Context-related biomarkers in endometrial cancer. A study with focus on obesity and genomic alterations

Abstract

Background: Endometrial cancer is the most common female pelvic gynaecologic malignancy in industrialised countries, and incidence has been increasing over the past decades. This has partly been ascribed to the increasing obesity epidemic seen worldwide, and particularly in affluent countries. Whereas increasing body mass index (BMI, kg/m2) is a known risk factor for endometrial cancer, less is known about its influence on tumour development and prognosis. Aims: The aim of this study was to increase the understanding about how contextrelated factors, including obesity (assessed by BMI and imaging methods) and genomic alterations (assessed by DNA ploidy status), are related to molecular tumour markers and outcome in endometrial cancer. By exploring gene and protein expression data from tumours arising in different settings, we aimed to shed light on potential contextrelated alterations, that may improve prognostication and represent relevant targets for therapy in future clinical trials. Materials and methods: For the studies included in this thesis (Paper I-IV), cohorts of patients treated for primary endometrial cancer at Haukeland University Hospital with thorough follow-up data and clinicopathological characterisation were used. For subsets of the patients, FFPE tissue was available for IHC analysis (Paper I-IV), fresh ethanol fixed tissue was used for DNA ploidy analysis (Paper II and III), and fresh frozen tissue was used for gene expression microarray (Paper II-IV) and RPPA analyses (Paper IV). Preoperative CT scans were used to study body fat distribution (Paper III). Results: High BMI was significantly associated with low FIGO stage, endometrioid histology and a high level of PR expression, but not ERα expression. Women with BMI≥25 had significantly better endometrial cancer survival compared to women with BMI<25 in univariable analysis, however not significant in multivariable analysis. Applying overall survival as outcome measure, increasing BMI independently predicted worse survival (Paper I). Aneuploidy was significantly associated with high age, high FIGO stage and high grade, non-endometrioid histology and ER/PR negativity, and independently predicted reduced survival. In ER/PR negative tumours, aneuploidy independently predicted recurrence and lymph node metastasis. A nine-gene prognostic ‘aneuploidy signature’, linked to low expression of chromosome 15q genes, was identified and validated in TCGA data (Paper II). Abdominal fat volumes were strongly positively correlated with BMI and waist circumference, and inversely correlated with liver density. High fat volumes and BMI were associated with low grade endometrioid tumours and PR and AR positivity, but not ERα positivity. The visceral fat percentage, VAV%, was not correlated with BMI or total abdominal fat volume, however, high VAV% was associated with high age and aneuploidy, and independently predicted reduced survival. Tumours from patients with low VAV% showed enrichment of gene signatures related to inflammatory and immunogenic signalling (Paper III). In endometrioid endometrial cancers, BMI was significantly correlated with a signature of hormone receptor expression, as well as PR and phospho-ERα (S118) levels. BMI was negatively correlated with RTK- and MAPK-pathway activation, and particularly phospho-MAPK (T202 Y204) level. In the subset of FIGO stage 1, grade 1-2 tumours, non-obese patients had significantly reduced survival compared to obese patients, associated with higher level of MAPK- and RTK-pathway activation. The obese patients had higher phospho-ERα (S118) levels, and showed enrichment of gene signatures related to oestrogen signalling, inflammation, immune signalling and hypoxia (Paper IV). Conclusions: BMI and imaging based estimates of obesity are associated with clinicopathological markers of less aggressive endometrial cancer (Paper I, III and IV). High BMI is associated with PR and AR but not ERα expression (Paper I, III and IV). Obese patients with endometrioid endometrial cancer have higher levels of phosphorylated ERα. Non-obese patients have higher levels of phosphorylated MAPK (Paper IV). High BMI is associated with improved DSS in univariable, but not multivariable analysis, and worse OS in multivariable analysis (Paper I). Increasing VAV% independently predicts reduced DSS (Paper III). Obesity is associated with improved DSS in patients with assumed excellent prognosis (Paper IV). Gene sets linked to inflammation and immune activation are enriched in tumours arising in patients with low VAV%, and equally in tumours arising in obese patients with FIGO stage 1, grade 1-2 tumours (Paper III and IV). DNA ploidy is a robust prognostic marker in endometrial cancer, and aneuploidy independently predicts reduced DSS. In patients with ER/PR negative tumours, aneuploidy independently predicts increased risk of lymph node metastases and recurrence (Paper II). A nine-gene aneuploidy signature is associated with reduced survival and low expression of chromosome 15q genes (Paper II).