Abstract PL08-03: Ovarian PDX models: Of Mice and (Wo)men…and personalized treatment

Abstract

Abstract Ovarian cancer is the most lethal gynecological malignancy. Multiple factors contribute to this poor outcome, including late stage of diagnosis, propensity of recurrence following initial remission, and poor responsiveness to chemotherapy following platinum-based therapies. Another major contributing factor is the heterogeneous nature of ovarian cancer, which is indeed multiple tumor type with varying responsiveness to chemotherapy. With no clear biomarkers, including mutational status to select initial chemotherapy options, treatment decisions are often empiric and not individualized. In addition, models for the development of novel therapies in ovarian cancer are poor and do not typically recapitulate the complex or individualized nature of this disease in our patients. To overcome this, we initiated the Mayo Avatar Program to develop treatment nave, orthotopic, intraperitoneal, patient-derived xenografts from all consenting patients undergoing primary debulking at the Mayo Clinic. We have now demonstrated the ability to engraft ovarian tumors at a high rate (∼75%), with the resultant PDX tumors demonstrating similarities to the source patient tumors histologically and molecularly. We have also demonstrated that the Avatar models are able to accurately predict responsiveness, or lack thereof, to platinum-based chemotherapy. We have now generated over 350 individualized models in all major histological subtypes to begin to understand determinant of sensitivity to standard cytotoxic chemotherapy, as well as novel targeted therapy. Using HRD analysis, we previously demonstated the usefulness of the ovarian PDXs in predicting responsiveness to the PARP inhibitor, niraparib (Haluska et al, ENA, 2014). This work demonstrated that PDXs that respond to niraparib are in the group of Avatars that have a HRD score greater than 42. Similarly, we demonstrated that Avatars that had expression of pregnancy-associated plasma protein-A (PAPP-A), benefitted from a therapeutic antibody targeting PAPP-A (Becker et al, Mol Cancer Ther 2015). Based on the ability of the Ovarian Avatars to predict responsiveness to chemotherapy retrospectively, a prospective clinical trial is now ongoing to determine if ovarian PDXs from patients can improve the our ability to individualize chemotherapy in patients with recurrent, platinum resistant ovarian cancer. Challenges continue to exist in the development and use of ovarian PDX modeling. These include, but are not limited to, protracted engraftment times and cost. In addition, the current state-of-the art for PDX modeling includes immunosuppressed hosts, such as NSG mice. This is a substantial barrier, as both innate and adaptive immunity have an impact on chemotherapy responsiveness. Additionally, immunotargeting agents, including checkpoint inhibitors, cell-based immunotherapy strategies and vaccines, have demonstrated promising early results, but cannot be optimally studies in immunosuppressed mice. The development of novel transgenic mouse strains expressing human cytokines to support the development and maintenance of the full human immune repertoire and xenografted tumors will hopeful begin to overcome these barriers. Citation Format: Paul Haluska. Ovarian PDX models: Of Mice and (Wo)men…and personalized treatment. [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference: Molecular Targets and Cancer Therapeutics; 2015 Nov 5-9; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2015;14(12 Suppl 2):Abstract nr PL08-03.