Single-nucleus chromatin accessibility reveals intratumoral epigenetic heterogeneity in IDH1 mutant gliomas

• Main Authors: Ruslan Al-Ali, Katharina Bauer, Jong-Whi Park, Ruba Al Abdulla, Valentina Fermi, Andreas von Deimling, Christel Herold-Mende, Jan-Philipp Mallm, Carl Herrmann, Wolfgang Wick, Şevin Turcan
• Format: Article
• Language: English
• Published: BMC 2019-12-01
• Series: Acta Neuropathologica Communications
• Online Access: https://doi.org/10.1186/s40478-019-0851-y

Abstract The presence of genome-wide DNA hypermethylation is a hallmark of lower grade gliomas (LGG) with isocitrate dehydrogenase (IDH) mutations. Further molecular classification of IDH mutant gliomas is defined by the presence (IDHmut-codel) or absence (IDHmut-noncodel) of hemizygous codeletion of chromosome arms 1p and 19q. Despite the DNA hypermethylation seen in bulk tumors, intra-tumoral heterogeneity at the epigenetic level has not been thoroughly analyzed. To address this question, we performed the first epigenetic profiling of single cells in a cohort of 5 gliomas with IDH1 mutation using single nucleus Assay for Transposase-Accessible Chromatin with high-throughput sequencing (snATAC-seq). Using the Fluidigm HT IFC microfluidics platform, we generated chromatin accessibility maps from 336 individual nuclei, and identified variable promoter accessibility of non-coding RNAs in LGGs. Interestingly, local chromatin structures of several non-coding RNAs are significant factors that contribute to heterogeneity, and show increased promoter accessibility in IDHmut-noncodel samples. As an example for clinical significance of this result, we identify CYTOR as a poor prognosis factor in gliomas with IDH mutation. Open chromatin assay points to differential accessibility of non-coding RNAs as an important source of epigenetic heterogeneity within individual tumors and between molecular subgroups. Rare populations of nuclei that resemble either IDH mutant molecular group co-exist within IDHmut-noncodel and IDHmut-codel groups, and along with non-coding RNAs may be an important issue to consider for future studies, as they may help guide predict treatment response and relapse. A web-based explorer for the data is available at shiny.turcanlab.org.