Diabetic nephropathy (DN) is one of the most important microvascular diseases in diabetes and has become the chief cause of the end-stage renal disease (1). The clinical incidence of DN is high, and its main features include proteinuria, increased serum creatinine (Scr), and decreased glomerular filtration rate (GFR) (2). Studies have revealed that the progression of DN is related to genetic factors as well as to hemodynamic and metabolic changes (3). At present, the primary diagnostic marker of DN is microalbuminuria (MA). However, there is still a controversy regarding whether the appearance of MA represents kidney damage and whether MA inevitably progresses to obvious proteinuria and chronic renal function decline. Therefore, it is necessary to identify other biological markers that can predict kidney damage.

Microarray technology and bioinformatics analysis have enabled the identification of genetic alterations at the genome level. In recent years, bioinformatics methods have been widely used to analyse microarray data to identify the differentially expressed genes (DEGs) in addition to various other analyses. In this study, we analysed multiple microarray datasets, including GSE30528, GSE47183, GSE104948, and GSE96804. The common DEGs were identified in GSE30528, GSE47183, and GSE104948 datasets. The hub genes were identified by protein–protein interaction (PPI) network analysis and ten algorithms of the cytoHubba plugin. The predictive capability of the biomarker was analysed by receiver operating characteristic (ROC) curve and logistic regression analyses. The correlation between the biomarker and other genes was analysed by Pearson’s rank correlation using R software. Furthermore, Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), and Gene Set Enrichment Analysis (GSEA) were used to determine the potential functions of the biomarker.

DN is the leading cause of end‐stage renal disease globally (11). Due to the extremely complex metabolic disorders that occur in patients with DN, once DN has reached the terminal stage, it is often more difficult to treat than other kidney diseases. Although many studies have investigated the pathogenesis of DN, it has not been clarified completely (12). Therefore, it is necessary to identify potential biomarkers for early diagnosis and targeted therapy of DN. This study used multiple microarrays to conduct bioinformatics analysis for identifying target genes and pathways involved in the occurrence and development of DN. The results of this study suggest that FN1 may play a key role in the pathogenesis of DN.

In our study, 55 overlapping DEGs were identified between renal glomerular tissues of DN patients and normal controls based on three microarray datasets. Through PPI network analysis and ten algorithms, four hub genes (all upregulated) were finally screened out. The results of logistic regression analysis revealed that FN1 had a significant relationship with DN (p < 0.05). ROC curve analysis showed that FN1 had a good predictive ability for DN, suggesting that FN1 plays an important role in the development of DN. To further understand the functions of FN1, we inferred its functions based on the genes that were the most closely associated with its expression, which is called “guilt of association.” The top 50 genes with the greatest absolute correlation coefficients in correlation analysis were selected for GO and KEGG enrichment analyses.

GO and KEGG enrichment analyses were used to explore the molecular mechanisms of the top 50 genes involved in the occurrence and development of DN as well as to determine the potential role of FN1 in the occurrence and development of DN. GO annotation result of the top 50 genes showed that ECM organisation, extracellular structure organisation, external encapsulating structure organisation, collagen fibril organisation, and cell−substrate junction assembly were chiefly enriched in BP, while CC included collagen-containing ECM, collagen trimer, fibrillar collagen trimer, banded collagen fibril, and complex of collagen trimers. It partly clarifies the intricacy of the pathogenesis of DN and that it destroys several cell components (13). In addition, MF of the top 50 genes mainly involved ECM structural constituent, collagen binding, integrin binding, glycosaminoglycan binding, and ECM structural constituent conferring tensile strength. The above results suggest that FN1 may be involved in these MFs and influence the progression of DN.

On performing KEGG enrichment analysis, we found that the top 50 genes were mainly involved in ECM–receptor interaction, protein digestion and absorption, focal adhesion, proteoglycans of cancer, complement and coagulation cascades, Chagas disease, amoebiasis, and pertussis. As the most significantly enriched pathway, circle charts showed that five genes (FN1, COL6A3, COL1A2, THBS2, and CD44) were enriched in ECM–receptor interaction. In addition, FN1 was significantly positively correlated with COL6A3, COL1A2, THBS2, and CD44. According to “guilt of association,” we can infer that the abnormal expression of these genes may be jointly involved in the pathogenesis of DN. GSEA also showed that FN1 plays a key role in ECM–receptor interaction. Therefore, we can infer that the co-expression of COL6A3, COL1A2, THBS2, CD44, and FN1 plays a key role in the ECM–receptor interaction pathway and promotes the progression of DN.

The ECM is made up of functional macromolecules and a complex mixture of structural components. It is well known that the ECM signalling pathway plays an important role not only in the morphogenesis of tissues and organs but also in the maintenance of cell and tissue structures as well as their functions (14, 15). ECM is also involved in growth, development, and wound repair, and its dynamic components can arbitrate a range of signals, in addition to regulating cell migration and proliferation (16). As an ECM protein-encoding gene, thrombospondin-2 (THBS2), encodes a secreted ECM glycoprotein, which can reflect the severity of tissue fibrosis (17, 18).

COL1A2 is a member of the type I collagen genes and encodes the proα2(I) chains. It has been reported that the upregulation of COL1A2 in diabetic kidney disease (DKD) can lead to increased renal fibrosis (19). COL6A3, a canonical collagen VI gene, encodes the collagen VI isoform. Collagen VI beaded microfibrils have been discovered in the ECM of almost all tissues (20, 21). Collagen VI is highly expressed in a range of cancers and promotes tumour growth and progression. It also influences the tumour microenvironment by augmenting the recruitment of macrophages and endothelial cells, thus enabling tumour inflammation and angiogenesis (22). However, its role in DN is not well understood. CD44 is a well-known cell-surface glycoprotein that is involved in diverse biological pathways, such as cell migration, proliferation, and lymphocyte activation (23, 24). Research has shown that CD44 plays an important role in the pathogenesis of experimental crescentic glomerulonephritis and collapsing focal segmental glomerulosclerosis (25). From PPI network analysis, we found that CD44 was associated with COL1A2, COL6A3, and FN1. This suggests that they all play a key role in the progression of DN.

FN1 encodes fibronectin, and there are two forms of fibronectin, soluble and insoluble, such as a glycoprotein present in a soluble dimeric form in the plasma, and dimeric or multimeric form at the cell surface and in the ECM. The encoded preproprotein is proteolytically processed to generate the mature protein. Fibronectin is related to cell adhesion and migration processes including embryogenesis, wound healing, blood coagulation, host defence, and metastasis (26). Some studies have shown that overexpression of FN1 promotes fracture healing by activating the TGF-β/PI3K/Akt signalling pathway (27). In a study on pre-eclampsia, it was found that excess FN1 can promote apoptosis and autophagy in vascular endothelial cells (28). Fibronectin is one of the important components of the ECM. Studies have revealed that FN1 plays a key role in glomerular sclerosis and fibrosis in chronic kidney disease (29, 30). The α5β1 integrin receptor and FN dimer combine to form an FN matrix, which can induce conformational changes in FN and promote the interaction of FN–FN to form new fibrils (31). Through continuous FN deposition, these fibrils grow into a stable insoluble matrix, and based on this, other ECM components are deposited (32). Therefore, the overexpression of FN1 plays a key role in the progression of fibrosis. Under mechanical stress, FN1 is upregulated in podocytes (33). Thus, we speculate that FN1 may play a key role in the disease progression of DN. Evidence from previous studies indicates that FN1 may be related to the development of DN and thus has the potential to be used as a diagnostic marker of DN. However, clinical studies are needed to verify the diagnostic value of FN1.

In summary, we investigated the key gene for DN using integrated bioinformatics analysis. We identified four candidate genes: FN1, C1QA, C1QB, and CD44. On further investigating the biological functions of FN1, we found that FN1 was positively correlated with THBS2, COL1A2, COL6A3, and CD44 and was involved in the development of DN by acting on the ECM receptor interaction pathway. THBS2, COL1A2, COL6A3, and CD44 are expected to be new research targets for DN, which may provide new directions for the diagnosis and treatment of DN. Consequently, further studies on these genes are warranted. Our findings can contribute to a better understanding of the pathogenesis of DN as well as the development of new therapeutic targets for DN. However, there are several limitations of our study. First, hub genes and pathways were identified based on small sample sizes by bioinformatics analysis. Second, our study did not include experiments to verify the expression of the identified hub genes in DN. All in all, the identified genes and related pathways can increase our understanding of the occurrence and development of DN. However, the biological functions and molecular mechanisms of the genes require further study.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/supplementary material.

(I) Conception and design: ZL and YZ. (II) Administrative support: YZ ,XG and JF. (III) Collection and assembly of data: ZL, JF, and ML. (IV) Data analysis and interpretation: ZL. (V) Manuscript writing: all authors. (VI) Final approval of manuscript: all authors.

This study was funded by the National Natural Science Foundation of China (81800675), Guangzhou medical key subject construction project of China (2021-2023) and The Program of Huadu District Science and Technology (No.20-HDWS-019).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.