MicroRNA-24/MODY gene regulatory pathway mediates pancreatic β-cell dysfunction

Abstract

Overnutrition and genetics both contribute separately to pancreatic β-cell dysfunction, but how these factors interact is unclear. This study was aimed at determining whether microRNAs (miRNAs) provide a link between these factors. In this study, miRNA-24 (miR-24) was highly expressed in pancreatic β-cells and further upregulated in islets from genetic fatty (db/db) or mice fed a high-fat diet, and islets subject to oxidative stress. Overexpression of miR-24 inhibited insulin secretion and β-cell proliferation, potentially involving 351 downregulated genes. By using bioinformatic analysis combined with luciferase-based promoter activity assays and quantitative real-time PCR assays, we identified two maturity-onset diabetes of the young (MODY) genes as direct targets of miR-24. Silencing either of these MODY genes (Hnf1a and Neurod1) mimicked the cellular phenotype caused by miR-24 overexpression, whereas restoring their expression rescued β-cell function. Our findings functionally link the miR-24/MODY gene regulatory pathway to the onset of type 2 diabetes and create a novel network between nutrient overload and genetic diabetes via miR-24.

FIG. 1.

Expression of miRNA-24 is increased in pancreatic islet cells. A: Levels of 13 miRNAs in isolated islets from 8-week-old db/db mice (black) were analyzed relative to controls (white). Among them, miR-127, miR-21, miR-30c, miR-375, and miR-7 were statistically significantly downregulated, whereas miR-24, miR-376a, miR-146a, and miR-34a were notably upregulated (P < 0.05). No changes were observed in miR-124a, miR-130a, miR-15a, and miR-181a. U6 small nuclear RNA was used as an internal control to normalize miRNA expression (*P < 0.05 or **P < 0.01 vs. control mice). B: Islets from 8- and 12-week-old db/db mice and controls were isolated, and the expression of miR-24 normalized to U6 was measured using qRT-PCR (*P < 0.05 or **P < 0.01 vs. control mice). The expression of miR-24 increased with age in the islets of db/db mice. Increasing expression levels of miR-24 and miR-34a in islets from HFD-fed mice (n = 5) compared with controls fed the standard diet (SD) (C) (*P < 0.05 or **P < 0.01 vs. SD), palmitate-induced islets (D) (*P < 0.05 or **P < 0.01 vs. ethanol), and in palmitate-treated MIN6 cells (E) (*P < 0.05 vs. ethanol) were observed by TaqMan qRT-PCR relative to corresponding controls. U6 detected by a TaqMan probe was used as an internal control. F: Levels of miR-24 were upregulated in MIN6 cells challenged for various times (12, 24, and 48 h) with (black) or without (white) palmitate (*P < 0.05 vs. ethanol).

FIG. 2.

Elevated miR-24 reduces cell viability and impairs β-cell function. A: Pre–miR-24 mimetics or pre-Neg at different concentrations (2, 10, or 50 nmol/L) were transfected into MIN6 cells for 48 h, when TaqMan qRT-PCR was carried out. Transfection caused an effective increase in miR-24 abundance in MIN6 cells. B: Overexpression of miR-24 for 48 h caused a decrease of cell viability in MIN6 cells, measured by the WST-1 assay. Transfection of 10 nmol/L miR-24 led to a decrease of cell number in the S phase (C) and to an increase in the G2 phase (D). The same results were detected with transfection of miR-24 at 50 nmol/L but not at 2 nmol/L, which was insufficient to induce cell cycle arrest. E and F: BrdU labeling was used to confirm the reduced DNA synthesis accompanying the elevation of miR-24. E: Representative images show BrdU and Hoechst stained cells, and at least 800 cells were counted. F: The BrdU labeling index is defined as the ratio of the number of BrdU⁺ nuclei to the total number of nuclei within the fields. G: Decreased cell proliferation was also detected in primary islets isolated from ICR mice. GSIS and KSIS assays were performed on MIN6 cells overexpressing miR-24 for 48 h, and the GSIS index (H) and KSIS index (I) were calculated. The results were similar to those in palmitate-treated cells. **P < 0.01 vs. pre-Neg.

FIG. 3.

miR-24 directly downregulates six transcription factors. A: 3′UTR sequences of the six transcription factors predicted to include miR-24 MREs were aligned with miR-24, and both WT and mutant sequences are listed. B: Luciferase reporter activities analyzed as above of WT gene promoters were significantly repressed by elevated miR-24, whereas those of the mutant gene promoters were reversed to the normal level (**P < 0.01 vs. vector; ##P < 0.01 vs. wt). C: mRNAs analysis showed that only Neurod1 and Pdx1 were slightly downregulated, but others were not altered (**P < 0.01 vs. actin). D: Conserved miR-24 binding sites in the 3′UTR of Neurod1 from mouse and human. E: Full-length 3′UTR sequence of mouse Neurod1 was inserted downstream of a reporter gene, and point mutation of the miR-24 seed pairing sequence was generated using the QuikChange Site-Directed Mutagenesis kit. miR-24 inhibited the luciferase activity of WT Neurod1, whereas those of the mutant construct and vector control were indistinguishable from each other (**P < 0.01 vs. vector+pre–miR-24; ##P < 0.01 vs. wt-Neurod1+pre–miR-24). F: miR-24 (50 nmol/L) apparently decreased the Neurod1 protein level starting from 48 h. G: miR-24 at both 10 and 50 nmol/L were sufficient to repress Neurod1 protein production.

FIG. 4.

Protein and mRNA level analysis of genes are functionally associated with miR-24. A: miR-24 was transfected into MIN6 cells for 48 h, and then total protein was extracted for analysis of Hnf1a, Neurod1, Pdx1, Parp1, Cdk4, Cyclind3, p27, p15, Pten, Cycind1, kir6.1, and β-tubulin by Western blot. Protein levels remaining in cells relative to the negative control were quantitatively analyzed by Quantity One 4.2.1 (Bio-Rad). B: A decrease of Cyclind3 was detected first at 24 h posttransfection of miR-24 (50 nmol/L) and that for Cdk4 was at 48 h. Meanwhile, the peak expression of p15 was at 24 h. C: Transfection with miR-24 at 10 and 50 nmol/L downregulated Cyclind3 and Cdk4 protein levels and upregulated the p15 level. D: Seven cell cycle–associated genes (Ccnd1, Ccnd2, Ccnd3, Cdk4, p15, p21, and p27) and genes of two components of ATP-sensitive potassium (K_ATP) channels (Kcnj8 and Kcnj11) were determined by qRT-PCR. Ccnd3, Cdk4, and p27 were decreased, whereas p15 and p21 were increased (*P < 0.05 or **P < 0.01 vs. actin). E: mRNA levels of Ccnd3 and Cdk4 were also downregulated in islets isolated from ICR mice (**P < 0.01 vs. actin).

FIG. 5.

Analysis of cellular phenotypes by silencing of Hnf1a and Neurod1. Silencing efficiencies of Hnf1a and Neurod1 were measured at mRNA (A) and protein (B) levels using qRT-PCR and Western blotting, respectively (**P < 0.01 vs. siNC). Results demonstrated that the siRNAs were effective. C: DNA synthesis was assessed by BrdU labeling. Hnf1a and Neurod1 were both able to inhibit DNA synthesis, reflecting the capacity to reduce proliferation. This inhibition was enhanced by knocking down both proteins (**P < 0.01 vs. siNC). D: GSIS assay was performed 48 h post-transfection of indicated siRNAs, and the released insulin was normalized to corresponding total insulin content. Basal and stimuli-induced insulin release were both repressed upon downregulation of Hnf1a or Neurod1 (*P < 0.05 or **P < 0.01 vs. siNC).

FIG. 6.

Overexpression of Neurod1 or Hnf1a blocked the cellular dysfunction induced by miR-24. Successful overexpression of Neurod1 restored the Cdk4 protein level (A), followed by the recovery of the GSIS index (B), KSIS index (C), and DNA synthesis (D) (**P < 0.01 vs. Neg+myc; ##P < 0.01 vs. miR-24+myc). Overexpression of Hnf1a also helped to rescue Cdk4 protein (E) as well as to restore the GSIS index (F), KSIS index (G), and DNA synthesis (H) (*P < 0.05 or **P < 0.01 vs. Neg+Adtrack; ##P < 0.01 vs. miR-24+Adtrack).

FIG. 7.

Knockdown of miR-24 expression rescued the GSIS defect in islets from HFD-fed mice. C56BL/6 mice aged 8 weeks were fed an HFD or standard diet (SD) for 10 weeks. Mice were fasted for 8 h before measuring body weight (A) and blood glucose (B). Islets from HFD-fed mice (n = 9) and SD-fed mice (n = 16) were isolated and cultured for 3 h before GSIS (C) was performed. HFD mice exhibited high body weight, hyperglycemia, and defective GSIS compared with SD mice. Secreted insulin was normalized to relative insulin content (**P < 0.01 vs. SD group). D: Islets from HFD mice were transfected with 100 nmol/L Cy3-labled Anti-Neg. After 48 h post-transfection, photographs of Cy3-labeled islets were acquired by fluorescent microscopy and used to quantify the transfection efficiency of miRNAs in primary isolated islets. After 3-h recovery in culturing medium, islets isolated from HFD mice were transfected with 100 nmol/L Anti-Neg or Anti-miR miRNA-24 inhibitor (Anti-miR-24) for 48 h, at which time qRT-PCR (E) and GSIS (F) were carried out. U6 was used as an internal control for miRNA analysis. Insulin secretion was normalized to relative insulin content (**P < 0.01 vs. Anti-Neg + SD; #P < 0.05 or ##P < 0.01 vs. Anti-Neg + HFD).

FIG. 8.

Oxidative stress induced miR-24 expression. MIN6 cells were precultured in DMEM with 5.5 mmol/L glucose for 24 h, and then cells were incubated with the indicated concentration of H₂O₂ for an additional 24 h or with glucose for an additional 48 h. Relative expression of miR-24 induced by H₂O₂ (A) or glucose (B) was analyzed by TaqMan qRT-PCR relative to corresponding controls. U6 detected by a TaqMan probe was used as an internal control. Values are the mean ± SEM of three individual experiments (*P < 0.05; **P < 0.01). C and D: The protein levels of Hnf1a and Neurod1 were observed by Western blotting. β-Tubulin was used as an internal control.