Glioblastoma (GBM) is one of the most common brain tumours and portends a poor prognosis; patient survival durations are typically <6 months for those with recurrent disease following surgery. Around 30% of newly diagnosed patients express the aberrant EGFR variant III (EGFRvIII), which is associated with unfavourable outcomes. Although brain tumours have an immunosuppressive microenvironment, T-cell-based therapies utilizing chimeric antigen receptors (CARs) have shown promise in this disease. These encouraging data prompted researchers to conduct a first-in-human clinical trial of a CAR-modified T cells directed at the EGFRvIII antigen in 10 patients with recurrent GBM.

In total, 10 patients received a single intravenous infusion of autologous anti-EGFRvIII CAR T cells, with assessment of target expression within the tumour. The manufacturing and infusion of the anti-EGFRvIII CAR T cells was feasible and safe, with no evidence of off-tumour toxicity or cytokine-release syndrome. Marcela Maus, senior author of the trial, comments on the main study findings: “targeting EGFRvIII was safe, without on-target or off-target toxicities. We found that the CAR T-cells readily infiltrate brain tumours, which is not trivial because of the blood–brain barrier. Importantly, levels of EGFRvIII expression in tumour cells either decreased or were eliminated following CAR-T-cell infusion.”

...more regulatory T cells, more PD-L1 expression, and other immunosuppresive markers present after CAR T-cell treatment compared with before

One patient had stable residual disease for >18 months of follow up. Seven patients had surgical intervention following anti-EGFRvIII CAR-T-cell infusion, which enabled tissue-specific analysis and in situ assessment of the tumour microenvironment. Notably, in five of the seven patients, the EGFRvIII antigen decreased, accompanied by an increase in the expression of inhibitory molecules compared with pre-infusion samples. The researchers showed that the infused product could traffic to the tumour, proliferate in situ, and exert anticancer activity that upregulated compensatory adaptive resistance mechanisms, such as increased expression of PD-L1 and IDO1. Collectively, these findings illustrate that the brain tumour microenvironment is not static; rather, many pathways are actived simultaneously in response to therapy. As Maus elaborates: “the brain tumour itself displayed an adaptive response, with increased immunosuppression. Moreover, there were more regulatory T cells, greater PD-L1 expression, and other immunosuppressive markers present after CAR T-cell treatment compared with before.”

In terms of what these findings indicate for the future, Maus laments: “not surprisingly, it is probably going to take more than a single antigen target to effectively eliminate GBM tumour cells: we will have to target more antigens, and block one or more of the immunosuppressive pathways that inhibit T-cell function.”