Dendritic cells (DC) are sentinels of the innate immune system that play an essential role in educating T lymphocytes how to respond to pathogen and tumor antigens. DC vaccines have induced cytotoxic T lymphocyte (CTL) responses against tumors experimentally and in human cancer patients. The rationale for using immunotherapy against glioblastoma multiforme (GBM), a destructive and lethal brain tumor, is the “seek and destroy” capacity of CTLs that can infiltrate the brain to kill tumor left behind after surgery. Tumor infiltrating T cells have been detected in GBM patients treated by DC vaccines, and extensively characterized by histological and molecular assays. Despite these encouraging results, immunotherapy is typically not curative for GBM patients.
The concentration of tumor infiltrating DC (TIDC) is positive prognostic factor for many cancers, yet their significance in brain tumors is poorly understood. Moreover, almost nothing is known about the temporal dynamics of DC trafficking into brain tumors, how that changes during immunotherapy, and the responsible mechanisms for inducing DC trafficking. Mounting evidence suggests that DC not only regulate anti-tumor T cell responses, but that some DC subsets can actually kill tumor directly. The central hypothesis of this grant is that the quantity of specific subsets of glioma-infiltrating DC significantly impacts the outcome of immunotherapy. This is a new hypothesis that may also be controversial. As such, testing this hypothesis will require unique animal models and in vivo imaging technologies in order to provide adequate supporting data.
EGFRvIII is a truncated, hyperactive version of the epidermal growth factor receptor (EGFR) that represents one of the few “glioma-specific” antigens expressed by tumor cells but not on normal tissue. We and others have targeted EGFRvIII using immunotherapy, and clinical trials are ongoing. Progress in this area of immunotherapy has been hindered by lack of a tractable animal model in which the tumor expresses the relevant antigen (e.g., EGFRvIII) but the host is tolerized to the antigen (as in human patients). We have recently developed a novel mouse model of glioblastoma that expresses human antigens using somatic cell gene transfer for evaluation of DC trafficking and immunotherapy.
The development of this glioma model is significant because it arises spontaneously and is more relevant than cell line-induced models that have failed to predict clinical outcomes. In this model, gliomas are induced that express renilla luciferase, and DCs can be adoptively transferred from firefly luciferase donors, thereby allowing both tumor and immune cells to be tracked by two-signal bioluminescent imaging in vivo. In Specific Aim 1, the trafficking of DC into glioma will be determined using two-signal bioluminescent imaging in mice treated by EGFRvIII vaccination. In addition, the trafficking of DC from the tumor into the draining lymph nodes will be determined and correlated to survival to define a quantitative relationship between trafficking and survival. These trafficking experiments will be corroborated by conducting a detailed analysis of TIDC using histology and flow cytometry. In a second set of experiments, chemokines IP-10, CCL21 and Flt3L will be overexpressed in the glioma using regulatable gene transfer technology. The conclusive role for these chemokines in DC trafficking will be shown using antibody-mediated interference.
The dogma in the DC field has been that “conventional” myeloid DC (mDC) are most important for priming anti-tumor immunity, and either T cells or natural killer (NK) cells are the effectors that directly kill tumor. Very recently, a novel subset of DC has been identified that is able to directly kill tumor cells, secrete high levels of interferon gamma (IFN-gamma), and also present antigen to naïve T cells. This novel subset of DC has been named “natural killer dendritic cells” (NKDCs). Several high-impact papers have demonstrated that NKDCs produce IFN-gamma and directly kill tumor to a greater extent than conventional NK cells. If these results are true, it significantly alters our understanding on the influence of DC in anti-tumor immunity. However, absolutely nothing is known about the behavior of NKDCs in the brain, or about their potential as a novel cellular therapy for glioma.
In Specific Aim 2, we will induce tumors in T cell-deficient mice, then adoptively transfer luciferase-labeled NKDC directly into the glioma and determine direct anti-tumor activity by bioluminescent imaging and survival. This experiment will also be done in immunocompetent mice to investigate if NKDC can prime adaptive anti-tumor immunity. While two-signal bioluminescent imaging is a powerful tool, this imaging may be difficult to translate to human studies. Therefore, we will also label NKDC with fluorescent quantum dots (Qdot). This approach has recently been used to locate the draining lymph nodes in large animals by in vivo imaging. Qdot-labeled NKDC will be infused directly into the tumor or intravenously. Trafficking will be measured by bioluminescent imaging as positive control, but then overlayed with Qdot fluorescence to detect draining lymph node involvement. The therapeutic effects of NKDC will be determined by measuring tumor growth and survival of tumor-bearing mice. Collectively, these experiments will provide new insights into the trafficking and anti-glioma activity of conventional DC and NKDC, with an eye on human clinical applications in the foreseeable future.