When Bruce Levine and his colleagues began attempting to engineer immune cells to fight cancer in the 1990s, few people thought they would succeed. “The wider community was very sceptical,” says Levine, an immunologist at the University of Pennsylvania Perelman School of Medicine in Philadelphia. When they presented their research at meetings, “we were in the last session on the last day, in a room no one could find”.

Today, these engineered immune cells, called CAR T cells, are among the most powerful therapies oncologists have to treat many types of blood cancer. And studies suggest that they might hold promise for brain cancer and other solid tumours, as well as autoimmune and other diseases. One research firm estimates that the value of the CAR-T-therapy market, expected to hit US$11 billion this year, will grow to nearly $190 billion by 2034.

But CAR-T therapies come with a serious downside — they are laborious to make and difficult to administer. After removing the immune cells, called T cells, from a person’s blood, physicians ship them off to a manufacturer, where technicians genetically engineer the cells to carry a specialized protein called a chimeric antigen receptor (hence ‘CAR T’) on their surface. The cells are grown and amplified into hundreds of millions more cells, frozen and returned to the hospital for re-infusion. Because of the complexity, only about 200 centres in the United States offer the therapy.

The race to supercharge cancer-fighting T cells

“This whole process, it’s just inefficient,” says Saar Gill, a haematologist and oncologist also at the Perelman School of Medicine. “If I’ve got a patient with cancer, I can prescribe chemotherapy and they’ll get it tomorrow.” With commercial CAR T, however, people have to wait weeks for treatment. That delay, along with the high cost of the therapy, plus the need for chemotherapy before people receive the CAR T cells, means many people who could benefit from CAR T never receive it. “We all want to get to a situation where CAR T cells are more like a drug,” says Gill.

Some biotechnology companies have an answer: alter T cells inside the body instead. Treatments that deliver a gene for the CAR protein to cells in the blood could be mass produced and available on demand — theoretically, at a much lower price than current CAR-T therapies. A single dose of commercial CAR-T therapy costs around $500,000. A vial of in vivo treatment might cost an order of magnitude less.

The idea has some high-profile proponents. The founders of Capstan Therapeutics, a company in San Diego, California, that is focused exclusively on in vivo cell therapies, include CAR-T pioneers Levine and Carl June, as well as Drew Weissman, who won a Nobel prize for his work on messenger RNA vaccines. CRISPR–Cas9 pioneer and Nobel prizewinner Jennifer Doudna has co-founded a separate company, Azalea Therapeutics in Berkeley, California, that is developing in vivo CAR T. And big pharma is taking notice. In March, the biopharmaceutical firm AstraZeneca agreed to pay up to $1 billion for the Belgium-based in vivo CAR-T company EsoBiotec, which launched its first human trial of an in vivo CAR-T therapy in January.

Although human trials have just begun, many researchers are excited about the potential of a simpler iteration of CAR T. “If it’s efficacious and safe, it could really challenge the current paradigm,” says Joseph McGuirk, a haematologist and oncologist who studies cellular therapies at the University of Kansas Medical Center in Kansas City. And “we need to challenge the current paradigm”.

Bringing the outside in

Many of the in vivo CAR-T therapies under development have taken pages from the ex vivo playbook. Similar to the approved therapies, in vivo approaches aim to destroy white blood cells called B cells, and so treat cancers that form in these cells. (CAR-T therapies destroy healthy B cells, too, but people can live without these cells.)

As with ex vivo, many in vivo therapies rely on an engineered version of a lentivirus to grab onto T cells and deliver the gene for the CAR protein into the cell’s genome (see ‘Made within’). But engineering cells inside the body is a tricky business. With ex vivo, the T cells have been removed from the body, so researchers don’t have to worry about introducing the gene into other cell types. But in the body, many cells share common receptors, so researchers have to find ways to specifically target T cells — or other immune cells that could join the fight.

“The stumbling block is, how do you get it to the right cell, the right place, right time?” says Michel Sadelain, a genetic engineer and director of the Columbia Initiative in Cell Engineering and Therapy at Columbia University in New York City and another CAR-T pioneer.

Each company has developed its own approach to solving this problem, and each has tweaked its vector in different ways. For example, Interius BioTherapeutics in Philadelphia, co-founded by Gill, is testing a vector that latches onto CD7, a protein found only on T cells and natural killer cells.

Umoja Biopharma in Seattle, Washington, is testing a lentiviral vector decorated with a protein that targets three receptors on T cells at once. The company has evidence from animal models that this strategy is more effective than targeting just one receptor, and hopes that it will more closely mimic what happens when a T cell becomes activated naturally, after an infection, for example.

If effective, such approaches could simplify manufacturing and get CAR-T therapy to more people more quickly. McGuirk notes that when the ex vivo CAR-T therapy Carvykti (ciltacabtagene autoleucel) was approved in 2022 for people with multiple myeloma, the manufacturer had limited production capacity, leading to long waiting times.

“We had 50 patients at one time on our wait list,” says McGuirk. “More than half of those patients died before we could get them a slot for manufacturing.”

Do cutting-edge CAR-T-cell therapies cause cancer? What the data say

In addition to simplifying production, there is another potential advantage to in vivo treatment. With ex vivo approaches, people receive chemotherapy before CAR-T treatment to eliminate unedited T cells and make space for the edited ones to proliferate. The idea is to create a blank slate. But with in vivo approaches, people would forgo the pre-treatment chemotherapy. “We don’t want to kill the very T cells that we hope to transduce,” says Gill. That would eliminate chemotherapy-related side effects, such as a greater risk of infections. It might also mean that people who are too ill to receive chemotherapy could still be eligible for CAR-T therapy.

It’s not entirely clear, however, whether in vivo CAR T delivered through a viral vector can eliminate all the side effects that come with ex vivo approaches. After infusion, CAR T cells multiply in the body and release chemicals that ramp up the immune system. In some cases, they can cause an inflammatory storm known as cytokine release syndrome. “It’s like a train that is running out of control,” says Adrian Bot, chief scientific officer at Capstan Therapeutics.

Some researchers suspect this might be less of a problem with in vivo approaches. That’s because the engineered T cells will be multiplying in the presence of other immune cells, including cells that could help to put the brakes on a runaway immune response, says Yi Lin, a haematologist and oncologist at Mayo Clinic in Rochester, Minnesota. As a result, the side effects might be milder, she says.

Even so, these in vivo approaches are unlikely to eliminate the concerns that have arisen in the past couple of years about secondary cancers caused by CAR-T therapies. Although the risk is low, many countries require CAR-T therapies to carry a warning label detailing this risk. The issue is that lentiviral vectors permanently integrate the CAR gene into the genome of the T cell. Each time the cell divides, it produces more CAR T cells. That means T cells might go on attacking their targets for long periods — years, in some cases.

Although that could be an advantage, researchers don’t have much control over where in the genome the lentivirus inserts the gene. If it lands in the wrong spot, it could prompt the formation of secondary cancers. Companies can screen for this problem with ex vivo therapies but not with in vivo.

Turning to RNA

Instead of viral vectors, Capstan and several other companies are turning to nanoparticles to ferry RNA into T cells. The RNA enters the cell’s cytoplasm, where its genetic information is used to make the CAR protein for only a short time. Although the therapy needs to be administered in multiple doses, it could prove to be safer. At the first sign of a serious problem, physicians could stop administering the treatment, and the CAR protein would go away in a few days or less.