UCLA

Glioblastoma (GBM) is the most common and lethal of all adult primary malignant brain tumors. For patients diagnosed with GBM, the median survival is 11-14 months despite advances in surgical resection techniques, chemotherapies, and radiation therapy (1). Alternate therapeutic strategies are being actively pursued to target GBM, with various immunotherapeutic modalities designed to generate an anti-GBM immune response showing considerable promise in preclinical models and clinical trials. To more effectively target GBM with these treatments, there has been an increasing appreciation of the numerous mechanisms involved in generating and maintaining the highly immunosuppressive tumor microenvironment in recent years. These studies have described a variety of microenvironmental and systemic factors that promote glioma cell evasion from the immune system. In light of these, it has become apparent that these factors must be understood and explicitly targeted to mount a successful immune response against GBM.

This thesis describes the utilization of two different immunotherapeutic strategies to target GBM. The first strategy created a novel GBM target by inducing NY-ESO-1 antigen expression with the demethylating agent, decitabine, and targeting that antigen with engineered T cells. Specifically, we utilized human GBM cell cultures to induce expression of the antigen. We evaluated NY-ESO-1 TCR-transduced T cell-mediated GBM tumor cytolysis in a series of in vitro cytotoxicity assays. Following this, we examined the application of this therapy using an intracranially-implanted xenograft model. Our studies demonstrated that decitabine could effectively upregulate NY-ESO-1 both in vivo and in vitro. Engineered T cells were able to induce tumor cytolysis in vitro and were able to traffic to and target tumor in vivo. Tumor-bearing mice receiving adoptive transfer of these engineered T cells demonstrated significantly increased survival over mice that received non-transduced T cells. By inducing expression of a novel target on GBM, we were able to generate a highly specific, anti-GBM immune response. This strategy represented a clinically translatable therapeutic technique for treating patients with GBM.

The second strategy focused on using existing GBM targets to generate an endogenous immune response in a syngeneic, immune competent mouse model. Briefly, we administered an autologous tumor lysate-pulsed dendritic cell (DC) vaccine to produce a glioma-specific immune response. In our studies, the vaccination appeared to be capable of inducing T cell infiltration into tumors; however, in large, established tumors, this infiltrating response was not sufficient to increase mouse survival and provide significant therapeutic benefit. We described the role of the negative costimulatory pathway, programmed death-1/ligand-1 (PD-1/PD-L1) in mitigating T cell activation and memory in a series of in vitro and in vivo studies. We noted that PD-1 blockade with PD-1 mAb was not sufficient to produce a T cell infiltrate. However, when administered with DC vaccination, PD-1 blockade activated the vaccine-generated T cell response in the tumor microenvironment. We found that T cells with PD-1 mAb were able to mediate significant tumor cytolysis when compared to T cells without PD-1 blockade in vitro. The adjuvant administration of PD-1 mAb with the DC vaccine resulted in significant survival benefit over DC vaccine alone in mice bearing large, established gliomas. Additionally, this dual treatment resulted in the increased expression of integrin homing and immunologic memory markers on T cells infiltrating tumor. These findings were corroborated in samples from patient GBM, with PD-1 blockade enhancing the T cell-mediated GBM cytolysis. Concerning this strategy, then, these findings provided us with a means to both generate and enhance a tumor-specific response.

While this second strategy proved effective, the mechanism underlying this PD-1/PD-L1-mediated suppression was not fully understood. As such, we proceeded to identify a PD-L1-expressing tumor infiltrating myeloid (TIM) cell population that appeared to dominantly regulate the PD-1/PD-L1 signaling mechanism. Importantly, we determined the role that these cells play in inhibiting the immune response using a series of in vitro and in vivo studies utilizing TIM depletion and PD-1 mAb treatment strategies. We found that depletion of TIMs in both human GBM cultures and murine glioma abolished PD-1/PD-L1-mediated inhibition of T cell activation. Targeting TIMs with colony stimulating factor-1 receptor inhibitor (CSF-1Ri) reduced the TIM population significantly and altered the remaining TIMs such that they demonstrated increased expression of chemotactic factors. While treatment with CSF-1Ri in conjunction with DC vaccine did not alter PD-L1 expression on remaining TIMs, we did note that there was increased TIL infiltration with this dual treatment significantly over DC vaccine alone. These findings suggested that TIMs exert inhibitory effects in the tumor microenvironment in a manner not restricted only to the PD-1/PD-L1 signaling mechanism. We found that the combined treatment of CSF-1Ri and PD-1 mAb with DC vaccination both increased TIL infiltration and activation in the tumor microenvironment. These findings were therapeutically relevant, with tumor-bearing mice receiving all three treatments showing a significant increase in survival over mice receiving each treatment alone. The studies outlined herein elucidated the role that TIMs play in dominantly mediating the PD-1/PD-L1 signaling mechanism to restrict TIL activation, as well as the ability to manipulate this population pharmacologically with clinically accessible agents.

In conclusion, this thesis demonstrates two distinct strategies to generate and enhance an immune response against GBM. In our first strategy, we utilized the adoptive transfer of engineered T cells to selectively target an antigen whose expression we artificially induced in GBM. This technique was largely effective. However, we were interested in directly targeting antigens already expressed by GBM. To that end, we described the utility of DC vaccination in generating an immune response. Further, we delineated the inhibitory mechanisms employed by TIMs in the tumor microenvironment and developed a therapeutic adjuvant to administer with DC vaccination. We confirmed the efficacy of these treatments in a series of in vitro and in vivo animal studies, and we recapitulated these findings in our novel, ex vivo human GBM studies. Together, the studies presented in this thesis represent an innovative approach to understanding and immunotherapeutically targeting the GBM microenvironment.