Identification and functional characterization of DEGs in GSE33455 and GSE55945
GEO microarray (GSE33455) was used to screen DEGs using limma package. Volcano plots displayed the distribution of the 20482 expressed genes (|Log2FC| > 1 and adj.p.value < 0.05 as the cutoff criteria, Figure 1A). A total of 540 DEGs were identified, of which 99 genes were upregulated and 441 genes were downregulated (Figure 1B). To further functionally analyze DEGs, GO term enrichment analysis including BP, CC and MF ontologies and KEGG pathway enrichment analysis were performed. For the upregulated genes, the DEGs were mainly enriched in SMAD protein signal transduction, BMP signaling pathway and regulation of transmembrane receptor protein serine/threonine kinase signaling pathway in the BP group. The KEGG results revealed that the DEGs were mainly enriched in adherens junction, TGF-beta signaling pathway, systemic lupus erythematosus, signaling pathways regulating pluripotency of stem cells and mitophagy–animal (Figure 1C; Table S1). For the downregulated genes, the DEGs were mainly concentrated in regulation of epidermal cell differentiation, response to virus, epidermis development, cellular response to hydrogen peroxide and cellular response to antibiotic in the BP group. The KEGG results revealed that the DEGs were mainly concentrated in NOD-like receptor signaling pathway, rheumatoid arthritis, IL-17 signaling pathway, NF-kappa B signaling pathway and TNF signaling pathway (Figure 1D; Table S2).
To screen out key genes contributing to docetaxel resistance, the GSE55945 dataset, consisted of 13 human prostate malignant tissues and 8 benign tissues, was used. A total of 1362 DEGs, including 80 upregulation and 1282 downregulation, were identified to be candidate key genes in prostate cancer tissues (Figure 2A, B). The results of GO terms of each ontologies and KEGG pathway enrichment analysis were shown in Supplementary Table 3, 4 and the top 5 were exhibited in Figure 2C, D. For the upregulated genes, the DEGs were mainly enriched in forebrain neuron fate commitment, cerebral cortex gabaergic interneuron differentiation, camera-type eye development, regulation of peroxisome proliferator activated receptor signaling pathway and negative regulation of oligodendrocyte differentiation in the BP group. The KEGG results revealed that the DEGs were mainly enriched in insulin secretion, linoleic acid metabolism, ferroptosis, arachidonic acid metabolism and calcium signaling pathway. For the downregulated genes, the DEGs were mainly concentrated in muscle system process, cell-substrate adhesion and regulation of blood pressure in the BP group. The KEGG results revealed that the DEGs were mainly enriched in cGMP-PKG signaling pathway, vascular smooth muscle contraction, hypertrophic cardiomyopathy, proteoglycans in cancer and Dilated cardiomyopathy. As we listed, the DEGs were involved in multiple biological processes and pathways, this revealed the possible mechanisms contributing to docetaxel resistance and tumorigenesis in CRPC, and it deserves further study.
Identification of key genes associated with docetaxel resistance in CRPC
Key gene candidates were identified through Venn diagram analysis. 1 overlapping upregulated gene and 49 overlapping downregulated genes were obtained in the intersection of the GSE33455 DEGs and the GSE55945 DEGs (Figure 3A and B). To further identify the key genes, we analysis the overall survival of the total 50 genes (overlapping genes of the upregulated gene and the downregulated genes) on UCSC Xena online analysis (497 samples from TCGA prostate cancer) (Figure S1; Figure 3C). We found that TUBB4A expression was negatively associated with the overall survival, and the expression level of SRPX and CSRP2 were positively associated with the overall survival (Figure 3C). In addition, we further verified the expression level of the key genes on GEPIA online analysis, and the results revealed that TUBB4A expression was significantly upregulated compared with normal prostate tissues, while the expression level of SRPX and CSRP2 were downregulated (Figure 3D). Altogether, we confirmed TUBB4A, SRPX and CSRP2 as the real key genes contributing to docetaxel resistance and tumorigenesis.
GSEA and validation of the function of TUBB4A, SRPX and CSRP2
To clarify the biological functions of TUBB4A, SRPX and CSRP2, GSEA was performed. Under the cut-off criteria Nominal P < 0.05, FDR < 0.05 and |NES| > 1, a total of 41 functional gene sets were enriched (Figure 4A; Table S5, S6, S7). The top 4 TUBB4A-regulated gene sets were “E2F targets”, “G2M checkpoint”, “MYC targets v1” and “MITOTIC spindle”. The top 4 SRPX-regulated gene sets were “epithelial mesenchymal transition”, “P53 pathway”, “hypoxia” and “estrogen response early”. Interestingly, the top 4 CSRP2-regulated gene sets were also “E2F targets”, “G2M checkpoint”, “MYC targets v1” and “MITOTIC spindle”.
Protein/gene interactions of TUBB4A, SRPX, CSRP2
To investigate the potential functions of TUBB4A, SRPX and CSRP2 in CRPC, GeneMANIA online analysis were employed. As shown in the GeneMANIA network, the 20 proteins/genes were highly associated with the key genes. Among them, PAICS, TUBA3E, HSPA5 and TUBB4B were the most related genes of TUBB4A. TNXB, CCDC80, KCTD13 and SRPX2 were the most related genes of SRPX. KAT14, AGPS, PIAS1 and ATF2 were the most related genes of CSRP2 (Figure 5A).