The type 2 diabetes-associated HMG20A gene is mandatory for islet beta cell functional maturity article

Mellado-Gil et al. Cell Death and Disease (2018) 9:279 DOI 10.1038/s41419-018-0272-z Cell Death & Disease ARTICLE Open Access Thetype2diabetes-associatedHMG20A gene is mandatory for islet beta cell functional maturity Jose M. Mellado-Gil 1 , Esther Fuente-Martín 1 , Petra I. Lorenzo 1 , Nadia Cobo-Vuilleumier 1 , Livia López-Noriega 1 , Alejandro Martín-Montalvo 1 , Irene de Gracia Herrera Gómez 1 , Maria Ceballos-Chávez 2 , Laura Gómez-Jaramillo 3 , Antonio Campos-Caro 3 , Silvana Y. Romero-Zerbo 4 , Júlia Rodríguez-Comas 5 , Joan-Marc Servitja 5,6 , Gemma Rojo-Martinez 4,6 , Abdelkrim Hmadcha 1,6 ,BernatSoria 1,6 , Marco Bugliani 7 , Piero Marchetti 7 , Francisco J. Bérmudez-Silva 4,6 ,JoseC.Reyes 2 , Manuel Aguilar-Diosdado 3,8 and Benoit R. Gauthier 1 Abstract HMG20A (also known as iBRAF) is a chromatin factor involved in neuronal differentiation and maturation. Recently small nucleotide polymorphisms (SNPs) in the HMG20A gene have been linked to type 2 diabetes mellitus (T2DM) yet neither expression nor function of this T2DM candidate gene in islets is known. Herein we demonstrate that HMG20A is expressed in both human and mouse islets and that levels are decreased in islets of T2DM donors as compared to islets from non-diabetic donors. In vitro studies in mouse and human islets demonstrated that glucose transiently increased HMG20A transcript levels, a result also observed in islets of gestating mice. In contrast, HMG20A expression was not altered in islets from diet-induced obese and pre-diabetic mice. The T2DM-associated rs7119 SNP, located in the 3′UTR of the HMG20A transcript reduced the luciferase activity of a reporter construct in the human beta 1.1E7 cell line. Depletion of Hmg20a in the rat INS-1E cell line resulted in decreased expression levels of its neuronal target gene NeuroD whereas Rest and Pax4 were increased. Chromatin immunoprecipitation confirmed the interaction of HMG20A with the Pax4 gene promoter. Expression levels of Mafa,Glucokinase, and Insulin were also inhibited. Furthermore, glucose-induced insulin secretion was blunted in HMG20A-depleted islets. In summary, our data demonstrate that HMG20A expression in islet is essential for metabolism-insulin secretion coupling via the coordinated regulation of key islet-enriched genes such as NeuroD and Mafa and that depletion induces expression of genes such as Pax4 and Rest implicated in beta cell de-differentiation. More importantly we assign to the T2DM-linked rs7119 SNP the functional consequence of reducing HMG20A expression likely translating to impaired beta cell mature function. Introduction Type 2 Diabetes Mellitus (T2DM) is a metabolic disease characterized by impaired insulin secretion and/or action in target organs that leads to elevations in blood glucose levels. Environmental factors as well as strong genetic components contribute to the pathogenesis of T2DM. Thus far, 100–120 susceptibility loci have been associated to T2DM by Genome Wide Association Studies (GWAS) 1–3 . Although functional defects remain to be © The Author(s) 2018 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a linktotheCreativeCommons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Correspondence: Benoit R Gauthier ([email protected]) 1 Department of Cell Regeneration and Advanced Therapies, Andalusian Center of Molecular Biology and Regenerative Medicine-CABIMER, Junta de Andalucia-University of Pablo de Olavide-University of Seville-CSIC, Seville, Spain 2 Department of Genome Biology, Andalusian Center of Molecular Biology and Regenerative Medicine (CABIMER) JA-CSIC-UPO-US, Seville, Spain Full list of author information is available at the end of the article These authors contributed equally: Jose M. Mellado-Gil, Esther Fuente-Martín. Edited by M. Agostini Official journal of the Cell Death Differentiation Association 1234567890():,; 1234567890():,; b DAPIINSHMG20A d c DAPIGLUCHMG20A DAPISMTHMG20A DAPIINSHMG20A DAPIGLUCHMG20A DAPISMTHMG20A a Fig. 1 (See legend on next page.) Mellado-Gil et al. Cell Death and Disease (2018) 9:279 Page 2 of 15 Official journal of the Cell Death Differentiation Association assigned, many of these loci point to primary defects in beta cell function rather than to insulin resistance 4 . In this context, several SNPs within the HGM20A gene (also known as iBRAF) have been associated to T2DM 5–8 .Of particular interest is the rs7119 SNP located within the 3′ UTR of the HMG20A transcript. In silico analysis revealed that rs7119 modifies a functional microRNA (miRNA) cis-regulatory element 9 . However, the molecular consequences of this SNP on the expression of HMG20A and/or its regulation and impact on islet physiology are still unknown. HMG20A is a member of the high mobility group (HMG) box-containing genes that binds to chromatin and exerts global genomic changes through establishing active or silent chromatin 10 . HMG20A is highly expressed in mature neurons and plays a key role in promoting neuronal differentiation during development 11 . In this context, HMG20A relieves the transcriptional repression imposed by the complex LSD1–CoREST histone demethylase 12 , which function is to silence neuronal genes in non-neuronal tissues through its interaction with the transcription factor REST. In analogy, epigenetic repression of the Rest gene in pancreatic precursors was shown to coincide with the activation of the core beta cell program 13 . Mature pancreatic beta cells do not express REST while forced expression results in inhibition of neuronal proteins of the insulin exocytotic machinery such as SNAP25 and SYNAPTOTAGMIN VII (SYT7) leading to impaired glucose-induced insulin secretion 14 . Interestingly, REST was shown to repress expression of key beta cell development genes, such as NEUROD and PAX4 15,16 . NEUROD is a bona fide target gene of HMG20A in neuronal cells 11 and mutations in this gene predispose individuals to maturity onset diabetes of the young 6 (MODY6) 17 . HMG20A is also implicated in epithelial-to- mesenchymal transition (EMT) through interaction with specific key regulators of this process such as SNAIL 18 . EMT is an example of cell plasticity and a key process during embryonic development, and together with the reverse transformation, the mesenchymal-to-epithelial transition (MET), are required for the formation of organs in the final destinations of embryonic migratory cells 19,20 . Both processes occur during pancreatic and islet development and require extensive reorganization of the epigenetic information of the cells. Remarkably, not only in development but also in response to different physiological demands, beta cells may de-differentiate in order to acquire plasticity capabilities and increase survival 21,22 , two processes that may implicate PAX4 23,24 . Based on these findings, we hypothesized that HMG20A may contribute to the regulation of key genes such as NeuroD and Pax4 which are essential for beta cell functional maturity as well as survival. Towards this goal, herein we investigate the expression profile and the gene regulatory function of HMG20A in pancreatic islets, as well as determine whether the SNP rs7119 associated with T2DM impacts HMG20A expression. We report that T2DM-associated SNP rs7119 leads to altered HMG20A expression, and that HMG20A regulates metabolismsecretion coupling genes as well as functional maturity of beta cells. Results HMG20A is expressed in pancreatic islets and transcript levels are decreased in islets from T2DM donors As a first step to assign a potential role of the HMG20A gene in pancreatic islet physiology, we determined its transcript levels in islets as compared to other tissues. Mouse pancreatic islets displayed comparable expression levels of Hmg20a to other organs such as adipose tissue (white and brown), brain and muscle whereas the liver displayed highest levels (Fig. 1a). HMG20A co-stained with INSULIN (beta cells), GLUCAGON (alpha cells) and SOMATOSTATIN (delta cells) whereas its expression was rarely detected in exocrine pancreas (Fig. 1b). A similar endocrine cell expression pattern was detected in human islets (Fig. 1c). We next assessed whether expression levels of HMG20A were altered in islets isolated from T2DM donors. The rational was to establish a correlation between HMG20A levels and the hyperglycemic environment that may be altered by T2DM-linked SNPs. HMG20A transcript levels were decreased by ~60 % in islets from T2DM patients as compared to islets purified from normoglycemic control donors (Fig. 1d). Metabolic stressors modulate HMG20A expression in islets As HMG20A transcript levels were decreased in T2DM islets, we reasoned that lipids and/or glucose might alter its expression. Towards addressing this premise, we used (see figure on previous page) Fig. 1 HMG20A is expressed in both mouse and human islets and is decreased in islets from T2DM donors. a HMG20A transcript levels were assessed by qPCR in islets, brain, liver, muscle, white adipose tissue (WAT), and brown adipose tissue (BAT) from mice (n= 6). Representative images of bmouse and chuman islets co-stained for HMG20A (green) along with INSULIN (INS), GLUCAGON (GLUC) or SOMATOSTATIN (SMT) (red). Nuclei are stained using DAPI (blue). Magnification 40×. White boxes define areas enlarged in panels bellow. dHMG20A mRNA levels were measured by qPCR in human islets isolated from normoglycemic (control) or type 2 diabetic (T2DM) organ donors (n= 7). Data are depicted as dot plots with means ± S.E.M. pvalues were determined using unpaired two-tailed Student's t-test. **p< 0.01 Mellado-Gil et al. Cell Death and Disease (2018) 9:279 Page 3 of 15 Official journal of the Cell Death Differentiation Association our mouse model of 45% high fat diet (HFD)-induced obesity and pre-diabetes to assess the contribution of lipids. These mice exhibit a typical 20% increase in body weight, develop glucose intolerance with the concomitant increases in insulin and leptin levels and decreased adiponectin levels 25 . Although these mice developed insulin resistance, glycaemia is normal due to increased insulin levels (Fig. 2a, b). Under these conditions, HMG20A protein levels were not significantly changed in islets as compared to levels in chow-fed control mice (Fig. 2c). In contrast, short-term exposure of INS-1E cells to palmitate resulted in a transient reduction in HMG20A transcript levels (Fig. 2d). We next assessed the direct action of high glucose on HMG20A levels in isolated islets. In human islets, HMG20A expression levels were transiently increased by approximately twofold at 72 h returning to basal levels by 96 h (Fig. 3a). A similar transient induction pattern, albeit attaining maximum levels at 48 h, was also observed in mouse islets and in INS-1E beta cells (Fig. 3b,c). Interestingly, a significant decrease was detected INS-1E cells at 72 h of high glucose exposure (Fig. 3c). It is noteworthy that expression levels of NEUROD, a bona fide target of HMG20A, mirrored those of HMG20A in human islets as well as in INS-1E (Fig. 3d,f), following the same tendency in mouse islets (Fig. 3e). In order to satisfy metabolic demand by fetus and mother, glucose metabolism is altered during pregnancy inducing metabolic stress that, if ill adapted, will lead to gestational diabetes. We therefore pondered whether Hmg20a expression levels were modulated during pregnancy. Consistent with this premise, transcript and protein levels of the Hmg20a gene were increased reaching a peak of maximum signal at 12.5 and 14.5 days post coitum, respectively, and thereafter declining (Fig. 4a–c). Pregnant females displayed mild glucose intolerance at 14.5 correlating with the peak in HMG20A protein levels (Fig. 4d). Taken together, these data suggest that increased glucose levels both in vitro and in vivo stimulate HMG20A expression. The rs7119 SNP variant within the 3′untranslated region (UTR) of human HMG20A reduces activity of a luciferase reporter gene Although GWAS have highlighted a panoply of SNPs correlating with T2DM, few have been assigned a functional consequence to the associated gene. Towards addressing this question, we assessed whether the human HMG20A SNPs associated to T2DM may impact HMG20A levels. Of particular interest is the SNP rs7119 located in the 3′UTR of the HMG20A transcript that is encoded in part by exon 10 and the entirety of exon 11 (Fig. 5c). The SNP rs7119 causes a “C”(reference allele “wt”)to“T”(“mut”allele) substitution. In silico analysis of Fig. 2 HMG20A protein expression is not altered in islets of high fat diet (HFD)-induced obesity and pre-diabetic mice. a Insulin tolerance test (ITT) and area under the curve (AUC) analysis in both control (CT) and HFD mice (CT, n= 10; HFD, n= 15). bFasting blood glucose levels in both control and HFD mice (CT, n= 10; HFD, n= 15). cRepresentative images of pancreatic sections from HFD and CT mice stained and quantified for HMG20A (n= 4). dHmg20a mRNA levels were assessed by qPCR in INS-1E cells cultured with 0.05 mM palmitate for 24, 48, 72 or 96 h (n= 3 independent experiments performed in triplicates). Data are represented as means + S.E.M. p values were determined by ordinary one-way ANOVA. *p< 0.05, **p< 0.01, ***p< 0.0001 Mellado-Gil et al. Cell Death and Disease (2018) 9:279 Page 4 of 15 Official journal of the Cell Death Differentiation Association the HMG20A 3′UTR revealed the binding site of a single miRNA (miR-571) to the wt allele whereas the “mut”allele results in the ablation of this miRNA binding site (Table 1). Interestingly expression of miR-571 was not detected in mouse islets as compared to the highly expressed miR375 (data not shown). Notwithstanding, the Fig. 3 HMG20A is regulated by glucose. a–cHMG20A and d–fNEUROD mRNA levels were assessed by qPCR in human (n= 3 donors performed in triplicates) and mouse islets (three independent preparations of pooled islets from three different isolation) as well as INS-1E cells cultured in 24mM glucose for 24, 48, 72, or 96 h (n= 3 independent experiments of all time points executed in triplicates). Data are represented as means + S.E.M. p values were determined by ordinary one-way ANOVA. *p< 0.05, compared to the control 0 time point Mellado-Gil et al. Cell Death and Disease (2018) 9:279 Page 5 of 15 Official journal of the Cell Death Differentiation Association c d a 0 10.5 12.5 14.5 17.5 DAPIINSHMG20A Detail Days p.c.: b Fig. 4 HMG20A is transiently increased in mouse islets during pregnancy. a Representative images of pancreatic sections from pregnant mice co-stained for HMG20A (green) and INSULIN (INS) (red). Nuclei were stained using DAPI (blue). Magnification 40×. White boxes define areas enlarged in panels below. p.c. post coitum. bQuantification of integrated fluorescence intensity for HMG20A in beta cells of islets from pregnant mice (n=3 mice per time point with 35-71 islets counted per time point). Data are depicted as dot plots with means ± S.E.M. cHMG20A mRNA levels in islets from pregnant mice (n= 5 mice) at 12.5 and 14.5 days of pregnancy. dOral glucose tolerance test (OGTT) and area under the curve (AUC) analysis of non-pregnant (NP) and pregnant (day 14.5 of pregnancy; D14.5) mice (NP, n= 8; D14.5, n= 6). Data are represented as means ± S.E.M. pvalues were determined by the unpaired two-tailed Student´s t-test (dfor AUC and c) or ordinary one-way ANOVA with Dunnet’s multiple comparison test (b). *p < 0.05, **p< 0.01, ****p< 0.0001, compared to the control 0 time point Mellado-Gil et al. Cell Death and Disease (2018) 9:279 Page 6 of 15 Official journal of the Cell Death Differentiation Association b rs7119 HMG20A DN A RN A GLuc CDS 3’UTR GLuc CDS 3’UTR wt mut d tss 1234567891011 SV40 promoter Promoter c a ef Fig. 5 The rs7119 SNP within the HMG20A 3′UTR decreases protein levels. Expression profile of putative miRNAs binding to the “mut”allele of the HMG20A transcript in ahuman 1.1E7 cells and bmouse islets. Data are presented respect to miR375 expression levels. cSchematic representation of the HMG20A gene depicting the promoter region (HMG20A promoter), transcription starting site (tss) and 11 exons. Gaussia luciferase (GLuc) reporter constructs harboring the 3′UTR corresponding to either wild type (wt) or rs7119 mutant (mut) allele under pSV40 promoter are shown below. dHMG20A mRNA levels assessed by qPCR in the human pancreatic beta cell line 1.1E7 cultured in 24 mM glucose for 24, 48, 72, or 96 h (n= 5 independent experiments performed in triplicates). eHuman pancreatic beta 1.1E7 and fRPE1 cell lines were transfected with the reporter construct bearing either wt or rs7119 mut variant of the HMG20A3′UTR. Both the Gaussia Luciferase activity and SEAP activity were determined 24 h post transfection (n= 3 independent experiments). Data are represented as means + S.E.M. pvalues were determined by ordinary one-way ANOVA (d) or unpaired Student's t-test (e,f). *p< 0.05 Mellado-Gil et al. Cell Death and Disease (2018) 9:279 Page 7 of 15 Official journal of the Cell Death Differentiation Association “mut”allele generates binding sites for six alternative miRNAs (Table 1) of which four, miR-143 (3p and 5p), miR-490-3p, miR-1246, and miR-1261, were expressed in the human 1.1E7 pancreatic cell line (Fig. 5a). In contrast mouse islets only expressed miR-143 and very low levels of miR-1246 (Fig. 5b). The latter findings suggest that the “mut”variant may alter expression of the HMG20A transcript due to the action of those miRNAs. To assess this premise, we generated two Gaussia luciferase (Gluc) reporter constructs harboring either the “wt”or “mut”3′ UTR of the human HMG20A transcript. Expression of these constructs was under the transcriptional regulation of the SV40 basal promoter (Fig. 5c). Transient transfections were then performed in the human pancreatic beta 1.1E7 and retinal pigmented epithelial RPE1 cell lines. The rational of using human cell lines was to maintain species-specific effects targeting the human 3′UTR of the HMG20A transcript, which is distinct than that of the mouse Hmg20a transcript. We initially confirmed that glucose also stimulated HMG20A expression in 1.1E7 beta cells (Fig. 5d). The “mut”variant displayed significantly less Gluc activity as compared to the construct bearing the “wt”variant 48 h post transfection in either cell lines (Fig. 5e, f). These results suggest that the diabetes-linked SNP rs7119 alters expression of the human HMG20A. HMG20A regulates the expression of beta cell genes involved in metabolism-secretion coupling and beta cell maturity To determine the specific contribution of HMG20A to islet beta cell function and the impact of its reduced levels in T2DM islets, we silenced HMG20A expression by RNA interference and then assessed transcript levels of key factors involved in beta cell identity and function. A 60% depletion of Hmg20a in INS-1E cells (Fig. 6a, b) resulted in a concomitant 60% inhibition of NeuroD transcript levels (Fig. 6c). Mafa, which is essential to maintain the beta cell phenotype, was decreased by 50% whereas Pdx1 was unaltered. Interestingly, Pax4 as well as Rest were increased by 1.8- and 1.5-fold, respectively, subsequent to Hmg20a silencing (Fig. 6c). We previously demonstrated that chronic overexpression of Pax4 leads to loss of beta cell identity characterized by decreased Mafa and Insulin expression 23 . We thus, pondered whether HMG20A could directly regulate Pax4 thereby maintaining its expression within a range that would not interfere with beta cell function. To validate this premise, chromatin immunoprecipitation (ChIP) experiments were performed using the islet specificPax4 gene promoter 24 . As expected H3K4me2, a marker for active promoters and enhancers specifically occupied this region (Fig. 6d) 26 . More importantly, HMG20A also occupied the Pax4 gene promoter region as well as exon 1 while control anti-IgG exhibited no binding (Fig. 6e). Since NEUROD and MAFA are involved in metabolism-secretion coupling in beta cell, genes implicated in this process were assessed after Hmg20a silencing. Glucokinase (Gck) and Insulin transcript levels were significantly repressed by 30% (Fig. 6f). In contrast, Pi3k and Glut2 levels were increased 1.8 and 2.6-fold, respectively (Fig. 6f). Transcript levels of Syt7 and Snap25, two factors involved in insulin granule membrane docking and fusion were not altered in Hmg20a-repressed INS-1E cells (Fig. 6f). The modulation of key HMG20A target genes was validated using a second independent siHMG20A (Fig. 6g). Taken together our results indicate that HMG20A coordinates expression of genes that will establish the mature beta cell phenotype. Glucose-stimulated insulin secretion is impaired by Hmg20a depletion We next assess the cellular and functional consequences of Hmg20a depletion in beta cells. Neither cell death nor proliferation was altered by HMG20A depletion in INS- 1E cells (Fig. 7a, b). In contrast, glucose-stimulated insulin secretion (GSIS) was blunted in these cells (Fig. 7c). Similarly, subsequent to a 40% depletion of Hmg20a in mouse islets (Fig. 7d), cell proliferation was unaltered while GSIS was drastically decreased as compared to control islets (Fig. 7e, f). Taken together, Hmg20a expression is essential for insulin secretion. Discussion The overall aim of the current study was to assign a functional role of the putative diabetes-linked HMG20A gene to islet physiology, and determine the pathophysiological consequence of the rs7119 SNP associated with this disease. We find that: (1) HMG20A is expressed in pancreatic islet beta cells, (2) HMG20A levels are decreased in T2DM islets, (3) high glucose stimulates HMG20A expression, (4) the T2DM-associated SNP within the 3′UTR of the HMG20A transcript decreases luciferase activity, (5) HMG20A regulates key genes in beta cell function and maturity, and (6) HMG20A TABLE 1 In silico identification of miRNAs binding sites within the rs7119 SNP located in the 3′UTR of HMG20A “C”reference alelle “T”risk allele (wt) (mut) hsa-miR-571 hsa-miR-1261 hsa-miR-1257 hsa-miR-1246 hsa-miR-586 hsa-miR-490 (5p, 3p) hsa-miR-143 (5p, 3p) Mellado-Gil et al. Cell Death and Disease (2018) 9:279 Page 8 of 15 Official journal of the Cell Death Differentiation Association a bsiCT siHMG20A DAPIHMG20A d e f c g Fig. 6 HMG20A regulates metabolism-secretion coupling genes in INS-1E cells. a Hmg20a was silenced by specific siRNA in INS-1E cells (n=6 independent experiments, each performed in triplicates). bRepresentative images of INS-1E cells immuno-stained for HMG20A (green) and DAPI (blue) confirming decreased protein level after treatment with siHMG20A. Magnification 40×. cBeta cell development and maturity genes, as Neurod, Pax4,Mafa, Pdx1, and Rest transcripts levels after silencing of Hmg20a in INS-1E cells (n= 4 independent experiments performed in triplicates). d H3K4m2 and eHMG20A occupancy of promoter, exon 1 and ATG regions of the Pax4 gene after chromatin immunoprecipitation using anti-H3K4m2 or anti-HMG20A antibodies, with IgG as a control for non-specific interactions (n= 4 independent experiments performed in triplicates). fBeta cell metabolism-secretion coupling genes transcripts levels after silencing of Hmg20a in INS-1E cells (n= 4 independent experiments, performed in triplicates). gAlteration in the expression of key genes was confirmed in INS-1E cells using a second siHMG20A (n= 4 independent experiments performed in triplicates). Data are represented as means + S.E.M. pvalues were determined by one-way ANOVA with Tukey’s multiple comparisons test (a, c, f, g). *p< 0.05, **p< 0.01, ***p< 0.001, ****p< 0.0001 Mellado-Gil et al. Cell Death and Disease (2018) 9:279 Page 9 of 15 Official journal of the Cell Death Differentiation Association