6 cells and then cloned into a pcDNA 3.1 vector. The primer sequences for PCR are listed in Supplementary Table 2.

The single-guide RNAs (sgRNAs) for the miR-455 promoter were constructed in a lentiCRISPRv2 puro vector according to the Zhang libraries (24). The lentiCRISPRv2 puro vector was digested with BsmBI. The sgRNA seed sequences are listed in Supplementary Table 3.

MIN6 cells (2 × 10⁵ cells/well) were transfected with 0.4 μg miR-455 promoter, 0.4 μg Egr2, and 0.1 μg constitutive Renilla expression plasmid, and luciferase activities were measured using a dual-luciferase reporter assay system (Vazyme, Nanjing, China) after transfection for 24 h. The complete 3′-untranslated region (UTR) of murine Cpeb1 containing either the wild-type (WT) or the mutated (MUT) miR-455 binding site was cloned behind the stop codon of the firefly luciferase open reading frame using specific primers. MIN6 cells were transfected with pmirPGLO reporter (100 ng) along with miR-455 or miR-455 inhibitor using Lipofectamine 2000 transfection reagent (Invitrogen). At 24 h posttransfection, dual luciferase reporter assays were performed using a luciferase assay system (Vazyme). The WT and MUT site primers are listed in Supplementary Table 2.

Insulin and KI-67 double-positive cells were detected following a standard intracellular staining procedure using Cytofix/Cytoperm solution (BD Biosciences). Insulin-allophycocyanin (APC) and KI-67-FITC antibodies were used for staining, and isotype antibodies were used as a negative control (NC). After double staining with insulin-APC and KI-67-FITC, the cells were analyzed with a flow cytometer (FACScan; BD Biosciences) equipped with FlowJo version 10 software (BD Biosciences).

Purified islets were incubated with 2 mg/mL collagenase II (Sigma-Aldrich) for 10 min, and insulin-positive cells were detected by following a standard intracellular staining procedure using Cytofix/Cytoperm solution. After staining with insulin-APC, the cells were analyzed via flow cytometry (BD FACSAria II SORP). Purified β-cell and non–β-cell fractions were collected for RNA extraction and qRT-PCR analysis.

miR-455, miR-455 inhibitor, sh-Cpeb1, and sh-Cdkn1b were inserted into a lentivirus-mouse insulin 2 promoter (MIP)-PGLV3/H1/puro vector (len-miR-455, len-anti-miR-455, len-shCpeb1, len-shCdkn1b). The indicated lentivirus (1 × 10⁹) was dissolved in 0.15 mL normal saline and infused into the pancreatic duct at 6 μL/min for 25 min using an R462 perfusion pump (RWD Life Sciences, Shengzhen, China), according to the procedure described by Xiao et al. (25). At 72 h after injection, islets were lysed to extract total RNA or protein to measure the overexpression efficacy. At 1 week postinjection, the mice were fed an HFD (60% energy from fat, D12494; Research Diets) for 20 weeks.

Mouse fasting blood glucose levels and fasting serum insulin levels were examined using a glucometer (OMRON, Kyoto, Japan) and an ELISA (ExCell Bio) after 12 h of fasting treatment. To perform the glucose tolerance tests, 1.5 g/kg glucose (Sigma-Aldrich) was injected i.p. into mice, whereas 0.75 units/kg insulin (Novolin R; Novo Nordisk, Bagsværd, Denmark) was injected i.p. into mice for insulin tolerance tests.

Total RNA was extracted using TRIzol (Invitrogen), as previously described (17). RT reactions were performed using a PrimeScript RT Reagent Kit (Takara Bio, Tokyo, Japan), and diluted cDNA was used for qRT-PCR analysis using a SYBR Premix Ex Taq II Kit (Takara Bio) with the appropriate primers listed in Supplementary Table 4. The relative expression of genes was determined using a comparative method (2^−ΔCT). miRNA and mRNA levels were normalized to the expression of small RNAs (sno234 and U6) or mRNA (Gapdh, Hprt, and Ppia), respectively. For miR-455 and U6, TaqMan probes (Ambion) were used to confirm our results.

Total cellular RNA was reversed transcribed using MultiScribe RT (Life Technologies) and an oligo (dT) anchor primer (5′-GCGAGCTCCGCGGCCGCGTTTTTTTTTTTTTTT-3′), and subsequent PCR was conducted with an anchor primer (5′-AAAAACGCGGGCCGCGGAGCTCGC-3′) and a specific primer for Cdkn1b (5′-GCCAATTATTGTTACACATT-3′) located near the 3′ end of the Cdkn1b 3′-UTR.

RNA immunoprecipitation (RIP) was performed using an EZMagna RIP Kit (EMD Millipore), according to the manufacturer’s protocol. Cells were lysed in complete RIP lysis buffer, and then 100 μL of whole-cell extract was incubated with RIP buffer containing magnetic beads conjugated with anti-Ago2 (Cell Signaling Technology) or anti-CPEB1 (Abcam) antibody or NC normal mouse IgG (Abcam). Furthermore, purified RNA was subjected to qRT-PCR analysis to demonstrate the presence of the binding targets using the respective primers. The primer sequences are listed in Supplementary Table 4.

Pancreata were fixed in 4% paraformaldehyde and embedded in paraffin, and the antigen in the cut sections was retrieved by boiling in 10 mmol/L Tris/EDTA (pH 9.0). Sections were permeabilized and blocked in PBS buffer containing 0.3% Triton X-100, 1% BSA, and 5% goat serum. Primary antibody (anti-KI-67, anti-insulin, antiglucagon) binding was performed overnight at 4°C, and incubation with secondary antibody was performed at room temperature for 1 h. The slides were analyzed using confocal laser scanning microscopy (CLSM) (LSM 700; Carl Zeiss) at ×20 or ×40 magnification. The antibodies are listed in Supplementary Table 5.

A Cy3-labeled miR-455 probe was designed and synthesized by GenePharma (Shanghai, China). For the fluorescence in situ hybridization assay, pancreata were fixed in 4% formaldehyde, permeabilized with 0.3% Triton X-100 for 15 min, and washed with PBS three times and once in 2× saline-sodium citrate buffer. Hybridization was carried out at 37°C for 16 h using DNA probe sets, followed by incubation with anti-insulin antibody overnight at 4°C and incubation with secondary antibody at room temperature for 1 h. Images were obtained with a CLSM and processed using ZEN imaging software.

Mouse islets (n = 200 islets/group) and MIN6 cells were lysed with radioimmunoprecipitation assay lysis buffer (Beyotime) containing 1% phenylmethylsulfonyl fluoride (Sigma-Aldrich). Then, Western blot analyses were conducted according to standard procedures using specific antibodies. The antibodies are listed in Supplementary Table 5.

MIN6 cells and primary islets were fixed in 4% paraformaldehyde and permeabilized and blocked in PBS buffer containing 0.3% Triton X-100 and 1% BSA. Primary insulin and KI-67 antibody binding was performed overnight at 4°C, and then incubation with secondary antibody was performed at room temperature for 1 h. Finally, the nuclei were stained with DAPI. Images were obtained using a CLSM at ×40 or ×20 magnification.

Pancreatic sections were stained with anti-insulin antibody and DAPI to determine β-cell mass. The pancreatic sections were scanned entirely using the 10× objective of a Zeiss LSM 700 microscope. β-Cell mass was calculated by multiplying the area of insulin-positive cells/pancreas area using ImageJ software (26).

MIN6 cells were seeded into 96-well plates (4 × 10⁴ cells/well) in 100 μL culture medium. A Cell Counting Kit-8 assay (Vazyme) was performed according to the manufacturer’s instructions.

All data represent at least three independent experiments and are shown as the mean ± SD. Comparisons between two groups were performed using Student t test and among multiple groups using ANOVA. Dunn multiple comparisons for one-way ANOVA and Fisher least significant difference for two-way ANOVA were used. P < 0.05 was considered significant. GraphPad Prism 7 software (GraphPad Software, San Diego, CA) was used for all calculations.

The data sets generated and/or analyzed during the current study are available from the corresponding authors upon reasonable request. The RNA sequencing raw data that support the findings of this study have been deposited in the National Center for Biotechnology Information Sequence Read Archive database (PRJNA731006; https://submit.ncbi.nlm.nih.gov/subs/sra/SUB9682154/overview).

To identify miRNAs potentially involved in β-cell mass expansion during obesity, we performed small RNA sequencing on total RNA from 8-week-old ob/ob mice, which are capable of long-term compensatory insulin hypersecretion (27). The characteristics of the animals used in this study are presented in Supplementary Fig. 1A and B. Using an absolute fold change of at least 1.0 and P < 0.05, we observed that the expression of 39 miRNAs was significantly upregulated and that the expression of 46 miRNAs was downregulated in ob/ob mice compared with WT mice. Among the increased miRNAs identified, the expression of miR-6932-3p (miR-6932) and miR-455-5p (miR-455) was the most significantly increased (Fig. 1A and Supplementary Table 1). miR-6932 expression was significantly lower than miR-455 expression in human islets, mouse islets, and MIN6 cells (Supplementary Fig. 1C). Therefore, we chose miR-455 for further analysis. Then, we measured the expression of miR-455 in the islets of ob/ob mice aged 4–12 weeks and observed an increase in miR-455 expression at 6 weeks of age with the onset of insulin resistance (Fig. 1B). Similarly, the pri-miR-455 transcript was also upregulated in the islets of ob/ob mice (Fig. 1C), indicating that miR-455 is regulated at the transcriptional level. We observed a similar increase in the expression of mature miR-455 in the islets of mice treated with a 15-week HFD, where miR-455 expression was significantly decreased, accompanied by an increase in glucose (Fig. 1D). We also found that miR-455 expression was dramatically increased in the islets of young db/db mice (Fig. 1E), all of which showed that this observation is not limited to one mouse model of obesity and insulin resistance. The expression of miR-455 was only significantly increased in islets, white adipose tissue (WAT), brown adipose tissue (BAT), and muscle in ob/ob and HFD mice compared with their respective controls (Fig. 1F). Interestingly, we also found that miR-455 levels were significantly increased in donors with T2D compared with donors without diabetes (Fig. 1G), and serum miR-455 expression levels were significantly correlated with BMI and HbA_1c in these subjects (Fig. 1H and I). Then, we found that miR-455 expression was approximately fourfold higher than that in exocrine glands, indicating that islets represent the main source of miR-455 in the pancreas (Fig. 1J and K). Moreover, we purified β-cells from islets via flow cytometric analysis, and the results revealed that miR-455 was enriched in purified β-cells (Fig. 1L). Overall, miR-455 expression in islets is increased in both dietary and genetic mouse models of obesity.

To determine the potential reason for the changes in miR-455 expression detected in the islets of obese mice, MIN6 cells, EndoC-βH1 cells, normal mouse islets, and human islets were exposed to pathophysiological concentrations of palmitate, glucose, and proinflammatory cytokines. The expression of miR-455 was increased in the presence of palmitate (0.5 mmol/L) (Fig. 2A–D) and a high glucose level (25 mmol/L) (Fig. 2E–H), but these changes were not observed in MIN6 cells or mouse islets incubated with proinflammatory cytokines (Supplementary Fig. 2A–C).

Next, we explored the potential molecular mechanism of miR-455 upregulation in the islets of obese mice. We predicted an miR-455 promoter region 2 kb upstream of the mouse and human miR-455 sequence and constructed three sgRNAs corresponding to this promoter in a lentiCRISPRv2 puro vector. Supplementary Fig. 2D and E show that miR-455 levels were decreased in MIN6 cells and human islets transfected with sgRNAs. These results revealed that the predicted promoter region can modify miR-455 expression. Then, MIN6 cells were exposed to 0.5 mmol/L palmitate for 48 h, and qRT-PCR was performed to confirm the expression levels of high-score transcription factors (score >15) of the miR-455 promoter, which were predicted by JASPAR. Among these candidates, Egr2, which has two potential binding sites in the mus-miR-455 promoter (Supplementary Fig. 2F), was mostly upregulated (Supplementary Fig. 2G). Interestingly, Egr2 also has one potential binding site in the hsa-miR-455 promoter (Supplementary Fig. 2F). Moreover, the mRNA and protein levels of EGR2 were upregulated in obese mouse islets (Fig. 2G and I) and MIN6 cells in response to palmitate treatment (Supplementary Fig. 2H) or high-glucose treatment (Supplementary Fig. 2I). The expression levels of EGR2 were also increased in human islets (Supplementary Fig. 2J and K) and EndoC-βH1 cells (Fig. 2K and L) exposed to palmitate and high glucose.

Egr2 is known as an important regulator of insulin resistance (28) and can be upregulated by palmitate in MIN6 cells (29). We next focused on studying whether obesity facilitates miR-455 upregulation by increasing Egr2 expression. As expected, in dual luciferase assays, MIN6 cells transfected with Egr2 overexpression plasmid exhibited a significantly higher Egr2 binding ability to the miR-455 promoter compared with the control, and after mutation of the R1 binding region, the binding ability was significantly decreased (Supplementary Fig. 2L). Moreover, Egr2 in the miR-455 promoter region was increased in the islets of obese mice (Fig. 2M and Supplementary Fig. 2M). A similar trend was obtained in MIN6 cells (Supplementary Fig. 2N and O), in EndoC-βH1 cells (Fig. 2N) and human islets (Fig. 2O) transfected with oe-Egr2. EMSA results revealed that the signal from the probe-protein-anti-EGR2 complex could be detected using an miR-455 probe in mouse islets (Fig. 2P), EndoC-βH1 cells (Fig. 2Q), human islets (Fig. 2R), and MIN6 cells (Supplementary Fig. 2P). Further research showed that palmitate-induced miR-455 upregulation in MIN6 cells and human islets was partially reversed by knockdown of Egr2 (Fig. 2S and T). These results indicate that the increased miR-455 expression in the islets of obese mice is mediated by upregulation of Egr2.

To explore the potential miR-455 function in β-cells, MIN6 cells were transfected with miR-455 mimic (miR-455) or miR-455 inhibitor (anti-miR-455). The knockdown and overexpression efficiencies were ∼80% and 220-fold, respectively (Supplementary Fig. 3A). Cell Counting Kit-8 assay results showed that cell proliferation was reduced by miR-455 silencing (Supplementary Fig. 3B). We observed that the upregulation of miR-455 in MIN6 cells resulted in an increase in the number of insulin-positive/Ki-67⁺ cells (Fig. 3A). Similar results were obtained in MIN6 cells (Fig. 3B), EndoC-βH1 cells (Fig. 3C), human islets (Fig. 3D), and mouse islets (Supplementary Fig. 3C). Next, we assessed whether miR-455 regulates MIN6 cell proliferation by evaluating cell-cycle status. Flow cytometry results indicated that miR-455 reduced the proportion of MIN6 cells in the G1 phase and increased the proportion in the S phase (Fig. 3E). Since cell-cycle progression is tightly regulated by cyclin-related proteins, we next tested the expression levels of cyclin-related proteins in MIN6 cells transfected with miR-455 or anti-miR-455. As shown in Supplementary Fig. 3D, miR-455 significantly decreased the expression level of CDKN1B, which is an inhibitor of the G1-S transition. The CDKN1B protein level was also decreased in MIN6 cells transfected with miR-455 (Fig. 3F). However, miR-455 did not significantly affect insulin synthesis, insulin content, or insulin secretion (Supplementary Fig. 3E–G) in MIN6 cells. Similar results were observed in human islets (Supplementary Fig. 3H–J).

In view of these findings, to verify whether ectopic expression of miR-455 also affects β-cell proliferation in vivo, 1 × 10⁹len-miR-455 cells or 1 × 10⁹len-anti-miR-455 cells were injected into 8-week-old male C57BL/6J mice through pancreatic intraductal infusion. We observed an ∼65-fold miR-455 upregulation in the islets that received len-miR-455 compared with those that received lentivirus-MIP-miR-NC (len-miR-NC) and an ∼70% decrease in miR-455 expression in the islets of len-anti-miR-455 mice compared with len-anti-NC mice (Fig. 3G), whereas the expression level of miR-455 was not significantly changed in other organs (Supplementary Fig. 3K). The len-miR-455 mice exhibited slightly improved glucose tolerance in the intraperitoneal glucose tolerance test (Fig. 3H). Similarly, the insulin levels in len-miR-455 mice improved after glucose injection compared with those in control mice (Fig. 3I), while miR-455 had no effect on glucagon in vivo (Supplementary Fig. 3L). The len-miR-455 mice showed a significant increase in both β-cell proliferation (Fig. 3J) and β-cell mass (Supplementary Fig. 3M). Moreover, the CDKN1B mRNA and protein levels were suppressed in the islets of len-miR-455 mice and vice versa (Supplementary Fig. 3N and Fig. 3K and L). Furthermore, no change was detected in TUNEL⁺ β-cells after ectopic expression of miR-455 (Supplementary Fig. 3O), indicating that cell death was not a primary mechanism underlying the effect on β-cell mass.

To address whether miR-455 mediates the compensatory expansion of β-cells during insulin resistance, 1 × 10⁹len-anti-miR-455 cells were injected into 8-week-old male C57BL/6 mice via pancreatic ductal infusion, and then, the mice were fed an HFD for 20 weeks (Fig. 4A). As shown in Fig. 4B, miR-455 was downregulated to 70% in the islets of len-anti-miR-455 mice compared with those receiving len-anti-NC even 20 weeks after injection. The len-anti-miR-455 treatment had no effect on cumulative energy intake (Supplementary Fig. 4A), body weight (Supplementary Fig. 4B), or body fat content (Supplementary Fig. 4C). However, inhibition of miR-455 slowly increased glycemia in random-fed mice after 20 weeks of HFD treatment (Supplementary Fig. 4D). The obesity-associated rise in serum insulin concentrations was slightly lower in HFD-fed animals treated with len-anti-miR-455 (Fig. 4C), and HOMA of insulin resistance indices in mice with miR-455 repression were increased (Fig. 4D). In accordance with this, glucose tolerance tests revealed impairment of glucose tolerance upon miR-455 knockdown (Fig. 4E), and insulin sensitivity was damaged upon downregulation of miR-455 (Fig. 4F). Counts of KI-67⁺ β-cells were lower in len-anti-miR-455 mice than in control animals (Fig. 4G). Moreover, we found that miR-455 knockdown markedly decreased β-cell mass compared with that in len-anti-NC mice (Supplementary Fig. 4E). These findings indicate that the miR-455 inhibitor resulted in damage to β-cell mass in diet-induced obese mice.

Furthermore, high expression of miR-455 in the islets of ob/ob mice was silenced by len-anti-miR-455 through pancreatic intraductal infusion (Supplementary Fig. 4F). Compared with len-anti-NC/ob mice, len-anti-miR-455/ob mice exhibited severe hyperglycemia and reduced systemic insulin levels (Fig. 4H and I) due to loss of β-cell mass and compensation (Fig. 4J and Supplementary Fig. 4G). Collectively, these data show that suppression of miR-455 expression in obesity model mice attenuated adaptive β-cell proliferation.

Most miRNAs function by suppressing their target gene expression, and we found an inverse expression pattern between miR-455 and Cdkn1b, which prompted us to verify whether Cdkn1b is a target gene of miR-455. Unfortunately, using a double luciferase reporter, we found no targeting relationship between miR-455 and Cdkn1b (Supplementary Fig. 5A).

To identify the potential miRNA targets of miR-455, in silico analysis was performed using TargetScan, miRDB, miRWalk, and starBase, which jointly predicted that two genes (Cpeb1 and Usp9x) may act as biological targets of miR-455 (Fig. 5A). Next, we verified the targeting relationship between the two genes and miR-455. Cpeb1 harbored an miR-455 binding site, which was conserved in humans, mice, and rats (Fig. 5B). We cloned a portion of the mouse 3′-UTR of Cpeb1 (1,316 nt) into a luciferase reporter construct and observed decreased activity in the presence of an miR-455 mimic (Fig. 5C). In addition, mutation of 8 nt in the binding site within the mouse UTR (at positions 1,289–1,296) from GGCACAUA to AATGTGCG abolished the inhibitory effect of miR-455 (Fig. 5C). Moreover, we conducted anti-Ago2 RIP in MIN6 cells transiently overexpressing miR-455. Endogenous Cpeb1 pulldown by Ago2 was specifically enriched in miR-455–transfected cells (Fig. 5D) and vice versa (Supplementary Fig. 5B). oe-miR-455 decreased CPEB1 expression in the MIN6 cells (Supplementary Fig. 5C). The mRNA and protein levels of CPEB1 were also significantly decreased in the islets of len-miR-455 mice but increased in the islets of len-anti-miR-455 mice (Fig. 5E and F and Supplementary Fig. 5D). Finally, silencing the expression of miR-455 in the pancreas of ob/ob mice increased CPEB1 expression (Fig. 5G and Supplementary Fig. 5E). Unfortunately, using the same technical means, we did not find a targeted relationship between miR-455 and Usp9x (Supplementary Fig. 5F).

To address the relevance of the miR-455-Cpeb1 interaction in the islets of obese mice, we measured the expression of Cpeb1 in pancreatic islets from ob/ob and HFD mice. As shown in Fig. 5H and I, the mRNA and protein levels of CPEB1 were significantly decreased, and similar trends were observed in MIN6 cells incubated with high glucose (Fig. 5J) or 0.5 mmol/L palmitate (Fig. 5K), indicating a targeted relationship between miR-455 and Cpeb1. We next verified whether miR-455 regulates β-cell proliferation by suppressing Cpeb1. We studied the role of Cpeb1 in β-cell proliferation, and the results showed that transfection with Cpeb1 siRNA (si-Cpeb1) markedly promoted MIN6 cell proliferation and that the effects of Cpeb1 knockdown were similar to those of miR-455 overexpression (Fig. 5L). Compared with cells transfected with si-Cpeb1, MIN6 cells transfected with si-Cpeb1 and anti-miR-455 exhibited significantly lower β-cell expansion ability (Fig. 5L and Supplementary Fig. 5G and H). Moreover, we found that len-shCpeb1–treated mice showed a higher positive KI-67 β-cell count and β-cell mass, while len-anti-miR-455 injection restored this effect (Fig. 5M and Supplementary Fig. 5I), suggesting that restoration of CPEB1 attenuated the β-cell proliferation effect of miR-455. Taken together, these findings indicate that miR-455 promotes β-cell mass expansion in a CPEB1-dependent manner.

As we previously verified, there is no targeting relationship between miR-455 and Cdkn1b, but miR-455 promotes adaptive pancreatic β-cell proliferation during obesity in a CDKN1B-dependent manner. This suggests that the regulatory effect of miR-455 on Cdkn1b is indirect. On the other hand, we confirmed that CPEB1 is a target gene of miR-455, and by controlling mRNA translation efficiency via the 3′-UTR, CPEB1 regulates many important biological processes, ranging from cell-cycle control (30) to regulation of insulin resistance (31,32). All these clues prompted us to hypothesize that CPEB1 might regulate Cdkn1b expression.

To better understand the relationship between CPEB1 and CDKN1B, we performed a bioinformatics search for target transcripts that could potentially be modulated by CPEB1. As expected, we found that Cdkn1b was one of the most promising predicted targets (Supplementary Fig. 6A). Since CPEB1 acts as a sequence-specific RNA binding protein (32), we analyzed the possible interaction between CPEB1 and Cdkn1b mRNA using RIP experiments. The RIP results revealed that the CPEB1 antibody, but not IgG, successfully pulled down Cdkn1b transcripts (Fig. 6A) and that overexpression of Cpeb1 in MIN6 cells enhanced this binding ability (Supplementary Fig. 6B). These data further confirmed that Cdkn1b mRNA is in the CPEB1 complex. In an attempt to determine the regulatory mechanisms underlying the interaction between CPEB1 and Cdkn1b, a sequence containing the WT or MUT 3′-UTR of Cdkn1b (Cdkn1b-WT and Cdkn1b-MUT, respectively) was cloned into a pGL3-control vector to construct a reporter system. When we cotransfected MIN6 cells with si-Cpeb1 and Cdkn1b-WT or Cdkn1b-MUT, we observed that oe-Cpeb1 strongly increased luciferase expression in the Cdkn1b-WT cotransfected group but not in the Cdkn1b-MUT group (Fig. 6B). These results support CPEB1 regulation of Cdkn1b mRNA through its 3′-UTR.

To further clarify the molecular mechanisms of CPEB1 in Cdkn1b regulation, we examined whether Cdkn1b mRNA stability was affected as a consequence of CPEB1 ectopic modulation. MIN6 cells transfected with oe-Cpeb1, si-Cpeb1, or vehicle were treated with actinomycin D (10 μg/mL). Consistent with the results of Galardi et al. (33), the ectopic modulation of CPEB1 did not change the half-life of Cdkn1b mRNA, indicating that CPEB1 action is not exerted through modulation of Cdkn1b mRNA stability (Supplementary Fig. 6C). Since CPEB1 was originally identified as a sequence-specific RNA-binding protein that promotes polyadenylation-induced translation (32), we asked whether the polyadenylation status of Cdkn1b mRNA was altered by CPEB1 modulation. We performed a PCR poly(A) tail length assay on endogenous Cdkn1b mRNA in MIN6 cells transfected with oe-Cpeb1, si-Cpeb1, or vehicle. The oe-Cpeb1 enhanced the polyadenylation of Cdkn1b mRNA, while si-Cpeb1–mediated knockdown of Cpeb1 decreased polyadenylation (Fig. 6C). Because poly(A) tail lengthening is expected to result in more efficient translation, we examined whether the CDKN1B level was altered by CPEB1 ectopic modulation. As shown in Fig. 6D and E, ectopic modulation of Cpeb1 changed the protein level of CDKN1B but not the mRNA level of Cdkn1b. Analysis of CDKN1B in the islets of len-shCpeb1 mice revealed a similar trend (Fig. 6F and G). All these observations indicate that CPEB1 regulates Cdkn1b expression by promoting elongation of the poly(A) tail and translation efficiency.

To verify that miR-455 can indirectly regulate Cdkn1b through CPEB1, we conducted a series of functional recovery experiments. Through RIP experiments, we found that the interaction between Cdkn1b and CPEB1 was significantly regulated by miR-455 (Fig. 7A and Supplementary Fig. 7A). As shown Fig. 7B, CDKN1B protein levels were decreased by Cpeb1 silencing, while anti-miR-455 restored CDKN1B expression. Analysis of CDKN1B in the islets of len-shCpeb1 or len-shCpeb1 and len-anti-miR-455 mice showed a similar trend (Fig. 7C). Next, we found that si-Cdkn1b enhanced MIN6 cell proliferation, while oe-Cpeb1 abrogated this effect (Fig. 7D and E and Supplementary Fig. 7B). Moreover, we verified that len-shCdkn1b promoted β-cell proliferation and expanded β-cell mass, while len-Cpeb1 partially restored this phenomenon (Fig. 7F and G and Supplementary Fig. 7C). These results support the role of miR-455 in promoting β-cell proliferation by inhibiting the expression of Cdkn1b via direct targeting of Cpeb1.

Obesity is frequently associated with diminished insulin sensitivity, which is normally compensated for by an expansion of the functional β-cell mass that prevents chronic hyperglycemia and the development of T2D. The molecular basis underlying compensatory β-cell mass expansion is largely unknown. In rodents, we found that β-cell mass expansion during obesity is associated with upregulated expression of miR-455, which was induced by Egr2. We also observed that oe-miR-455 increased β-cell proliferation both in vitro and in vivo. Mechanistically, we observed that the miR-455/CPEB1 pathway inhibits translation of Cdkn1b mRNA and finally leads to lower levels of CDKN1B in β-cells, in turn significantly promoting β-cell adaptive proliferation (Fig. 8). These findings point to a major role for miR-455 in compensatory β-cell mass expansion during obesity.

Ob/ob mice model increased islet mass that is induced by severe insulin resistance and have been extensively studied as a model of T2D (34). In our study, we used RNA sequencing to detect differentially expressed miRNAs in the islets of ob/ob mice. Our results revealed that 85 miRNAs were significantly changed of which 65 had been identified as differentially expressed in previous research (35,36), including miR-455, miR-184, miR-375, and miR-122a. Yang et al. (37) found that miR-455 expression was increased in the serum of patients with insulin resistance. However, the mechanism of obesity-induced miR-455 upregulation and whether miR-455 can regulate β-cell function remain unclear. Therefore, we chose miR-455 for further analysis. First, to identify the possible causes of the variation in miR-455 expression in the islets of obese mice, we used a bioinformatics approach to search for putative transcription factors that could bind to the miR-455 promoter, among which Egr2 was found to be upregulated in response to palmitate treatment in primary islets and in ob/ob mouse islets. Our results are consistent with a report by Hayes et al. (38) in which the expression of Egr2 was found to be associated with T2D via genome-wide association studies. Accumulating evidence has revealed that palmitate and glucose can induce increased expression of Egr2 under insulin resistance (39,40). Here, we show that the Egr2 expression level was elevated in the pancreatic islets of obese mouse models, in MIN6 cells, and in human islets incubated with 0.5 mmol/L palmitate and 25 mmol/L glucose. Luciferase reporter, EMSA, and chromatin immunoprecipitation analyses showed that Egr2 can directly combine with the promoter of mus-miR-455 and hsa-miR-455 to induce an increase in miR-455 expression.

Here, we show that obesity induced miR-455 upregulation in islets and that the levels of miR-455 were increased in the muscle, WAT, and BAT of obese mice. miR-455 has previously been shown to activate AMPKa1 and then promote the brown adipogenic program and mitochondrial biogenesis, suggesting that miR-455 is a potential therapeutic for treating obesity (41). miR-455 expression has been evaluated in muscle with induced inflammation (42). Fang et al. (43) confirmed that miR-455 can ameliorate lipid metabolic disorders of the liver in db/db mice by inhibiting SOCS3. All these studies indicate that miR-455 may play a significant role during diabetes. However, despite this body of evidence, thus far, no studies have investigated the role of miR-455 in pancreatic β-cells. In line with previous studies showing that miR-455 plays a critical role in cell proliferation (44,45), we observed that miR-455 overexpression induced β-cell proliferation in vivo and in vitro. Furthermore, we demonstrate that β-cell–specific knockdown of miR-455 through pancreatic ductal infusion markedly aggravated HFD-induced insulin resistance resulting from a decrease in β-cell hyperplasia. Moreover, in line with islet expression analysis in insulin-resistant models, β-cell–specific knockdown of miR-455 (len-anti-miR-455) via pancreatic ductal infusion in ob/ob mice decreased the β-cell mass and KI-67 incorporation rate. The level of miR-455 was simultaneously increased in the islets of different types of obese mouse models characterized by β-cell mass expansion, suggesting a general role for miR-455 in this important compensatory mechanism. Interestingly, our in vivo experimental results revealed that insulin secretion and blood glucose level did not substantially differ between len-miR-455 and control mice. Although many studies have explored whether miRNAs play an important role in obesity and diabetes (10,46), miRNAs simply act as regulators of gene expression via posttranscriptional regulation. Thus, the function of miRNAs might not trigger pleiotropic effects. For example, Tattikota et al. (35) found that miR-184 regulates β-cell proliferation but has no effect on insulin secretion, and Jacovetti et al. (12) showed that miR-338-3p and miR-451 regulate β-cell proliferation but do not significantly affect insulin content or glucose-induced insulin secretion.

Based on a report by Shoshan et al. (47), miR-455 contributes to melanoma growth and metastasis by targeting CPEB1. In our study, we verified that miR-455 can directly target CPEB1 in β-cells, which regulate many important biological processes, ranging from cell-cycle control to insulin resistance (30,31), by controlling mRNA translation efficiency via the 3′-UTR (32,48). In agreement with these findings, we discovered that CPEB1 promotes Cdkn1b translation efficiency through elongation of the 3′-UTR of Cdkn1b mRNA. Consistent with the results of Galardi et al. (33), CPEB stimulates Cdkn1b expression by promoting elongation of the poly(A) tail and translation efficiency. Here, our results reveal that Cpeb1 overexpression facilitates Cdkn1b translation, while miR-455 restored this effect. Previous studies have shown that CDKN1B plays a negative role in the proliferation of β-cells, which is consistent with our findings that the expression level of CDKN1B is downregulated in compensatory proliferative β-cells (49,50). Altogether, these results support a role of miR-455 in promoting β-cell adaptive proliferation by inhibiting the expression of Cdkn1b via direct targeting of Cpeb1.

In summary, we demonstrate the effects of nutritional stress on miR-455, which is involved in compensatory β-cell mass expansion during obesity. Our results reveal that the miR-455/CPEB1/CDKN1B axis is required for obesity-induced adaptive pancreatic β-cell proliferation. These findings provide novel insights into the mechanism of compensatory β-cell expansion in response to increased insulin demand. Detailed knowledge of the mechanisms controlling the level and activity of miR-455 during obesity may pave the way to new therapeutic strategies, with the ultimate goal being prevention of T2D by promoting β-cell mass regeneration.

Acknowledgments. The authors thank Rui Liang (Organ Transplant Center, Tianjin First Central Hospital, Nankai University, Tianjin, China) for providing human islets.

Funding. This work was supported by the National Natural Science Foundation of China (grants 82070801 and 82100858), open fund of the State Key Laboratory of Pharmaceutical Biotechnology, Nanjing University (grant KF-GN-202103), “111” project (B16046), Priority Academic Program Development of Jiangsu Higher Education Institutions, China Postdoctoral Science Foundation (grant 2020M671661), Jiangsu Province Science Foundation for Youths (grant BK20200569), and Jiangsu Province Postdoctoral Research Fund (grant 1412000016).

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

Author Contributions. Q.H., J.M., Yuh.L., Y.Y., and Yue.L. performed the experiments, Y.P., Y.Z., L.L., and D.L. analyzed data. J.C., F.Z., and L.J. designed the project, interpreted the data, and wrote the manuscript. L.J. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.