240 OncoTrack: Methods for Systematic Next Generation Oncology Biomarker Development

proteomic data on cancers from patient material provides hypotheses regarding potential novel drug targets. Laboratory platforms to validate whether target modulation would provide a clinical benefit are often highly reductionist, using long-established cell lines growing in 2 dimensions in vitro. These models do not reflect the complexity of a tumour in situ, where biochemical pathways are wired with connections to the complex tumour environment provided by stromal, immune and endothelial cells. PREDECT has the goal of comparing the pathological and molecular profiles of novel in vitro platforms with those of human tumours. In a stepwise manner, material from in situ tumours will be “deconstructed”, for example through tissue slices, 3-D heterogenously complex models (with stroma etc), simple 3D models, complex 2D models to simple 2D models and the profiles of each platform compared in the resting state and when perturbed by drugs or RNA interference. “Reconstruction” of complexity will also be attempted and profiled. Proteomics of models of breast, prostate and lung cancer will be compared in a Work Package based on a centralised TMA platform and Web-based microscopy and, together with transcriptional profiling, a systems approach will interrogate the fidelity of in vitro platforms to represent human tumors. Because clinical material presents both logistic and quality problems for ongoing and intense studies of target validation, PREDECT aims to use material from genetically engineered mouse models, and some “advanced” xenografts, with pathology and molecular profiles which closely match cohorts of human tumours. PREDECT hopes to provide more appropriate platforms both for target validation and subsequent preclinical studies which will replace a current cascade of tests which are poorly predictive of clinical activity.