Background: Predictable periodontal regeneration following periodontal disease is a major goal of therapy. The objective of this proof of concept investigation was to evaluate the ability of cementoblasts and dental follicle cells to promote periodontal regeneration in a rodent periodontal fenestration model.Methods: The buccal aspect of the distal root of the first mandibular molar was denuded of its periodontal ligament (PDL), cementum, and superficial dentin through a bony window created bilaterally in 12 athymic rats. Treated defects were divided into three groups: 1) carrier alone (PLGA polymer sponges), 2) carrier + follicle cells, and 3) carrier + cementoblasts. Cultured murine primary follicle cells and immortalized cementoblasts were delivered to the defects via biodegradable PLGA polymer sponges, and mandibulae were retrieved 3 weeks and 6 weeks post‐surgery for histological evaluation. In situ hybridization, for gene expression of bone sialoprotein (BSP) and osteocalcin (OCN), and histomorphometric analysis were further done on 3‐week specimens.Results: Three weeks after surgery, histology of defects treated with carrier alone indicated PLGA particles, fibrous tissue, and newly formed bone scattered within the defect area. Defects treated with carrier + follicle cells had a similar appearance, but with less formation of bone. In contrast, in defects treated with carrier + cementoblasts, mineralized tissues were noted at the healing site with extension toward the root surface, PDL region, and laterally beyond the buccal plate envelope of bone. No PDL‐bone fibrous attachment was observed in any of the groups at this point. In situ hybridization showed that the mineralized tissue formed by cementoblasts gave strong signals for both BSP and OCN genes, confirming its nature as cementum or bone. The changes noted at 3 weeks were also observed at 6 weeks. Cementoblast‐treated and carrier alone‐treated defects exhibited complete bone bridging and PDL formation, whereas follicle cell‐treated defects showed minimal evidence of osteogenesis. No new cementum was formed along the root surface in the above two groups. Cementoblast‐treated defects were filled with trabeculated mineralized tissue similar to, but more mature, than that seen at 3 weeks. Furthermore, the PDL region was maintained with wellorganized collagen fibers connecting the adjacent bone to a thin layer of cementum‐like tissue observed on the root surface. Neoplastic changes were observed at the superficial portions of the implants in two of the 6‐week cementoblast‐treated specimens, possibly due in part to the SV40‐transformed nature of the implanted cell line.Conclusions: This pilot study demonstrates that cementoblasts have a marked ability to induce mineralization in periodontal wounds when delivered via polymer sponges, while implanted dental follicle cells seem to inhibit periodontal healing. These results confirm the selective behaviors of different cell types in vivo and support the role of cementoblasts as a tool to better understand periodontal regeneration and cementogenesis. J Periodontol 2004;75:154‐161.