Characterization of mismatch-repair/microsatellite instability-discordant endometrial cancers

Abstract

Background: Mismatch-repair (MMR)/microsatellite instability (MSI) status has therapeutic implications in endometrial cancer (EC). The authors evaluated the concordance of testing and factors contributing to MMR expression heterogeneity.

Methods: Six hundred sixty-six ECs were characterized using immunohistochemistry (IHC), MSI testing, and mut-L homolog 1 (MLH1) methylation. Select samples underwent whole-transcriptome analysis and next-generation sequencing. MMR expression of metastatic/recurrent sites was evaluated.

Results: MSI testing identified 27.3% of cases as MSI-high (n = 182), MMR IHC identified 25.1% cases as MMR-deficient (n = 167), and 3.8% of cases (n = 25) demonstrated discordant results. A review of IHC staining explained discordant results in 18 cases, revealing subclonal loss of MLH1/Pms 1 homolog 2 (PMS2) (n = 10) and heterogeneous MMR IHC (mut-S homolog 6 [MSH6], n = 7; MLH1/PMS2, n = 1). MSH6-associated Lynch syndrome was diagnosed in three of six cases with heterogeneous expression. Subclonal or heterogeneous cases had a 38.9% recurrence rate (compared with 16.7% in complete MMR-deficient cases and 9% in MMR-proficient cases) and had abnormal MMR IHC results in all metastatic recurrent sites (n = 7). Tumors with subclonal MLH1/PMS2 demonstrated 74 differentially expressed genes (determined using digital spatial transcriptomics) when stratified by MLH1 expression, including many associated with epithelial-mesenchymal transition.

Conclusions: Subclonal/heterogeneous MMR IHC cases showed epigenetic loss in 66.7%, germline mutations in 16.7%, and somatic mutations in 16.7%. MMR IHC reported as intact/deficient missed 21% of cases of Lynch syndrome. EC with subclonal/heterogeneous MMR expression demonstrated a high recurrence rate, and metastatic/recurrent sites were MMR-deficient. Transcriptional analysis indicated an increased risk for migration/metastasis, suggesting that clonal MMR deficiency may be a driver for tumor aggressiveness. Reporting MMR IHC only as intact/deficient, without reporting subclonal and heterogeneous staining, misses opportunities for biomarker-directed therapy.

Plain language summary: Endometrial cancer is the most common gynecologic cancer, and 20%-40% of tumors have a defect in DNA proofreading known as mismatch-repair (MMR) deficiency. These results can be used to guide therapy. Tests for this defect can yield differing results, revealing heterogeneous (mixed) proofreading capabilities. Tumors with discordant testing results and mixed MMR findings can have germline or somatic defects in MMR genes. Cells with deficient DNA proofreading in tumors with mixed MMR findings have DNA expression profiles linked to more aggressive characteristics and cancer spread. These MMR-deficient cells may drive tumor behavior and the risk of spreading cancer.

Keywords: Lynch syndrome (LS); biomarker; endometrial cancer; microsatellite instability (MSI); mismatch-repair (MMR) deficiency; transcriptome.

Figure 1:

Examples of subclonal and heterogeneous MMR IHC staining. A) Case 8 subclonal MLH1 and PMS2 expression with 40% intact staining (5x magnification). B) Case 18 Heterogeneous MSH6 expression with 60% intact staining with strong diffuse MSH2 expression in the corresponding area (5x magnification). C) Case 12 Heterogeneous MSH6 expression with 5% intact staining (5x & 10x magnification) with strong diffuse MSH2 expression in the corresponding area (5x magnification).

Figure 2:

Examples of notable IHC staining patterns. A) Case 19: H&E staining demonstrating MSI morphology with tumor-infiltrating lymphocytes (TIL) within the stroma (#), intratumoral TIL (arrow), and peritumoral inflammatory response (star); B) Case 19: Intact MLH1 staining (5x magnification). C) Case 9: Original tumor with subclonal MLH1/PMS2; D) Case 9: Aortic lymph node metastasis with distinct foci of intact staining and complete loss MLH1/PMS2 (staining for MLH1 shown at 5x magnification).

Figure 3:

Characterization of metastasis and recurrence of tumors with MMR heterogeneity. (A-C) Example of digital spatial transcriptomics performed on cases with subclonal expression. A) Case 8 with subclonal MLH1 and PMS2 expression in primary tumor. Highlighted regions (B & C) as an example of selected regions of interest (ROI). B) Fluorescently tagged morphology markers used to display region of tumor with deficient MLH1 expression. Yellow indicates a pan-cytokeratin antibody (AE-1/AE-3), Green is a nucleic acid antibody (Syto-83), Blue indicates MLH1 fluorescent antibody, and Red indicates a CD-3 antibody. C) Fluorescently tagged morphology markers used to display region of tumor with intact MLH1 expression. D) Example of pelvic disease recurrence from Case 8 with complete loss of MLH1/PMS2 expression. E) Differential gene expression data between cells with intact MLH1 expression and deficient MLH1 expression. Red and green data points with adjusted p-value demonstrating significantly different gene expression. Labeled genes associated with epithelial-mesenchymal transition. F) Recurrence-free survival (RFS) of the entire cohort by MMR expression, including significantly worse RFS of cases presented in this manuscript with MMR intratumoral heterogeneity.