ID #555 Small molecule drug screening, chemoinformatics and integrative transcriptomics identify targetable MYC-dependent therapeutic vulnerabilities in high-risk medulloblastoma
Louisa Taylor, Beth Poole, Dean Thompson, Gemma Llargués-Sistac, Janice Law, Michael Tseung, Melanie Beckett, Samantha Jepson-Gosling, Janet Lindsey, Alaide Marcavallo, Colin Kwok, Carlos Nery, Dan Astley, Harvey Che, Gareth Veal, Robert Wechsler-Reya, Laura Donovan, Rebecca Hill, Louis Chesler, John Anderson, Steven CliffordAbstract
Group 3 medulloblastomas with MYC-amplification (MYC-MBG3) do not respond to standard-of-care regimens and represent one of the worst prognosis childhood brain tumours. MYC is not directly targetable with small molecule inhibitors. The systematic identification of compounds which instead target complementary MYC-dependencies represents a promising approach to rationally re-design effective therapeutic strategies.
We derived three MYC-regulable cell avatars from cellular models of MYC-MBG3 (iD425, iD283, iHD-MB03). Transcriptomic analysis, alongside a primary patient tumour cohort (n = 321), identified a common MYC-dependent signature shared between patients and avatars, enriched for proliferative and DNA damage repair processes. Parallel high-throughput drug screens (>500 small molecule inhibitors) revealed differential MYC-dependent sensitivities, consistent across cell avatars, to 80 small molecules spanning 25 drug classes. Bespoke chemoinformatics (predictive compound selectivity, blood-brain-barrier penetration, clinical development stage) integrated with MYC-dependent transcriptomics, identified seven proceedable target dependencies, including CHEK1, AURKA and PLK1. Pharmacological and genetic validation short-listed therapeutic leads and associated small molecules, which were advanced to single agent efficacy trials in a spontaneous in vivo model of MYC-MBG3 (GTML). Inhibition of AURKA (with alisertib) and CHEK1 (with prexasertib) was found to significantly extend survival and reduce tumour burden. To anticipate and circumvent the development of cellular resistance, alisertib and prexasertib were next trialled in combination. In a GTML ex vivo 3D model, the alisertib/prexasertib combination demonstrated strong efficacy, additivity and striking MYC(N)-dependency. In a Myc-dependent in vivo allograft model (Myc-DNp53), alisertib/prexasertib combination therapy significantly increased survival. Pharmacokinetic studies confirmed alisertib/prexasertib central nervous system penetration.
Our findings demonstrate dual-AURKA/CHEK1 inhibition exerts a robust anti-tumour effect in MYC-MBG3 that warrants clinical evaluation. Importantly, we present an integrative screening pipeline which encompasses human tumour analysis and avatar-based screening to uncover actionable therapeutic vulnerabilities in MYC-MBG3, which is readily adaptable to other high-risk medulloblastoma disease features.