Skip to Main Content

CAR-T cell therapy has been a boon for treating blood cancers. Since the technology was first brought to the clinic, CAR-T has offered patients months or years of life after they had exhausted all other treatment options and would have died within weeks.

“It’s been incredible,” said Marcela Maus, an immunologist and cell therapist at Mass General Cancer Center. “We’ve seen patients who had multiple lines of therapies and progressed after all of those, [then] get CAR-T and go into long-term remission.”

But CAR-T does have hefty limitations, and scientists like Maus are researching ways to overcome some of its major shortcomings. These issues have prevented CAR-T from being used safely and effectively outside of leukemia and myeloma, and even patients who have responded spectacularly to CAR-T usually see their cancers return. The therapies are also still incredibly costly and carry risks, including a reaction known as a cytokine storm that can be life-threatening.


Potential solutions to these problems are still in the early stages, but scientists are beginning to get a vision of what the future of CAR-T cell therapy might look like. It could involve synthetic biology to engineer a more advanced cell, or engineering other parts of the T cell to make it work better in the challenging environment around a tumor.

“The field is growing tremendously,” Maus said. “Different people are working on different issues then, ideally, the data kind of decides what’s going to be the next big thing.”


Here’s a look at what experts see as some of the most promising approaches.

The controllable CAR

Current CAR-T cells use their CAR, or chimeric antigen receptor, to identify and kill cancer cells. These are synthetic proteins that bind to a specific target, like a protein on a cell surface membrane, and then activate the T cell to kill any cell carrying this target.

Armed with a CAR, T cells become pros at killing cancer cells that have their target, but they’ll also kill normal cells that happen to carry the protein, too. Once a CAR-T cell is in the body, there isn’t much a clinician can do to rein it in if it starts causing a lot of toxicity.

“Once we let the CAR out, they’re like teenage kids,” Maus said. “You can maybe watch, but you can’t really control them. So, there’s some desire to be able to turn them on or off at will.”

So, researchers are also trying to create CAR-T cells that they can manually activate or deactivate. As a group, these are known as controllable CARs, and most work by engineering an additional genetic circuit in the CAR-T cell. In theory, clinicians should be able to activate a switch on the genetic circuit that induces the CAR-T cell to activate their CAR and express it on the T cell’s surface membrane, thereby activating the receptor. Then, after a while, the T cell will disarm.

“The goal is really getting our hands back on the steering wheel for a bit,” Maus said.

There are several ways that synthetic biologists are doing this. In one example, researchers engineered a CAR with a protein switch that activates the receptor in the presence of blue light. In another example, researchers added a gene to CAR-T cells that force it to create its CAR and express it on the cell surface, thereby activating it, only in the presence of ultrasound radiation.

“That way, it can be focused into a specific location,” said Peter Yingxiao Wang, a synthetic biologist at University of California, San Diego, who works on controllable CARs. “When the light or ultrasound is on the tumor locally, they can activate the CAR gene to trigger killing. Anywhere else, the CAR T-cells will be benign.”

The idea is that the clinician can focus the light or ultrasound onto the tumor to get CAR-T cells to begin killing there. Once that signal is turned off, the CARs should disarm or slowly degrade and deactivate the CAR-T cell’s killing function. This way, even if the CAR does kill healthy tissue, the damage will theoretically be limited to the area around the tumor.

“But this is an infant field right now,” Wang added. “A lot of these studies are just proof of concepts to show that they’re technically achievable. If you want to move to clinical trials, all of the components must be optimized.”

Scientists also must show that they’re truly safe in humans, and that keeping the damage to a smaller surface area will be enough to outweigh the risks in treating tumors located near vital organs like the heart.

The logic-gated CAR

Other researchers are working on developing new CARs that can function like a biomolecular computer, able to make simple logical decisions to target cancer cells. Conventional CARs can cause dangerous toxicity because they only use one protein to identify cancer cells, and it may be impossible to discover the perfect target that exists only on cancer cells and not at all on healthy cells.

“You can never uniquely define cancer or any other healthy tissue just by one marker,” explained Wilson Wong, a synthetic biologist at Boston University. “It just doesn’t work. It’s like trying to find a person and saying, ‘he has black hair.’ It’s like, oh, my God, you’ll never find him.”

But it might be possible to distinguish cancer cells from healthy ones by looking at multiple proteins on a single cell. So, researchers like Wong have begun building more advanced CAR T-cells that use genetic circuits that only activate a CAR under more complex conditions, like the presence of several specific proteins that aren’t often seen in combination on healthy cells.

In this sense, the CAR is making a logical decision like basic Boolean computing, and synthetic biologists call this technique logic-gating.

“There’s a lot of cool genetic circuits you can build,” said Yvonne Chen, a synthetic biologist at UCLA. “One can think of conditional systems to obliterate cancer cells. One can build OR-gates, AND-gates, and NOT-gates, and layer them on top of one another.”

Although, Chen added, a drawback of logic-gating is that by increasing the complexity of the system, you might also be increasing the chance something goes wrong. “It’s important not to overcomplicate the design. Sophisticated circuits are exciting, but sometimes the solution itself causes problems. For example, for an AND-gate, you also make it easier for the tumor to escape. If the tumor loses either target A or B, it escapes from therapy,” she said.

The armored CAR

Another issue with conventional CAR-T therapy is that after a while, T cells can simply stop working. Solid tumors, like lung or pancreatic cancer, often have strategies to defend themselves from immune system attacks, including those from CAR-T cells. That makes it harder for CAR-T cells to treat solid tumors and can provide an opening for the tumor to return or progress.

So, researchers like Chen are working on “armoring” the CAR T-cell against the hostile signals in the microenvironment around a solid tumor. One of these signals is called TGF-beta, a protein which can help shut down T cell activity and help cancer cells avoid death and detection from the immune system. Chen was able to create a CAR cell that is not only resistant to TGF-beta, but can actually subvert the signal and become more deadly when it encounters TGF-beta.

“Instead of being dysfunctional, they become activated,” Chen said. “That actually converts a tumor defense mechanism into a stimulatory signal for our T cells and tells them, you’re in an environment where you’re likely to encounter a tumor cell. Get ready.”

Other scientists are working to keep CAR-T cells — which can lose power over time — functional for longer. “Even with a good antigen, the T cells rapidly lose function,” said Shivani Srivastava, an immunologist at the Fred Hutchinson Cancer Research Center who works on this problem. “If you trigger a T cell or CAR over and over again, that causes the cell to become exhausted rather than turning into a memory cell or something else.”

In one case, Stanford immunologist Crystal Mackall engineered a CAR-T cell that takes breaks before returning to work. She did this by creating a transient CAR that can be turned on or off. “It can enhance [the T cell’s] function and limit how exhausted they are by giving them periodic rest,” Srivastava said. “That’s a really interesting strategy in principle.”

But most of the tactics that scientists have tried so far in the realm of armored CAR-T cells haven’t worked in the long term, Srivastava said. You need a strategy that can help the CAR T-cells persist long enough to eradicate the cancer and prevent its return, which might be a lifelong project for the immune system.

“We’ll have to find the right combination that will be durable,” she said. “Often we can find strategies that enhance function for only a short period of time.”

New cell types

Some future approaches might see T cells abandoned altogether. Scientists are slapping synthetic receptors on new or different cell types, such as natural killer cells. One company, called CoImmune, is putting CARs on a synthetic cell called a CIK cell, or cytokine-induced killer cell.

“This is a novel cell type. They don’t occur in nature,” explained Charles Nicolette, the biotech’s chief executive.

They’re made by taking white blood cells and growing them while exposing them to certain immune molecules called cytokines. The advantage of creating new cell types is that biologists can combine certain useful traits from other immune cells, Nicolette said. For example, CIK cells could have the NK cell’s natural ability to distinguish normal cells from malignant ones and the CAR T-cell’s enhanced ability to kill.

One day, UCLA’s Chen hopes to take this concept even further. To her, the ideal cancer-killing cell would not be derived from anything biological, but “a completely artificial cell.”

“Instead of taking a cell from a patient, but rather build a completely defined, minimal cell that can do what we want and nothing else. It cannot evolve. Cannot mutate. Then, self-destruct when you don’t want it there,” she said. But, she added, creating synthetic cells like that would be unimaginably challenging, and it might not be possible to create a cell that’s both persistent but also unchangeable.

Still, a scientist can dream.

Create a display name to comment

This name will appear with your comment