Abstract
Immunohistochemistry for mismatch repair proteins MLH1, MSH2, MSH6, and PMS2 is an effective screen to detect individuals at risk for Lynch syndrome. College of American Pathologists guidelines stipulate that protein expression should be reported as present versus absent, as most patients with germline mutations in a mismatch repair gene have complete loss of protein expression in tumor cells. A similar approach is employed to screen for cancer patients eligible for immune checkpoint blockade. This "all or none" interpretive approach ignores substantial evidence that mismatch repair may be more finely regulated by other mechanisms. We have observed clinically that MSH6 expression is variable, even in carcinomas that are overall considered positive for MSH6 expression. A proof-of-principle study was therefore designed to more rigorously quantify the protein expression of MSH6 and its binding partner, MSH2, using image analysis applied to age-matched endometrioid grade 2 subsets that were either mismatch repair intact or MLH1-deficient due to MLH1 gene methylation. In both endometrioid groups, MSH6 expression was significantly lower than MSH2 expression. MSH6 expression increased in higher grade, mismatch repair intact serous carcinomas, but it was still significantly lower than that for MSH2. MSH2 expression was consistently high across the 3 different tumor groups. These results suggest that MSH6 expression is subject to wide fluctuations in expression, even when overall its expression is considered intact. While such fluctuations are likely not relevant for Lynch syndrome screening, they may be more impactful when considering patients eligible for immune checkpoint blockade.