Genetics and Genomics of Endometrial Cancer

Abstract

Endometrial cancer (EC) is the most common gynecological cancer among women in the developed world and is hypothesized to arise from excess estrogen exposure from established risk factors like estrogen-only hormone therapy and obesity. EC is divided into the common “estrogen-dependent” endometrioid subtype and the rare “estrogen-independent” non- endometrioid subtype. However, this broad categorization of EC is not sufficient based on evidence for EC heterogeneity. Furthermore, family history and hereditary syndromes also increase risk, suggesting a genetic component. This dissertation examines the genetic and genomic architecture of EC to provide insight into its etiology and heterogeneity. In Chapter 1, a four-study EC meta-analysis of 4,907 cases and 11,645 controls in women of European ancestry is presented. Four loci reached genome-wide significance. One novel susceptibility locus at 6p22.3 was identified and two previously discovered loci at 6q22.31 and 13q22.1 were confirmed. Genes near the 6p22.3 locus are implicated in malignancy and poor prognosis in many cancers, highlighting the potential importance of this region to general cancer susceptibility. In Chapter 2, we conduct an exome-wide association study of EC. Using a new, commercially-developed exome array comprising ~260,000 putative functional exonic variants, we genotyped a multiethnic population of 3,067 women (1,169 EC cases and 1,898 controls) from the Epidemiology of Endometrial Cancer Consortium to test whether rare variants in coding regions are associated with endometrial cancer risk. No variants reached global significance in this study. Larger studies are needed to detect associations between rare exonic variants and EC. In Chapter 3, we combined targeted next-generation sequencing from archival EC tissue with clinical, immunohistochemical, and epidemiologic data for a comprehensive characterization of EC in 37 women from the Nurses’ Health Study. Mutations most frequently occurred in TP53, PTEN, and PIK3CA. TP53 mutations were seen in the majority of tumors that were p53 abnormal. Low grade correlated with frequency of PTEN and PIK3CA mutation. The archival EC tissue had mutation profiles consistent with previous studies, supporting use of targeted sequencing panels on archival tissue for mutation detection. Our comprehensive annotation of EC tumors demonstrates the utility of integrating many data types to reveal differences between tumors.