Originally published 2017-01-28.
Can a drug make teeth repair themselves instead of needing a filling? In 2017, a team at King's College London showed that, in mice, it can at least help. They placed a small collagen sponge soaked in tideglusib, a drug once tested for Alzheimer's disease, into tiny holes drilled into mouse molars. Within about six weeks the teeth had laid down enough new dentine to fill the injury, and the sponge had broken down.
That is a genuinely interesting result, but it is not a cure for tooth decay and it is not available at the dentist. Here is what the study did, how natural tooth repair works, and what the findings do and do not mean.

What the King's College study actually did
The research, led by Paul Sharpe with Vitor Neves, Rebecca Babb and Dhivya Chandrasekaran, appeared in the journal Scientific Reports under the title "Promotion of natural tooth repair by small molecule GSK3 antagonists." The team tested three drugs that block an enzyme called glycogen synthase kinase 3 (GSK-3): BIO, CHIR99021 and tideglusib.
In adult mice, the researchers drilled into the upper first molars until the soft pulp at the centre was exposed. They then laid a collagen sponge soaked in one of the drugs against the exposed pulp and sealed the tooth with a standard glass ionomer cement. For comparison, some teeth received a sponge with no drug, and others received mineral trioxide aggregate (MTA), a material dentists already use to protect exposed pulp.
At four and six weeks, scans and tissue samples showed more mineral at the injury site in the drug-treated teeth. On average the authors reported roughly twice as much as the plain sponge and about 1.7 times as much as MTA. By six weeks the new dentine had largely filled the defect, and the pulp stayed alive.
The size of the holes
It helps to picture the scale. Mouse teeth are tiny, and the injuries in this work were well under a millimetre across, often described as about 0.14 mm. Human cavities are usually far larger, so the result is a proof of principle rather than a ready treatment.
How a tooth repairs itself
A tooth is not a dead block of mineral. Under the hard outer enamel sits dentine, the calcified tissue that makes up most of a tooth's bulk. Inside that is the pulp, which holds nerves, blood vessels and cells that can build more dentine.
When decay or a knock threatens the pulp, the tooth responds by laying down a thin protective layer of so-called reparative dentine. That natural defence works for small injuries but cannot close a large cavity, which is why dentists remove decayed material and pack the space with an artificial filling.
Sharpe summed up the motivation in an interview with The Guardian: "The tooth is not just a lump of mineral, it's got its own physiology. You're replacing a living tissue with an inert cement." Fillings work well, but a tooth that rebuilds its own tissue could, in theory, stay healthier for longer.
Where GSK-3 comes in
GSK-3 acts as a brake on a signalling system called the Wnt pathway, which helps stem cells divide and mature. By blocking GSK-3, the drugs release that brake. In the study, a marker of Wnt activity rose about threefold in pulp cells within a day of treatment, consistent with the idea that the drug woke up the tooth's own repair cells.
Why an Alzheimer's drug?
Tideglusib was developed as a treatment for neurodegenerative disease. It was tested in people with progressive supranuclear palsy and Alzheimer's disease, but those trials were stopped in 2011 and 2012 because it did not work well enough.
For the dental researchers, that history was useful. A drug that has already been given to people in clinical trials comes with existing safety data, which could, in principle, shorten the path to testing in dentistry. The collagen sponges used in the study are also already approved for clinical use. Neither point guarantees a fast route, but both remove some early hurdles.
What the headlines got wrong
Coverage at the time often said a drug could "regenerate teeth" and end fillings. A few corrections are worth making.
- It did not regrow whole teeth. The drug boosted dentine inside an existing tooth. Enamel, the outer layer, is not living tissue in adults and does not grow back.
- It was tested in mice only. The sample sizes were small, roughly six to nine teeth per time point, and the authors note the study was not blinded.
- "Complete" repair was a slight stretch. The paper's results describe almost complete filling of the defect at six weeks.
- It is not a substitute for dental care. No tideglusib tooth treatment is available to patients.
Ben Scheven, an oral cell biologist at the University of Birmingham who was not involved, welcomed the approach, telling The Guardian that "dentistry is not only about filling and drilling, but also about keeping the teeth healthy."
Other efforts to rethink the filling
The King's College team was not alone. Around 2015, a group in the UK reported work on a "pulp cap" material intended to encourage stem cells to form dentine, and an Australian study suggested that early decay could often be halted with fluoride varnish and better home care before drilling was needed. Approaches that catch decay early, rather than repair large cavities, are already part of modern preventive dentistry.
For its own approach, the King's College team said further animal work, starting with rats, would have to come before any trial in people. Until such trials report, GSK-3 tooth repair remains an experimental idea rather than a treatment option.
What you can do for your teeth now
While the science develops, the most reliable ways to avoid fillings remain unglamorous: brush twice a day with fluoride toothpaste, cut down on frequent sugary snacks and drinks, clean between your teeth, and keep regular dental check-ups so problems are caught while they are small.
If you have tooth pain, sensitivity or a visible hole, talk to a dentist. Early decay can sometimes be managed without drilling, but only a professional examination can tell.
Sources: Neves et al., "Promotion of natural tooth repair by small molecule GSK3 antagonists", Scientific Reports (2017); Tideglusib (Wikipedia); Dentin (Wikipedia); quotes from Hannah Devlin's report in The Guardian (9 January 2017); first reported here via ScienceAlert.

COMMENTS