Increased activity in cell division-induced inhibition in response to metal ions
From the TCGA PRAD dataset, we identified 1380 DEGs, including 906 upregulated and 474 downregulated genes (see Table 1, Fig. 1). GO analysis results indicated that upregulated genes primarily participate in biological processes such as "GO:0051301cell division," "GO:0000281mitotic cytokinesis," and "GO:0000278mitotic cell cycle." On the other hand, downregulated genes are mainly involved in processes like "GO:0010273detoxification of copper ion," "GO:0071280cellular response to copper ion," and "GO:0071294cellular response to zinc ion" (see Table 2, Fig. 2).
Based on the functional annotation results, we collected four relevant gene sets. Utilizing the "GSVA" algorithm from the R package, we quantified scores and conducted differential comparisons. The results indicate that lymph node metastasis patients exhibit significantly higher GSVA scores for cell division and the mitotic cell cycle compared to non-metastatic patients, with p-values of 0.0003 and 0.0015, respectively. The GSVA scores for cellular response to copper ion and zinc ion in lymph node metastasis patients are significantly lower than those in patients without lymph node metastasis, with p-values of 1.2e-08 and 4.8e-05, respectively (see Fig. 3). These findings suggest that patients with lymph node metastasis display heightened activity in cell division while exhibiting suppressed responses to metal ions.
Filtering DNA Methylation-Regulated DEGs in Patients with Lymph Node Metastasis
Subsequently, we utilized the "ChAMP" R package to analyze the DNA methylation data detected in the TCGA PRAD dataset. We identified 18,498 DMPs, including 1,636 with increased methylation levels and 37,993 with decreased methylation levels, covering a total of 8,155 genes. Based on the DNA methylation data from the GSE220910 dataset, we filtered 81,009 DMPs, comprising 43,016 with increased methylation levels and 37,993 with decreased methylation levels, covering a total of 15,185 genes. Combining the previously identified 1,380 DEGs from TCGA PRAD, we ultimately obtained 263 DEGs related to 382 DMPs (refer to Table 3 and Fig. 4).
To validate whether the aforementioned DEGs are regulated by DNA methylation, we separately calculated the Pearson correlation coefficients (PCC) between the 263 DEGs and their corresponding 259,531 methylation probe beta values. Correlation tests were conducted, resulting in 13 relationships with PCC < -0.3 and test p-values < 0.01. Among these, LTA (cg16243606), DOK3 (cg20138861), TNFRSF25 (cg00834988), and CHRM1 (cg23890832) exhibited PCC coefficients of -0.4092, -0.4111, -0.4054, and − 0.4598, respectively, with p-values < 0.01 (Table 1 and Fig. 5).
Table 1
13 DEGs Regulated by DNA Methylation
DMP | DEGs | PCC | cor.test pValue |
cg16243606 | LTA | -0.4092 | 1.57E-21 |
cg11261261 | GFI1 | -0.3844 | 5.55E-19 |
cg00842595 | ARHGAP15 | -0.3390 | 7.36E-15 |
cg14730097 | LAT | -0.3020 | 5.78E-12 |
cg02741305 | COL11A2 | -0.3106 | 1.35E-12 |
cg20138861 | DOK3 | -0.4111 | 9.95E-22 |
cg00834988 | TNFRSF25 | -0.4054 | 4.01E-21 |
cg23890832 | CHRM1 | -0.4598 | 2.05E-27 |
cg27587826 | LENG8 | -0.3022 | 5.68E-12 |
cg14519392 | ENGASE | -0.3378 | 9.28E-15 |
cg12969193 | IL21R | -0.3013 | 6.53E-12 |
cg17429587 | BMPER | -0.3338 | 1.99E-14 |
cg09641955 | KIF2C | -0.3429 | 3.49E-15 |
Aberrant expression of LTA, DOK3, TNFRSF25, and CHRM1 is associated with the prognosis of PCa lymph node metastasis
We observed a significant upregulation of LTA, DOK3, and TNFRSF25 genes in patients with lymph node metastasis, while CHRM1 showed a significant downregulation, with fold change (FC) values of 1.3316, 1.1148, 1.1095, and 0.8813, respectively (refer to Table 3). The p-values were 2.47e-07, 5.42e-07, 0.0003, and 7.46e-08, respectively (see Fig. 6). Corresponding probe methylation levels can be seen in Figure S1.
To further validate the potential of the above genes in predicting clinical outcomes, we integrated patient data from 498 cases with recorded gene expression and prognosis information. Prognostic analyses were conducted for LTA, TNFRSF25, DOK3, and CHRM1. Patients were stratified into high and low expression groups based on the expression levels of LTA (threshold: 0.1787). The constructed survival model yielded a Hazard Ratio (HR) of 2.1620 [95% CI: 1.3930–3.3570] with a test p-value of 0.0004 (see Fig. 7A). Additionally, focusing on lymph node metastasis patients, 81 cases were categorized into high and low expression groups using a threshold of 11.2949. The resulting survival model had an HR of 2.4840 [95% CI: 0.8405-7.3400] with a test p-value of 0.0890 (see Fig. 7B). The same approach was applied to TNFRSF25, DOK3, and CHRM1 genes (see Fig. 7C-D, Fig. 8). Kaplan-Meier survival analysis revealed that, regardless of lymph node metastasis occurrence, upregulation of LTA, TNFRSF25, and DOK3 expression was associated with a shortened Progression-Free Interval (PFI) in PCa patients, while CHRM1 exhibited the opposite effect. Additionally, we supplemented the prognosis analysis for the probes corresponding to these four genes (see Figure S2, Figure S3).