For the roughly 8.7 million people worldwide living with type 1 diabetes, daily life revolves around a problem their bodies cannot solve: the immune system has destroyed the beta cells that produce insulin, and nothing available today can bring them back. A pair of new mouse studies, along with advancing clinical programs from two biotech companies, suggest that replacement cells grown in the lab may finally be closing that gap.
In the more novel of the two studies, published in Frontiers in Endocrinology and indexed on PubMed (PMID 39279997), researchers took non-endocrine cells, engineered them to produce human insulin, loaded them onto tiny scaffold beads called microcarriers, and implanted them into mice whose beta cells had been chemically destroyed. The transplanted cells did something that matters enormously: they lowered blood glucose after the mice ate, meaning the cells could sense rising sugar and respond, not just leak insulin at a flat rate.
A design built to last
What sets this experiment apart is the engineering under the hood. Previous attempts to force cells into making insulin often relied on powerful viral promoters to switch on the insulin gene. The problem is that cells tend to recognize those promoters as foreign and shut them down within weeks or months, a phenomenon called gene silencing. The Frontiers team used a different architecture: an internal ribosome entry site, or IRES, which piggybacks on the cell’s own protein-making machinery rather than introducing an outside switch. In theory, that makes the insulin output harder for the cell to mute.
The theoretical advantage is real, but the published data covers a limited observation window. The paper does not yet show whether insulin production holds steady for months, whether the engineered cells grow in uncontrolled ways, or how the mouse immune system treats them over time.
Stopping the immune attack
Building a reliable insulin-producing cell is only half the challenge. In type 1 diabetes, the immune system is the original villain, and it will attack transplanted cells just as it attacked the originals. A second mouse study, published in the Journal of Clinical Investigation (DOI 10.1172/JCI190034), tackled that problem directly.
That team combined hematopoietic stem-cell transplants with islet-cell transplants after what the paper describes as “gentle” conditioning using a CD117 antibody. Traditional conditioning regimens borrow from chemotherapy and carry serious toxicity. The antibody-based approach clears space in the bone marrow for new immune stem cells without the same collateral damage. In the mice, the strategy both reversed existing autoimmune diabetes and prevented it from coming back, with durable results reported across the study’s endpoints.
The distinction between the two papers is important. The Frontiers study answers the question: Can we build a cell that reliably makes insulin? The JCI study answers a different one: Can we stop the body from destroying transplanted cells? A working human therapy would almost certainly need to solve both problems at once.
Industry is already in the clinic
While academic labs refine the science in mice, two publicly traded companies have moved related approaches into or toward human testing. Vertex Pharmaceuticals has disclosed three programs in SEC filings: zimislecel (VX-880), a stem-cell-derived islet therapy that has already been administered to human patients in clinical trials, with early data showing restored insulin production in some participants; VX-264, which encases similar cells in a protective device; and a hypoimmune gene-editing program designed to make transplanted cells invisible to the immune system.
Sana Biotechnology, in its fiscal year 2025 annual report, describes first-in-human plans for its candidate UP421, with C-peptide production (a direct marker of new insulin being made inside the body) as an endpoint. Sana is also developing a scalable stem-cell-derived product called SC451.
Both companies’ filings include extensive risk factors and cautionary language, a legal requirement that also serves as a reality check. When a filing says a therapy “may” restore insulin function, that phrasing is chosen deliberately. Timelines are contingent on trial results, manufacturing scale-up, and regulatory decisions that have not yet been made.
Crucially, neither Vertex nor Sana references the IRES-based construct from the Frontiers paper. Their platforms use differentiated stem-cell-derived islet cells or proprietary engineering, not the promoter-free gene architecture the academic team developed. Any connection between the mouse work and these commercial pipelines is, as of June 2026, speculative.
What could still go wrong
The gap between a successful mouse experiment and a safe human therapy remains wide, and several specific uncertainties deserve attention.
Hypoglycemia risk. Engineered cells that produce insulin continuously or in excess could drive blood sugar dangerously low. The IRES system ties insulin output to the cell’s overall translational activity, which generally tracks metabolic demand, but whether that correlation is precise enough to prevent glucose crashes during exercise, fasting, illness, or use of other medications has not been tested outside controlled lab conditions.
Immune rejection in diverse patients. The JCI study’s conditioning protocol worked in inbred mouse strains that are genetically identical. Human immune systems are wildly diverse. Translating antibody-based conditioning to real patients introduces variables that preclinical work cannot fully model, including infection risk during the conditioning window, off-target effects on blood cell production, and the challenge of dosing an antibody regimen for a chronic disease rather than a one-time procedure.
Durability. Neither mouse study demonstrates years-long function, which is the minimum bar a human therapy would need to clear to justify its complexity and cost over existing insulin-pump and continuous-glucose-monitor systems that, while burdensome, keep many patients alive and reasonably well.
Manufacturing and cost. Even if the biology works, producing personalized or semi-personalized cell therapies at scale is an unsolved industrial problem. Vertex’s early clinical work required immunosuppressive drugs to protect transplanted cells, adding cost and side effects. Encapsulation devices and hypoimmune editing aim to eliminate that need, but neither has yet proven it can do so reliably in large patient populations.
Where the field stands now
Cell-based insulin replacement has been a goal since the Edmonton Protocol demonstrated in 2000 that transplanted cadaveric islets could free some patients from insulin injections, at least temporarily. The limiting factors then were supply (donor pancreases are scarce) and immune rejection (patients needed heavy immunosuppression). A quarter-century later, the supply problem is being addressed by stem-cell differentiation and cell engineering, and the rejection problem is being attacked from multiple angles: encapsulation, gene editing, and gentler immune conditioning.
The two mouse studies and the corporate clinical programs represent different bets on how to clear the remaining hurdles. The IRES construct offers a potentially silencing-resistant way to produce insulin. The CD117 conditioning protocol offers a less toxic way to reprogram immunity. Vertex’s and Sana’s programs offer the manufacturing infrastructure and regulatory experience needed to move any of these ideas into patients.
None of these approaches alone is sufficient to cure type 1 diabetes in people today. But taken together, they map a plausible route from concept to clinic. The next decisive evidence will not come from more mouse experiments. It will come from carefully designed human trials that measure durability, safety, and quality of life over years, not weeks. For millions of people who check their blood sugar before every meal, that timeline cannot arrive soon enough.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
