The seven CAR-T cell therapies the US Food and Drug Administration has approved over the past decade all follow an elaborate pattern: draw blood, ship it to a specialised facility, isolate T cells, modify them genetically, grow enough of them, ship them back, infuse. Weeks, dedicated equipment, laboratory staff — and a price among the highest in medicine.

A team at MIT has now shown that much of that could be skipped. Rather than engineering immune cells outside the body, the researchers injected lipid nanoparticles that deliver genetic instructions directly to natural killer (NK) cells inside the body. The cells become CAR NK cells: engineered to recognise and destroy diseased T cells.

Why NK cells rather than T cells

The choice is not a detail. Using CAR T cells against malignant T cells — T cell lymphomas, for instance — creates a problem: the engineered cells and their targets are both T cells, so the weapons can attack each other. NK cells belong to a different lineage and can be aimed at diseased T cells without that self-destruction.

That opens two fronts at once: cancers of T cells, and autoimmune diseases in which misdirected T cells attack healthy tissue.

How the engineering works inside the body

The particles are made of fatty molecules and carry messenger RNA — the instructions for a chimeric antigen receptor. Their surfaces are decorated with antibody fragments that bind receptors on NK cells, steering the cargo to the right address. The receptor is swappable: it can recognise CD19, a protein on B cells, or CD3, found on T cells.

After injection, the NK cells produce the receptor themselves and become hunters of the chosen target.

What happened in mice

- About 80% of the targeted NK cells were reprogrammed. - In an aggressive melanoma model, three of five treated animals reached complete remission. - In humanised mice carrying T cell lymphoma, tumours were eliminated entirely after five injections — comparable to CAR NK cells manufactured the conventional way outside the body. - In autoimmune hepatitis, the treatment eliminated the autoreactive T cells and damaged liver tissue healed. That took a single injection.

The senior author is Daniel Anderson, a professor in MIT's Department of Chemical Engineering and a member of the Koch Institute for Integrative Cancer Research and the Institute for Medical Engineering and Science. The lead authors are postdoc Yuang Chen and graduate students Julia Monkovic and Haseeb Mughal; Robert Langer, MIT's David H. Koch Institute Professor, also contributed. The work is published in Nature Biotechnology and was funded by Sanofi, the Marble Center for Cancer Nanomedicine and a Koch Institute Support (core) Grant from the National Cancer Institute.

What it would change — and what is still missing

An infusion product instead of bespoke cell manufacturing changes the economics: no apheresis centre, no weeks-long logistics chain, no per-patient production run. In principle a clinic without transplant infrastructure could administer it.

But the results come from mouse models. No CAR NK therapy has been approved by the FDA, and the open questions are not small: how long the reprogramming lasts, how many doses a lipid nanoparticle can be given before the immune system reacts against it, how precise target recognition stays in a human body, and what happens when healthy cells carry the same marker.

The striking part is the architecture. The drug is a set of instructions, not a finished cell product — which makes the patient's body the factory, and moves the hard problem from logistics to biology.