Reprogramming the Tumor Microenvironment to Improve Immunotherapy of Glioblastoma by Co-Targeting VEGF and Ang2

SUMMARY Glioblastoma (GBM) is a universally fatal brain tumor. Immune checkpoint blockers (ICBs), such as anti- programmed cell death-1 protein (aPD1), alone or in combination with bevacizumab - an anti-vascular endothelial growth factor antibody (aVEGF), failed to improve survival in phase III clinical trials in GBM. This failure is, in part, due to the formidable barriers that the GBM tumor microenvironment (TME) creates. First, GBM tumor cells are highly proliferative and invasive with low neoantigen load, and thus, can easily evade immune surveillance. Second, GBM vessels are abnormal, and thus, they create a leaky, hypoxic and edematous TME and limit the delivery of drugs and the access of antitumor immune cells such as cytotoxic T lymphocytes (CTLs) into the tumor resulting in a cold CTL-excluded TME. Moreover, limited number of CTLs that accrue within GBM TME are dysfunctional. In contrast, pro-tumor immune cells such as regulatory T cells (Tregs) and “M2-like” macrophages preferentially accumulate in GBM. Unlike CTLs, Tregs and “M2-like” macrophages do not require intact vessels for trafficking to the tumor and thrive and proliferate in the GBM TME. Third, our pilot studies indicate that aPD1 aggravates vascular abnormalities and inflammatory responses in GBM, causing toxicities. Collectively, these features give rise to a strongly immunosuppressive TME in GBM that resists both the standard of care (SoC) and immunotherapy. Our preclinical and clinical studies and those of others indicate that angiopoietin-2 (Ang-2) can shorten the duration of aVEGF-induced vascular normalization. Thus, we hypothesize that normalizing tumor vasculature by co-targeting angiopoietin-2 (Ang-2) and aVEGF (abbreviated as aA2V) can both overcome resistance to aPD1 and reduce toxicities in patients with GBM. We will test if aA2V+aPD1 can durably normalize tumor vessels and improve their function. At cellular level, we will test if aA2V+aPD1 can repair the dysfunctional endothelial cells to express adhesion molecules, which are required for CTL cell trafficking, convert them to non-canonical antigen presenting cells to present tumor antigens to CTLs, and collectively result in improved CTL infiltration and function (Aim 1). We will further determine the involvement of antibody- dependent cell cytotoxicity (ADCC) in aPD1-induced adverse events and their alleviation by aA2V (Aim 2). Finally, we will use our newly developed surgical model that faithfully recapitulates GBM therapy in mice including SoC (surgery and chemo radiation) to test whether combining SoC with aA2V+aPD1 can promote durable responses (longer survival and memory responses) (Aim 3). Our findings will provide unprecedented insights into the mechanisms of resistance to immunotherapy in GBM, establish a novel strategy to overcome this resistance while abrogating putative adverse effects, and directly inform the design of clinical trials of GBM patients with combination aA2V+aPD1 therapy and SoC.