Abstract Medulloblastoma (MB) is a heterogeneous disease comprised of four molecular groups: Wingless (MB-WNT), sonic hedgehog (MB-SHH), Group 3 (MB-Grp3) and Group 4 (MB-Grp4). Relapse (rMB) after multimodal treatment (surgery, radiotherapy, chemotherapy) occurs in 30-40% of patients and is typically fatal. Despite this, progress understanding the biology of rMB is limited, resulting in a paucity of biology-driven trials in rMB. We have undertaken multi-omic characterisation of 120 rMBs and their diagnostic counterparts, revealing putative drivers of rMB. Their functional validation has potential to underpin biology-driven therapies for rMB. Putative driver genes (n 150) from our ‘omic discovery set with acquired/enriched alterations in rMB were assessed using RNAi screening.shRNAs (n = 5/gene) were introduced into high-risk MB cell-lines (n = 3) and radiotherapy applied (1-3Gy, weekly). Changes in hairpin abundance, relative to baseline/controls, were used to identify genes with roles in tumour fitness/survival. Cell cycle regulator, epigenetic modifier and transcriptional co-factor genes were among those implicated in medulloblastoma fitness. Clinical tractability of targeting these genes was used to prioritise individual rMB-driver validation. Nuclear exporter XPO1 was prioritised as a candidate rMB driver; mutations occur in MB-SHH and are maintained at relapse. The role for XPO1 in medulloblastoma growth/survival was validated using MB XPO1 knock-down models. Importantly, XPO1 inhibitors such as selinexor are available and in clinical use. While established in the adult setting for refractory-myeloma, selinexor is also being trialled in paediatric malignancies including gliomas. We show selinexor enhances radiotherapeutic effects in vitro concordant with its significance under radiotherapeutic pressure in RNAi screens. Ongoing work will assess selinexor in in vivo models of high-risk and rMB to understand whether selinexor has potential to prevent rMB. In summary, we demonstrate that focused screening can identify and prioritise targetable rMB-drivers functionally relevant to disease biology, for pre-clinical assessment with a view to future clinical exploitation.
Beckett et al. (Tue,) studied this question.