What was found
Kim, Jung, and colleagues at Yonsei University dissected E13 and E15 mouse mandibular first-molar tooth germs and separated the dental mesenchyme into lingual and buccal halves at the cap and bell stages. Bulk RNA sequencing showed that, at the cap stage, the lingual half was enriched for pattern-specification, odontogenesis, and proliferation genes, while the buccal half was enriched for stem-cell-development, regeneration, and neural-crest differentiation genes (Kim et al. 2025, Results; Fig. 1-3). The authors validated selected markers by RT-qPCR and in situ hybridization.
The functional test came in two parts. First, when whole, lingual-only, or buccal-only cap-stage tooth germs were cultured for two days and then transplanted under the kidney capsule for four weeks, whole germs and lingual fragments formed calcified tooth tissue with amelogenin-positive ameloblasts, whereas buccal fragments formed only surrounding bone (Kim et al. 2025, Results; Fig. 5). Second, in recombination experiments, lingual mesenchyme paired with either lingual or buccal epithelium produced calcified teeth, while buccal mesenchyme paired with either epithelium did not (Kim et al. 2025, Results; Fig. 6).
Positional memory survives dissociation
To test whether the lingual-buccal difference is an imposed tissue-context effect or a cell-intrinsic property, the authors dissociated lingual and buccal mesenchymal cells into single cells, reaggregated them separately or in mixed combinations, and recombined them with wild-type dental epithelium. Using Wnt1CreERT2;R26RtdTomato lineage tracing, they showed that lingual-derived cells contributed to dentin sialoprotein-positive odontoblasts and dental pulp, while buccal-derived cells contributed to fibrillin-positive periodontal-ligament cells and Sp7-positive osteoblasts in the surrounding tissue (Kim et al. 2025, Results; Fig. 7). The positional identity therefore persisted through complete dissociation and reaggregation.
The Wnt-BMP axis
The signaling distinction centered on R-spondin/Wnt activity in the lingual compartment and BMP-inhibitor activity in the buccal compartment. When cultured alone, lingual mesenchyme upregulated Rspo1, Rspo2, and Rspo4, while buccal mesenchyme upregulated Grem1, Chordin, Chrdl2, and Noggin. In a microslide chemotaxis assay, lingual cells migrated toward R-spondin1 conditioned medium and buccal cells migrated toward Noggin protein (Kim et al. 2025, Results; Fig. 8-10). The authors propose that Wnt activation favors tooth formation and BMP-inhibitor activity favors surrounding tissue formation along the lingual-buccal axis.
What it means for bioengineered tooth germs
For the bioengineered tooth germ program, the paper adds a positional-information requirement to the engineering problem. Reconstituting a tooth germ from dissociated cells is not just a matter of mixing the right epithelial and mesenchymal populations; the mesenchyme must retain or be re-established with the correct lingual-buccal organization. The finding that cells remember their position after dissociation is encouraging for reaggregation approaches, but it also implies that engineering protocols must preserve or recreate the Wnt-low/BMP-inhibitor-high buccal and Wnt-high lingual pattern to produce both the tooth and its supporting apparatus.
What it does not show
The work is in embryonic mouse tooth germs, not human cells, and the transplants matured under the kidney capsule rather than in a functional jaw. The study does not demonstrate eruption, occlusion, innervation, or long-term function. It also does not establish how transferable the lingual-buccal positional code is to induced pluripotent stem cell-derived or adult-derived dental cells, which is the cell-sourcing problem the field still faces.
Where we differ from the coverage
Press reports framed the result as mapping the cells that “know how to grow teeth” or as a step toward regrowing human teeth. The paper itself does not claim human translation; it identifies positional information in mouse embryonic dental mesenchyme. The path from this developmental insight to a human protocol for growing a third dentition remains speculative and long.
Provenance: grounded in the open-access full text of Kim et al. 2025, International Journal of Oral Science 17:67, DOI 10.1038/s41368-025-00391-7. Method and sourcing standard at /method/.