What was tested
Mariam Kakabadze and colleagues at Tbilisi State University and Tbilisi State Medical University built a three-part pulp substitute and tested it in two animal models (DOI 10.62347/rjpq4886, American Journal of Stem Cells, 15(1):13-26, Epub 15 February 2026, a small specialist journal). The parts: porous Cultispher microcarriers loaded with bone marrow stem cells (BMSCs) from Lewis rats, freeze-dried at minus 40 degrees C and 10 to 15 Pa and stored at room temperature; a commercially available fibrinogen-thrombin gel (Tisseel) to carry and fix them; and, for the ectopic arm, longitudinally sectioned porcine molars decellularized by a 0.1 to 1 to 10 percent sodium dodecyl sulfate gradient (24 hours each) plus Triton X-100 and DNase/RNase, leaving under 2 percent of initial DNA after 72 hours (n = 5 per group, 10 percent SDS significantly better than 1 percent at 48 hours, p less than 0.0001). In the ectopic arm, 20 rats received the cell-loaded constructs inside decellularized tooth slices under the skin. In the orthotopic arm, 6 sheep had partial pulpotomy (about 7 to 8 mm of coronal pulp removed), the canals filled with rehydrated freeze-dried rat BMSC microcarriers plus fibrinogen and thrombin, and the access sealed with Biodentine and composite. Rats were harvested at 5, 14, 30 and 60 days, sheep at 10, 20 and 30 days. The work was funded by the Shota Rustaveli National Science Foundation of Georgia (grant YS-23-1748).
What they found
All in vivo observations were descriptive. In rats, the fibrin gel degraded within about 12 days and was replaced by vascularized connective tissue; by four weeks the constructs held organized, vascularized, partially mineralized fibrovascular tissue with a capillary network, and no acute inflammation. In sheep, at about two weeks the canal lumen held basophilic pulp-like tissue with new vessels of varying diameters; by three weeks a compacted cell layer morphologically resembling early odontoblasts lined the dentin; and by the 30-day endpoint a continuous layer of mature odontoblast-like cells sat beneath a continuous predentin zone, with a dentinal bridge formed and high VEGF and Ki-67 expression on immunostaining. Notably, by 20 days the transplanted BMSCs were virtually undetectable on the microcarriers: the authors’ own mechanism is that the freeze-dried cells work as paracrine stimulators, through preserved biomolecules and extracellular vesicles, recruiting the host’s own cells, rather than by surviving and engrafting. They cite their prior work that freeze-dried BMSCs retain more than 53 percent of their paracrine factors.
What it does not show
There is no quantitative histological scoring at all: the methods state that no scoring system was applied, and the animal numbers (20 rats, 6 sheep) with single descriptive endpoints per timepoint cannot support efficacy claims. No functional test was done: no blood flow measurement, no innervation assay, no dentinogenic activity beyond morphology, no biomechanics. The “odontoblast-like” label is a shape judgment, not a lineage or functional demonstration. The orthotopic cells are rat-derived in sheep, a xenogeneic pairing, and the paper does not report immunosuppression. The viability of freeze-dried cells after rehydration is asserted through citations to other groups’ work, not measured here; on the data presented, this construct behaves as a cell-free paracrine product. And the model is surgical pulp amputation in healthy young sheep, not infected or necrotic pulp, which is the condition regenerative endodontics actually faces.
Where it sits in the field
For the pulp and dentin repair program, this is the cell-based wing, one step closer to a deliverable than yesterday’s in vitro BMP-7 scaffold paper: it shows a storable, off-the-shelf preparation that in one large-animal pulpotomy model produced vascularized pulp-like tissue with a predentin border within 30 days. The freeze-drying angle is the real contribution, since bankable, room-temperature products are the practical bottleneck for cell therapies, and the paracrine-not-engraftment mechanism, if confirmed with quantitative and functional assays, would put this in the same family as secretome and extracellular-vesicle approaches already in the clinical literature. It qualifies rather than strengthens the program’s human tier: one group, one journal, descriptive histology. The current field assessment stands at /field/.
Where we differ from the coverage
We found no press or popular coverage of this paper to differ from. Against the paper’s own framing, we hold back two inferences. First, the title and abstract promise “stem cells,” but the authors’ own results section has the cells disappearing within 20 days and the mechanism section describes a paracrine, biomolecule-and-vesicle product; the honest summary is that a freeze-dried cell-derived preparation, not a living graft, drove the observed tissue. Second, the conclusion claims “robust efficacy,” while the study is explicitly descriptive throughout, with no quantified outcome measure; robust is not a characterization this design can support.
Provenance: grounded in the full text of record (the journal’s full-text PDF for article AJSC0169871), retrieved 19 October 2026; every number above is taken from the article’s abstract, methods, results text and figure captions. Method and sourcing standard at /method/.