A protein coating painted onto extracted human teeth is reported to regrow a thin layer of enamel-like mineral in the laboratory, rebuilding enamel's microscopic structure and recovering much of its hardness and stiffness. That is the central claim of "Biomimetic supramolecular protein matrix restores structure and properties of human dental enamel," by Abshar Hasan and colleagues, received 8 March 2025 and accepted 1 October 2025. No patient was treated. The author list runs to 25 names ending with Alvaro Mata, and includes José Carlos Rodríguez-Cabello, Stefan Habelitz and Avijit Banerjee. Shuo Chen, the dossier's other named figure, is not on it.
What problem the paper addresses
Once Dental Enamel is lost, the body cannot replace it. The paper ties enamel's toughness to its composition and its architecture: tiny crystals lined up and bundled in set ways, with a thin prism-free skin at the surface over a deeper region organised into prisms. In the authors' telling, earlier repairs fail on one count or another. Several older treatments can lay down mineral, but not enamel's ordered structure. Engineered composites need fabrication conditions far from those in the body. Bioinspired peptide and ion treatments contain toxic ingredients such as triethylamine, take from 10 minutes to 12 hours to apply, or rebuild only one region of enamel. The team set out to make a coating that could rebuild every region in a way a clinic could use.
How the coating is said to work
During normal tooth development the protein amelogenin starts out disordered and then assembles into fibrils rich in β-sheet structure, which guide crystal growth. The paper's idea is to imitate that scaffold with an Elastin-like Recombinamers (ELRs) molecule, an engineered protein made of three repeating parts: a water-avoiding VPGIG motif, a water-loving VPGKG motif and an acidic segment taken from the salivary protein statherin. The authors report that adding calcium ions (1.5 mM) and letting the solution dry makes the molecules stack into fibrils 15–40 nm wide and several micrometres long. Without calcium, no fibrils formed.
X-ray scattering gave spacings of 4.7 Å and 10 Å, a pattern typical of stacked β-sheets. Another signal at 48.7 Å was read as the width of a single filament. Computer simulations had VPGIG forming each filament's core while calcium bound to the statherin segments. From these results the authors built a model in which each fibril of about 300 Å contains roughly six filaments. They note that 48 Å filaments have been reported in the enamel matrix of embryonic cattle. The calcium bound in the fibrils is also meant to act as a starting point for new crystals. This is the paper's version of the Biomimetic Enamel Mineralization approach.
What was actually done
All tooth work used human molars extracted for clinical reasons, under University of Nottingham ethics approval FMHS 313-0721. Tooth sections 0.5–1.0 mm thick were roughened with 37% phosphoric acid to mimic erosion. They were then coated with a crosslinked ELR solution and left in a mineralising solution at 37 °C, usually for 10 days. That solution was supersaturated with fluorapatite and carried more fluoride than saliva does, and the new crystals are fluorapatite. The authors report detecting fluorine in the new crystals but not in the native enamel, and they credit it for some results, including better acid resistance. They also report that the new layer's chemical composition is similar to native enamel's. To approach mouth conditions, the team repeated runs in artificial saliva with 1 ppm fluoride, in an enzyme-supplemented saliva stirred with steel balls, and in whole saliva from three donors for two weeks. A later formulation swapped the original solvents (DMF/DMSO with an HDI crosslinker) for ethanol-water with glutaraldehyde, and the coating was then reported to set in 3–4 minutes.
What the paper reports
The mechanical figures use a few measures. Stiffness (E, Young's modulus) describes how little the material bends under heavy force, as in chewing. Hardness (H) describes how well it resists dents and scratches. Wear strength combines friction and how fast material is worn away, and the paper presents it as a measure of resistance to fatigue failure. All are in gigapascals (GPa). The durability results come from laboratory tests that simulate brushing and chewing, not from use in real mouths.
- Structure. New crystals grew outward from the original enamel crystals in the same orientation, a process called epitaxial growth. The new layer copied prismatic, inter-prismatic and aprismatic patterns. On bare dentine it formed an enamel-like layer joined to collagen fibrils. The mineral layer was about as thick as the coating, up to about 10 μm.
- Hardness and stiffness on prismatic enamel. Acid-etched enamel measured E = 36.9 ± 14.3 GPa and H = 1.1 ± 0.6 GPa. After treatment it measured 76.3 ± 18.7 GPa and 3.1 ± 0.8 GPa, against 80.7 ± 18.3 GPa and 3.4 ± 0.9 GPa for native enamel.
- Wear strength. This rose from 91.6 to 171.3 GPa, above native enamel's 153.9 GPa.
- Fracture toughness. This went from 0.7 to 1.1 MPa·m^0.5, against 1.2 for healthy enamel.
- Durability. After 60 minutes of continuous electric brushing in the laboratory, said to equal about one year of use, treated enamel measured E = 68.6 ± 16.4 GPa and H = 2.76 ± 0.8 GPa, which the authors call similar to native enamel's 70.2 and 3.0 GPa. The legend for that panel, which covers native and treated enamel after 15 and 60 minutes of brushing, marks one significant difference (p = 0.0055) without saying which pair was compared. Wear after two weeks of simulated chewing at 75 N, said to equal about 3.5 years, was reported as similar to native enamel: volume loss of 3.5 ± 1.4 mm³ against 2.9 ± 1.0 mm³, and height loss of 30.5 ± 8.9 µm against 29.1 ± 9.3 µm. In acid at pH 4.0, treated enamel kept more stiffness than native enamel, which the authors attribute to fluorine in the new mineral.
- Layer on dentine. The enamel-like layer measured E = 58.3 ± 16.7 GPa and H = 1.4 ± 0.3 GPa, which the authors call comparable to the native enamel range (E = 50–90 GPa, H = 2.5–4 GPa). They attribute the lower hardness to a difference in how the nanocrystals are oriented on native enamel and on remineralised dentine.
- Donor saliva. Treated enamel measured E = 74.6 ± 15.3 GPa and H = 2.7 ± 1.1 GPa.
What the evidence does and does not show
The study is entirely laboratory work on sectioned extracted teeth. Its groups are small. Friction, wear strength and fracture toughness used n = 3 samples per group, and the chewing tests used n = 8. Indentation data were pooled from 3 independent experiments, and some electron-microscopy proof of crystal continuity rests on n = 2. Etching with phosphoric acid in a few seconds or minutes is a stand-in for real erosion and decay. The methods describe digesting the coating with elastase for 10 days at 37 °C after mineralisation to expose the new crystals. The authors say the coating slowly degrades during mineralisation and that the enzyme step speeds this up; the paper does not show how the coating would be handled or cleared in a patient.
The authors are open about the limits. They write that their tests "do not fully recreate the complexity of the in vivo oral environment", and they state that confirming regeneration of natural enamel would need work in living mouths, which they plan for the future. They also note that some studies dispute whether enamel and dentine are crystallographically continuous. The paper's only commercial comparison is the coating's own hardness and stiffness against Duraphat varnish; it does not test any remineralising product's performance on teeth. Even so, the authors conclude that the coating's capability "far surpasses current commercial dental remineralisation alternatives". That comparison is asserted rather than measured, since no such product was tested head to head. Cell tests on three cell lines showed the coating was compatible, but only in vitro.
How it fits the dossier
The paper's details bear directly on the press and patent claims weighed in Enamel Regrowth Claims vs. Evidence: The Press Summary, the Abstract and the Patent Text. Its mineralisation periods run from 10 days to two weeks, though the paper itself does not say whether these are what the press phrase "within weeks" refers to. Its numbers come from laboratory tests on tooth sections, not from patients. Its material is a statherin-bearing ELR, which differs from the amelogenin systems in the dossier's Chinese and US patents. The paper does not say whether it relates to the polypeptide-and-calcium patent applications such as US2025145663A1 — Substrates Comprising Elastin-Like Polypeptides and Calcium Ions. For contrast, Amelogenin Peptide TRAP Against Fluoride: A 2025 In Situ Test on Early Enamel Lesions (Bai et al.) tested a peptide on enamel softened and then remineralised in volunteers' mouths. That design comes closer to real use, though it measures a much narrower outcome.
Source notes & attribution
- https://pmc.ncbi.nlm.nih.gov/articles/PMC12586552/