Mismatch repair protein expression in 1049 endometrial carcinomas, associations with body mass index, and other clinicopathologic variables

Abstract

Objective: Links between obesity, with its attendant estrogen abnormalities, and the endometrial carcinoma (EC) DNA Mismatch Repair Protein (MMR) system have recently been proposed. We investigated relationships between body mass index (BMI) and clinicopathological correlates including MMR expression in a large single institution EC cohort.

Methods: Clinical and pathological databases from 2007 to 2012 were used to identify consecutive hysterectomy specimens with EC. Univariate and multivariate analyses were used to explore relationships between BMI, age, stage, tumor type and immunohistochemical results for MLH1, PMS2, MSH2 and MSH6.

Results: 1049 EC were identified. Overall, BMI was higher amongst women with normal MMR (p=0.002). However, when stratified by age and specific MMR, statistically significant differences localized exclusively to women <50years old with loss of MSH2 and/or MSH6 (p=0.003 and p=0.005 respectively). Higher BMI correlated with endometrioid FIGO 1 and 2 tumors (p<0.001) and with stage 1a (p<0.001). Conversely, MMR abnormalities did not show significant associations with stage (p=0.302) or histologic grade (p=0.097).

Conclusions: BMI showed statistically significant associations with MMR expression, tumor grade and stage amongst 1049 consecutive EC. Obesity correlates with lower grade and stage EC. A link between BMI and maintenance of the MMR system is not supported by our data because the only statistically significant association occurred in women <50years old with MSH2 and/or MSH6 abnormalities where Lynch syndrome related cases are expected to cluster.

Keywords: BMI; Body mass index; DNA mismatch repair protein; Endometrial cancer; MMR.

Figure 1

BMI data stratified by MMR protein absence (any protein absent, and individual MLH1, PMS2, MSH2, or MSH6 absent), compared to a control group of women with all MMR proteins present. Comparison demonstrates that women with EC MMR intact have a higher BMI than those with any abnormal MMR protein (p=0.002). This largely holds true when stratified by the individual abnormal protein lost (remaining comparisons). Boxes: 25–75^th percentile; whiskers: range; central line: mean; central dot: median.

Figure 2

Women <50 years of age who lack MSH2 and/or MSH6 in their EC specimens have markedly lower BMIs than their <50 year counterparts with all proteins intact (p=0.003 and p=0.005, respectively). Boxes: 25–75^th percentile; whiskers: range; central line: mean; central dot: median.