Endometrial Cancer Characteristics and Risk of Recurrence

The Role of Epigenetic Silencing of MLH1

Lindsay E. Borden, MD; Tonja M. Locklear, PhD; Douglas J. Grider, MD; Janet L. Osborne, MD; Erin J. Saks, MD; Fidel A. Valea, MD; David A. Iglesias, MD


Am J Clin Pathol. 2022;157(1):90-97. 

In This Article

Materials and Methods

A retrospective study was conducted among all women who were diagnosed with endometrial adenocarcinomas treated at a single institution from January 2012 through December 2017. These women underwent universal testing for microsatellite instability by ancillary immunohistochemical stains for the MMR proteins MLH1, MSH2, MSH6, and PMS2, per institutional standards. Formalin-fixed, paraffin-embedded blocks were cut at a 4-μm thickness. Immunohistochemical staining was performed using the following antibodies: MLH1 (clone M1; Ventana), MSH2 (clone G219-1129; Ventana), MSH6 (clone 5P93; Ventana), and PMS2 (clone A16-4; Ventana). Study data were collected and managed using REDCap electronic data capture tools.[3,4] The study was approved by our institutional review board (protocol 2472).

MMR proteins are normally present in all cells but are especially detectable in cells that are replicating, such as endometrial carcinoma. In January 2012, our institution implemented a protocol for universal screening of all endometrial cancers to detect tumor expression of MMR proteins. The presence of MMR proteins is noted by nuclear positivity on immunohistochemistry. Immunohistochemical stains are interpreted as positive when 5% or more of the carcinoma is nuclear positive regardless of any variability in staining intensity. PMS2 is uniformly weaker in intensity in our laboratory compared with MLH1, MSH2, and MSH6 nuclear positivity. Cytoplasmic and membranous expression is not considered in the interpretation. A negative nuclear stain indicates a deficient state, which can be due to complete loss of the protein or due to a nonfunctioning protein. MLH1 is a heterodimer with PMS2; MSH2 is a heterodimer with MSH6. Thus, losses of proteins usually, but not always, occur in pairs. Such MMR protein losses are usually acquired (somatic) when involving MLH1 and PMS2 proteins. A representative example of this staining pattern is shown in Figure 1. Loss of MLH1 protein is most commonly due to hypermethylation of its promoter region and also results in loss of PMS2 protein. When MLH1 protein expression is lost, the tissue is tested for MLH1 promoter hypermethylation to determine if the MLH1 protein loss is germline or somatic. Germline loss of MLH1 protein and its heterodimer, PMS2, is diagnostic of Lynch syndrome. On occasion, MLH1 protein staining demonstrates aberrant nuclear dot-like positivity, obscuring true loss of expression Figure 2. However, all such cases demonstrate complete loss of PMS2 protein and are positive for MLH1 promoter hypermethylation, confirming sporadic acquired loss of MLH1 function by hypermethylation. In rare studies showing subclonal loss of MLH1 and PMS2 within the carcinoma, all are found to have MLH1 promoter hypermethylation in the area with the loss of MLH1 and PMS2 proteins.[5] Loss of MSH2 and MSH6 proteins is most always germline, another mechanism in Lynch syndrome. However, there are recent reports of rare biallelic somatic mutations in MMR protein genes, including MSH2 and MSH6, that are not germline.[6] MSH2 and MSH6 proteins form the heterodimer MutSa, which recognizes DNA base mismatches and thus initiates repair. While MSH6 protein is not known to dimerize with any other protein, sometimes MSH2 protein dimerizes with MSH3 when MSH6 is mutated.[7] The result is the rare isolated absence of MSH6 protein. No cases of subclonal loss of expression of MSH6 and/or MSH2 were found in our study. All cases of suspected Lynch syndrome were referred for genetic testing.

Figure 1.

Representative staining pattern showing concurrent loss of MLH1 and PMS2 in International Federation of Gynecology and Obstetrics grade 1 endometrioid adenocarcinoma (×10). A, H&E. B, MLH1 absent. C, MSH2 present. D, MSH6 present. E, PMS2 absent.

Figure 2.

Nuclear dot-like positivity of MLH1 in International Federation of Gynecology and Obstetrics grade 1 endometrioid adenocarcinoma, ×10.

Cases were included in the database if complete MMR testing results were available and subsequent MLH1 hypermethylation testing results were also available, if indicated. Patients were excluded if they had nonendometrioid histology or were diagnosed with a synchronous primary malignancy with or without concurrent treatment. Clinical and pathologic data were abstracted by chart review and included patient demographics, tumor histology and grade, staging characteristics, adjuvant treatment received (vaginal brachytherapy, pelvic radiation, chemotherapy, or hormone therapy), and time to recurrence and/or death as well as treatment received for recurrence. Patients were censored at the date of last contact or death. Patients were then categorized into one of three groups based on MMR testing: intact MMR expression (MMR+), probable MMR mutation (MMR–), or hypermethylation of MLH1 (hMLH1+).

Early stage patients were further categorized into risk groups, adapted from criteria presented in Gynecologic Oncology Group (GOG)–99.[8] High-intermediate risk was defined as one or more of the following risk factors: grade 2 or 3 histology, greater than 50% depth of invasion, or presence of lymphovascular space invasion (LVSI). To be characterized as high-intermediate risk, patients younger than 50 years required all three risk factors, patients between the ages of 50 and 69 years required two risk factors, and patients 70 years or older required one risk factor. The remaining patients were subdivided into two groups: low risk and low-intermediate risk. Low risk was defined as grade 1 or 2 histology with no myometrial invasion. Low-intermediate risk was defined as early stage patients who met neither low-risk nor high-intermediate-risk criteria.

Fisher-Freeman-Halton test and Kruskal-Wallis test were used to compare characteristics across groups. A multiple logistic regression was used to predict recurrence by group in early stage (stage I and II) patients, controlling for select demographic, disease, and treatment characteristics. Kaplan-Meier product limit and log rank test along with a simple Cox proportional hazards regression were used to compare differences in survival among groups. SAS 9.4 (SAS Institute) was used for all analyses.

In accordance with the journal's guidelines, we will provide our data for the reproducibility of this study in other centers if requested.