Skip to content

Third-Generation CAR T-cell Therapy for Glioblastoma Is Effective in Animal Model

Key findings

  • Researchers at Massachusetts General Hospital have created an EGFRvIII-specific third-generation chimeric antigen receptor (CAR) that expresses the co-stimulatory molecules CD28 and OX40
  • The researchers also generated artificial antigen-presenting cells (aAPCs) that express an EGFRvIII domain as well as the costimulatory molecules CD32, CD80 and CD137L
  • When co-cultured with the aAPCs, the CAR T cells persisted and robustly proliferated under demanding conditions in long-term cultures
  • Mice implanted with single low doses of the CAR T cells and the aAPCs survived as long as 121 days, and the treatment resulted in focal tumor tissue destruction

Chimeric antigen receptor (CAR) T-cell adoptive immunotherapy has been successful in patients with hematologic malignancies, and now it is being explored for solid tumors, including brain tumors. CARs are exciting because they directly recognize cell surface antigens independently of the major histocompatibility complex (MHC), which makes them universal for all patients and resistant to tumor escape by MHC downregulation.

One of the key factors in CAR T-cell therapy is to select the target antigen carefully. Most antigens targeted by CARs are not tumor-specific. They are expressed in normal tissues in the body as well as by tumors. Targeting them can cause significant off-target toxicities.

Bob S. Carter, MD, PhD, chief of neurosurgery at Massachusetts General Hospital, and colleagues are focusing on CAR T-cell therapy for glioblastoma that is directed against the epidermal growth factor variant III (EGFRvIII). This mutation increases glioma proliferation, invasion and therapeutic resistance, and it is not expressed in normal brain tissue.

Most recently, the research group developed a third-generation EGFRvIII CAR that incorporates costimulatory molecules to boost the activation and persistence of T cells. They also generated artificial antigen-presenting cells (aAPCs) that express additional costimulatory molecules.

In PLOS ONE, the research team reports that combination therapy with the new EGFRvIII CAR T cells and the aAPCs had an antitumor effect in mice implanted with human glioblastoma.

CAR Construction and Creation of EGFRvIII-specific aAPCs

The researchers' specific method was to link EGFRvIII with CD28 and OX40. They designated this third-generation construct as G3-EGFRvIII CAR.

Next, the team genetically engineered two aAPC lines:

  • One that expressed the extracellular domain of EGFRvIII, which they designated EGFRVIIIΔ654
  • One that expressed EGFRvIII and the co-stimulatory antigens CD32, CD80 and CD137L, which they designated CD32-80-137L-EGFRVIIIΔ654

Results of CAR T-cell Exposure to aAPCs

To determine the efficiencies of various cell lines, the researchers co-cultured G3-EGFRvIII CAR T cells with irradiated EGFRVIIIΔ654, CD32-80-137L-EGFRVIIIΔ654 or CD32-80-137L aAPCs. Within three weeks, the T cells had proliferated by 15- to 25-fold. Proliferation was greatest in the culture with CD32-80-137L-EGFRVIIIΔ654 aAPCs.

The researchers then assessed G3-EGFRvIII CAR T cells under more demanding conditions. They stimulated them with IL-2 supplementation or hygromycin selection every week for up to six weeks. Once again, proliferation of T cells was greatest when they were co-cultured with CD32-80-137L-EGFRVIIIΔ654 aAPCs.

A Single Treatment Extends Survival

To assess their new treatment in vivo, the researchers mixed human glioblastoma cells with three groups of T cells: G3-EGFRvIII CAR, their original EGFRvIII CAR (designated G1-EGFRvIII CAR) and a control group. Each group of T cells was injected into peripheral blood mononuclear cells and cocultured with CD32-80-137L-EGFRVIIIΔ654 aAPCs.

Immunodeficient mice were then intracranially injected with 25,000 glioblastoma cells plus 25,000 effector cells (G3-EGFRvIII CAR, G1-EGFRvIII CAR or control). The G3 mice survived substantially longer than the G1 and control mice: average survival was 48.5, 32 and 34 days, respectively. One of the G3 mice survived for 90 days and another for 121 days.

In brain specimens from GI and control mice, T cells only rarely persisted in the tumor area, but in all G3 mice, there were multiple T-cell clusters in several tumor areas. Moreover, focal destruction of tumor tissue was evident in G3 mice and not in the other two groups.

Potential Antitumor Therapy

Dr. Carter's group notes that in 2012, a U.K. researcher set out two criteria for future success of CAR T-cell adoptive immunotherapy: smaller CAR T-cell doses for safety and practicality, and development of T cells to expand in vivo in a controlled manner and persist for longer periods. They conclude that this study meets both criteria.

In animals, EGFRvIII has been detected in up to 30% of glioblastoma tumors, the researchers comment. They believe their new therapeutic strategy may complement the standard of care for patients with glioblastoma.

Learn more about the Stephen E. and Catherine Pappas Center for Neuro-Oncology

Refer a patient to the Department of Neurosurgery at Mass General


As he reflects on the first year in his new position, Dr. Bob Carter discusses what made him come back to Mass General, recent research projects he's worked on and more.


In this video, Bob Carter, MD, PhD, chief of Neurosurgery Service, discusses an active research program at Massachusetts General Hospital that's involved in developing Car T-cell therapy for brain tumors.