DIPG-25. GENETIC ALTERATIONS TARGETING THE MAPK PATHWAY CONFERS PRECLINICAL SENSITIVITY TO TRAMETINIB IN A CO-CLINICAL TRIAL IN DIPG Free

The survival of children with DIPG remains dismal, and new treatments are desperately needed. The development of patient-specific in vitro and in vivo models represents one such opportunity, however the time taken to establish such models and the rapid disease progression has been thought to limit the utility of such an approach. We sought to explore a co-clinical trial model for DIPG patients enrolled in an ongoing biopsy-stratified study in order to identify rational therapeutic options with individualised preclinical evidence as to their efficacy. To date we have established novel patient-derived in vitro cultures from biopsy specimens of 11 patients, in both 2D (laminin matrix) and 3D (neurosphere) conditions, as well as orthotopic xenografts in vivo, with a high concordance in their molecular profile compared to the original tumour specimen (methylation BeadArray, exome, RNA sequencing). Cells were screened against a series of common and bespoke FDA-approved drugs based upon previous evidence in DIPG and/or the specific molecular alterations found in the patient sample. We identified a high degree of in vitro sensitivity to the MEK inhibitor trametinib (GI50 23-312nM) in samples which harboured genetic alterations targeting the MAPK pathway, specifically the non-canonical BRAF_G469V mutation and those affecting PIK3R1(N564D, H450E_VF_ins), with assessment of tumour volume in vivo by MRI. Allelic imbalance of PIK3R1_N564D by stochastic selection was observed in two independent cultures from the same patient showing a greater response to trametinib (6.4-fold) as well as a decreased sensitivity to dasatinib in the mutant-enriched culture (38-fold). RNAseq of the cultures revealed differential gene expression associated with hypoxia and interferon signalling linked to drug sensitivity. These data show the feasibility in generating patient-specific, testable hypotheses that may be clinically translated in a subset of patients, and we are currently exploring parallel resistance modelling to further inform novel treatment strategies at tumour progression.