How CAR T Cell Therapy Works, and Why Its Risks Come Built In

CAR T cell therapy takes your own T cells out, rewrites them in a laboratory to carry a synthetic receptor aimed at a marker on cancer cells such as CD19 or BCMA, then returns them to your bloodstream to hunt the tumor. It is often described as a living drug because the infused cells keep dividing and working inside the body. The catch is that its two most feared complications, cytokine release syndrome and a neurologic syndrome called ICANS, are not random side effects. They are the predictable overflow of the very immune reaction the treatment is designed to ignite.

Key points#

• A chimeric antigen receptor (CAR) is a lab-built protein that lets a T cell recognize a cancer marker directly, without the usual antigen-presentation step.
• CD19 marks most B cell cancers and BCMA marks plasma cells, which is why these two targets anchor the approved products.
• Cytokine release syndrome (CRS) and ICANS arise from the same intense immune activation that makes the therapy effective.
• Interleukin-6 sits at the center of CRS, which is why the IL-6 blocker tocilizumab became a standard antidote.
• In 2025 the FDA removed the formal risk-management program for these products, reflecting a maturing safety record rather than disappearing risk.

Reprogramming a T cell to find cancer#

A normal T cell is a careful inspector. It reads fragments of protein displayed on other cells through the major histocompatibility complex and reacts only when the presentation is right. A CAR shortcuts that whole ritual.

According to the National Cancer Institute, a chimeric antigen receptor is a single engineered protein assembled from parts that never share one molecule in nature. It threads through the T cell membrane. The outward-facing end is built from antibody fragments that clamp onto a chosen antigen on the tumor surface. The inward-facing end wires straight into the T cell's machinery for killing and multiplying, so the instant the receptor grabs its target, the cell is switched on. Because recognition no longer depends on antigen presentation, the response can be strikingly fast.

Getting those cells ready is a logistical feat. A team collects your T cells, sends them to a manufacturing facility where a viral vector inserts the CAR gene, grows the edited cells into the hundreds of millions, and ships them back for infusion. The National Cancer Institute estimates the full cycle at roughly three to five weeks.

Choosing a target that spares the patient#

The whole design lives or dies on the target. The ideal antigen is crowded on the cancer and scarce on any tissue the body cannot spare. Two targets have carried the field so far. CD19 studs the surface of most B cell cancers, which is why it underpins the products approved for B cell acute lymphoblastic leukemia and several non-Hodgkin lymphomas. BCMA, short for B cell maturation antigen, flags plasma cells and is the target in multiple myeloma.

Six autologous CD19 or BCMA products have reached the United States market. The CD19 group includes tisagenlecleucel, the first approved in 2017 for young patients with acute lymphoblastic leukemia, along with axicabtagene ciloleucel, brexucabtagene autoleucel, and lisocabtagene maraleucel. The BCMA group holds idecabtagene vicleucel, cleared in March 2021, and ciltacabtagene autoleucel, cleared in February 2022, both for relapsed or refractory multiple myeloma. In December 2025 the label for lisocabtagene maraleucel was widened to cover relapsed or refractory marginal zone lymphoma. There is a built-in cost to targeting CD19 and BCMA: healthy B cells and plasma cells wear the same markers, so the therapy predictably wipes them out and leaves you more prone to infection.

One engine, two kinds of harm#

The signature dangers of CAR T therapy are easiest to understand as a single mechanism running too hot. A 2021 review in the Journal of Experimental and Clinical Cancer Research traced the chain of events step by step, and it explains why the harms shadow the benefit so closely.

Cytokine release syndrome#

When CAR T cells lock onto tumor cells, they release perforin and granzymes that push the cancer cell into an inflammatory form of death called pyroptosis. The dying cells leak alarm signals, molecules such as HMGB1, ATP, and free DNA, into the surrounding tissue. Those signals rouse macrophages, which the review names as the true engine of the syndrome. The activated macrophages flood the body with inflammatory cytokines, among them interleukin-1, interleukin-6, interferon-gamma, and GM-CSF.

Interleukin-6 is the ringleader. It switches on the cells lining blood vessels through JAK-STAT signaling, and those vessel-lining cells answer by making still more interleukin-6, a self-amplifying loop. The clinical picture climbs from fever and low blood pressure toward organ dysfunction. That single pathway is why tocilizumab, an interleukin-6 blocker first developed for rheumatology, became a frontline remedy, frequently combined with corticosteroids.

ICANS, when the brain is caught in the crossfire#

The neurologic syndrome, formally immune effector cell-associated neurotoxicity syndrome, usually arrives just after the peak of cytokine release, a timing clue that the two share a root. The same review describes circulating cytokines activating the vessel-lining cells inside the brain and prying open the tight junctions of the blood-brain barrier. Once that barrier gives way, inflammatory cells and mediators seep into the central nervous system, and downstream products such as glutamate and quinolinic acid can overstimulate neurons.

There is also a target-specific twist. The review notes that CD19 appears on pericytes, the support cells that hug the brain's small vessels, so CD19-directed cells can hit this tissue through an on-target, off-tumor effect. In the clinic ICANS shows up as confusion, trouble finding words, tremor, and in severe cases seizures. The National Cancer Institute lists corticosteroids as the usual treatment, with anakinra, an interleukin-1 blocker, held in reserve when steroids fall short.

A safety story that keeps evolving#

Oversight has shifted as the field has gathered experience. In a safety communication dated June 26, 2025, the FDA lifted the Risk Evaluation and Mitigation Strategies that had governed every approved CD19 and BCMA autologous product since launch. The agency stated that adverse-event reporting for cytokine release syndrome and neurologic toxicity had stayed stable, and that current labeling could convey the risks adequately given how practiced the hematology and oncology community had become at spotting and managing them.

The updated labeling still asks teams to watch patients for at least two weeks, with daily monitoring for at least the first week, to keep you near a healthcare facility for those two weeks, and to tell you not to drive for two weeks after infusion. The FDA framed the change as trimming strain on the healthcare system without loosening safe administration.

None of that erases the danger. It reflects a hard-won familiarity with complications that are, by design, fused to the mechanism. That link is the whole lesson: the force that clears a refractory cancer is the same force that inflames the body, and both pour out of one engineered cell.