Quick answer
A phase I trial at the Third Affiliated Hospital of Sun Yat-sen University in Guangzhou gave mesenchymal stem cells grown from gum tissue to 18 adults with type 1 diabetes, in what the team describes as the first randomised, double-blind, placebo-controlled trial of the approach. Participants had four infusions three weeks apart and were followed for 48 weeks. Five of the nine people who received the cells (55.6%) showed some improvement in residual beta-cell function, and one stopped injected insulin after the first infusion while keeping good glucose control [1][2]. Results were published in Signal Transduction and Targeted Therapy on 1 October 2026 [2].
Key takeaways
- 18 adults with type 1 diabetes were randomised 1:1 to gum-derived mesenchymal stem cells (GMSCs) or placebo; each had four intravenous infusions three weeks apart, with 48 weeks of follow-up [1][2].
- C-peptide — a marker of how much insulin the body still makes — rose from baseline in the treated group, and 55.6% (5 of 9) had measurable improvement in residual beta-cell function [2].
- One participant stopped exogenous insulin after the first infusion and kept good glucose control through follow-up. That is one person, not a rate [1][2].
- No serious adverse events were reported, and the treatment did not add to the risk of hypoglycaemia or other adverse events [2].
- The proposed mechanism is that the cells push the immune T cells that attack the pancreas into an exhausted state — supported by single-cell sequencing and animal work, not yet proven in people [2].
The condition in context: a disease that starts early and lasts a lifetime
Type 1 diabetes is an autoimmune disease: the immune system mistakes the insulin-producing beta cells in the pancreas for a threat and destroys them. Once they are gone, the body cannot make insulin, and insulin has to be injected or pumped for life. Insulin keeps people alive and well, but it does not stop the underlying attack, and it does not restore the cells that were lost — which is why protecting what remains has been a long-standing goal.
| Region | What the data shows | Source |
|---|---|---|
| Worldwide | About 9.15 million people living with type 1 diabetes — 1.81 million of them under 20 (19.8%), 6.28 million aged 20–59 (68.6%) and 1.06 million aged 60 or over (11.8%). | IDF Global Diabetes Map 2025, as cited in a Shanghai Securities News disclosure, 8 July 2026 [3] |
| China | About 599,000 people with type 1 diabetes, of whom 117,000 are under 20. | IDF Global Diabetes Map 2025, same citation [3] |
| United States | In the SEARCH for Diabetes in Youth study, prevalence of type 1 diabetes at ages 0–19 was 2.00 per 1,000 and incidence 23.6 per 100,000 person-years — measured in 2001 and 2002–2005 respectively, and higher than rates reported in many European countries. | Bell RA et al., Diabetes Care 2009;32(Suppl 2):S102–S111, via CDC Stacks [4] |
Those numbers count different things in different years, so treat them as scale rather than comparison. The practical point is that type 1 diabetes usually begins in childhood or young adulthood and then lasts for decades — which is why a therapy that slows the loss of beta cells, even modestly, would matter a great deal if it holds up in larger trials [3].
What happened
The trial was run at the Third Affiliated Hospital of Sun Yat-sen University by the endocrinology team of Chen Yanming together with the group of Zheng Songguo of Shanghai Jiao Tong University School of Medicine's Songjiang Institute and its Institute of Cell and Gene Therapy, under a national key research and development programme [2].
Eighteen people with type 1 diabetes were randomised 1:1 to GMSCs or placebo. Each had an intravenous infusion every three weeks, four times in total, and was followed for 48 weeks. The trial is registered as ChiCTR1800018339 [1][2][5].
The results were accepted by Signal Transduction and Targeted Therapy, a journal with a reported impact factor of 81.2, and published online on 1 October 2026. Tang Qi, Zhou Li, Fan Liwen and Zhang Wei are listed as co-first authors; Chen Yanming and Zheng Songguo as equal corresponding authors [2].
Mesenchymal stem cells can dampen immune responses, and they have been taken from bone marrow and umbilical cord tissue before. Gum tissue is easier to reach, the cells multiply faster in the laboratory, and they are less likely to provoke an immune reaction against themselves — so a small piece of gum can, in principle, yield a treatment dose with less invasive collection than bone marrow [5]. Previous work in animals suggested they could protect pancreatic islets, but there was no randomised human evidence in type 1 diabetes before this trial [2].
The case as reported
Accounts of individual participants come from the reports of the teams involved; they are the investigators' account, not an independent review.
Among the nine people who received the cells, one stopped injected insulin after the first infusion and, over the period of follow-up, maintained what the team describes as ideal glucose control — a level of control usually reached with insulin, not without it [1][2]. Across the treated group, daily insulin requirements fell from baseline, and C-peptide levels rose. The reporting does not say how long the effect lasted beyond the 48-week follow-up, or whether the person who stopped insulin later restarted it [2].
Why it matters, and what it does not settle
| What it tells you | What it does not tell you |
|---|---|
| A randomised, placebo-controlled design in humans is a real step up from animal work, and the safety signal over 48 weeks was clean [2]. | Whether the improvement in beta-cell function lasts, or translates into fewer complications years later. |
| Nine treated patients is enough to see a signal, not enough to know the true response rate. | How many people would benefit in routine care. 55.6% is 5 out of 9. |
| The proposed mechanism — inducing exhaustion in the CD8+ T cells that attack the pancreas — is supported by sequencing and animal data [2]. | That the mechanism is the actual reason for any benefit in people; that needs larger trials. |
Availability and limits
- This is a phase I trial. Phase I is primarily about safety. It is not evidence that the treatment works, and it is not an approved therapy.
- 18 people. Small trials produce unreliable effect sizes, and a result this size often shrinks in larger studies.
- "Improvement" is not remission. Higher C-peptide and lower insulin needs are meaningful, but none of the reporting claims participants were cured or that insulin was stopped in more than one case.
- The treatment is not available outside a trial. We found no published statement that GMSC therapy for type 1 diabetes can be obtained as routine care, in China or elsewhere.
- Follow-up was 48 weeks. Durability beyond that is unknown.
What to ask before pursuing any stem cell treatment
| Question | Why it matters |
|---|---|
| Is the treatment part of a registered trial, and what is the registration number? | Unregistered "stem cell therapy" sold to medical travellers is a known problem worldwide. |
| What phase is it, and how many patients have been treated? | Phase I says little about whether it works. |
| What are the measured outcomes, and over how long? | C-peptide is a surrogate marker, not a measure of how you feel or of long-term complications. |
| What does it cost, and is it covered by insurance or research funding? | Experimental cell therapies are usually self-funded. |
| Would a non-resident be eligible, and is English support available? | Trial protocols usually have strict inclusion criteria and local follow-up requirements. |
Questions patients ask
What are gum-derived mesenchymal stem cells?
Mesenchymal stem cells are cells that can develop into several tissue types and that dampen immune activity. These ones are grown from a small sample of gum tissue. Compared with bone marrow or umbilical cord sources, gum tissue is easier to sample, the cells multiply faster in the laboratory and they are less likely to trigger an immune reaction [5].
Does this mean type 1 diabetes can be reversed?
No. In this trial, C-peptide rose and insulin needs fell in the treated group, and one of nine people stopped injected insulin during follow-up [1][2]. That is a signal in 18 patients over 48 weeks, not a reversal, and not an approved treatment.
What is C-peptide and why is it measured?
C-peptide is a fragment released when the body makes insulin, so it is used as a marker of how much insulin a person still produces themselves. A rise in C-peptide suggests the surviving beta cells are working better. It is a laboratory measure, not a clinical outcome [2].
Is this treatment available now?
We found no published statement that it is available outside a clinical trial. It is registered as ChiCTR1800018339 [5]. Anyone considering stem cell treatment should check the registration and ask what phase the study is in.
How is this different from stem cell treatments offered abroad?
The distinction is evidence: this trial was randomised, double-blind and placebo-controlled, with a published protocol number and peer-reviewed publication. Clinics that sell unproven stem cell infusions usually cannot show either [2][5].