TMIC-19. Engineered T cells against the glycoprotein EFEMP1/fibulin-3 target glioblastoma cells in vitro and in vivo

Abstract Glioblastoma (GBM) tumors remain a challenge for immunotherapy due to their heterogeneity, low mutational burden, and ability to suppress immune responses. An important but largely disregarded contribution to immunosuppression is the ability of GBM cells to change the composition of the neural environment by secreting extracellular matrix molecules that affect the accumulation and function of immune cells. We have demonstrated that antibodies against the secreted glycoprotein EFEMP1/fibulin-3, which is upregulated in the pericellular ECM (glycocalyx) of GBM cells but absent from normal brain, block immunosuppressive signaling induced by this protein, allow immune reactivation, and disrupt tumor progression. Here, we describe the anti-tumor effects of T cells engineered for stable expression of a second-generation chimeric antigen receptor derived from our leading anti-fibulin-3 antibody. Anti-fibulin-3 CAR-T cells showed increased release of IFNgamma, IL-2, and granzyme B when exposed to fibulin-3-secreting GBM cell lines and tumor stem cells, but not in presence of fibulin-3-negative cells. Both CAR-T cell activation and proliferation were also triggered by purified fibulin-3. Fibulin-3 induced upregulation of Src/Erk/MAPK signaling in CAR-T cells as well as activation of NFAT-dependent transcription. None of these effects were observed in control T cells or in absence of fibulin-3. Anti-fibulin-3 CAR-T cells efficiently killed GBM cells even at a low effector:target ratio (1:1); this effect was inhibited by an NFAT-inhibitor peptide (11R-VIVIT) and anti-perforin antibody, suggesting conventional cytotoxic mechanisms. Furthermore, anti-fibulin-3 CAR-T cells, delivered locally, persisted in the tumor and significantly extended the survival of mice carrying intracranial GBM xenografts. The same result was observed in mice carrying subcutaneous GBM xenografts, treated with IV injections of CAR-T cells. Given the widespread expression of fibulin-3 in the core and perivascular areas of GBM tumors, our results highlight this cell-surface matrix protein as an attractive target to complement current approaches for GBM immunotherapy.