Beta-Cell Replacement and Regeneration: Building New Cells

Why beta-cell replacement is hard, in plain English

The pancreas does not regrow its insulin-producing cells. Once the immune attack has destroyed enough of them, the body cannot put them back. That is one of the things that makes T1D the condition it is. Beta-cell replacement is the part of the field that asks a different question: if the body cannot rebuild what it has lost, can we put cells in from the outside that take over the job?

Two problems sit immediately in the way. The first is supply. Where do new beta cells come from in the numbers needed? The second is the immune system, which has not stopped recognising beta cells as something to attack. New cells dropped into a person with T1D face the same immune environment that took the originals. Every route in this part of the pipeline is solving for those two problems at once. The Insel 2015 staging consensus (Diabetes Care) gave the prevention end of the field its shared language; the replacement end is still developing the equivalent.

Solving supply pushes the field from cadaveric donor islets toward stem-cell-derived islets that can be manufactured at scale. Solving the immune problem pushes toward either suppressing the recipient’s immune system (which works but brings its own burden) or hiding the new cells from the immune system (which is harder but, if it works, removes the burden). Three routes have emerged from those two problems, and where each route sits today is what the next section is for.

Three routes, three trade-offs

Each of the three routes is doing something different to source cells, and something different to manage the immune response. The diagram lays them side by side; the rest of this section is what each one means in practice.

Route 1 is cadaveric islet transplantation. The cells come from a deceased donor’s pancreas; usually multiple donors are needed for one recipient because the yield from each donor is limited. The cells are infused through the portal vein into the liver, where they engraft. Because the cells are from another person, the recipient stays on chronic immunosuppression, the long-standing baseline for islet transplantation; the regimen is similar to a solid-organ transplant. Eligibility is restricted, in current UK and US practice, to adults with severe hypoglycaemia unawareness who have already optimised their AID and CGM and still have repeated severe hypos that put them at risk. It is real, and it is reachable through specialist services in selected centres; access is limited by donor supply and by the burden of lifelong immunosuppression.

Route 2 keeps the immunosuppression but replaces the cell source. Stem-cell-derived islets can be manufactured at scale, which solves the supply ceiling that has kept Route 1 small. The Vertex VX-880 / zimislecel programme is the lead candidate (Reichman 2025, NEJM, phase 1/2, 12 participants, all with at least 12 months follow-up); ten of the twelve participants no longer required exogenous insulin at the data cut-off, on chronic immunosuppression. The phase 3 trial is now under way. The immune burden in Route 2 is the same as Route 1; the gain is in scalability of supply.

Route 3 is the route that tries to remove the immunosuppression altogether. Two strategies are being tested. The first is gene-editing the cells so the immune system does not recognise them; the second is putting the cells inside a physical device that lets glucose and insulin through but blocks immune cells. Carlsson and colleagues (2025, NEJM) reported the first peer-reviewed first-in-human result on the gene-editing route: the UP421 study, a single participant, gene-edited cadaveric islets transplanted into the forearm, no immunosuppression, with C-peptide production and glucose responsiveness sustained at 12 months. Encapsulation routes (ViaCyte / CRISPR Therapeutics, Sernova Cell Pouch) are in earlier development. Earlier-phase encapsulation work (including the discontinued Vertex VX-264 in March 2025) is the field’s reminder that this route is genuinely hard.

What the Vertex result actually shows, and what it does not

The Reichman 2025 result is the most clinically advanced piece of evidence in the field, and it deserves to be read carefully. Twelve adults with T1D and severe hypoglycaemia unawareness received a single intraportal infusion of zimislecel under continuous immunosuppression. All twelve restored some endogenous insulin secretion. All twelve achieved time in range above 70% and HbA1c under 7%. Ten of the twelve no longer required exogenous insulin at the data cut-off (October 2024, with at least 12 months follow-up on every participant). Insulin requirements across the cohort fell by an average of 92%. For a person whose primary problem was severe hypoglycaemia unawareness, those numbers are real and they are large.

What the result does not show is also worth being plain about. Twelve participants is a small cohort; phase 3 is the next step and it is the trial that will tell us how the result generalises to a wider, less-selected population. The follow-up is around a year on most participants; T1D durability is a question that is answered in decades, not in years, and we do not yet know how long the engrafted cells will keep functioning. The immunosuppression is chronic, with the trade-offs that brings: cancer surveillance, infection risk, drug interactions, the burden of taking medicines daily for life. For an adult living with severe, dangerous hypos that AID and CGM have not solved, that trade-off can be net-positive; for a child diagnosed at four with a working hybrid closed-loop system, it is a different calculation entirely.

The honest framing is that VX-880 is a serious step forward in beta-cell replacement under chronic immunosuppression for severely hypo-unaware adults. It is not yet a cure for the broader T1D population, and it has not been positioned by the field as one. The phase 3 trial will tell us where the population edge sits.

Why Carlsson 2025 UP421 matters, and why n=1

The Carlsson UP421 result sits at a different point on the road. It is a single-participant first-in-human study, conducted at Uppsala University Hospital, that used gene-edited cadaveric islets transplanted into the forearm without any immunosuppression. The peer-reviewed twelve-week findings were published in the New England Journal of Medicine in 2025 with an accompanying editorial: cells alive on PET-MRI imaging, circulating C-peptide, glucose-responsive insulin secretion, and no safety issues. The twelve-month follow-up, confirmed by the sponsor’s corporate update in March 2026, reports continued cell survival, continued C-peptide production, and no safety issues at fifty-two weeks.

What that result proves is the mechanism. The Hypoimmune Platform gene-editing approach can keep allogeneic human islet cells alive, glucose-responsive, and immunologically invisible in a person with T1D for at least a year, without immunosuppression. The fundamental concept works in a human body. That is genuinely hopeful, and it is the result the field has been working toward for a long time.

What the result does not yet prove is how it generalises. n=1 means one participant; we do not know whether the next ten people with the same procedure will have the same outcome. We do not know whether the result holds at five years or ten. We do not know whether the same approach using stem-cell-derived islets (the next-generation candidate, SC451, with phase 1 trials expected in 2026) will perform the same way as the cadaveric-derived UP421 cells. The honest reading is that this is a hypothesis the field is now testing, not yet a therapy. It belongs in the conversation as proof of mechanism, with the qualifier that proof of mechanism is the point at which the next generation of trials begins, not the point at which a clinic opens.

What the family conversation looks like today

For an adult with severe hypoglycaemia unawareness whose AID and CGM have not restored safety, cadaveric islet transplantation is real, it is an established clinical service in selected centres, and the route in is a referral from the diabetes team to a specialist transplant service. The trade-off is the lifelong immunosuppression, which is the conversation that the transplant service is set up to have in detail. For the same population, Vertex VX-880 (zimislecel) is in active phase 3 trials and the route in is research enrolment through the team; eligibility is narrow, but the conversation is allowed and the team is set up for it.

For everyone else, the family conversation about beta-cell replacement is not yet a clinical conversation; it is a research-pipeline conversation. The technology is moving; the pivotal trials over the next few years will tell us how fast it reaches the point where commissioners and clinicians can offer it more broadly. In the meantime, what helps the family is keeping the rest of the management as clean as possible: the AID system tuned, the CGM data trusted, the structured education in place, and the team’s contact details on the fridge for when the question is clinical rather than speculative. Asking the team where the trial pathways live near you is a fair question; the team will know what is open and where, and the answer changes year on year.

References for this part