PO-200 3D In vitro cultures of PDX-derived tumour for anti-cancer drug discovery

Introduction Patient-derived xenograft (PDX) models emulate tumour complexity and heterogeneity, and reflect the variation in pathologic and genetic diversity within the population. As such, these models are a valuable source in preclinical drug development. We employ hydrogel embedding as a 3D culture technique to generate short term cultures of dissociated PDX material. This miniaturised setup allows economical screening of drug libraries, testing of drug combinations and pre-selection of relevant models for in vivo follow up. Material and methods In a high throughput approach, small amounts of tumour material are seeded in 384 well plates. Following exposure to multiple concentrations of standard-of-care chemotherapeutics, small molecules, targeted therapies, antibodies and antibody-drug-conjugate, the cultures are fixed and stained with cellular markers at the experimental endpoint. A 3D stack acquisition of each well is obtained, followed by high content image analysis using an in-house developed analysis platform, OminerTM. This enables measurement of relevant features such as tumour volume, tumour invasion, nucleus size and fraction of apoptotic cells; hence detailed tumour culture responses can be quantified, generating dose-dependent profiles for selected features. Results and discussions We successfully developed >50 cultures derived from breast, stomach, pancreatic, colon, bladder and lung cancer. Each model has unique growth characteristics and morphologic differences are prominent both between the different tissues of origin and similar pathologies. Histologic characterisation of the 3D tumour cultures show hallmarks of the original tumour, such as retention of relevant biomarkers. We present drug response data for various cancer indications, including multiple lung and breast cancer pathologies. Conclusion We demonstrate phenotypic variation amongst different PDX models and differential responses to anti-tumour therapies. This method facilitates the comparison and in depth interrogation of responses to both established, biosimilar and novel cancer drugs - and allows follow up in the same model in vivo.