Loss of DNA mismatch repair proteins in prostate cancer

Recent studies have suggested an increased risk of prostate cancer in men with Lynch syndrome driven by germline mutations in mismatch repair (MMR) genes. However, the incidence and clinical implication of MMR deficiency in sporadic prostate cancers remain poorly understood. We immunohistochemically stained for MLH1, MSH2, MSH6, and PMS2 in a set of tissue microarray consisting of 220 radical prostatectomy specimens and evaluated the relationship between loss of their expression and available clinicopathological features. MLH1, MSH2, MSH6, and PMS2 were lost in 2 (0.9%), 6 (2.7%), 37 (16.8%), and 27 (12.3%) prostate cancers, respectively. Loss of at least 1 MMR protein was identified in 50 (22.7%) cases. There were no statistically significant associations between MMR deficiency and patient age, family history of prostate cancer, Gleason score, or pT/pN stage. Nonetheless, the levels of preoperative prostate-specific antigen (PSA) were significantly (P = .015) higher in patients with MMR deficiency (mean ± SD: 9.12 ± 9.01 ng/mL) than in those without abnormal MMR (5.76 ± 3.17 ng/mL). There were 15 (6.8%) cases showing loss of at least 2 MMR proteins, which was not significantly associated with PSA level or tumor grade/stage. Additionally, 5 and 2 cases showed losses of at least 3 MMR proteins and all 4 proteins, respectively. Kaplan-Meier analysis revealed no significant associations between loss of MLH1 (P = .373), MSH2 (P = .348), MSH6 (P = .946), or PMS2 (P = .681), or at least 1 (P = .477), 2 (P = .486), or 3 (P = .352) MMR proteins and biochemical recurrence. Further analyses of the data on programmed death-ligand 1 (PD-L1) expression previously stained in the same set of tissue microarray demonstrated associations between loss of ≥2 MMR proteins and a higher rate of PD-L1 expression in cancer cells (17.2% vs 5.2%; P = .033) as well as between cases showing both loss of ≥1 MMR protein(s) and PD-L1 expression in tumor-infiltrating immune cells vs a higher risk of biochemical recurrence (P = .045). MMR protein loss was seen in a subset of prostate cancers. Interestingly, it was associated with significantly higher levels of PSA. Moreover, immunohistochemical detection of MMR proteins together with other proteins, such as PD-L1, might be helpful in predicting tumor recurrence following radical prostatectomy.