Type 2 diabetes mellitus (T2DM) is a metabolic disorder characterized by hyperglycemia caused by a relative lack of insulin. The number of cases of T2DM are increasing worldwide and it has become an important health concern. According to a survey by the International Diabetes Federation, the prevalence if T2DM is expected to exceed 10% of the global adult population by 2040 (1). The main causes of T2DM are insulin resistance and an insulin secretion defect. The majority of patients with insulin resistance are obese and exhibit symptoms, such as slight fatigue and thirst in the early stages of T2DM (2). Research suggests that alterations in multiple genes and signaling pathways are involved in regulating the development of T2DM. However, a lack of research on the precise molecular mechanisms of T2DM progression limits the treatment efficacy of the disease at present. Therefore, understanding the molecular mechanisms of T2DM occurrence and development is of utmost importance for non-invasive diagnosis and targeted therapy in the future.

Pancreatic cells, particularly β-cells, play an important role in the occurrence and development of T2DM (3). Initial-phase insulin secretion dysfunction in pancreatic β-cells is the primary feature of T2DM. A number of studies have explored the factors that contribute to impaired pancreatic β-cell function in T2DM, including endoplasmic reticulum stress, lipotoxicity, mitochondrial dysfunction, oxidative stress, low inflammation of islets and glucotoxicity (4,5). For example, Park et al (6) found that the deposition of islet amyloid polypeptide (IAPP) upregulated the Fas receptor of pancreatic β-cells and initiated an apoptotic cascade. Although pancreatic β-cell damage is observed in patients with T2DM, the mechanisms and signaling pathways involved remain unknown. Bioinformatic analysis of microarrays enables the overall analysis of differentially expressed genes (DEGs) in the development of T2DM. Marselli et al (7), gathered pancreatic β-cell samples from tissues by laser capture microdissection (LCM) and detected DEGs between patients with T2DM and normal donors. However, the interactions between DEGs, signal pathway interaction networks and transcription factors (TFs) warrant further comprehensive analysis.

In the present study, we selected GSE20966 from the Gene Expression Omnibus (GEO) database, and used the limma package in R software to screen DEGs. Subsequently, we analyzed the Gene Ontology (GO) functions and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways associated with the resulting DEGs. Moreover, a protein-protein interaction (PPI) network of the DEGs was established and TFs were selected. We also identified core genes by a comprehensive analysis.

The rapid development of high-throughput sequencing technology allows biological problems to be addressed by gene sequencing. Currently, high-throughput sequencing is beginning to be widely used to find candidate genes for numerous diseases. Since the pathogenesis of T2DM is not clear, comprehension of the molecular mechanisms of T2DM is required for non-invasive diagnosis and targeted therapies.

Laser capture microdissection (LCM) technology has been utilized to extract samples. LCM obtains target cells directly from frozen or paraffin-embedded tissue sections without destroying the surrounding tissue morphology. It is often utilized to accurately separate individual cells from tissue (11). The accurate separation of pancreatic β-cells from heterogeneous tissue is a prerequisite for continuous and meaningful bioinformatic analysis. Marselli et al gathered pancreatic β-cells from patients with T2DM and healthy controls by LCM for microarray analysis. In this study, we extracted the expression data from GSE20966, and identified 208 upregulated and 121 downregulated DEGs by bioinformatic analysis. To further investigate the interactions between the DEGs, GO function and KEGG pathway enrichment analyses were performed.

The GO analysis indicated that the upregulated DEGs were primarily involved in the defense response, regulated exocytosis, and the acute inflammatory response, while the downregulated DEGs were primarily associated with regulation of ion transport, heart contraction and process and the regulation of signaling. Insulin resistance and pancreatic β-cell structural dysfunction are caused by an inflammatory response and lead to the development of T2DM (12,13). Moreover, the knockdown of the Na⁺/K⁺ ATPase subunit FXYD2 in the ion transport pathway improved mouse pancreatic β-cell proliferation and glucose tolerance (14,15). Furthermore, the enriched KEGG pathways of the upregulated DEGs were involved in pancreatic secretion and the complement and coagulation cascades. The complement pathway is primarily related to host defense and inflammation (16). The expression of C1q and C5a in the complement pathway are associated with diabetic vascular complications, such as diabetic retinopathy and diabetic nephropathy (17,18). The downregulated DEGs were primarily involved in carbohydrate digestion and absorption, insulin secretion and the TLR pathway. Solute carrier family 2 member 2 (SLC2A2) was associated with carbohydrate digestion, absorption pathway and the insulin secretion pathway. SLC2A2, also known as glucose transporter 2 (GLUT2), is the transmembrane carrier protein that transports glucose primarily in the liver and blood (19). SLC2A2 can detect the presence of extracellular sugar and signal to regulate insulin secretion by the pancreatic β-cells (20). Mutation of the SLC2A2 gene conveys a high risk for the conversion of impaired glucose tolerance (IGT) patients to T2DM patients (21). The TLR pathway can promote the synthesis of many cellular activity factors, adhesion molecules and inflammatory factors, ultimately affecting immunity and the inflammatory response (22). The TLR pathway is associated with the pathogenesis and development of diabetes (23). Intercellular adhesion molecule-1 (ICAM-1) exhibits an upregulated expression in the retinal e ndothelial cells of diabetic retinopathy rats. Through the accumulation and adhesion of leukocytes, ICAM-1 can cause retinal vascular damage and destroy the blood-retinal barrier (24). Rajamani and Jiala found that TLR2 and TLR4 expression were upregulated in hyperglycemia-induced human microvascular retinal endothelial cells (HMVRECs), and activated NF-κB to produce biomediators of inflammation and ICAM-1 (25). Therefore, these pathways are most likely to be important in the development of diabetes and diabetic complications. Additional studies are required to identify all the DEGs in T2DM.

The primary function of TFs is to bind a specific site of a gene and regulate the expression of the target gene in the cell. In this study, the TFs, HNF1A, STAT3 and GR, were shown to be potential modulators of T2DM. HNF1A belongs to the HNF1 homeobox family, and HNF1A is essential for the regulation of pancreatic β-cell differentiation. The I27L polymorphism of HNF1A decreases β-cell mass or impairs function and leads to a high risk of T2DM (26,27). HNF1A is a key transcription factor mediating the expression of dipeptidyl peptidase-4 (DPP4) and angiotensin converting enzyme 2 (ACE2) in pancreatic β-cells, both of which may have therapeutic potential for T2DM (28,29). Notably, ACE2 expression activates STAT3, shown herein to be an important TF. The activation of the JAK/STAT3 pathway has been reported to encourage the development of vasculopathy in T2DM (30). However, Tiano et al concluded that the activation of STAT3 signaling inhibits the synthesis and accumulation of fatty acids in pancreatic cells of diabetic mice induced by a high-fat diet, thus preventing pancreatic β-cell damage (31). In addition, STAT3 signaling activation has been demonstrated to enhance the function of insulin secretion (32). This indicates that STAT3 is a double-edged sword in the development of T2DM, preventing the pancreatic cells from further damage and maintaining the secretion of insulin, while aggravating the development of vascular complications. GR, also known as NR3C1, is the receptor bound by cortisol and other glucocorticoids. Recent studies have demonstrated that the Bcl1 polymorphism in intron 2 of GR is associated with insulin resistance and hyperinsulinemia, although the underlying mechanisms remain unclear (33,34). Therefore, additional research is required to explore the role of GR in T2DM insulin resistance.

In this study, we also constructed a PPI network for the DEGs, and there were 4 genes with a connectivity level ≥10: ALB, REN, C3 and CCL2. ALB was the DEG with the highest degree of connectivity. The serum glycated ALB level in patients with T2DM is related to coronary artery disease (CAD). As the level of serum glycated ALB increases, the presence and severity of CAD increases (35,36). Rodiño-Janeiro et al also reported that the elevated expression of glycated ALB upregulated NAPDH oxidase in vitro, and the enhanced oxidative stress may mediate diabetic vasculopathy (37). REN is important in the development of diabetic vasculopathies, such as diabetic retinopathy and diabetic nephropathy (38,39). A subsequent study using biopsy samples from diabetic patients revealed that renin plays a role in diabetic vascular disease by activating a renin-angiotensin system (40). In this study, we found that the gene expression of C3 (log₂ FC=1.44, P-value <0.01) was upregulated in pancreatic β-cells of patients with T2DM. In the complement signaling pathway, C3 is a key protein in both the alternative pathway and the classical pathway. Elevated rates of diabetes and insulin resistance are closely related to increased serum C3 (41–43). The main feature of T2DM is pancreatic β-cell damage and dysfunction, due to a shift of β-cell status from proliferative to apoptotic (44). Notably, a recent study demonstrated that C3 stimulated intracellular calcium and ATP levels by activating the C3/C3aR signaling pathway, and increased glucose-dependent insulin secretion and protection against apoptosis (45). Dos Santos et al also reported that C3 silencing led to apoptosis under normal physiological conditions and following exposure to cytokines. The addition of exogenous C3 prevents cytokine-induced apoptosis in β cells through C3-mediated activation of the AKT signaling pathway and inhibition of c-Jun N-terminal kinase activity (46). Therefore, it is conceivable that elevated C3 may have a protective effect on apoptosis in T2DM, indicating that C3 may be a potential therapeutic target for T2DM. CCL2 belongs to the CC chemokine family, and is also known as monocyte chemoattractant protein 1 (MCP1). In vitro studies on patients with proliferative diabetic retinopathy have demonstrated that the level of CCL2 is significantly increased in patients compared with healthy controls (47). Liu et al found that hyperglycemia may affect hypomethylation of the CpG site in the CCL2 promoter region, and enhanced differential expression of serum CCL2 was important in the occurrence and development of vasculopathy in T2DM (48). Recent studies have proposed that the CCL2 2518A/G polymorphism is associated with diabetic retinopathy in T2DM; as the number of G alleles increased, the prevalence of diabetic retinopathy was elevated (47,49).

In the present study, we identified two core genes, SERPING1 and ANPEP. SERPING1, also known as the C1-inhibitor (C1INH), is a protease inhibitor that belongs to the SERPIN superfamily. The STAT3 signaling pathway regulates the expression of SERPING1. The function of SERPING1 is to inhibit activation of both the classical pathway and the lectin pathway to reduce production of C3 convertase. Notably, the expression levels of SERPING1 and C3 were both upregulated in this study, indicating that the activation of the complement system may be achieved by the alternative pathway. The accumulation of C3 in vivo is probably caused by the effect of SERPING1 inhibiting C3 convertase. We hypothesized that the STAT3 signaling pathway stimulates the expression of SERPING1 and promotes the accumulation of C3 to produce an anti-apoptotic effect in pancreatic β-cells. Moreover, the SERPING1-mediated regulation of the complement pathway may inhibit the inflammatory response in pancreatic β-cells. Thus, SERPING1 may play a dual role, an anti-inflammatory one, while maintaining an anti-apoptotic effect. Further studies are required to confirm this hypothesis. ANPEP, a broadly specific aminopeptidase, is associated with a number of cellular process, including cell proliferation, apoptotic differentiation, angiogenesis, and chemotaxis (50). The results presented in Fig. 5 suggest that HNF1A simultaneously regulates the expression of ANPEP and ALB, and that there is an interaction between ALB and ANPEP. Expression of ALB stimulates the production of reactive oxygen species (ROS) by the NADPH enzyme to activate oxidative stress (37). The expression levels of both ANPEP and REN in our study are upregulated. Both of them activate the renin-angiotensin signaling pathway to stimulate ROS generation. Pancreatic β cells are more sensitive to ROS, and so this can lead to direct damage of pancreatic β cells and promote apoptosis. Therefore, we suggest that HNF1A-mediated ANPEP and ALB expression may accelerate pancreatic β-cell damage and insulin resistance through oxidative stress. Pedersen et al compared the direct overlap between heterogeneous islet diabetes-associated genomes by genome-wide association studies (GWAS) to establish ANPEP as a diabetes susceptibility gene (51). Locke et al found that ANPEP had a significant allelic expression imbalance by comparing the allelic expression of RNA and DNA from islets of diabetic and non-diabetic individuals. This suggests that ANPEP is a pathogenic gene for T2DM (52). However, the precise mechanisms of action of ANPEP in T2DM remain unknown, and further research is required to confirm this hypothesis. In our further studies, we aim to validate the selected SERPING1 and ANPEP genes in T2DM tissue samples from patients and animal models.

In conclusion, in the present study, we conducted a thorough bioinformatics analysis of DEGs by GSE20966 data screening and identified several genes implicated in the development and progression of T2DM. A total of 329 genes were identified, of which SERPING1 and ANPEP are probable core genes of T2DM. This study reveals a series of valuable genes for further research into the non-invasive diagnosis and targeted therapy of T2DM. However, bioinformatics analyses merely indicate a general direction for further research. To confirm the functions of DEGs in T2DM, molecular biology experiments are required.