TMOD-05. GENOME-WIDE DNA METHYLATION PROFILE: A POWERFUL STRATEGY TO RECAPITULATE HETEROGENEITY OF PEDIATRIC BRAIN TUMORS IN PRIMARY CELL LINES

BACKGROUND: Development of in vitro models of pediatric brain tumors (pBT) is instrumental for both understanding the contributing oncogenic molecular mechanisms and identifying and testing new therapeutic strategies. Primary cell lines should be established and managed to prevent epigenetic and genetic alterations and thus recapitulating the original tumor. DNA methylation (DM) is a stable epigenetic modification, altered in cancer and recently used to classify tumors. We aim to apply DM and Copy Number Variation (CNV) profiling to characterize pBT primary cell lines and tumors. METHODS: We investigated 34 pBT tissues from different histology paired to 52 their derived primary cultures in both 2D and 3D conditions, as stem-cells or in serum-supplemented medium, and both short and long-terms in culture. We studied 18 additional pBT-derived cell-lines, 9 organoids, 5 commercial cell-lines, and 122 pBT tissues from the same histological categories, as controls, for a total of 240 genome-wide DM profiles. We analyzed DM and CNV profiles by using Illumina EPIC-arrays. By means of a bump hunting strategy, we identified differentially methylated regions in faithful vs unfaithful cell lines, and performed a functional characterization using over-representation analysis. Results The 69% (25/36) of cells at early passages retained genetic alteration and the same DM patterns of the original tumors, with no differences related to 2D/3D methods or the presence of serum in media. The 70% (24/34) of primary cell lines analyzed at later passages (>5 or >14 days in culture) diverged from the primary tumor, the totality of those cultured with serum. All divergent cells clustered together acquiring common deregulated epigenetic signature induced by serum culture media, 2D methods and longer time in culture. CONCLUSIONS: We have shown that global DM profiles, along with CNV analysis are useful tools to detect the recapitulation of pBT-derived primary cell-lines from the original tumor. Whatever subgroups tested, our results suggest that in vitro models should be passaged as little as possible to retain the epigenetic and genetic alterations of the tumors and thus to be considered relevant for basic and translational biology.