The significance of clinico-pathological and molecular sub-groups in Malignant Rhabdoid Tumours

Abstract

Introduction Malignant Rhabdoid Tumours (MRT) are aggressive early childhood tumours characterised by biallelic inactivation of SMARCB1. Having the potential to arise in an array of distinct tissues (CNS-located atypical teratoid rhabdoid tumours, ATRT; extra-cranial rhabdoid tumours, ECRT) they are often treated as distinct entities therapeutically and in clinical/biological studies although emerging sub-groups of MRT have provided new understanding of the disease heterogeneity. Lack of consensus on sub-group number and biology can be seen as a hurdle to future studies. Methods Gene expression and methylation array profiling of primary MRT was performed on clinico-pathologically annotated tumour profiles from UK cancer centres and combined with published MRT data in a meta-analysis. To characterise the common biological features of MRT, regardless of location, differential expression, methylation, gene and pathway analyses were compared to other paediatric embryonal tumour expression and methylation profiles (i.e. Medulloblastoma, Ewings Sarcoma, Rhabdomyosarcoma, Wilms tumour and Neuroblastoma). Survival analysis was carried out on UK MRT samples to identify novel DNA methylation markers associated with disease outcome. Based on evidence suggesting immune system involvement in at least one MRT subgroup, “methylCibersort” a novel in-silico method was used to estimate immune cell infiltration in a large cohort of pan-CNS tumours including MRT. Results Clustering all MRT together recapitulates the subgroups observed in ATRT alone; broadly overlapping with recently published ATRT and ECRT subgroup models. A putative expanded subgrouping model encompassing all MRT highlights additional heterogeneity and defines novel subgroup characteristics. Subgroup differences were shown to better explain differences in MRT biology than tumour location alone. Survival analysis identified a number of novel survival associations with DNA methylation state. Immune infiltration estimation using methylCibersort identified differences in immune interactions across a large dataset of different CNS tumours, and presented novel prognostic feature. Conclusion MRT is a complex disease owing both to the rarity of the tumour, resulting in lack of comprehensive genomic profiling, and heterogeneity observed in the tumour biology. This thesis presents evidence to support the definition of MRT as a related tumour type with differences arising due to disease subgroups. In addition, a meta-analysis comparing published subgrouping schemes seeks to direct future research by providing a subgroup consensus encompassing all MRT. Novel survival associations and immune infiltration estimates provide new avenues for further research.