What was tested

Caroline Anselmi, Sepideh Aminmansour, and colleagues, led by Marcos Bottino’s group with collaborators at the University of Michigan, University of Florida, UNESP Araraquara, and UFMG, report in Biomimetics (volume 11, issue 8, article 518, published 23 July 2026) a head-to-head comparison of three single-polymer electrospun scaffolds: polycaprolactone (PCL), polydioxanone (PDO), and photocrosslinked gelatin-methacryloyl (GelMA) (Anselmi et al. 2026, DOI 10.3390/biomimetics11080518). Each scaffold was seeded with three human dental or orofacial mesenchymal stem cell types: commercial dental pulp stem cells (DPSCs), primary periodontal ligament stem cells (PDLSCs), and primary alveolar bone mesenchymal stem cells (aBMSCs). The work was funded by NIH/NIDCR grant R01DE031476 and sits in the pulp-dentin repair program (/programs/pulp-dentin-repair/) as scaffold-level primary research. Everything is in vitro; there is no animal or human data.

How the materials behaved

All three scaffolds produced smooth, randomly oriented fibers in the 200 to 600 nm range, with PDO fibers the thickest and GelMA the thinnest (p at most 0.0023, n = 150 fibers) (same source, Results). The materials diverged sharply in bulk behavior. Hydrated tensile strength was about 4 MPa for PCL and PDO (not significantly different, p = 0.1140) versus about 1 MPa for GelMA (p < 0.0001), with PDO the stiffest and GelMA the least stiff (p at most 0.0001), and PCL the most extensible (p at most 0.0014). Water contact angle was about 133 degrees for hydrophobic PCL, under 90 degrees for both PDO and GelMA (p < 0.0001 between them). Over three months in PBS, PCL retained roughly 95 percent of its initial mass while GelMA retained only about 14 percent, with PDO intermediate; the degradation comparison is reported with confidence intervals rather than pairwise tests.

What the cells did

Metabolic viability (alamarBlue, days 1, 3, and 7, n = 8) favored the hydrophilic polymers: for DPSCs, both PDO and GelMA exceeded PCL at days 3 and 7 (p at most 0.005). Mineralized matrix (Alizarin Red S at 14 and 21 days under osteo-odontogenic differentiation medium, n = 6) is the load-bearing result: at 21 days GelMA showed the highest mineral deposition for DPSCs (p < 0.0001; PCL versus PDO not different, p = 0.4873) and also for PDLSCs (p < 0.0001), where GelMA additionally gave the highest day-7 viability (p at most 0.014). For aBMSCs, GelMA gave the highest day-7 viability (p at most 0.032) but the mineralization difference among polymers was not significant, so the headline effect does not generalize across all three cell types.

The trade-off

The biologically best scaffold is the physically weakest. GelMA’s superior viability and mineralization for DPSCs and PDLSCs comes with the lowest tensile strength (about 1 MPa versus about 4 MPa), the lowest extensibility, and the fastest degradation (about 14 percent mass remaining at three months versus about 95 percent for PCL). The authors state the design implication plainly: polymer choice sets a biological-mechanical trade-off that matters for load-bearing versus pulp-space applications, and they call for 3D constructs, co-culture, vascularization, and mechanical-loading validation before any clinical inference (same source, Discussion).

What it does and does not show

This is a materials comparison in cultured cells, not a pulp-capping or dentin-bridge model. It shows that polymer identity modulates dental stem cell response in vitro, with GelMA leading on viability and mineralization for two of three cell types at defined timepoints. It does not show blend-ratio optimization (each polymer was spun under its own parameters, so composition is confounded with processing, as the authors concede), nor lineage-specific differentiation: there is no qPCR, no ALP activity, and no protein-level confirmation, so the assays cannot establish odontoblast identity. Donor numbers are not stated, cell spreading was scored qualitatively, and there is no true empty control. We also note the cell types were cultured at differing serum concentrations, which complicates cross-type comparison.

Where we differ from the coverage

We found no press or popular coverage of this paper to differ from. Framing that describes the result as tuning mineralization by “composition” overstates it: the study compared three discrete, differently processed polymers, not blend ratios, and one of the three cell types showed no mineralization effect at all.

Provenance: full open-access text read at MDPI and PubMed Central (PMC13510016, CC BY); all numbers above were verified against the article text by two independent analysts. Exact viability and mineralization percentages appear only in figures and were not quoted. Method and sourcing standard at /method/.