Mechanisms of resistance to bispecific T-cell engager therapy for ovarian cancer

5571 Background: Patients with recurrent ovarian cancer (OC) invariably develop multidrug resistance and have poor prognoses. Bispecific T-cell engager (BiTE) immunotherapy represents an innovative therapeutic modality where cytotoxic T-cells are activated and recruited to tumor cells using a predefined tumor-associated antigen. Clinical trials evaluating BiTEs in OC are currently in progress with promising results, and exploration of potential mechanisms of resistance to BiTEs could provide meaningful insight into disease biology and combination strategies. Methods: After obtaining informed consent, experiments were performed on patient samples collected under an IRB-approved tissue-banking protocol. Samples were collected before patients started clinical trials and at disease progression. Serum, peripheral blood mononuclear cells (PBMCs), and malignant ascites were collected when available for translational analysis. Immune cell subsets were evaluated for activation and exhaustion via flow cytometry, and tumor cells were subjected to western blot analysis and IHC. Ex vivo cytotoxicity assays were performed using BITEs, patient tumor samples, OC cell lines (immortalized), and patient-derived and healthy donor PBMCs. Patient serum samples were analyzed by multiplex ELISA to determine cytokine and chemokine signatures. Results: Serum and PBMC samples from 10 patients and ascites from 2 patients were used for these analyses. Cells derived from patient ascites after progression on BiTEs were primarily comprised of tumor cells and myeloid-derived suppressor cells (MDSCs). These tumor cells had lost expression of the target antigen recognized by the BITE and upregulated markers as N-cadherin and vimentin with concordant downregulation of E-cadherin consistent with epithelial to mesenchymal transition (EMT). Ascites immune subset analysis showed a predominance of CD11B+, CD163+, and PD-L1+ cells. Compared to PBMCs from healthy donors, patient PBMCs showed higher frequencies of CD4+ TIM3+ LAG3+ cells consistent with T-cell dysfunction. Compared with healthy donors, serum from patients revealed higher levels of myeloid recruiting and activating chemokines and cytokines (CCL3, CXCR3, CXCL10, and IL18) and low immunostimulatory cytokines (IL-7 and IFN-α). Coculture of patient tumor cells with BiTEs and T-cells ex vivo showed diminished cytotoxicity, consistent with the loss of target antigen expression. Patient PBMCs retained cytotoxicity against OC cell lines ex vivo, suggesting preserved cytotoxic capacity. Finally, T-cells cocultured in the presence of patient serum displayed suppressed cytotoxicity against OC cell lines. Conclusions: Resistance to BiTE therapy is multifactorial, primarily involving tumor-associated antigen loss, EMT, and an immunosuppressive tumor microenvironment. Combination treatment approaches to mitigate these factors might increase response rates.