PREDICT-BC

Background. Biliary tract cancers (BTC) are aggressive malignancies in which systemic therapy offers only a modest survival advantage. The discovery of a long tail of driver genetic alterations (GAs) has pushed the field to evolve towards precision medicine. Among highly recurrent GAs, much interest was drawn initially by FGFR2 fusions, IDH1 and BRAF mutations and ERBB2 amplifications, because of the immediate availability of genotype matched single agent therapeutics (GMSAT). Recently, the clinical development of KRAS-targeted drugs and a PRMT5 inhibitor, active against a metabolic vulnerability linked to MTAP loss, have coopted the highly prevalent KRAS mutations and bi-allelic MTAP inactivation in the list of actionable GAs. However, GMSAT show limited efficacy, due to heterogeneous response rates (primary resistance) and lack of prolonged benefit in responding patients (secondary resistance).
Hypothesis. The BTC field is confronted with lack of pre-clinical models that recapitulate the complexity of genetics/biology inherent to BTC and the heterogeneity of clinical responses to GMSAT. Overcoming this bottleneck is key to developing translational investigations on combination therapies able to increase GMSAT efficacy.
Aims/Methods. A biobank of molecularly and clinically annotated patient-derived xenografts (PDX) (˜60) and organoids (PDO) (˜40) is already available to our consortium. We will expand this biobank through the involvement of the three participating clinical sites, predictably adding ˜ 30 pre-clinical models per year. Focusing on FGFR2, ERBB2, IDH1, BRAF, KRAS and MTAP GAs (incidence > 50% of BTC patients), we will define responses to GMSAT, which will be aligned to those of donor patients. We will subsequently validate the biobank suitability for translational studies. Selected PDO models with primary resistance to GMSAT will be assayed in unbiased fully automated drug library screenings, to capture drugs capable of combating resistance when co-administered with GMSAT. A second approach will focus on models that exhibit robust but short-lived responses to GMSAT, in absence of emerging resistance mutations. Here, through single-cell profiling of the transcriptome and chromatin accessibility landscape, we will define tumour cell states associated to the minimal residual disease (MRD) and nominate actionable vulnerabilities inherent to MRD. Candidate drugs emerging from the outlined analyses will be tested in combination with GMSAT. Further studies will address mechanisms underpinning the superior activity of newly defined drug combinations and interrogate the potential impact on immune-tumor interaction in a co-culture model.
Expected results/Impact. Our work will generate an unprecedented biobank of patient-derived pre-clinical BTC models. These models will be validated in translationally focused pre-clinical studies designed to identify drugs capable of significantly increasing the therapeutic benefit of co-administered GMSAT.