J Obes Metab Syndr 2023; 32(1): 64-76
Published online March 30, 2023 https://doi.org/10.7570/jomes22010
Copyright © Korean Society for the Study of Obesity.
1Department of Biochemistry, All India Institute of Medical Sciences (AIIMS), Jodhpur; 2Indian Institute of Technology (ITT)-Madras, Chennai; 3School of Humanities, Indira Gandhi National Open University (IGNOU), New Delhi; 4Department of Endocrinology and Metabolism, All India Institute of Medical Sciences (AIIMS), Jodhpur; 5Department of General Surgery, All India Institute of Medical Sciences (AIIMS), Jodhpur, India
Department of Biochemistry, All India Institute of Medical Sciences (AIIMS), Medical College Building, Basni Industrial Area Phase 2, Jodhpur, Rajasthan, 342005, India
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background: Growth differentiation factor-15 (GDF-15) is involved in insulin resistance and diabetes. In this study, we determine the associations of GDF-15 with miR-181b-5p, miR-330-3p, mothers against decapentaplegic homolog 7 (SMAD7), and insulin resistance in visceral adipose tissue (VAT) and peripheral blood mononuclear cells (PBMCs) in type 2 diabetes mellitus (T2DM) patients.
Methods: Sixty patients, equally divided into those with T2DM and non-diabetic controls, were recruited for gene expression analysis. Protein-protein interaction (STRING), target prediction (miRNet), and functional enrichment were conducted accordingly.
Results: Our study showed that VAT and PBMCs had similar expression profiles, where GDF-15 and miR-181b-5p were upregulated, whereas SMAD7 and miR-330-3p were downregulated. Serum GDF-15 could differentiate between T2DM and non-diabetic patients (P<0.001). Target prediction revealed a microRNA (miRNA)-messenger RNA regulatory network, transcription factors, and functional enrichment for the miRNA that suggested involvement in T2DM pathogenesis.
Conclusion: VAT GDF-15 is associated with insulin resistance and is possibly regulated by miR-181b-5p, miR-330-3p, and SMAD7 in T2DM.
Keywords: Diabetes mellitus, type 2, Growth differentiation factor 15, Intra-abdominal fat, MicroRNAs, SMAD7
Type 2 diabetes mellitus (T2DM), an all-compassing health issue, is a complex, multifactorial disease responsible for several chronic comorbidities, including cardiovascular and renal complications. According to the 2021 International Diabetes Federation (IDF) Diabetes Atlas, 537 million adults aged 20 to 79 years have diabetes, and the number is predicted to increase to 783 million by 2045.1 Insulin resistance (IR) and β-cell dysfunction are the primary mechanisms behind T2DM pathogenesis. The development of IR and impaired glucose tolerance is further attributed to an increased visceral fat depot, an independent risk factor for T2DM and cardiometabolic disorder.2,3 According to recent findings, unhealthy adipose tissue (AT) expansion results in deposition of M1 macrophages and excess production of pro-inflammatory cytokines.4 However, the interaction of excess visceral adipose tissue (VAT) with obesity-related metabolic complications like IR and T2DM is not fully understood.
Growth differentiation factor 15 (GDF-15), belonging to the transforming growth factor-β (TGF-β) superfamily, is expressed in various human tissues and macrophages. GDF-15 is responsible for energy intake and maintenance through the gut-brain axis.5 Elevated level of GDF-15 is observed in individuals with diabetes and obesity. Further, elevated serum GDF-15 level is positively associated with the risk of diabetes and IR.6-8 GDF-15 is also released from adipocytes; as a stress-inducible cytokine, it is upregulated through several inflammatory proteins,9 increasing its level in AT and circulation. Mothers against decapentaplegic homolog 7 (SMAD7) is a protein that negatively regulates the TGF-β/SMAD signalling pathway by forming hetero-complexes with receptor-activated SMADs (R-SMADs) and directly inhibiting their activity.10 Conditional expression of SMAD7 enhances IR by reducing insulin (INS) expression and β-cell-related nuclear factors in pancreatic β-cells,11 enhancing IR. Furthermore, SMAD7 inhibits murine protein serine/ threonine kinase 38 (MPK38) activity, significantly contributing to obesity.12
MicroRNAs (miRNAs) are implicated in diabetic pathology. Various miRNAs have been considered potential circulatory markers and therapeutic targets in T2DM.13 Transcription factors and miRNA can mutually regulate each other and jointly regulate their shared target genes to form feed-forward loops, playing decisive roles in various pathophysiological processes.14 In white adipose tissue (WAT) of mice, miR-181b improved glucose homeostasis and INS sensitivity by enhancing INS-mediated protein kinase B (AKT)-phosphorylation.15 Circulatory miR-330-3p is increased in gestational diabetes patients with an aggressive diabetic phenotype and worse outcomes.16 MiR-330-3p is downregulated in the VAT of T2DM patients.17 Moreover, its target genes are significant modulators of β-cell proliferation and INS secretion.16
We hypothesized that the deregulated expression of miRNAs miR-181b-5p and miR-330-3p and of SMAD7 and GDF-15 in VAT might contribute to the development of IR in T2DM individuals. The serum GDF-15 level was quantified, and the VAT and peripheral blood mononuclear cells (PBMCs) of the study population were analyzed to explore the associations between GDF-15, SMAD7, miRNA expression levels, and IR indices. We integrated the experimental findings
This study was conducted following the principles of the Declaration of Helsinki for medical research involving human subjects, and ethical approval was obtained from the Institutional Ethics Committee (IRB approval no. AIIMS/IEC/2018/636). Informed consent was obtained for all participants before inclusion into the study.
Patients awaiting laparoscopic abdominal surgery in the Department of General Surgery at a tertiary care centre were recruited by convenience sampling. Subjects aged 18 to 60 years willing to participate by donating fasting venous blood and VAT samples were included in the study. Thirty T2DM patients were recruited after examining their clinical history and medical records from the outpatient clinics of the Endocrinology and Metabolism and General Surgery departments. Patients with diagnosed diabetic complications, acute or chronic infective conditions, endocrine disorders other than diabetes, current pregnancy, or receiving drug therapy were excluded. We recruited non-diabetic individuals as controls after investigating their diabetic profiles.
Fasting venous blood samples were collected in ethylenediaminetetraacetic acid (EDTA), clot activator, and fluoride vacutainers before anaesthesia induction and the surgical procedure and immediately were processed to obtain serum and plasma. The biochemical parameters were analyzed in a Clinical Chemistry Analyzer AU680 (Beckman Coulter). Fasting INS was analyzed by chemiluminescence assay in Advia Centaur XP (Siemens Healthineers). The same samples were stored at –80 °C, and serum GDF-15 was analyzed in batches by a commercially available sandwich enzymelinked immunosorbent assay (ELISA) kit (Cat. no. #EHGDF15; Thermo Fisher Scientific).
IR indices were assessed following the homeostatic model assessment (HOMA), the HOMA2 calculator (available at: https://www.dtu.ox.ac.uk), total cholesterol to high-density lipoprotein (HDL) ratio, triglyceride to HDL ratio, and triglyceride glucose (TyG) index.
Total RNA was isolated from PBMCs using Trizol (RNA-XPress Reagent; HiMedia Laboratories) following the manufacturer’s instructions. VAT was collected during surgery from the same patient and transferred to the laboratory in 1% phosphate buffer saline in a sterile container. Total RNA was subsequently isolated following the protocol as previously described.18 RNA samples with 260 of 280 and 260 of 230 ratios ≥1.8 were considered for further downstream processes. Total RNA was reverse transcribed using a miScript II RT Kit (cat. no. 218160; Qiagen) as per the manufacturer’s instructions. HiFlex Buffer was used to prepare cDNA and quantify mature miRNAs and messenger RNAs (mRNAs).
A CFX96 Real-Time System and CFX Manager Software (Bio- Rad) were used for real-time expression analysis of miR-181b-5p (Hs_miR-181b_1, cat. no. MS00006699), miR-330-3p (Hs_miR- 330-3p_2, cat. no. MS00031738), SMAD7 (cat. no. 330001 PPH 01905C), and GDF-15 (cat. no. 330001 PPH01935C) using miScript Primer Assays and RT2 qPCR Primer Assay (Qiagen) (Table 1). A miScript SYBR Green PCR Kit (cat. no. 218075; Qiagen) was used. A 10 μL volume of the reaction mixture was prepared for each sample, and real-time polymerase chain reaction (RT-PCR) was performed on a 96-well reaction plate under the following conditions: initial activation step at 95 °C for 15 minutes, followed by 40 cycles of 15 seconds at 94 °C, 30 seconds at 55 °C, and 30 seconds at 70 °C. All samples in the experimental and control groups were analysed in duplicate, and the mean value was used for analysis. The mean cycle threshold (Ct) values were normalized against the internal control genes glyceraldehyde-3-phosphate dehydrogenase (GAPDH; for mRNA) (cat. no. 330001 PPH00150F) and RNA, U6 small nuclear (RNU6; for miRNA) to calculate gene expression according to the ΔΔCt method.
For protein-protein interaction (PPI), we used the Search Tool for the Retrieval of Interacting Genes/Proteins database (STRING v11.0) (https://string-db.org/) and used the search terms GDF-15, SMAD7, and INS as the list of input proteins to obtain functional interpretations of their interacting proteins in
We searched Gene Expression Omnibus (GEO) datasets with the following keywords: “Type 2 diabetes,” “diabetes,” “DM,” “T2DM,” “visceral adipose tissue,” “VAT,” “omental adipose tissue,” and “insulin resistance.” The inclusion criteria were studies involving human tissue analysis and those using expression profiling by array. The exclusion criteria were series, sample, or platform studies; analysis on
The three datasets GSE54350, GSE16415, and GSE20950 were selected for analysis. The samples collected for GSE54350 (diabetes, n=6; non-diabetes, n=6) were AT and macrophage, whereas only omental AT was gathered for the other two datasets, GSE16415 (diabetes, n=5; healthy controls, n=6) and GSE20950 (INS-resistant, n=10; INS-sensitive, n=10).
The functional enrichment of miR-181b-5p and miR-330-3p was predicted using miRNA Enrichment Analysis and Annotation (miEAA) (https://ccb-compute2.cs.uni-saarland.de/mieaa2/), TAM 2.0 (http://www.lirmed.com/tam2/), and MIENTURNET (http://userver.bio.uniroma1.it/apps/mienturnet/), all interactive tools for network-based analysis of miRNA-target interactions. The thresholds for the minimum number of interactions and the false discovery rate (FDR) were 2 and 0.05, respectively. The miRTarBase database was chosen to visualize the miRNA-target interaction network. The filter was set according to the “Strong” interaction type.
Data were analyzed in Microsoft Excel and R programming platform (version 3.6.3) using RStudio. Variables were tested for normality by density plot, Q-Q plot, and Shapiro-Wilk significance test. Differences between groups were tested using a two-sample t-test, and Fisher’s exact test was used for categorical variables. For correlation analysis, Spearman’s rank correlation test was used. A twotailed
Among the 60 study subjects, 34 were men and 26 were women. As expected, significantly higher body mass index (BMI), waist-hip ratio (WHR), fasting blood sugar (FBS), glycated hemoglobin (HbA1c), total cholesterol, and triglycerides were observed in the diabetic group (Table 2). Serum GDF-15 level was significantly elevated in the diabetic group compared to the controls (Supplementary Figure 1A). Receiver operating characteristic (ROC) curve analysis revealed that serum GDF-15 had an area under the curve (AUC) of 0.82 for T2DM (Supplementary Figure 1B).
INS sensitivity and %β-cell function were lower in T2DM cases than in controls, with HOMA-β (controls: 92.78 [66.97 to 109.04]; T2DM: 46.09 [20.13 to 69.85];
Serum GDF-15 was significantly and positively correlated to age, BMI, WHR, HbA1c, and triglycerides (Supplementary Table 1). The PBMC as well as VAT GDF-15 expression were significantly correlated with FBS (
The expression of miR-330-3p in PBMCs and VAT was significantly and positively correlated with the respective SMAD7 expression (PBMC: ρ=0.855,
In PBMCs, GDF-15 (fold change 2.75±1.76,
The PPI network yielded 10 predicted functional partners for the three input proteins based on known interactions from curated databases, laboratory experiments, and text mining (Figure 1C). These proteins were AKT1, insulin-like growth factor 1 receptor (IGF1R), insulin receptor substrate 1 (IRS1), insulin receptor (INSR), IRS2, IGF1, yes-associated protein 1 (YAP1), neural precursor cell expressed developmentally downregulated gene 4-like (NEDD4L), SMAD specific E3 ubiquitin protein ligase 1 (SMURF1), and SMURF2. Based on the predicted data from miRNet, miR- 181b-5p interacts with IRS2, IGF1R, SMAD7, and YAP1. MiR- 181b-5p also targets the transcription factors Finkel-Biskis-Jinkins (FBJ) murine osteogenic sarcoma (FOS), Yin Yang 1 (YY1), CCAATenhancer binding protein-β isoform (C/EBPβ), and metastasis-associated 1 (MTA1). On the other hand, miR-330-3p is involved with the genes SMURF2 and YAP1 and the transcription factors tumor protein p53 (TP53) and YY1.
INSR binding, SMAD binding, and protein binding for molecular function (Supplementary Figure 2A), as well as INSR complex, transcriptional regulator complex, and protein-containing complex for cellular components were among the top categories (Supplementary Figure 2B). The top categories in the biological processes were positive regulation of glucose import, regulation of glucose metabolic process, and INSR signalling pathway (Supplementary Figure 2C). Disease-gene association showed the network to be involved in hyperglycaemia, diabetes mellitus, glucose metabolism disease, and glucose intolerance (Supplementary Figure 2D). IR, regulation of lipolysis in adipocytes, mammalian target of rapamycin signalling pathway, and T2DM were the significant Kyoto Encyclopaedia of Genes and Genomes (KEGG) pathways (Supplementary Figure 2E). The network genes observed in the microarray datasets, CCAAT enhancer binding protein beta (CEBPB), IGF1R, INSR, IRS1, IRS2, NEDD4L, SMURF2, and YAP1, were significantly and differentially expressed in the VAT of T2DM patients (Figure 1D).
The miRNA-target enrichment analysis according to miRTar- Base revealed 11 target genes for miR-181b-5p and miR-330-3p (adjusted
In this study, we report the expression of GDF-15, SMAD7, miR181-5p, and miR-330-3p in the VAT and PBMCs of a T2DM population.
GDF-15 may regulate the pathogenesis of T2DM via SMAD signalling, and there may exist a possible crosslink between INS signalling proteins and the TGF-β/SMAD pathway, which was predicted in our
Both miR-181b and miR-330-3p have been implicated in diabetes mellitus in animal and human studies. The significance of circulating miRNA is two-fold, as they are potential non-invasive markers and mediators of crosstalk between multiple signalling pathways and act differently in physiological and pathological conditions.20 The findings of this study suggest that miR-181b-5p and miR-330- 3p participate in the development of T2DM through an interconnected network. The miR-181 family is a critical regulator of AT function. Sun et al.15 studied AT endothelial cells and found that miR-181b reduced endothelial dysfunction in epididymal WAT, suggesting that miR-181 inhibitors can be a novel therapeutic approach to ameliorate IR. Copier et al.21 worked on a diabetic cardiomyopathy mouse model and found that miR-181-5p was downregulated two-fold, suggesting it as a biomarker (Table 3). However, miR-181b overexpression contributes to the development of diabetic nephropathy by targeting tissue inhibitor of metalloproteinase 3 (TIMP3).22 In this study, miR-181b-5p was upregulated in VAT and peripheral blood with 5.31- and 2.10-fold change, respectively.
Dysregulated gut microbiota-miR-181 axis induces obesity and IR, where the miR-181 family is not only elevated in patients with obesity, but also promotes fat mass gain and WAT inflammation.23 This explains the upregulated miR-181b-5p in VAT and peripheral blood found in our study. GDF-15 and miR-181b-5p in PBMCs and GDF-15 in VAT were also significantly correlated with HOMA-%β and HOMA2-%β, implying a role in diminishing β-cell function.
MiR-330-3p in diabetes has primarily been explored in gestational diabetes mellitus (GDM).14,24,25 In GDM, increased level of miR- 330-3p was associated with a better response to treatment and was strongly associated with poor glycaemic control.24 Plasma miR- 330-3p was increased by 11.1-fold in GDM compared to non-diabetic pregnant women. E2F1 and cell division control protein 42 (CDC42), targets of miR-330-3p, are involved in β-cell growth and proliferation and glucose-dependent INS secretion. Our
The bioinformatic analysis revealed transcription factors for miR-181b-5p, namely FOS, YY1, MTA1, C/EBPβ, and TGFB1. The products of the FOS gene family form the AP-1 transcription factor complex, which mediates a hyperglycaemia-induced activation of the human TGF-β1 promoter.31 C/EBPs are a family of C/ EBPβ is increased under cytokine stimulation. Rahman et al.32 further studied C/EBPβ in obesity-induced inflammation and observed that it induces pro-inflammatory genes in macrophages and adipocytes, possibly explaining its connection to the development of IR. The YY1 protein, a common transcription factor to both miR-181b-5p and miR-330-3p, has been widely studied in diabetes and acts by repressing TGF-β1. It is a potential target for fasting hyperglycaemia and treatment of diabetic nephropathy.33 The TGF-β signalling pathway has been implicated in childhood obesity in a recent
GDF-15 is involved in the pathogenetic mechanism of multiple disorders like diabetes, obesity, and cancer.38,39 It has shown great promise as a diagnostic and prognostic biomarker in various cancers.40,41 In T2DM, serum GDF-15 can act as a predictive adjunct marker because of its reasonable diagnostic accuracy. GDF-15 Fc fusion proteins have shown promise as therapeutic agents in obesity- related disorders and comorbidities, including T2DM.3 Moreover, an in-silico analysis pointed towards possible functionalities of GDF-15 and SMAD7 in an interconnected pathway with miR- 181b-5p and miR-330-3p towards developing IR and T2DM. Exploring and validating the regulatory pathways for these miRNAs will undercover new candidates for clinical testing and therapeutic targeting for hyperadiposity and related complications. Furthermore, connecting the dots between VAT and INS changes in obesity and T2DM will provide a better understanding of the pathogenesis of metabolic disorders. The differential expression (upregulation and downregulation) of the mRNA and miRNA in VAT were also reflected in the blood, which implies that the molecular interactions in VAT relevant to IR also are occurring in the blood. This can also be attributed to blood gene expression being influenced by other tissues.
To conclude, the current study reports, for the first time, the association of GDF-15 with SMAD7 and the regulatory miRNAs, miR-181b-5p and miR-330-3p, in Indian patients with T2DM. The upregulated GDF-15 and miR-181b-5p and downregulated SMAD7 and miR-330-3p in PBMCs and VAT of T2DM patients implicate their role in IR, as supported by the
The authors declare no conflict of interest.
The primers used for real-time polymerase chain reaction
|Primers for mRNA|
|Assay name||Cat. no.||Unigene no.||RefSeq accession no.||Reference position|
|Primers for miRNA|
|Assay name||Position||Mature miRNA ID||miScript primer assay ID|
mRNA, messenger RNA; SMAD7, mothers against decapentaplegic homolog 7; GDF-15, growth differentiation factor-15; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; miRNA, microRNA; RNU6, RNA, U6 small nuclear.
Baseline demographic and biochemical characteristics of the study population
|Variable||Controls (n = 30)||T2DM (n = 30)|
|Age (yr)||46.54 (40.50–54.75)||52.58 (45.25–58.00)||0.087|
|Woman||12 (40.00)||14 (46.67)|
|Man||18 (60.00)||16 (53.33)|
|Body mass index (kg/m2)||23.72 (21.88–26.15)||28.28 (26.00–31.57)||< 0.001†|
|Waist-to-hip ratio||0.88 (0.86–0.92)||0.93 (0.91–1.00)||< 0.001†|
|Fasting blood sugar (mg/dL)||98.00 (92.50–107.75)||141.00 (108.00–179.75)||< 0.001†|
|HbA1c (%)||5.65 (5.5–6.0)||7.45 (6.7–8.2)||< 0.001†|
|Insulin (mIU/L)||8.75 (5.59–12.64)||8.79 (4.03–11.67)||0.442|
|Total cholesterol (mg/dL)||166.00 (150.25–188.50)||200 (177.75–213.75)||0.008‡|
|Triglycerides (mg/dL)||103.50 (78.75–118.00)||161.00 (129.25–186.25)||< 0.001†|
|HDL-C (mg/dL)||35.00 (32.00–42.00)||40.5 (35.25–46.75)||0.068|
|LDL-C (mg/dL)||110.50 (90.25–129.00)||131.00 (113.25–154.75)||0.067|
|GDF-15 (pg/mL)||689.57 (562.42–817.80)||1,362.19 (955.14–1,958.35)||< 0.001†|
Values are presented as median (interquartile range) or number (%). The probability of the Wilcoxon rank sum test with continuity correction between the two groups was used for nonparametric numerical data, whereas the probability of two-sample t-test was used for parametric data after testing for equality of variances.
*Fisher’s exact test was used for qualitative and categorical data; Statistical significance indicated at †
T2DM, type 2 diabetes mellitus; HbA1c, glycosylated hemoglobin; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; GDF-15, growth differentiation factor-15.
Comparison of miR-181b-5p and miR-330-3p expression from this study with that of existing studies and their roles in pathways related to deranged insulin signalling and diabetes mellitus
|MicroRNA||Relative expression||Signalling pathways/targets/sites||Experimental model||Reference|
|181b||Increased||TIMP3||Diabetic patients (n = 54) with and without nephropathy||Zhu et al.22|
|181b||Reduced||AKT-eNOS-NO signalling||Epididymal adipose tissue endothelial cells in obese mice||Sun et al.15|
|181b-5p||Reduced||Circulation||C57BL/6 male mice with diabetic cardiomyopathy||Copier et al.21|
|181b||Reduced||NF-κB||Diabetic, pre-diabetic, and healthy control human subjects||Dehghani et al.20|
|181 family||Increased||Gut microbiota-miR-181-fat axis||C57BL/6 obese mice||Virtue et al.23|
|330-3p||Increased||Circulation||GDM and non-diabetic pregnant patients||Sebastiani et al.16|
|330-3p||Increased||Circulation||GDM patients and non-diabetic controls||Pfeiffer et al.24|
|330-3p||Increased||Circulation||GDM and non-diabetic pregnant patients||Xiao et al.25|
|181b-5p||Increased||Circulation, VAT||T2DM and non-diabetic controls||Current study|
|330-3p||Reduced||Circulation, VAT||T2DM and non-diabetic controls||Current study|
TIMP3, tissue inhibitor of metalloproteinase 3; AKT, protein kinase B; eNOS, endothelial nitric oxide synthase; NO, nitric oxide; NF-κB, nuclear factor kappa-light-chain-enhancer of activated B cell; GDM, gestational diabetes mellitus; VAT, visceral adipose tissue; T2DM, type 2 diabetes mellitus.