Teaching the immune system to stop
New cells aren't enough: the immune system remembers beta cells and attacks them again. Hiding the cells is one answer. Teaching the immune system tolerance is the other, and much of the progress comes from other autoimmune diseases.
A cure needs two things: new insulin-producing cells, and an immune system that doesn’t destroy them. The second is harder than it sounds. Even a perfectly matched organ isn’t safe. In the 1980s, surgeons in Minnesota transplanted part of a pancreas from identical twins into their twin with type 1 diabetes, with little or no immunosuppression. Within 6–12 weeks the grafts stopped working, and biopsies showed the immune system had selectively destroyed the beta cells, and only them (Sibley et al., 1985; Sutherland et al., 1989). There was no rejection, since the tissue was genetically identical. The autoimmunity simply came back. The immune system remembers.
Two different answers
Hide the cells. Sana’s approach removes the HLA molecules T cells look for and adds CD47, a “don’t eat me” signal. It doesn’t change the immune system at all. In Uppsala’s first patient, the edited cells escaped both rejection and the autoimmune attack for at least 14 months (more).
Teach the immune system tolerance. If the immune system stops attacking beta cells, it protects the body’s own remaining cells too, and the disease itself is gone. There are four main routes:
| Route | Examples | Where it stands (November 2026) |
|---|---|---|
| Antigen-specific treatment, to teach tolerance to one target | Diamyd (GAD), oral insulin | Both failed in their latest large trials (2026) |
| Regulatory T cells, the immune system’s own brakes | GentiBio, City of Hope, PolTREG, Nanjing, MUSC bodyguard cells | First engineered versions entering human trials; no efficacy results yet |
| Resetting the immune system | CAR-T in lupus and other diseases, CD7 CAR-T in China | Striking results in lupus; in type 1 diabetes, only small safety studies |
| Slowing the attack | Teplizumab, baricitinib, low-dose ATG | Teplizumab approved for delaying onset; others in phase 2–3 |
Why other diseases matter
Type 1 diabetes doesn’t have to solve tolerance alone. Lupus, systemic sclerosis and myositis drive much of the work on CAR-T and regulatory T cells, and the results transfer.
The most striking is the immune “reset”. CAR-T cells, first developed for blood cancers, wipe out the B cells; when the immune system rebuilds, it seems to forget its autoimmunity. In the German CASTLE trial, 9 of 10 people with severe lupus were in remission at 24 weeks, all off steroids and other immune drugs (Nature Medicine, 2026). There was no control group, and the patients needed chemotherapy first.
Type 1 diabetes has a precedent. In Brazil, people aged 13–31 with newly diagnosed type 1 diabetes had their immune system wiped with high-dose chemotherapy and rebuilt from their own stem cells. 20 of 23 became insulin-free; 12 stayed so for a mean of 31 months (Couri et al., JAMA 2009). The concept worked, but the treatment was far too harsh to offer widely.
An open question is whether a B-cell reset is enough here. Type 1 diabetes is driven mainly by T cells, though B cells help direct the attack. That’s one reason Chinese teams are testing CAR-T aimed at T cells (CD7), which is a much stronger intervention.
The next generation may make cells unnecessary: CAR-T made inside the body, with genes delivered straight to T cells, possibly without chemotherapy. That would matter for children. It has reached first human studies in lupus, not yet in type 1 diabetes.
The honest picture
The field has passed proof of principle but hasn’t reached routine use. It’s concentrated in a few countries, a few diseases and early phases. And the bar for side effects is higher than in cancer: today’s CAR-T needs chemotherapy, specialised manufacturing and hospital care.
Where it could lead
The long-term picture is a combination: a gentle way to reset or calm the immune system, followed by new beta cells that then don’t need to be hidden. That is the second road to a real cure, next to cells the immune system can’t see.
Timing matters for the treatments that slow the attack. They work in the months around diagnosis, or before symptoms, which is why screening matters. For someone diagnosed years ago, cell replacement is the relevant road.