A mostly overlooked component of CAR-T cells has a surprisingly strong effect on the cancer-fighting cells’ behavior, scientists reported on Tuesday, including in ways that might affect their safety and efficacy.
The component is called the co-stimulatory domain, and the two CAR-T therapies approved last year to treat forms of leukemia and lymphoma — Yescarta and Kymriah — use different ones.
A T-cell (blue) finds and destroys a cancer cell (green).
Although the authors of the new study, in Science Signaling, are careful not to say one co-stimulatory domain is better than the other, their analyses of CAR-Ts in test tubes concluded that those with the co-stimulatory molecule CD28 attacked cancer cells more quickly and more intensely than those with the co-stimulatory molecule 4-1BB. The latter is “slower burning and more gentle,” said lead author Alex Salter of the Fred Hutchinson Cancer Research Center.
But in mice with lymphoma, they found, the 4-1BB CAR-T cleared cancer cells more effectively. The 4-1BB version also had higher expression of genes for what’s called T cell memory, which lets T cells live longer and maintain persistent anti-cancer effects.
Novartis’ Kymriah, approved for B-cell acute lymphoblastic leukemia and non-Hodgkin lymphoma, uses 4-1BB, while Gilead’s Yescarta, approved for diffuse large B cell lymphoma, uses CD28. There are no peer-reviewed studies comparing them head-to-head in patients, and in long-ago animal studies neither CD28 nor 4-1BB CAR-Ts were consistently superior to the other.
The blitzkrieg behavior the scientists found with CD28 CAR-Ts might play a role in the out-of-control immune response, called cytokine release syndrome, that is a common and sometimes lethal side effect of CAR-Ts, they said. (Both Yescarta and Kymriah carry warnings about that.) The slow-and-steady behavior of 4-1BB CAR-Ts, the scientists added, might be better at preventing the relapses that some cancer patients suffer.
Although senior author Dr. Stanley Riddell of the Fred Hutch suggested that CAR-T makers “consider modifying CD28 CAR-Ts to make them less intense,” he said it would be “a big leap” to infer anything about patient outcomes from the mouse and test tube results.
CAR-Ts are T cells that have been genetically engineered to produce a receptor (the CAR, or chimeric antigen receptor) that binds to tumor cell surface antigens. The CAR protrudes from the engineered T cell. Almost never discussed is a molecule attached to the CAR and just inside the T cell: the co-stimulatory domain. The first generation of CAR-Ts (invented in academic labs but never developed into drugs) didn’t even have one, said Jennifer Brogdon, director of exploratory immuno-oncology at Novartis.
But in 2002 scientists led by Dr. Michel Sadelain of Memorial Sloan Kettering Cancer Center added the CD28 co-stimulatory domain to the CAR-Ts they were developing, reporting that it amplifies CAR-Ts’ anti-cancer activity. That discovery, Sadelain said, “is what made the field move forward.”
When Novartis began collaborating with the University of Pennsylvania to develop CAR-T therapy, Brogdon said, the academic scientists were testing both CD28 CAR-Ts and 4-1BB CAR-Ts in lab animals. “They observed enhanced persistence of CAR-Ts with the 4-1BB,” she said, and those eventually became Kymriah.
The faster, stronger anti-cancer activity that the Hutch scientists report with the CD28 CAR-T might be desirable for some cancers, while the slower, more persistent 4-1BB might be better for others. Both Novartis and Gilead are actively testing their CAR-Ts against additional blood cancers and, in Gilead’s case, solid tumors. But although Gilead’s Yescarta uses CD28, in a collaboration with Pfizer it is testing Pfizer’s monoclonal antibody utomilumab, which binds to and stimulates 4-1BB, in lymphomas, while Pfizer itself is testing the 4-1BB agonist in solid tumors. According to lab animal data, the antibody can boost T cell activity.
“We believe this research highlights the potential to use different co-stimulatory domains to address different patient populations in order to optimally activate critical aspects of T cell function,” Peter Emtage, senior vice president of cell therapy research at Gilead’s Kite unit, said in a statement to STAT, adding that Kite is “focused on developing and refining our T cell co-stimulatory domains.”
The new findings add to emerging evidence in favor of 4-1BB over CD28 presented at this year’s annual meeting of the American Society of Clinical Oncology. There, scientists from the Beijing-based biotech startup ImmunoChina Pharmaceutical Co. reported results from a clinical trial of 47 patients with B-cell acute lymphoblastic leukemia. They received either CD28 CAR-Ts or 4-1BB CAR-Ts.
All 28 patients receiving 4-1BB CAR-Ts had an “objective response,” meaning fewer malignant cells after treatment; 15 of the 17 patients receiving CD28 CAR-Ts did. In addition, blood levels of CAR-Ts in the 4-1BB patients were “significantly higher” than in the CD28 patients, which might bode well for continued efficacy and therefore, for staying cancer-free longer. Almost all the patients (45) suffered cytokine release syndrome, but the five with the most severe form were in the CD28 group. The ImmunoChina scientists wrote, “4-1BB CAR-T cells show enhanced safety, efficacy, and expansion” compared to CD28 CAR-Ts, suggesting they would be a “superior therapeutic strategy.”
The new study’s findings can be used “going forward to design safer and more effective” CAR-Ts, Riddell said, including reducing cytokine storms and maintaining remission longer. In addition to Gilead and Novartis’ ongoing research on optimizing co-stimulatory molecules for their CAR-Ts, academic labs like Sadelain’s are also in hot pursuit. “You can anticipate new and better designs to emerge in the near future,” he said. “Stay posted.”
Leave a Reply