EPM2A acts as a protective factor in prostate cancer, evidence from a real-world patient cohort

Abstract

EPM2A encodes a dual specificity phosphatase and has been proven to be a potential biomarker in several cancers but has not been mentioned in prostate cancer (PCA). We investigated the prognostic and therapeutic value of EPM2A in PCA. The TCGA-PRAD cohort was collected to evaluate the differential expression, prognostic value, immunocyte infiltration and drug sensitivity of EPM2A in PCA. We constructed a nomogram model to predict the recurrence probability for PCA patients. Immunohistochemistry was used to validate the different transcript levels of EPM2A between tumor and normal tissues. A real-world AHMU-PC cohort was employed for validation. The results showed decreased expression of EPM2A in 95.65% of tumor tissues and was related to their prognosis, especially PCA (p = 0.008, HR = 0.57, 95% CI: 0.371–0.863). Further multiple analysis by adjusting clinical features revealed that EPM2A acted as an independent prognostic factor (p = 0.014, HR = 0.589, 95% CI: 0.386–0.898). Pathway enrichment analysis showed variable signaling activation between high EPM2A expression patients (HEXP) and low EPM2A expression patients (LEXP). The HEXP group contained higher infiltration of immunocytes than the LEXP group, as well as high levels of PD-1, PD-L1 and PD-L2, while LEXP patients were more sensitive to cisplatin, paclitaxel and bicalutamide therapy. The nomogram containing the EPM2A group, T stage and Gleason score showed a preferable prognostic value (AUC = 0.755; Hosmer‒Lemeshow, p = 0.486). In validation, we confirmed the lower transcript level of EPM2A in PCA than in normal tissues (120.5 ± 2.159 vs. 138.3 ± 1.83, p = 0.035) and correlated it with the expression level of PD-1 (R = 0.283). Among the 66 patients from the AHMU-PC cohort, we further validated the function of EPM2A in PCA patients. HEXP patients had longer recurrence-free survival times (1207 ± 110 vs. 794.2 ± 97.02, p = 0.0063) and favorable prognoses (HR: 0.417, 95% CI: 0.195–0.894, p = 0.0245). Collectively, we identified the prognostic value of EPM2A in PCa via a bioinformatics method. Patients with higher EPM2A may be more sensitive to immunotherapy, and patients with lower EPM2A were more suitable for bicalutamide, cisplatin and paclitaxel therapy.