Immune Monitoring of Vaccine Quality and Persistence of Specific T Cell Responses in Five AML Patients Receiving Extended Dendritic Cell Vaccination Under Compassionate Use

High rates of early disease relapse are observed in patients with acute myeloid leukemia (AML). Due to poor health status, many patients cannot tolerate intensive chemotherapy and/or stem cell transplantation, resulting in a high unmet medical need for new therapy options. Clinical benefit of SCT is associated with immune responses that can control residual leukemia. To replace SCT in non-eligible patients, we designed an autologous dendritic cell (DC) vaccine approach that is given 4 times weekly at the beginning (immunization phase) followed by booster vaccines at week six and then monthly, with the intention to induce immune responses that delay or prevent relapse. The DCs secrete bioactive IL-12, but not IL-10, allowing activation of innate and adaptive responses. T cell responses are directed to two target antigens expressed in AML: Wilm's tumor-1 (WT-1) and preferentially expressed in melanoma (PRAME). Prior to initiation of a Phase I/II study implementing this vaccine approach (EudraCT No.: 2014-003520-44; clinicaltrials.gov No.: NCT02405338), five patients were treated under compassionate use (CU) using DC vaccines prepared according to an approved GMP manufacturing protocol.

Here we report immune monitoring studies, exploring quality of vaccines and specificity and persistence of T cell responses. DC vaccines produced in batch lots were cryopreserved in multiple aliquots for thaw prior to application. Extensive characterization showed DCs to be of mature phenotype, with high levels of positive co-stimulatory molecules and fewer negative regulatory molecules. Antigen-loading by electroporation of in vitro transcribed RNA, led to full length intracellular protein expression in most cells. Chemokine-directed migration was measured for all DC preparations. A novel dual-color ELISpot assay showed that DCs of each patient secreted IL-12 but not IL-10 after CD40 ligand stimulation. Despite extensive chemotherapy and impaired hematopoiesis, high quality DC vaccines with essential functions were easily generated from monocytes of these patients.

T cell responses to target antigens were assessed with two assays. Intracellular interferon-gamma production was measured in PBL acquired at various time points during treatment. Standard ELISpot assays independently confirmed T cell responses to both antigens. In addition, responses to non-immunizing antigens (hTERT and survivin) were detected in some cases, demonstrating antigen-spreading during treatment. T cell responses to one or both targets were found in 4/5 patients; thus, both self-antigens were immunogenic when presented by polarized mDCs. Persistent vaccination allowed maintenance of T cell responses over extended periods of time in 4/5 patients. Interestingly, antigen-specific responses to over-lapping peptides were only found in two patients when anti-PD-1 blocking antibody was present in the stimulation cultures, thereby enabling T cells to produce cytokine after antigen-specific stimulation.

To date 3/5 patients are alive without current signs of relapse. Two patients received DC vaccines for 22 and 30 months and have been followed for 53 and 41 months respectively, since completion of chemotherapy. The third patient had signs of relapse and vaccination was halted after 10 months. This individual was then given allogeneic SCT and remains in CR after 35 months. The two remaining patients are deceased. One patient who showed no T cell activity relapsed after 12 months of vaccination and died at month 20 during chemotherapy. The fifth patient was vaccinated for 24 months without signs of relapse but died due to unrelated heart disease. The ongoing clinical study using extended DC vaccination will provide more information whether persistent vaccination against these self-antigens contributes to immune responses that prevent disease relapse in AML patients.