ABSTRACT
Malignant rhabdoid tumors (MRTs) are extremely rare and highly aggressive pediatric cancers classically defined by biallelic loss of the SMARCB1 gene, with rare involvement of SMARCA4. However, the molecular mechanisms leading to this loss are not yet fully understood. MRTs occur predominantly in infants, with the highest incidence in children under one year of age. Clinically, they are characterized by early metastatic dissemination and dismal outcomes, with 5-year event-free survival rates below 20%. There are currently no curative therapies for these patients.
Here, we performed integrated genomic, transcriptomic, and epigenomic profiling of 16 patient-derived MRT models including intracranial, renal, and soft tissue origins. While SMARCB1 deficiency was ubiquitous, we observed substantial heterogeneity in the mechanisms driving its inactivation. Only two tumors harbored detectable coding single-nucleotide variants in SMARCB1; the predominant mechanisms involved large-scale deletions and broad loss-of-heterozygosity (LOH) on chromosome 22, with extensive LOH in tumors lacking point mutations or focal deletions, consistent with allelic loss as a frequent “second hit.” In contrast, SMARCA4 remained intact across all models, reinforcing the mutual exclusivity of SMARCB1 and SMARCA4 alterations. Structural analyses revealed extensive variation, including more than 400 events per tumor on average and candidate gene fusions such as AHI1:MYB, whereas alterations in TP53 and BRCA1/2 genes were infrequent. Transcriptomic and epigenomic profiling showed heterogeneity driven by tissue of origin, disease progression, and therapeutic response, with subtype-specific programs and epigenetic modulation of DNA repair and immune-related genes (SLFN11, MGMT, LIF) linked to treatment sensitivity.
Collectively, our findings refine the molecular definition of MRTs, showing that while SMARCB1 loss remains the foundational driver, tumor behavior is further shaped by structural variation, impaired DNA repair pathways, and dynamic epigenetic landscapes. These integrated changes contribute to tumor heterogeneity, progression, and differential therapeutic vulnerabilities. Beyond advancing mechanistic understanding and identifying candidate biomarkers for patient stratification, our multi-omics dataset represents a valuable resource for the research community, supporting future studies and efforts to improve clinical management of this highly aggressive pediatric malignancy.