Abstract A22: Feasibility and efficacy of a human-derived plasma 3D culture (huP3D) model to predict clinically effective drug treatment concentrations

Despite significant improvements in research and development in the cancer field, about 95% of oncology drugs in clinical trials fail to receive FDA approval. Compelling evidence suggests that three-dimensional (3D) cell culture techniques provide useful advantages for modeling drug sensitivity; however, there are still challenges, including vast differences in cell growth rates, inconsistent use of drug metrics across different cellular platforms, costly and relatively time-consuming approaches, and findings that may not translate to humans due to the incorporation of non-native environments. A biochemically characterized human blood plasma matrix is used as the framework upon which the human-derived plasma 3D culture (huP3D) exists. Breast cancer (BCa) cell lines representing different molecular subtypes (MCF7, ZR-75-1, MDA-MB-453, SK-BR-3, and MDA-MB-231) were screened for commonly used chemotherapies (capecitabine, carboplatin, paclitaxel, docetaxel, epirubicin, methotrexate, and cyclophosphamide) in 2D and huP3D models after 7 days of treatment by flow cytometry. Relevant drug metrics for half maximal concentrations of inhibition (IC50), response (EC50), proliferation inhibition (GI50), and growth rate inhibition (GR50) were calculated. Literature searches of phase I/II studies for the clinical effective drug concentration (steady state concentration, Css) of the same chemotherapeutic drugs on patients, and preclinical IC50 on 2D or other 3D models, were performed to determine the ability of 2D, other 3D, and huP3D cultures to predict clinical efficacy. BCa cell lines in huP3D were able to synthetize relevant tumor microenvironment components including extracellular matrix proteins (collagens, laminin, and fibronectin), extracellular vesicles, and cytokines. Moderate (paclitaxel, capecitabine, methotrexate, cyclophosphamide) to strong (epirubicin, docetaxel, carboplatin) effects on BCa cell survival were found in 2D cultures, in addition to a heterogeneous response to the chemotherapeutic drugs in the BCa cell lines tested, with ZR-75-1 being the most resistant to the different drugs. 2D effective drug concentrations did not correlate with the literature search Css on patients (R2=0.05). On the contrary, BCa cells cultured in huP3D showed increased drug resistance to all the chemotherapeutic drugs and a better correlation with the clinical data than 2D and other 3D models. In particular, the GR50 drug metric showed the strongest correlation with clinical Css values (R2=0.87). Our results showed the feasibility and efficacy of the huP3D drug response metrics to predict clinically effective therapies better than current preclinical models. Our model presents a reproducible, human-derived, high-throughput in vitro 3D tissue culture system for modeling drug sensitivity and clinical efficacy. A critical value of the model would be to focus on the development of more personalized models using cells and plasma components from the same patient. Citation Format: Kristin Calar, Simona Plesselova, Somshuvra Bhattacharya, Pilar de la Puente. Feasibility and efficacy of a human-derived plasma 3D culture (huP3D) model to predict clinically effective drug treatment concentrations [abstract]. In: Proceedings of the AACR Special Conference on the Evolving Landscape of Cancer Modeling; 2020 Mar 2-5; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2020;80(11 Suppl):Abstract nr A22.