Gene Expression and Mutation Analysis of Early-Stage Recurrent Endometrial Cancer

Abstract

Purpose/Objective(s)As currently under investigation in PORTEC-4a, an integrated molecular model of endometrial cancer is postulated to risk stratify women to better identify those more likely to benefit from adjuvant radiation therapy following hysterectomy for early-stage disease. Stratification of tumor biology is based on several features, including: p53 mutation, POLE mutation, CTNNB1 mutation, microsatellite instability (MSI), LICAM expression and the presence of LVSI. TCGA data evaluating a large cohort of early and advanced stage endometrial tumors suggests that gene expression profiles may also distinguish between favorable and unfavorable biology of endometrial tumors. We aimed to determine whether recurrent early-stage endometrial tumors have distinct gene-expression patterns from early-stage tumors that did not recur.Materials/MethodsUsing the TCGA-UCEC dataset available via the Genomic Data Commons, 394 Stage I-II tumors’ gene expression data were evaluated and analyzed using R (R Core Team, 2020). Of these cases, 17 were confirmed as relapsed endometrial cancer (10 endometrioid and 7 serious/mixed) and 370 were tumors without documentation of a recurrence. Gene expression analysis was performed with limma, gene set variation analysis with GSVA, and gene enrichment with Metascape. The frequency of several common mutations was evaluated in the recurrent tumor subset.ResultsThere were 860 differentially expressed genes in recurrent versus not recurrent early-stage tumors identified from the TCGA-UCEC dataset. The most significant of which include DDIT3, THAP12, EMSY, CLNS1A, and AAMDC (log[Fold Change] > 1000, adjusted P-value < 0.001). Differentially expressed genes show pathway enrichment in GO Biologic Processes and Reactome Gene Sets, specifically, mitochondrion organization, proteolysis, organelle disassembly, translation, response to endoplasmic reticulum stress, proteinases, cellular component disassembly, covalent chromatin modification, and intrinsic apoptotic signaling pathway. Non-overlapping mutations within the recurrent tumors include, 9 of 17 recurrent tumors with p53 mutation, 3 of 17 with CTNNB1 mutation, 4 of 17 with PIK3R1, PIK3CR and PIK3CB mutation, and 1 of 17 a mutation in POLE.ConclusionOf recurrent early-stage tumors within the TCGA-UCEC dataset, of which as least a majority classify as intermediate or unfavorable according to PORTEC-4a criteria, differential gene expression in recurrent tumors compared to not recurrent tumors demonstrate relative increased expression of genes involved in catabolic and cell cycle processes. PORTEC-4a does not employ gene expression signatures to distinguish amongst favorable, intermediate, and unfavorable subtypes of early-stage endometrial tumors. However, further investigation may prove gene expression incorporation into stratification beneficial. As currently under investigation in PORTEC-4a, an integrated molecular model of endometrial cancer is postulated to risk stratify women to better identify those more likely to benefit from adjuvant radiation therapy following hysterectomy for early-stage disease. Stratification of tumor biology is based on several features, including: p53 mutation, POLE mutation, CTNNB1 mutation, microsatellite instability (MSI), LICAM expression and the presence of LVSI. TCGA data evaluating a large cohort of early and advanced stage endometrial tumors suggests that gene expression profiles may also distinguish between favorable and unfavorable biology of endometrial tumors. We aimed to determine whether recurrent early-stage endometrial tumors have distinct gene-expression patterns from early-stage tumors that did not recur. Using the TCGA-UCEC dataset available via the Genomic Data Commons, 394 Stage I-II tumors’ gene expression data were evaluated and analyzed using R (R Core Team, 2020). Of these cases, 17 were confirmed as relapsed endometrial cancer (10 endometrioid and 7 serious/mixed) and 370 were tumors without documentation of a recurrence. Gene expression analysis was performed with limma, gene set variation analysis with GSVA, and gene enrichment with Metascape. The frequency of several common mutations was evaluated in the recurrent tumor subset. There were 860 differentially expressed genes in recurrent versus not recurrent early-stage tumors identified from the TCGA-UCEC dataset. The most significant of which include DDIT3, THAP12, EMSY, CLNS1A, and AAMDC (log[Fold Change] > 1000, adjusted P-value < 0.001). Differentially expressed genes show pathway enrichment in GO Biologic Processes and Reactome Gene Sets, specifically, mitochondrion organization, proteolysis, organelle disassembly, translation, response to endoplasmic reticulum stress, proteinases, cellular component disassembly, covalent chromatin modification, and intrinsic apoptotic signaling pathway. Non-overlapping mutations within the recurrent tumors include, 9 of 17 recurrent tumors with p53 mutation, 3 of 17 with CTNNB1 mutation, 4 of 17 with PIK3R1, PIK3CR and PIK3CB mutation, and 1 of 17 a mutation in POLE. Of recurrent early-stage tumors within the TCGA-UCEC dataset, of which as least a majority classify as intermediate or unfavorable according to PORTEC-4a criteria, differential gene expression in recurrent tumors compared to not recurrent tumors demonstrate relative increased expression of genes involved in catabolic and cell cycle processes. PORTEC-4a does not employ gene expression signatures to distinguish amongst favorable, intermediate, and unfavorable subtypes of early-stage endometrial tumors. However, further investigation may prove gene expression incorporation into stratification beneficial.