In vivo gene therapy: teaching the body to manufacture its own CAR-T cells against autoimmune diseases

Why this news matters

Can the body be taught to fight autoimmune diseases on its own? A team from Wuhan University of Science and Technology (China) has just taken a striking step in that direction. Their work, presented on 2 September 2026 in the New England Journal of Medicine, describes a technique that combines two cutting-edge technologies — CAR-T cells and gene therapy — to enable the body itself to produce the cells it needs to 'reboot' its immune system.

Autoimmune diseases arise when the immune system, which should protect us, mistakes its target and attacks the body's healthy cells. In many of these conditions, white blood cells called B lymphocytes produce antibodies directed against our own tissues. The idea here is to eliminate those 'misprogrammed' B lymphocytes so that the body replaces them with new ones that no longer attack the organism.

What CAR-T cells are and why they are so difficult to manufacture

CAR-T cell therapy involves extracting T lymphocytes (immune cells) from the patient and genetically modifying them in the laboratory so that they recognise and destroy a specific target: tumour cells or, in the case of autoimmune diseases, harmful B lymphocytes. This technology has already demonstrated its efficacy against certain cancers and, more recently, in some severe autoimmune diseases.

The problem is that manufacturing these cells is complex and very costly: the cells must be extracted, modified outside the body (ex vivo), multiplied, and then re-infused — all within highly specialised facilities. This complexity enormously limits access to these treatments.

The innovation: letting the body manufacture its own CAR-T cells

The approach taken by the Chinese researchers bypasses that entire laboratory process. Using a virus as a 'messenger', they introduce genetic instructions directly into the patient's T lymphocytes, inside the body. Those instructions teach the cells to recognise the protein CD19, which is present on the surface of B lymphocytes. The result is that the modified T lymphocytes destroy the B lymphocytes.

In other words: instead of bringing in ready-made CAR-T cells from outside, the body itself is turned into the 'factory'. If this strategy is confirmed, it could democratise therapies that today are reserved for very few centres.

Results in 16 patients: precision and control

The team tested their method in 16 patients with various autoimmune diseases, seven of them with multiple sclerosis. After a single intravenous infusion of the virus, the participants' bodies began mass-producing these new CAR-T cells.

Two findings stand out for their safety implications. First, the system was highly precise: 99% of the genetically modified cells were T lymphocytes — that is, precisely the cells intended to be modified. Second, the gene therapy does not appear to have caused dangerous mutations — the DNA fragments carried by the virus did not insert into sensitive regions of the genome — and the viruses were eliminated rapidly by the body without triggering an excessive immune reaction.

An immune system 'reboot'

The proliferation of CAR-T cells caused rapid depletion of B lymphocytes. Within two months, these had been completely replaced by new B lymphocytes that were no longer programmed to attack the body. It is, in a sense, like reinstalling the operating system of a corrupted computer.

This immune 'reboot' rapidly improved patients' condition. Six months after treatment, those with multiple sclerosis showed better cognitive and motor function and felt less fatigued. A more detailed clinical analysis also revealed reduced neuronal damage (multiple sclerosis damages the myelin sheath surrounding axons) and reduced activation of microglia, the immune cells of the brain.

Safety: manageable side effects

None of the patients experienced lasting serious side effects from the treatment. The most frequent event — observed in 11 of the 16 patients — was an inflammatory response between one and three weeks after the viral infusion, but it was temporary and lasted no more than two weeks.

The authors conclude that generating CAR-T cells in vivo with a viral vector is associated with 'manageable' side effects and could be effective for treating neurological autoimmune diseases. They also highlight its principal advantage: by avoiding the complex production of these cells outside the body, this technique has the potential to make CAR-T therapy a more accessible and adaptable intervention.

Why caution remains essential

The results should be kept in proper perspective. This is an initial study, with very few patients, no control group, and only a few months of follow-up. We do not yet know how long the benefit lasts, nor what the long-term safety profile is of introducing genetic instructions via a virus. The authors themselves note that larger and longer studies will be needed to properly evaluate the efficacy and safety of this strategy.

In medicine, a promising announcement is not equivalent to an available treatment. Between a first proof of concept and routine clinical practice there almost always lies several years of rigorous trials.

Clínica Valorian's analysis

What makes this work fascinating is not only the clinical result, but the shift in strategy: rather than manufacturing complex cells in the laboratory, the body is taught to manufacture them itself. If confirmed, that idea could transform access to therapies currently limited by their cost and complexity — and not only in oncology, but also in autoimmune diseases affecting millions of people.

At the same time, we maintain the caution that rigour demands: 16 patients is a very small number, and gene therapy raises legitimate questions about long-term safety. We will follow the upcoming trials closely. What is today a hopeful proof of concept will need to demonstrate, with more patients and longer follow-up, that it is also safe and durable.