AML Cell Vaccines Co-Expressing CD80 and IL-15/IL-15 Receptor Alpha Induce Activation and Cytolytic Activity in Post Remission Autologous Patient PBMC Ex Vivo

There is a critical need for more effective therapy for acute myelogenous leukemia (AML). Although many patients achieve remission, most relapse with poor outcomes. Even after allogeneic Stem Cell Transplantation (SCT), 30-50% of patients relapse due to the persistence of residual disease. To address the poor immunogenicity of AML cells and the diminished immune responsiveness of patients, our candidate autologous AML vaccine is lentivirally engineered, in each patient's leukemic cells, to express CD80, IL-15, and IL-15 Receptor alpha (IL-15Rα). In prior studies in a syngeneic 32Dp210 murine AML model, CD80-mediated co-stimulation of T-cells combined with immune activation by the IL-15/IL-15Rα heterodimer showed unprecedented synergy in induction of anti-leukemic cytolytic activity (Shi, Y. et al, 2018). This was observed in both ex vivo co-culture and in vivo where vaccinated leukemic mice had >80% cure rates. No local skin, organ, or systemic toxicity was observed, nor was there evidence of systemic cytokine release of IL-6 or TNFα after SC or IV injection of up to 10 8 transduced irradiated AML cells. We confirmed the feasibility of producing patient-derived AML vaccines by transduction of 16 independent AML samples with a tri-cistronic lentiviral vector (TLV) that contains human CD80, IL-15 and IL-15Rα. Transduction levels were 11-71% of cells (median 38.6%). To define the minimum transduction level required for PBMC activation and to assess synergy of co-expressed human CD80, IL-15, and IL-15Rα, allogenic U937 leukemia cells were initially used as stimulators. Transduced U937 (U937-TLV) had high-level surface expression of CD80 and IL-15, secreted IL-15 (7 ng/ml/24 hours from 2 x 10 6 cells/ml) and activated CD3+ T-cells from an AML patient (Fig.1). Mixtures of irradiated U937-TLV with non-transduced U937 were created at fixed ratios (100%, 80%, 40%, 20%, 10%, 5%, 0%) for overnight co-culture with patient PBMC. At 24 hours, the T-cells were analyzed for activation by measurement of the frequency of CD69+ CD4 or CD8 T cells (Fig. 1), normalized to expression of unstimulated PBMC (0%) and the percentage of maximal CD69 expression with 100% U937-TLV (100%). Background levels of activation due to the presence of allogenic U937 were negligible. Co-culture with as little as 10% transduced U937-TLV reliably activated patient T-cells. To assess the synergy of CD80, IL-15 and IL-15Rα expression, parallel experiments were performed with PBMC co-cultured in IL-15 containing supernatants from U937-TLV cells (Fig. 1). The frequencies of activated T-cells were significantly higher after co-culture with U937/U937-TLV cells than after stimulation with IL-15-containing supernatants from similar ratios of U937/U937-TLV, confirming the synergy of CD80 and IL-15/IL-15Rα in the transduced cells. To better, model the clinical setting, we assessed induction of immune responses of patient T cells to autologous transduced AML. PBMC were stimulated with transduced or non-transduced autologous AML cells vs stimulation with allogeneic U937-TLV, or with anti-CD3/CD28 beads to define maximal stimulation. Negative controls included culture of PBMC alone. All patients had T-cell activation, as measured by induction of CD69, HLA-DR and CD95 (Fas) expression, although there was heterogeneity in the nature of responses, e.g., disparate induction of the markers in individual patients (Fig. 2A and B). Induction of cytotoxic effector pathways was confirmed by detection of CD178 (FasL) and perforin expression (Figure 2C and D). Overall, all patients' PBMC had the capacity to mount T-cell responses of similar magnitude to both allogeneic U937-TLV and autologous vaccine. These studies establish that autologous AML cells transduced with CD80, IL-15 and IL-15Rα can elicit specific anti-leukemic T-cell responses, even in the face of prior lymphodepleting chemotherapy. A strength of this autologous vaccine strategy is that it is agnostic to which AML proteins are immunogenic for each patient. Although uniformly detected, there was heterogeneity in the induction of activation markers and effector pathways, which may reflect host and/or disease-related differences. The mechanisms underlying differences in the nature of responses in patients will be important to understand and will provide the basis for future immune correlative studies for our Phase 1 vaccine trial in transplant ineligible AML patients. Figure 1 Figure 1. Disclosures Kohn: Lyrik Therapeutics: Membership on an entity's Board of Directors or advisory committees; MyoGene Bio: Membership on an entity's Board of Directors or advisory committees; ImmunoVec: Membership on an entity's Board of Directors or advisory committees; Pluto Immunotherapeutics: Membership on an entity's Board of Directors or advisory committees; Allogene: Membership on an entity's Board of Directors or advisory committees; UC Regents: Patents & Royalties; Bluebird Bio: Membership on an entity's Board of Directors or advisory committees; Sangamo Biosciences: Membership on an entity's Board of Directors or advisory committees.