Atypical teratoid/rhabdoid tumors (ATRTs) with SMARCA4 mutation are molecularly distinct from SMARCB1-deficient cases

Abstract

Atypical teratoid/rhabdoid tumors (ATRTs) are very aggressive childhood malignancies of the central nervous system. The underlying genetic cause are inactivating bi-allelic mutations in SMARCB1 or (rarely) in SMARCA4. ATRT-SMARCA4 have been associated with a higher frequency of germline mutations, younger age, and an inferior prognosis in comparison to SMARCB1 mutated cases. Based on their DNA methylation profiles and transcriptomics, SMARCB1 mutated ATRTs have been divided into three distinct molecular subgroups: ATRT-TYR, ATRT-SHH, and ATRT-MYC. These subgroups differ in terms of age at diagnosis, tumor location, type of SMARCB1 alterations, and overall survival. ATRT-SMARCA4 are, however, less well understood, and it remains unknown, whether they belong to one of the described ATRT subgroups. Here, we examined 14 ATRT-SMARCA4 by global DNA methylation analyses. We show that they form a separate group segregating from SMARCB1 mutated ATRTs and from other SMARCA4-deficient tumors like small cell carcinoma of the ovary, hypercalcemic type (SCCOHT) or SMARCA4 mutated extra-cranial malignant rhabdoid tumors. In contrast, medulloblastoma (MB) samples with heterozygous SMARCA4 mutations do not group separately, but with established MB subgroups. RNA sequencing of ATRT-SMARCA4 confirmed the clustering results based on DNA methylation profiling and displayed an absence of typical signature genes upregulated in SMARCB1 deleted ATRT. In summary, our results suggest that, in line with previous clinical observations, ATRT-SMARCA4 should be regarded as a distinct molecular subgroup.

SMARCA4‐deficient rhabdoid tumours show intermediate molecular features between SMARCB1‐deficient rhabdoid tumours and small cell carcinomas of the ovary, hypercalcaemic type

“…Taken together, the above results, using different sample sets and statistical methods, confirm observations from a recent study [22] and suggest that ECRT SMARCA4 have distinct DNA methylation profiles from ECRT SMARCB1 and are more closely related to SCCOHT than to ECRT SMARCB1 .…”

“… * References: (a) Holdhof et al , 2021 [22]; (b) Saunders et al , 2020 [23]; (c) Schneppenheim et al , 2010 [15]; (d) Chun et al , 2016 [30]. …”

“…While SMARCB1 inactivation is a well‐described, recurrent feature of ECRT and ATRT, the alternative SMARCA4 inactivation is only seen in a subset of ATRTs [18,22] and is exceedingly rare in ECRTs. To our knowledge, the cases presented herein with their clinical, pathological, and/or molecular data represent the largest series of ECRT SMARCA4 reported to date.…”

“…Clinical data for the five cases were provided by local physicians. Clinical data of these five ECRT SMARCA4 were pooled with data from seven previously published ECRT SMARCA4 cases with confirmed SMARCA4 genomic alterations [15,22,30]. Details about the origin, tumour type, and analysis performed on each sample are provided in supplementary material, Tables S1 and S2.…”

“…In 2010, Schneppenheim et al reported a germline truncating pathogenic variant in SMARCA4 in two siblings affected by ATRT and ECRT, respectively [15]. Including that initial description, at least 20 SMARCA4‐deficient ATRTs have since been described, but only eight cases of SMARCA4‐deficient ECRT (ECRT SMARCA4 ) have been published, with SMARCA4 mutational status only reported for five of those cases [15–23]. Thus, a systematic description of the clinical and molecular characteristics of ECRT SMARCA4 is lacking.…”