Over the past few years, the treatment of some hard-to-treat blood cancers has been revolutionized by therapies based on engineered T cells, which leverage the patient’s own immune system to destroy cancerous cells. But until recently researchers haven’t had much luck developing these T-cell therapies—called CAR T—for solid tumors, which make up the vast majority of cancer diagnoses. 

The lack of progress has been disappointing for many in the field. “I think there was a little bit of pessimism some time ago when CAR T for solid tumors didn’t cure all the patients,” says Marcela Maus, director of cellular immunotherapy at the Massachusetts General Hospital Cancer Center.

New trial results, however, suggest that scientists are finally making some headway with next-generation CAR T therapies. Last week, BioNTech presented preliminary results from a clinical study of one called BNT211 at the European Society for Medical Oncology conference in Madrid. The team treated 44 people with solid tumors, mostly ovarian and germ-cell cancers, with varying doses of CAR T cells and, in some cases, a vaccine to help boost the therapy. Among 38 people for whom there was enough data to assess how well the treatment worked, 45% responded, meaning their tumors shrank or disappeared altogether. The presentation focused on a different group of 27 participants who received a higher dose of the treatment. In that group, the researchers saw an even better response rate: nearly 60%. But there were also more serious side effects. 

This is just one of the hundreds of CAR T therapies in clinical trials. Researchers are working  on ways to make CAR T more potent, more precise, and safer. “We’re learning, we’re advancing, and I think it is starting to work in solid tumors,” Maus says. “I’m very hopeful that this is going to be a dramatically useful therapy.”

Targeting systems 

T cells are immune cells that help the body fight infections by destroying diseased cells or recruiting other immune cells to attack. Unfortunately, they have a hard time recognizing cancer cells. CAR T treatments offer a workaround.

To create these therapies, technicians harvest T cells from a patient’s blood. Then they genetically engineer the cells to carry a receptor called a chimeric antigen receptor, or CAR, which can bind to a protein on the surface of the cancer cell. Next they grow these engineered cells in the lab until they number in the millions, and reinfuse them back into the body. When the cells encounter the protein they’re designed to recognize, they activate and start destroying the cancer cells. “They are very much a living drug,” says Andrew Jallouk, a hematologist and oncologist at Vanderbilt University. 

One of the key challenges in using this approach against solid tumors has been finding the right protein to target. “This is what the whole field is really after. How do you find the right antigen?” says Travis Young, vice president of biologics at Calibr, an institute within Scripps Research that is focused on drug discovery and development. 

Some of the proteins that would make the best targets are also found on vital tissue. So there is a risk that T cells would attack healthy cells in the process of targeting a tumor. That’s exactly what happened in one trial 15 years ago, when researchers engineered T cells to target HER-2, a common surface protein in many breast cancers. One patient went into respiratory distress minutes after receiving the therapy and died five days later. The T cells recognized low levels of HER-2 on her lung cells and waged an attack on the wrong tissue.

BioNTech has avoided this issue by targeting a unique protein called Claudin-6, which is present in fetal tissue and some types of cancer but not in healthy adult tissue. 

Another option is to make the T cells smarter. By engineering T cells with multiple receptors, researchers can create cells that turn on only when certain conditions are met—a kind of biological logic gate.

For example, researchers can create cells that require the presence of two antigens to activate (an “and” gate), or cells that activate in the presence of either receptor (an “or” gate). “You can create multiple inputs to your cell, just like the computer would do,” Young says. The T cell can then use that logic to decide whether it’s encountering a tumor cell or a normal cell. It’s more akin to the way T cells work naturally: they have multiple inputs and negative and positive feedback loops. 

Arsenal Bio is one of the companies pursuing this “logic gate” approach. In January, Arsenal launched a clinical trial to test a CAR T therapy for ovarian cancer. 

But sometimes there isn’t a unique protein or set of proteins available for the treatment to zero in on. In that case, if tumor-specific targets don’t exist, it might be possible to add them. In October, a team of researchers from Columbia University reported in Science that they had developed a CAR T therapy that relies on engineered bacteria to tag tumors. The researchers tweaked a strain of E. coli to carry green fluorescent protein and injected the bacteria into mice. The bacteria accumulated in the animals’ tumors. Then they injected the mice with T cells targeting that green protein. “We’re painting the tumors green, and the T cells can ‘see’ green,” says Rosa Vincent, a synthetic biologist and PhD student at Columbia, who was first author on the study.

Why the bacteria accumulate only in tumors isn’t entirely clear. But Vincent suspects that it has to do with the tumor microenvironment. “Because it’s so immunosuppressed, it’s the perfect, permissive environment for the bacteria to grow,” she says. “You only need one cell and it’ll grow exponentially. Whereas if it gets deposited in healthy tissue, the immune system will clear it immediately.” This strategy isn’t yet ready for clinical trials, but the team is already thinking about how to move the research forward. Humans are more sensitive than mice to toxins found on the surface of E. coli. So “the major risk is going to be sepsis and toxic shock,” she says. “But there are so many engineering strategies that we can use to reduce the toxicity of the strains.”

A natural “off” switch

Harnessing the immune system to fight cancer is a double-edged sword. The T cells need to be powerful enough to destroy malignant cells. But if they are too strong, they can release so many inflammatory molecules that they prompt a  whole-body inflammatory response, which can be deadly. This problem, called cytokine release syndrome, happens even with approved CAR T therapies. In mild cases, the syndrome feels like the flu, with muscle aches, body aches, and fever. But in severe cases, this rampant inflammation can be dangerous. 

Striking a balance between efficacy and toxicity has been a persistent challenge for CAR T therapies, and BioNTech has yet to find the right mix. More than half the participants in last week’s study experienced cytokine release syndrome. Most of the events were mild, but there were two more serious cases of the syndrome, including one patient  who experienced acute respiratory distress and spent time in intensive care. But the high rate of this problem is, ironically, “sort of a good sign,” says Maus. It shows the therapy is working.

Making sure the T cells are only targeting cancer cells helps make CAR T therapies safer, but physicians would also like to be able to rein in the T cells if they start to cause damage.

Young and his colleagues at Calibr have developed a switchable CAR T therapy that requires an antibody to activate the T cells. First, researchers administer the antibody, which binds to cancer cells. Next, they infuse the T cells, which become activated when they bind to the antibody. “The CAR T cells, in the absence of the antibody, don’t target anything,” Young says. And because the antibody doesn’t stick around for more than a few days, “the CAR T cells will have a natural ‘off’ to them.” That allows the researchers to pull back on a treatment if there are adverse effects.

A test of time

At BioNTech, researchers are trying to address another chronic issue with CAR T therapies: durability. The engineered T cells don’t always last long enough to fully eradicate cancer from the body. By combining the CAR T cells with an mRNA vaccine, they hope to improve their staying power. BioNTech’s mRNA vaccine delivers  instructions for making the same antigen the T cells target: Claudin-6. The more antigen is around, the more excited the T cells get. Cancer cells carry Claudin-6 too, of course, but the microenvironment in solid tumors can hamper the activity of T cells. “By the time that CAR T cells get in there and they have all this immunosuppressive stuff happening to them, they may not expand very well,” Jallouk says. The  vaccine should ensure that T cells encounter Claudin-6, become activated, and replicate well right away. 

The preliminary results presented in Madrid indicate that this approach may work. In the group that didn’t receive the vaccine, “by day 50 the majority of cells are not seen anymore,” said John Haanen, a cancer researcher from the Netherlands Cancer Institute who presented the results. The CAR T cells had better staying power in patients who received the vaccine. Many of them still had CAR T cells present 90 days out. “Now, whether that will translate into better efficacy compared to not having the vaccine—I think we need a bit more data to tell,” Jallouk says. “But I think it’s a reasonable approach to try to enhance expansion and persistence.”

Eventually the company plans to launch a phase 2 trial to test the therapy in more patients. “There are a lot of companies working in this field, and there are a lot of new technologies that are being tried,” Jallouk adds. Even trials that aren’t “a roaring success” can provide valuable lessons, he says: “I have a lot of hope that eventually we’ll get a formula that will work well in solid tumors.”