Abstract
BACKGROUND: Atypical teratoid rhabdoid tumors (ATRTs) remain a significant clinical challenge in pediatric neurooncology. Despite the simplicity of the genomic profile, mostly characterized by the functional loss of SMARCB1 and in rare cases SMARCA4, ATRTs are molecularly and epigenetically diverse, with three main molecular subgroups identified: ATRT-TYR, ATRT-SHH and ATRT-MYC. METHODS: To address the lack of effective mechanism-of-action based treatments and comprehend the biology of these subgroups, we conducted 10X single-nucleus transcriptome (n = 12) and multiome (transcriptome + ATAC) (n = 7) sequencing on tumor samples (ATRT-TYR = 6, ATRT-SHH = 9, ATRT-MYC = 4). Validation datasets included 10X single-cell transcriptomics (n = 1), 10X single-nucleus transcriptomics of cell lines (n = 5), and single-cell (n = 3) and single-nucleus (n = 9) SMART-seq transcriptomics. Separation of malignant cells from tumor microenvironment (TME) cells was achieved by comparison of expression profiles against literature-based marker genesets of normal cell types. Tumor cell populations were characterized based on enrichment in fetal cell types and recurrent tumor programs across cancer types, coupled with non-negative matrix factorization (NMF) and differential expression analyses. RESULTS: ATRT-TYR exhibited choroid plexus-like and cilia-like cell populations, ATRT-SHH displayed radial glia-like, OPC-like and NPC-like cell populations and ATRT-MYC mainly demonstrated a mesenchymal-like cell population. Each subgroup represented unique differentiation trajectories. However, we identified a population of intermediate neuronal precursor-like cells (IPC-like) shared across all subgroups, expressing markers of pluripotent stem cells and cell cycle. RNA + ATAC transcription factor (TF) enrichment analysis for IPC-like cells revealed TFs linked to each subgroup’s differentiation trajectory, suggesting IPC-like cells as a putative ground-state tumor cell population driving differentiation of each subgroup. Functional validation of these findings in ATRT organoid models is currently ongoing. CONCLUSIONS: This study elucidates defined ATRT subgroup-specific differentiation trajectories that may provide guidance for the development of subgroup-based therapies.