Abstract 4435: Genome-scale DNA methylation changes delineate uterine leiomyoma subgroups

Abstract

Abstract Symptomatic uterine leiomyomas (ULs) occur in as many as 20% of women. These bening lesions are a major burden to women's health by causing health complications such as abnormal bleeding, pain and infertility, and they are a common cause of hysterectomy that is at present the only curative treatment. Similar to other solid tumors, genetic and epigenetic changes are likely to have value as predictors of UL clinical outcome, as well as response to treatment. Recent findings from us and others have provided evidence for existence of at least three distinct molecular UL subgroups, each displaying a characteristic genetic driver aberration; MED12 mutation, HMGA2 activation or FH mutation. We conducted DNA methylation analysis of bisulfite sequencing data from 59 tumors and 38 normal myometrium samples on targeted genomic regions (84.5Mb) such as CpG islands, shores and shelves, GENCODE promoters and DNaseI hypersensitive sites from ENCODE data using Agilent SureSelect methyl-seq target enrichment system. The studied samples consisted of 17 MED12-mutated, 11 HMGA2-upregulated and 6 FH-mutated tumors. Whole-genome sequencing and expression array profiling had been performed previously from the same specimens (Mehine, Kaasinen, et al. 2014). We developed a data processing pipeline that makes use of: (i) Trim Galore tool for quality control of sequencing data (ii) Bismark for alignment and DNA methylation calling (Krueger & Andrews 2011). Hierarchical clustering of all detected CpG sites revealed that MED12- and FH-mutated tumors cluster according to the driver changes, and HMGA2-upregulated tumors cluster in two separate branches. Global hypermethylation profile was detected in FH-mutated tumors, which is compatible with previous reports in paragangliomas (Letouzé et al. 2013). Four tumors with COL4A5/COL4A6 aberration did not display uniform DNA methylation pattern and four tumors without known genetic driver changes clustered among normal myometrium samples; one of these tumors harbors a somatic TP53 rearrangement created through massive chromothripsis rearrangement event. As myometrium samples showed uniform global DNA methylation by clustering clearly separate from majority of the tumors, statistically powerful comparison of UL subgroups with different genetic background to myometrium samples was enabled. Furthermore, integration of differentially methylated regions with significant expression changes allowed us to characterize genes which are epigenetically activated or silenced in different UL subgroups. All ULs displayed increased methylation of genes enriched in developmentally related biological processes. Our results provide strong evidence that the known UL subgroups are distinguishable by global DNA methylation profiling. The role of the identified epigenetically activated or silenced genes should be studied further to allow better understanding of the biological processes related to UL development and clinical outcome. Citation Format: Eevi Kaasinen, Amjad Alkodsi, Miika Mehine, Hanna-Riikka Heinonen, Netta Mäkinen, Mervi Aavikko, Kati Kampjärvi, Minna Taipale, Pia Vahteristo, Rainer Lehtonen, Lauri A. Aaltonen. Genome-scale DNA methylation changes delineate uterine leiomyoma subgroups. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 4435.