Armoring CAR T cells to prevent self-sabotage and better fight cancer

In Lewis Carrol’s “Through the Looking Glass,” protagonist Alice races with the Red Queen and is unable to gain a lead despite her best efforts. “Now, here, you see, it takes all the running you can do, to keep in the same place. If you want to get somewhere else, you must run at least twice as fast as that!” the Queen yelled.

Biologists have co-opted this story to explain why organisms need to constantly adapt in the evolutionary race for survival—dubbing it the Red Queen Hypothesis. And though this analogy has been used most commonly in the context of parasite-host interactions, it also captures the ongoing battle between cancer treatments and tumor evolution: Each time researchers develop a powerful new therapy, tumors evolve new ways to evade, resist, or neutralize it.

Now, a cross-disciplinary team at the University of Pennsylvania has identified a striking example of such biological ingenuity and a way to fight back. Led by Wei Guo of the School of Arts & Sciences, the researchers revealed a mechanism by which solid tumors sabotage chimeric antigen receptor (CAR) T cell therapy by turning the body’s own immune warriors against each other. Their findings are published in Nature Cancer.

“This is one of the more insidious ways solid tumors evade immune attack,” Guo says. “The field had seen CAR T cell fratricide before, but this is the first time we’re seeing that it can happen remotely—through signaling molecules—without direct cell-to-cell contact.”

CAR T cells are genetically engineered to seek out and destroy cancer cells by recognizing specific tumor antigens. Guo explains that they’ve shown “great success against blood cancers, but for solid tumors like in pancreatic cancer, progress has been limited.”

The researchers had previously discovered that solid tumors release small extracellular vesicles (sEVs)—tiny, lipid membrane-bound packages that ferry cargo between cells, including those that are loaded with tumor antigens. When CAR T cells engulf these vesicles, they absorb the tumor markers onto their own surfaces, Guo explains. As a result, other CAR T cells mistake their fellow cancer fighters for cancer cells, launching fatal attacks. This misidentification leads to fratricide and drastically reduces the efficacy of the therapy.

The team’s latest findings show the scale and stealth of the tumor’s strategy says Serge Fuchs, of the School of Veterinary Medicine and long-time collaborator of Guo. “A T cell can be compromised even before it reaches the tumor. It can be intercepted and reprogrammed through vesicle uptake alone. That expands the landscape of immune evasion tremendously," he says.

The study also uncovered the cellular trigger for this deceptive vesicle release. Once CAR T cells are introduced into the tumor environment, they release TNF-alpha, a cytokine that in turn stimulates the tumor to ramp up EV secretion. The research team identified RAB27A, a small GTPase protein, as a key regulator in this process. Its activated form drives the release of antigen-rich vesicles, effectively creating a feedback loop in which the immune system's attack unwittingly causes its own undoing.

However, the team found that by engineering CAR T cells to overexpress a protein called serpinB9 inside the cells, they were able to make the CAR T cells resistant to fratricide.

SerpinB9 naturally inhibits granzyme B, the enzyme T cells use to kill targets. In this context, it protects them from being killed by each other. “It’s like giving them body armor,” says Guo. When combined with anti-PD-1 checkpoint inhibitors, a common adjunct in immunotherapy, the armored CAR T cells showed enhanced tumor control in mouse models.

Fuchs notes that these modifications represent a kind of second-generation CAR [chimeric antigen receptor] design. He says, “Not only do the cells carry a targeting mechanism, but they also have an onboard system to resist enemy.”

For coauthor Ravi Radhakrishnan, professor of bioengineering at the School of Engineering and Applied Science, these findings highlight the value of foundational, systems-level science in driving translational breakthroughs.

“The field has been focused on vesicles as a biomarker or diagnostic tool, especially in the context of liquid biopsy,” he says. “But stepping back to ask, ‘How does this all work at the cellular and molecular level?’ has opened up entirely new therapeutic strategies.”

“This is the kind of science that happens when you get people together who speak different scientific languages but are fluent in the same big questions,” says Fuchs. “And in this case, it’s a question cancer has been asking for a long time: How do I survive? We’ve just figured out one more way it answers that—and how we might silence it.”

Wei Guo is the Hirsch Family President's Distinguished Professor in the Department of Biology in the School of Arts & Sciences at the University of Pennsylvania.

Serge Y. Fuchs is the Elizabeth and William Whitney Clark Professor of Oncology in the Department of Biomedical Sciences in the School of Veterinary Medicine at Penn.

Ravi Radhakrishnan is the Herman P. Schwan Department Chair of Bioengineering in the Department of Bioengineering and a professor of chemical and biomolecular engineering at Penn’s School of Engineering and Applied Science.

Authors include Penn Arts & Sciences’ Wenqun Zhong, Zhiyuan Qin, Jingbo Yang, and Ziyan Yu; Nils W. Engel, Yi Fan, Carl H. June, Leyuan Ma, Neil C. Sheppard, Xiaowei Xu, and Dongdong Yan from the Perelman School of Medicine at Penn.

This work was supported by National Institutes of Health (Grant R35 GM141832.)