Work

Integratome analysis of adipose tissues reveals abnormal epigenetic regulation of adipogenesis, inflammation, and insulin signaling in obese individuals with type 2 diabetes

Year: 2021

Type: article

Abstract: Obesity and type 2 diabetes (T2D) often but, not invariably, coexist. We examined the transcriptomes and methylomes of subcutaneous adipose (SAT) and visceral adipose tissues (VAT) samples from 25 obese individuals (36% men; mean±standard deviation (SD) body mass index: 39.1 ± 4.6 kg/m2; age: 38.6 ± 11.8 years; 12 had T2D; Table S1) with or without T2D during metabolic surgery. By integrating these datasets with public tissue-specific regulatory networks, we revealed perturbations in adipogenesis, inflammatory, and insulin signaling pathways in obese individuals with T2D with validation using multiple external databases. The whole transcriptome profiles identified only a few differentially expressed genes (DEGs) in T2D, particularly in T2D-SAT (Figure 1A; Table S4). This finding accorded with previous reports showing only modest differences in gene expression between T2D and control subjects1. Other studies also identified DEGs implicated in glucose and insulin metabolism mainly in VAT compared with SAT2. These low levels of expression emphasize the need to use an integrated approach to identify these complex gene networks. Despite these differences in DEG in SAT and VAT, enrichment analysis indicated dysregulation of cell metabolism and inflammation in both SAT and VAT in obese individuals with T2D (Figure 1B,C). Global methylation levels, as determined by their relative distances to CpG islands, transcription start sites (TSSs), histone marks, enhancer regions, and other annotations, were similar between VAT and SAT in both T2D and control individuals (Figure S2A-E). However, transcription factor binding sites (TFBSs) in T2D-VAT showed a differential distribution of methylation compared to T2D-SAT (Figure 2A,B). Additionally, 19 and 31 differentially methylated regions (DMRs) were detected in T2D-SAT and T2D-VAT, respectively (Figure 2C; Table S5). We discovered a novel hypomethylated region in the promoter of LCLAT1 in both T2D-VAT and T2D-SAT samples (Figure 2D). According to ENCODE data and ChromHMM analysis, this hypomethylated region could facilitate transcription factor (TF) binding and activate gene expression. We also found hypermethylation spanning the 5'UTR of HOXA3 specific to T2D-SAT, accompanied by a depletion of its coding mRNA levels in individuals with T2D (Figure 2E). Simply integrating DEGs with DMRs cannot fully elucidate the complex biological networks implicated in T2D and obesity. Thus, we used tissue-specific regulatory networks to discover epigenetically dysregulated gene modules in adipose tissues and their associations with T2D (Supplemental materials). We detected three modules in T2D-SAT and five modules in T2D-VAT by integrating transcriptomes, methylomes, and tissue-specific regulatory networks (Figure 3A, 3C, 3E; Figure S3A-F). There were 19 genes common to the T2D-SAT and T2D-VAT modules, with an enriched functional annotation of transcriptional regulation (Figure 3B). These findings suggested dysregulated biological pathways shared by SAT and VAT in obese individuals with T2D. Among the 82 T2D-VAT module genes, some were implicated in circadian rhythm, disruption of which could contribute to the development of T2D3, while many others were known genes associated with T2D. For example, slight downregulation of TFEB might result in reduced adipogenesis known to be associated with an increased risk of T2D4. Interestingly, we identified a HOX gene-enriched module in T2D-SAT (Figure 3C-D, Supplemental materials). The distribution and pattern of HOX genes differed between the upper and lower body which might explain the different prognostic significance of VAT (predominate in the upper body) and SAT (predominate in the lower body)5. Given the association of VAT with cardiovascular disease and T2D risk, and the protective effects of SAT6, the identification of HOX genes as a major linking biomarker provides new insights regarding the causal role of adipogenesis in T2D. In this HOX gene-enriched T2D-SAT module, we identified multiple genes with differential expression and methylation, with several examples highlighted (Figure 3D). Consistent with the previous studies,7 we found downregulated trends of HOXD9 and MEOX2, and upregulation of PRRX1. Along with APCDD1 and SPI1, PRRX1 could inhibit PPARγ-mediated adipocyte differentiation and adipogenesis. On the other hand, cooperative expression of HOXD9, MME, SPI, and TLR4 might impair insulin signaling and secretion accompanied by obesity-induced inflammatory responses8, 9. Using EpiMap10, the rs34872471 genetic signal overlapped with the adipose tissue-specific enhancer region nearest to the TCF7L2 promoter in T2D patients (Figure 3C). Other genes in the module were potentially novel T2D markers, such as ASPA, an interactor of TCF7L2, which was hypermethylated with downregulation in T2D-SAT. The novel T2D-SAT-specific 5'UTR of HOXA3 hypermethylated region (Figure 2E) was related to all HOXA3-regulated genes (Figure 3A), supporting their roles in the epigenetic regulation in T2D. Taken together, this HOX gene-enriched module may participate in inhibiting PPARγ-mediated adipocyte differentiation and adipogenesis, and impairing insulin signaling and secretion accompanied by obesity-induced inflammatory responses. In another T2D-SAT-specific module (Figure 3E-F), we identified novel T2D biomarkers. The expression levels of ACACB, ELF1, IL1RL1, and SPI1 were confirmed by qPCR validation in additional T2D-SAT samples (Figure S4). ELF1, IL1RL1, NR3C1, and TFCP2 were known to reduce adipocyte differentiation which can lead to abnormal glucose metabolism and inflammation, while APBB1IP and FBN1 were predicted to be involved in these biological pathways. Taken together, this module might provide a novel epigenetic pathway regulating insulin signaling through adipocyte differentiation and inflammatory responses in obese patients with T2D. We identified 161 potential biomarkers in these networks and DMRs which were independently validated in at least one external database relevant to comorbidity, druggability, expression quantitative trait loci (eQTL), genome-wide association studies (GWAS), TFBSs, TFs, or the T2D integratome (T2Di) (Figure 4A-B; Table S6-S7). Of these 161 modular biomarkers enriched in multiple databases, 73.9% were validated in at least one external dataset and 48.4% were TFs. Amongst the biomarkers shared by both tissues, 7 were potential drug targets. By integrating differential gene expression and methylation levels in SAT and VAT collected during metabolic surgery from obese T2D and non-T2D individuals with tissue-specific regulatory networks, we found multiple epigenetic regulatory networks in both SAT and VAT associated with obesity in T2D. These findings confirmed current knowledge regarding the pathophysiological roles of different adipose tissues in insulin resistance, inflammation, and development of T2D whilst revealing novel relationships not detectable by single-layered analysis. This study was partially supported by the Research Grants Council of Hong Kong General Research Fund (GRF) grant 14102014 to TFC, the Research Grants Council Theme-based Research Scheme grant (T12-402/13N) and Research Impact Fund (R4012-18) to RCWM and JCNC. ACSY, JWL are co-founders of Codex Genetics Limited. NJ is working at Codex Genetics Limited. JCNC and RCWM are co-founders of GemVCare. The research protocol and informed consent forms were approved by the Joint Chinese University of Hong Kong-New Territories East Cluster Clinical Research Ethics Committee (Ref no. CRE-2011.131) in accordance with the Declaration of Helsinki. All individuals gave written informed consent. TFC, JCNC, and RCWM conceived and directed the study. APSK and CCHL managed the clinical research ethics and patient consent. SKHW and EKWN performed the adipose tissue biopsies. HML and YH performed the experiments. NJ, ACSY, and JWL performed the bioinformatic analyses. NJ, TFC, and JCNC interpreted the data. NJ drafted the manuscript. TFC and JCNC revised the manuscript. All authors read and approved the final version. RNA sequencing data are deposited at the NCBI Sequence Read Archive (SRA) database with the accession number PRJNA679994. DNA methylation data are deposited at the Gene Expression Omnibus (GEO) database with the accession number GSE162166. RNA sequencing data are deposited at the NCBI Sequence Read Archive (SRA) database with the accession number PRJNA679994. DNA methylation data are deposited at the Gene Expression Omnibus (GEO) database with the accession number GSE162166. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Source: Clinical and Translational Medicine

DOI: 10.1002/ctm2.596