Adenovirus Mart-1–engineered Autologous Dendritic Cell Vaccine for Metastatic Melanoma

Dendritic cells (DCs) are the key physiologic stimulators of naive and primed cells to an antigen.1,2 Many types of myeloid DCs have been tested clinically with objective clinical responses, both partial and complete. Many current immunotherapy efforts use myeloid DCs pulsed with synthetic T-cell epitopes.3–6 Such strategies are limited by identification of specific human leukocyte antigen (HLA)-restricted epitopes, binding strength, and/or kinetics of peptide binding, as well as the paucity of well-characterized major histocompatibility complex (MHC) class II-restricted epitopes. DCs that have been genetically engineered to express a full-length tumor antigen can stimulate T cells to a variety of class I and class II epitopes.7–9 This provides for continuous synthesis of the transgene10,11 and a polyclonal CD8 and CD4 T-cell response, which may lead to stronger, broader T-cell activation. MART-1/Melan-A (MART-1) is an immunogenic melanosomal-derived self antigen expressed in normal and malignant pigmented cells, including >90% of melanomas.12,13 The proximal MART-1 promoter was characterized by our group.14 For the most common MHC class I HLA type HLA-A*0201 (A2.1), the immunodominant peptide (MART-127-35) has been identified and extensively studied, although other peptides, including MART-153-62 may also be processed and presented.15–17 This immunodominant MART-127-35 peptide has been shown to bind other HLA-A subtypes of the A2 superfamily (A2.2, 2.4, 2.5, 2.6, 2.9, and A69.1, and also HLA-B35 and −B45), hence patients who are not HLA-A2.1+ may be able to mount CD8+ T-cell responses to this antigen.17–20 MHC class II-restricted DR421 and DQ522 DR11 and DR52 epitopes have also been described.23 We reported the results of a phase 1 dose-escalation clinical trial testing vaccination with autologous MART-127-35 peptide-pulsed DC.4 The goal of this study was to determine the most favorable dose and route of peptide-loaded DC administration. Immunization with the single MART-127-35 peptide pulsed onto granulocyte macrophage-colony stimulating factor (GM-CSF)/interleukin (IL)-4–differentiated DCs resulted in peptide-specific immune activation at all the doses and both the routes were tested, although 107 DCs administered intradermally (ID) seemed to be superior for both the immune and clinical activity. Of the 18 patients enrolled (10 with measurable disease), there was 1 complete response (CR) and 2 disease stabilizations (SD). We observed that the frequency of circulating MART-127-35–specific CD8+ T cells did not correlate with the clinical outcome. However, vaccine-induced acquisition of CD8+ and CD4+ T-cell responses to antigens not encoded by the vaccine but expressed by the tumor, an immunologic phenomenon termed determinant spreading, was observed exclusively in the patient with a CR. A subsequent phase 2 study of MART-127-35 peptide-pulsed DC, in which 10 stage II to IV melanoma patients received the same MART-127-35 peptide-pulsed autologous DC vaccine (ID, 107 DC/injection), produced a CR in just 1 patient with immunologic findings consistent with determinant spreading to tyrosinase.24 As in the earlier phase 1 trial, most of the other 9 subjects were successfully immunized against MART-1, but did not have evidence of determinant spreading. Taken together, these phases 1 and 2 data provide supporting evidence that the immunologic phenomenon of determinant spreading may play an important role in clinical response to immunotherapy. We have studied adenovirus (AdV)-mediated engineering of DC with full-length tumor antigens for CD8+ and CD4+ T-cell activation.7,10,25,26 The benefits of this strategy include antigen-specific activation of both CD8+ cytotoxic and CD4+ helper T cells, AdV-induced intermediate level maturation of DCs27 (Vujanovic, ’08 submitted), long-term antigen expression by the DCs (days and weeks instead of hours for peptides28), potent activation of CD8+ T cells specific for immunodominant and subdominant epitopes from a tumor antigen,9 and broad activation of tumor antigen-specific CD4+ T cells which produce the Th1 cytokines interferon (IFN)-γ and IL-2.29 Together, these data indicate that adenovirally-transduced DC (AdV/DC) are a potent vehicle for T-cell activation. We have also conducted studies in murine melanoma models, testing this strategy and defining the mechanism.26,30–33 Recent literature has identified crosstalk between DCs and natural killer (NK) cells as important for shaping the immune response. DCs, which take up pathogenic material and become activated, can activate NK cells. Conversely, NK cells can improve DC function through both cytokine secretion and cell-cell contact.34–37 In murine models testing E1-deleted adenovirus encoding the cDNA for MART-1/Melan-A (AdVMART1)-transduced DC vaccines, we have identified a critical role for NK cells.38 NK function has not yet been studied in human antigen-engineered DC clinical trials. Herein we report the first clinical trial with an antigen engineered DC-based vaccine. This antigen-engineered DC vaccine was able to activate both CD8+ and CD4+ T cells specific for the MART-1 melanoma antigen in a subset of patients.