A platform of patient-derived microtumors identifies therapeutic vulnerabilities in ovarian cancer.

Abstract

BackgroundIn light of the frequent development of therapeutic resistance in cancer treatment, there is a strong need for personalized model systems representing patient tumor heterogeneity, while enabling parallel drug testing and identification of appropriate treatment responses in individual patients. Using ovarian cancer as a prime example of a heterogeneous tumor disease with complex microenvironment and high recurrence rates, we developed a 3D preclinical ovarian cancer model comprised of patient-derived microtumors (PDM) and autologous tumor-infiltrating lymphocytes (TILs) for identification of treatment vulnerabilities and validation of anti-cancer drug efficacy using immunohistochemistry, immune cell phenotyping, functional assays and protein profiling analyses. MethodsPDM and TILs were isolated from fresh primary ovarian cancer tissue specimen using mechanical disruption and limited enzymatic digestion and were subsequently cultured in suspension in defined media in the absence of serum. The heterogeneous cellular composition of isolated PDM as well as autologous TILs was analyzed by FFPE immunohistochemistry and multi-color flow cytometry, respectively. For in-depth protein profiling of PDM we established Reverse Phase Protein Array (RPPA) analyses of >110 total and phospho-proteins. Treatment efficacy in response to chemotherapeutics as well as immunotherapeutic compounds was assessed in PDM and PDM-TIL co-cultures using a functional viability assay in microplate format. ResultsThe enzymatic digestion of primary ovarian cancer tissue and suspension culture in defined serum-free media allowed fast and efficient recovery of patient-derived microtumors (PDM). Immunohistochemical analyses demonstrated histopathological comparability of ovarian cancer PDM with corresponding patient tumor tissue. Reverse Phase Protein Array (RPPA)-based analyses of >110 total and phospho-proteins enabled the identification of patient-specific sensitivities to standard, platinum-based therapy and thereby the prediction of potential treatment-responders. Finally, combining PDM and autologous TILs for individual efficacy testing of immune checkpoint inhibitor treatment demonstrated the potential for patient-specific enhancement of cytotoxic TIL activity by this therapeutic approach. Conclusion3D patient-derived ovarian cancer microtumors represent a preclinical, ex vivo tumor model that reflects intertumoral heterogeneity and represent the cellular complexity of individual patient tumors. Combining protein pathway analysis and anti-cancer drug efficacy testing of PDM enables drug mode-of-action analyses and therapeutic sensitivity prediction within a clinically relevant time frame after surgery. Follow-up studies in larger cohorts are currently under way to further evaluate the applicability of this platform to support clinical decision-making and personalizing cancer treatment.