Abstract Embryonal Tumors with Multilayered Rosettes (ETMR) are aggressive infant tumors with dismal survival due to lack of effective treatment options and nearly ubiquitous occurrence of relapse. Despite identification of chromosome 19 miRNA cluster (C19MC) amplifications as proposed driver in ∼90% of tumors, our comprehension of the distinct cellular populations that form an ETMR and their evolution over time is still limited. To identify transcriptionally and/or genetically distinct cell populations within ETMRs, we generated scRNA and scMethylation sequencing profiles from ten patient samples. Our initial studies identified that ETMRs consist of transcriptionally distinct populations that form rosette- and neuropil-like structures (manuscript in press). These populations lacked expression of C19MC within more differentiated neuropil-like cells while C19MC was highly expressed in rosette-like cells, suggesting a role in promoting stemness in ETMRs. To extend on this finding, we applied scMethylation to study the epigenetic and genetic substructure of the tumors, leveraging the capacity of the method to inform both. By generating high-resolution copy-number profiles for each individual cell within the tumors we could resolve genetic heterogeneity and infer lineage trees. While structural alterations involving C19MC were detected in every cell, multiple samples showed subclonal amplification of C19MC. In contrast, chromosome 2 gain, which occurs in ∼60% of all tumors, is present in every cell, suggesting a role in early tumorigenesis prior to C19MC amplification. By comparing paired diagnosis and relapse samples, we identified outgrowth of cells harboring C19MC amplification and in general more primitive phenotypes in relapsed ETMRs compared to cells at diagnosis. Taken together, we identified that while amplification of C19MC is a highly common event in ETMR, the regulation and distribution of the proposed driver is heterogeneous throughout the tumors. Expression of this event is associated with more primitive cell populations and likely provides a selective growth advantage upon relapse.
Lambo et al. (Fri,) studied this question.