Abstract 5015: Innovative and predictive models against breast cancer

Abstract Many reports support that xenografts from patient-derived xenografts (PDXs) in mice or rats well recapitulate the molecular diversity, cellular heterogeneity, and histology seen in patient tumors. However, shortcomings including limited clinical diversity of the PDXs, absence of influence of human microbiome, absence of drug human metabolism and reduced immune system restrain the predictive values of these PDX models. To set-up a holistic integration of these criticisms, we have associated efforts from public hospitals, academic groups, biotechs and private pharmaceutical companies with the financial support of the French Ministry of Industry. Patient surgical specimens from different breast tumor subtypes, among 9 other cancer pathologies, are collected to establish large collections of PDXs in mice. In addition, in vitro primary cultures of cells from these samples are conducted to establish a collection of cell lines from the stromal and the tumoral compartments. The 40 established breast PDX models represent the 4 major molecular clinical subtypes among which 37% are triple negative, 33% are Luminal B HER2-, 22% are Luminal B Her2+ and 8% Her2+. The models are evaluated for ex vivo and in vivo sensitivities to 4 clinically relevant anticancer drugs, for histology, immune infiltrates, and molecular characteristics including the analysis of gene polymorphism, RNAseq gene expression and NGS exome sequencing of 112 genes. The metagenomic sequencing of the gut microbiota from stools collected in both patients and tumor-bearing mice before and after chemotherapeutic treatments improves our knowledge on the role of microbiota in cancer progression and treatment. Secondly, to increase the predictability of the models, we are generating mice with humanized liver showing distinct drug pharmacokinetic profiles as compared with parental mice. Finally, the humanization of the immune system in mice is developed by several approaches including the use of induced pluripotent stem cells from cancer patients. These tumor collection and model characterization were performed under harmonized procedures within the consortium, allowing high quality material and results. All model characteristics are being compiled in a web-based database for efficient features search and data mining. We will present the first characterized breast cancer models and will discuss their usefulness and chance to bring benefit to patients via this holistic strategy developed within the IMODI initiative. Citation Format: Olivier Duchamp, Severine Tabone-Eglinger, Caroline Mignard, Isabelle Goddard, Alain Bruno, Maia Chanrion, Loreley Calvet, Olivier Degoul, Kevin Dhondt, Liliane Goetsch, Julie Miralves, Christophe Lautrette, Cindy Pensec, Françoise Le Vacon, Isabelle Treilleux, Laurent Arnould, Philippe Vaglio, Grégoire Prevost, Olivier Tredan. Innovative and predictive models against breast cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 5015. doi:10.1158/1538-7445.AM2017-5015