Metabolic footprinting of response to treatment in ovarian cancer ex vivo models Rita Mendes1,2,*, Gonçalo Graça3, Fernanda Silva4, Teresa F Mendes1,2, Ana CL Guerreiro1,2, Patrícia Gomes-Alves1,2, Ana Félix4,5, Catarina Brito1,2, Inês A Isidro1,2 1IBET, Instituto de Biologia Experimental e Tecnológica, Oeiras, Portugal; 2Instituto de Tecnologia Química e Biológica António Xavier, Universidade Nova de Lisboa, Oeiras, Portugal; 3Department of Metabolism, Digestion and Reproduction, Imperial College London, UK 4CEDOC-FCM-NOVA, Portugal; 5IPOLFG, Portugal *firstname.lastname@example.org Abstract: Diagnosing chemoresistance and predicting patient outcome is still a major challenge in oncology, namely in ovarian carcinoma (OvC). Tumor progression and the sensitivity of cancer cells to drug treatments are deeply influenced by the complex interactions that occur within the tumor microenvironment, including metabolic crosstalk. We have recently developed a long-term OvC patient-derived explant (PDE) ex vivo model, which retains features of the original tumor microenvironment. Here, our goal is to use this model to determine drug efficacy by uncovering metabolic signatures of treatment responses. OvC-PDEs were challenged with the standard-of-care drug combination of carboplatin and paclitaxel, or the single agents, weekly, over two weeks. Cell death was longitudinally evaluated by lactate dehydrogenase leakage assay and culture supernatants were analysed by untargeted LC-MS-based metabolic footprinting. Multivariate data analysis was used to explore metabolic patterns in response to different treatments. Carboplatin induced the lowest cytotoxic response, followed by paclitaxel and the combination. The latter presents the highest variability among cases. Analysis of metabolic footprints by principal component analysis revealed different treatment response groups, supporting the existence of characteristic metabolic signatures. Discriminant features between drug treatment groups were ranked by partial least squares discriminant analysis for subsequent identification of corresponding metabolites. Overall, this study provides a proof-of-concept for drug response evaluation by metabolic footprinting, paving the way to uncover potential biomarkers of drug response and resistance.