What the study is

Huen Li, Nianzuo Yu, Haofeng Liu and colleagues at Jilin University’s Hospital of Stomatology and its mathematics department, with corresponding authors Bei Chang, Xiaoduo Tang, Junhu Zhang and Ting Ye, published “A Wet-and-Dry Research Model: Unveiling Biomechanics in Odontoblast Polarization” in the International Endodontic Journal on 2 June 2026 (doi:10.1111/iej.70188; free full text at PMC13569296). The target is a known gap in the pulp-dentin repair route: regenerated dentin is only tubular, and therefore only functional, if the cells secreting it establish the polarized, columnar shape of odontoblasts, and the mechanical cues that drive that shape change were poorly characterized. The group’s answer is a “wet-and-dry” platform. The wet side prints adhesive micropatterns in PEGDA hydrogel by photolithography, shaped as a rectangle plus a triangular extension meant to mimic an odontoblast cell body and its process, and seeds single human dental pulp stem cells (hDPSCs, isolated from routinely extracted impacted third molars of healthy adults, passages 3 to 5) onto them. The dry side is a vertex-based computational model of the same adhesion-and-spreading process, whose force predictions are checked against the experiments. The platform is reported against the PRILE and PRIASE 2021 guidelines, with at least three independent experiments and at least 30 cells per group.

What the geometry did

Shape was not a detail here, it was the variable. Across patterns of equal 2700 square micrometers, raising the aspect ratio from 1:1 to 1:4 progressively increased every one of six polarization metrics, which track Golgi position, nuclear displacement and nuclear deformation. Differentiation markers followed the same gradient: the alkaline-phosphatase-positive fraction rose from 12.02 plus or minus 1.89 percent on 1:1 patterns to 25.04 plus or minus 1.73 percent on 1:4 patterns (p less than 0.0001), and DSPP fluorescence rose with it. Mechanically, elongated cells showed more vinculin at focal adhesions (mean fluorescence 10.94 plus or minus 1.06 on 1:1 versus 18.59 plus or minus 0.85 on 1:4, p less than 0.0001), more myosin II, and more nuclear YAP1. The vertex model independently predicted that traction force peaks concentrate at the cell periphery, most prominently at the tips of elongated cells, while adhesion forces spread more broadly, with both rising as aspect ratio grows, matching the measured morphology. The caveat: this is single human cells on a synthetic 2D substrate over about a week, not dentin formed in a tooth.

The mechanism chain, and the tooth germ check

The inhibitor experiments build one causal chain: disrupting actin with cytochalasin D or microtubules with nocodazole collapsed the geometry-induced polarization; blocking the RhoA-ROCK axis with Y-27632 collapsed stress fibers, reduced nuclear YAP1 and cut the odontogenic readouts; blocking YAP directly with verteporfin did the same, placing YAP downstream of RhoA in the authors’ model. That is the vinculin-RhoA-YAP axis of the paper’s conclusion. The ex vivo leg gives it tissue context: E14.5 mouse mandibular first molar tooth germs cultured with Y-27632 developed reduced tissue dimensions and a disorganized odontoblast layer, with polarization metrics falling accordingly. One honest wrinkle the authors report themselves: Rho GTPase inhibition disrupted Golgi positioning without abolishing odontoblast process formation, so the two events are regulated in parallel rather than in strict sequence. The computational model is explicitly phenomenological: it reproduces shape and force distribution but does not simulate molecular recruitment such as vinculin accumulation.

Boundary and what it changes

This is in vitro human cell work plus ex vivo mouse tissue; no dentin regeneration was shown in an animal, and the authors flag missing live-cell tension sensors and the limited physiological realism of the patterns as next steps. The pulp-dentin repair program page already records the human autograft trial and preclinical small-molecule dentin repair; what this paper adds is not evidence upward in that tier structure but a named, druggable mechanotransduction axis and a concrete design dial, scaffold geometry, that future pulp-capping and regeneration materials could be tuned against. Read next to this site’s September 24 record of YAP suppression in dental epithelial progenitors under orthodontic load, it completes a useful symmetry: the same YAP-centered mechanosensing machinery now has an experimental foothold on both the epithelial and the mesenchymal side of the tooth, in models that are each a step away from a whole, loaded jaw. The field assessment does not move on this record; it sharpens a mechanism inside the repair route.

Provenance: grounded in the full text of Li et al. (2026), International Endodontic Journal 59(10):2103-2121, doi:10.1111/iej.70188, retrieved via PubMed Central (PMC13569296) on 26 September 2026. See /method/.