Analysis of key genes and their functions in placental tissue of patients with gestational diabetes mellitus
This study was aimed at screening out the potential key genes and pathways associated with gestational diabetes mellitus (GDM).
The GSE70493 dataset used for this study was obtained from the Gene Expression Omnibus database. Differentially expressed genes (DEGs) in the placental tissue of women with GDM in relation to the control tissue samples were identified and submitted to protein-protein interaction (PPI) network analysis and subnetwork module mining. Functional enrichment analyses of the PPI network and subnetworks were subsequently carried out. Finally, the integrated miRNA–transcription factor (TF)–DEG regulatory network was analyzed.
In total, 238 DEGs were identified, of which 162 were upregulated and 76 were downregulated. Through PPI network construction, 108 nodes and 278 gene pairs were obtained, from which chemokine (C-X-C motif) ligand 9 (CXCL9), CXCL10, protein tyrosine phosphatase, receptor type C (PTPRC), and human leukocyte antigen (HLA) were screened out as hub genes. Moreover, genes associated with the immune-related pathway and immune responses were found to be significantly enriched in the process of GDM. Finally, miRNAs and TFs that target the DEGs were predicted.
Four candidate genes (viz., CXCL9, CXCL10, PTPRC, and HLA) are closely related to GDM. miR-223-3p, miR-520, and thioredoxin-binding protein may play important roles in the pathogenesis of this disease.
KeywordsGestational diabetes mellitus Differentially expressed genes Protein-protein interaction network Integrated regulatory network Transcription factors
Cytochrome P450, family 1, subfamily A, polypeptide 1
Differentially expressed genes
Forkhead box O1
Gestational diabetes mellitus
Kyoto Encyclopedia of Genes and Genomes
POU class 1 homeobox 1
Protein tyrosine phosphatase 1B
Protein tyrosine phosphatase, receptor type C
Zinc finger E-box-binding homeobox 1
Expectant mothers with gestational diabetes mellitus (GDM), a common pregnancy complication, have an increased risk of developing type 2 diabetes mellitus . Over the past 20 years, the prevalence of GDM has doubled, affecting approximately 10% of pregnancies in the USA [2, 3]. Babies born to mothers with GDM are typically at a high risk for macrosomia, neonatal cardiac dysfunction, neonatal hypoglycemia, stillbirth, childhood obesity, and type 2 diabetes mellitus [4, 5, 6]. Given the worldwide prevalence and adverse outcomes of GDM, there is an urgent need to grasp the pathophysiology and pathogenesis of the disease .
Previous studies have suggested that GDM is caused by enhanced insulin resistance and pancreatic beta (β)-cell dysfunction , involving genes that are related to insulin signaling, insulin secretion, maturity-onset diabetes of the young, and lipid and glucose metabolism, to name a few [8, 9]. Subsequently, it was found that inflammatory pathways , metabolic disorder , oxidative stress , and vitamin D concentrations  were also related to GDM. Furthermore, some genetic alterations, such as those of the genes encoding β3-adrenergic receptor  and transcription factor 7-like 2 polymorphism , were also found to be associated with GDM. Moreover, GDM results in major changes in the expression profiles of placental genes, with a significant increase in markers and mediators of inflammation . Recently, several microarray studies have verified that the cytochrome P450, family 1, subfamily A, polypeptide 1 (CYP1A1), estrogen receptor 1 (ESR1) , fibronectin 1 (FN1), and leptin (LEP)  genes were essential for the pathogenesis of GDM. However, because the genes related to GDM have not yet been fully identified, the biological processes underlying the pathogenesis of this disease remain unclear.
In this study, the gene expression profiles of placental tissue from women with GDM were compared with those of matched normal placental tissue by microarray analysis, to screen out differentially expressed genes (DEGs) in GDM. The identified DEGs were then submitted to Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway and Gene Ontology (GO) enrichment analyses to explore the crucial pathways of GDM. Additionally, a protein-protein interaction (PPI) network was constructed and subnetwork module mining was performed to seek out the candidate disease genes. Finally, microRNAs (miRNAs) and transcription factors (TFs) that target the candidate DEGs were identified and analyzed. The results from this study may lay the groundwork for future research on the pathogenesis of GDM.
The gene expression dataset GSE70493, which is based on the GPL17586 [HTA-2_0] Affymetrix Human Transcriptome Array 2.0 [transcript (gene) version] platform, was downloaded from the National Center for Biotechnology Information’s Gene Expression Omnibus database (http://www.ncbi.nlm.nih.gov/geo/). This dataset comprised 63 placental tissue specimens collected from 32 cases of GDM and 31 matched pregnancies without maternal complications.
The probe-level data (CEL files) were converted to expression estimates by the Puma  and Oligo  packages in R, and the original expression dataset was processed into expression values using the robust multi-array average algorithm  with the default settings implemented in Bioconductor. The DEGs were identified with the limma  software package according to the expression values of the sample probes, and only those with a p-value of less than 0.01 were selected and annotated for further analysis.
GO and KEGG pathway enrichment analyses
To assess the functions and significantly enriched pathways of the DEGs, ClusterProfiler  was used to identify the overrepresented GO terms in the biological process (BP), cellular component (CC), and molecular function (MF) categories, as well as the KEGG pathway categories. The hypergeometric distribution threshold for these analyses was a p-value of < 0.05.
PPI network construction and subnetwork module mining
The Search Tool for the Retrieval of Interacting Genes (STRING, ver. 10.0, https://string-db.org/)  database was used to analyze functional interactions between the DEGs and other genes, under a confidence score threshold of > 0.4. The PPI network was established using Cytoscape (ver. 3.3.0, http://www.cytoscape.org/) . Then, the topology of the network was analyzed, and the hub nodes in the network were obtained by calculating the average degree of each node. The average degree is the average number of edges connecting all the nodes in the network, measured by three indexes: degree centrality , betweenness centrality , and closeness centrality .
Molecular Complex Detection (MCODE)  is an automated method for searching molecular complexes with similar functions in large protein interaction networks. The MCODE (ver. 1.4.2, http://apps.cytoscape.org/apps/mcode) plugin of Cytoscape was used to analyze the subnetwork modules with similar functions in the original PPI network. Then, GO and KEGG pathway analyses of the subnetwork modules were performed to evaluate their functions.
Prediction of miRNAs and transcription factors that regulate the DEGs
The TFs associated with the DEGs were predicted by the position weight matrices from TRANSFAC and JASPAR in the Enrichr database , under the hypergeometric distribution threshold of p < 0.01. miRNAs associated with the DEGs were predicted by miRTarBase in the Enrichr database, under the hypergeometric distribution threshold of p < 0.01. On the basis of the miRNA–DEG regulatory network and TF–DEG regulatory network, the integrated DEG–miRNA–TF regulatory network was constructed using Cytoscape.
Analysis of the differentially expressed genes
Functional enrichment analyses
Through KEGG pathway enrichment analysis, the top 20 enriched pathways were identified on the basis of the p-value (Fig. 2b). The significantly enriched categories included pathways for allograft rejection, graft-versus-host disease, type I diabetes mellitus, autoimmune thyroid disease, inflammatory bowel disease, the Toll-like receptor signaling pathway, and Th1 and Th2 cell differentiation. The genes associated with those pathways were HLA, CXCL9, CXCL10, and PTPRC.
PPI network and subnetwork module analyses
Node genes measured in three indexes (top 15)
Prediction of miRNAs and transcription factors that regulate the target genes
miRNAs and transcription factors (TFs) regulating DEGs
GDM describes the condition of abnormal sugar metabolism or potential decreased glucose tolerance before pregnancy and is confirmed during pregnancy [30, 31, 32]. It is a complex disease, being influenced by many factors such as the environment, society, and genes . Moreover, genetic studies have suggested that multiple genes are involved in the disease . In our study, DEGs in GDM and their enriched functions were screened out via bioinformatic analysis, and four key genes (viz., HLA, CXCL9, CXCL10, and PTPRC) were identified to be crucial to the disease. Moreover, miR-223-3p, miR-520, and TBP were found to be strongly linked to those DEGs, indicating their importance in GDM.
CXCL9 and CXCL10 are categorized as “inflammatory” chemokines. Shimada and coworkers postulated that the binding of CXCL10 to CXCR3 played a crucial role in the suppression of pancreatic β-cell proliferation . Besides this, CXCL10 could interact with Toll-like receptor 4 to continuously activate c-Jun N-terminal kinases and protein kinase B (Akt), induce the cleavage of p21-activated protein kinase 2, and switch the Akt signal from proliferation to apoptosis, resulting in the suppression of pancreatic β-cell proliferation . The present study demonstrated that CXCL10 was significantly enriched in the Toll-like receptor signaling pathway, leading us to speculate that it is a key gene that participates in the pathogenesis of GDM by regulating the progress of the Toll-like receptor signaling pathway. Although CXCL9 has similar functional and structural characteristics as CXCL10, it was reported that CXCL9 could not bind to Toll-like receptor 4 . In this study, CXCL9 was significantly enriched in the cytokine signaling pathway and may thus play a critical role in the pathogenesis of GDM by regulating the inflammatory pathway.
HLA, the gene for the human MHC, plays a pivotal role in the antigen presentation of extracellular and intracellular peptides and the regulation of immune responses . Compared with other regions of the human genome, the MHC genes on chromosome 6 are more associated with the susceptibility to common diseases like diabetes, and indeed many reports have shown that HLA gene variants are related to the predisposition to type 1 diabetes mellitus . Additionally, although type 2 diabetes mellitus is not an autoimmune disease or associated with the HLA gene, there is evidence that genes in the HLA region might have an influence on the genetic susceptibility to this metabolic disorder . Importantly, Steinborn and colleagues found that GDM was related to an increased humoral immune response against HLA-class II antigens . Our study highlights the importance of HLA in the progression of GDM, during which the gene is downregulated, and emphasizes that the autoimmune response is significantly associated with the disease pathogenesis.
PTPRC (CD45) has an essential role in lymphocyte development, antigen receptor signal transduction, and modulation of the signals emanating from integrin and cytokine receptors . In diabetes mellitus, protein tyrosine phosphatases act as negative regulators of insulin signal transduction . A previous study demonstrated that the homozygous deletion of protein tyrosine phosphatase 1B (PTP1B) in myocytes enhanced both the insulin-dependent activation of insulin receptor autophosphorylation and the tyrosine phosphorylation of insulin receptor substrates, and increased insulin sensitivity . Moreover, it was shown that the expression of PTPRC was related to residual β-cell function in type 1 diabetes mellitus . Our results reveal that PTPRC is likely to be a key gene that impacts GDM.
Because miR-223 was found to be significantly dysregulated in GDM, it was selected as a potential circulating biomarker for this disease . Additionally, as a stress-related miRNA, miR-223 negatively regulated the cryopyrin-encoding gene NLRP3 and subsequently interleukin-1 beta production . In our study, production of the TFs zinc finger E-box binding homeobox 1 (ZEB1) and Forkhead box O1 (FOXO1) was regulated by miR-223-3p. FOXO1, a target of insulin signaling, regulates metabolic homeostasis in response to oxidative stress. The interaction of FOXO1 with β-catenin could attenuate the WNT signaling pathway, which is involved in lipid metabolism and glucose homeostasis . Besides this, FOXO1 was targeted by miR-520 h and miR-520 g-3p, which were speculated to influence insulin sensitivity in human white adipose tissue through their predicted effects on glucose metabolism . ZEB1, a zinc finger TF, is associated with placental development. It was reported that ZEB1 cooperated with FOXO members to suppress B-lymphocyte proliferation . TBP is a universal eukaryotic TF. It was found that the enhancement of TBP-2 expression caused impairment of glucose-induced insulin secretion and insulin sensitivity . In the present study, TBP was found to regulate many HLA genes (HLA-DQA1, HLA-F, and HLA-DQA2), implying its indispensable role in GDM.
In conclusion, four immune-related DEGs of GDM (viz., HLA, CXCL9, CXCL10, and PRPTC) appeared to be associated with not only the autoimmune process but also residual β-cell function. miR-223-3p, miR-520 (i.e., miR-520 h and miR-520 g-3p), and TBP regulated most of the DEGs, especially cellular metabolism-related genes (FOXO1 and ZEB1). These results provide new insights into the mechanisms of GDM pathogenesis.
Conception and design of the research: YW; acquisition of the data: HY; analysis and interpretation of the data: YW; statistical analysis: FL; drafting of the manuscript: YW; revision of the manuscript for important intellectual content: XS. All authors read and approved the final manuscript.
Ethics approval and consent to participate
Consent for publication
The authors declare that they have no competing interests.
- 7.Wang YH, Wu HH, Ding H, Li Y, Wang ZH, Li F, Zhang JP. Changes of insulin resistance and beta-cell function in women with gestational diabetes mellitus and normal pregnant women during mid- and late pregnant period: a case-control study. J Obstet Gynaecol Res. 2013;39:647–52.PubMedCrossRefPubMedCentralGoogle Scholar
- 15.Pagan A, Sabater-Molina M, Olza J, Prieto-Sanchez MT, Blanco-Carnero JE, Parrilla JJ, Gil A, Larque E. A gene variant in the transcription factor 7-like 2 (TCF7L2) is associated with an increased risk of gestational diabetes mellitus. Eur J Obstet Gynecol Reprod Biol. 2014;180:77–82.PubMedCrossRefGoogle Scholar
- 18.Liu X, Gao Z, Zhang L, Rattray M. Puma 3.0: improved uncertainty propagation methods for gene and transcript expression analysis. BMC Bioinformatics. 2013;14:1471–2105.Google Scholar
- 26.Cukierski WJ, Foran DJ. Using Betweenness centrality to identify manifold shortcuts. Proc IEEE Int Conf Data Min. 2008;1:949–58.Google Scholar
- 29.Chen EY, Tan CM, Kou Y, Duan Q, Wang Z, Meirelles GV, Clark NR, Ma'ayan A. Enrichr: interactive and collaborative HTML5 gene list enrichment analysis tool. BMC Bioinformatics. 2013;14:1471–2105.Google Scholar
- 37.Bodis G, Toth V, Schwarting A. Role of human leukocyte antigens (HLA) in autoimmune diseases. Rheumatol Ther. 2018;7:018–0100.Google Scholar
- 43.Nieto-Vazquez I, Fernandez-Veledo S, de Alvaro C, Rondinone CM, Valverde AM, Lorenzo M. Protein-tyrosine phosphatase 1B-deficient myocytes show increased insulin sensitivity and protection against tumor necrosis factor-alpha-induced insulin resistance. Diabetes. 2007;56:404–13.PubMedCrossRefGoogle Scholar
- 48.Capobianco V, Nardelli C, Ferrigno M, Iaffaldano L, Pilone V, Forestieri P, Zambrano N, Sacchetti L. miRNA and protein expression profiles of visceral adipose tissue reveal miR-141/YWHAG and miR-520e/RAB11A as two potential miRNA/protein target pairs associated with severe obesity. J Proteome Res. 2012;11:3358–69.PubMedCrossRefGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.