Identification and Validation of Hub Genes and Construction of miRNA-Gene and Transcription Factor-Gene Networks in Adipogenesis of Mesenchymal Stem Cells

Abstract

Background: Adipogenic differentiation stands as a crucial pathway in the range of differentiation options for mesenchymal stem cells (MSCs), carrying significant importance in the fields of regenerative medicine and the treatment of conditions such as obesity and osteoporosis. However, the exact mechanisms that control the adipogenic differentiation of MSCs are not yet fully understood.

Materials and methods: We procured datasets, namely GSE36923, GSE80614, GSE107789, and GSE113253, from the Gene Expression Omnibus database. These datasets enabled us to perform a systematic analysis, including the identification of differentially expressed genes (DEGs) pre- and postadipogenic differentiation in MSCs. Subsequently, we conducted an exhaustive analysis of DEGs common to all four datasets. To gain further insights, we subjected these overlapped DEGs to comprehensive gene ontology enrichment and Kyoto Encyclopedia of Genes and Genomes pathway analyses. Following the construction of protein-protein interaction (PPI) networks, we meticulously identified a cohort of hub genes pivotal to the adipogenic differentiation process and validated them using real-time quantitative polymerase chain reaction. Subsequently, we ventured into the construction of miRNA-gene and TF-gene interaction networks.

Results: Our rigorous analysis revealed a total of 18 upregulated DEGs and 12 downregulated DEGs that consistently appeared across all four datasets. Notably, the peroxisome proliferator-activated receptor signaling pathway, regulation of lipolysis in adipocytes, and the adipocytokine signaling pathway emerged as the top-ranking pathways significantly implicated in the regulation of these DEGs. Subsequent to the construction of the PPI network, we identified and validated 10 key node genes, namely IL6, FABP4, ADIPOQ, LPL, PLIN1, RBP4, ACACB, NT5E, KRT19, and G0S2. Our endeavor to construct miRNA-gene interaction networks led to the discovery of the top 10 pivotal miRNAs, including hsa-mir-27a-3p, hsa-let-7b-5p, hsa-mir-1-3p, hsa-mir-124-3p, hsa-mir-155-5p, hsa-mir-16-5p, hsa-mir-101-3p, hsa-mir-21-3p, hsa-mir-146a-5p, and hsa-mir-148b-3p. Furthermore, the construction of TF-gene interaction networks revealed the top 10 critical TFs: ZNF501, ZNF512, YY1, EZH2, ZFP37, ZNF2, SOX13, MXD3, ELF3, and TFDP1.

Conclusions: In summary, our comprehensive study has successfully unraveled the pivotal hub genes that govern the adipogenesis of MSCs. Moreover, the meticulously constructed miRNA-gene and TF-gene interaction networks are poised to significantly augment our comprehension of the intricacies underlying MSC adipogenic differentiation, thus providing a robust foundation for future advances in regenerative biology.

Figure 1

Box plots of GSE36923 (a), GSE80614 (b), GSE107789 (c), and GSE113253 (d) after normalization. The vertical axis is the name of the samples, while the horizontal axis stands for the values of expression. The black line stands for the median of data and represents the normalization degree. After normalization, the black line in each group was almost collinear, which indicates an excellent degree of normalization.

Figure 2

Volcano plots of differentially expressed genes of GSE36923 (a), GSE80614 (b), GSE107789 (c), and GSE113253 (d). The abscissa is log2 (FC), and the ordinates are −log10 (P-value). The red dots stand for the upexpressed genes, and the blue dots stand for the downexpressed genes, while the gray dots represent genes not differentially expressed.

Figure 3

Heat maps of the differentially expressed genes (top 50 upregulated and downregulated genes) of GSE36923 (a), GSE80614 (b), GSE107789 (c), and GSE113253 (d). Red represents a high expression, and the deeper the red color, a higher expression value. Blue represents low expression, and a deeper blue color, a lower expression value.

Figure 4

The Venn diagram of overlapped upregulated (a) and downregulated (b) differentially expressed genes of four datasets.

Figure 5

Gene ontology (GO) enrichment results and Kyoto Encyclopaedia of Genes and Genomes (KEGG) pathway analysis of overlapped upregulated (a) and downregulated (b) differentially expressed genes (DEGs). Bubble charts show the enrichment of DEGs in GO results and signaling pathways. The x-axis label represents the gene ratio, and the y-axis label represents GO terms. The size of the circle stands for gene count, and the color of the circle stands for adjusted P value.

Figure 6

Protein–protein interaction network. The circle represents genes, and the line indicates the interactions among genes. A thicker line stands for a higher edge confidence, and a larger node size stands for a higher degree.

Figure 7

Oil red O staining and validation of 10-node genes using real-time quantitative polymerase chain reaction. The staining with Oil red O shows a comparison between adipogenic human bone marrow-derived mesenchymal stem cells (b) and controls (a). The expression levels of FABP4, ADIPOQ, LPL, PLIN1, RBP4, ACACB, and G0S2 were significantly increased (c–i), while the expression levels of IL6, NT5E, and KRT19 were significantly decreased (j–l) after adipogenic induction. P-values were calculated using the two-tailed Student's t-test, and statistical significance was considered when  ^∗P < 0.05. Scale bar = 200 μm.

Figure 8

The miRNA–target gene interaction network. The circle represents genes, and the line indicates the interactions among genes. A thicker line stands for a higher edge betweenness, and a larger node size stands for a higher degree.

Figure 9

The TF–target gene interaction network. The circle represents genes, and the line indicates the interactions among genes. A thicker line stands for a higher edge betweenness, and a larger node size stands for a higher degree.