Because MSC treatment is being introduced more widely as a clinically available therapy, the method of administration must be considered to better mitigate risk. Although for a number of other factors also need consideration, including cell source, cell donor condition, cell population, and timing of MSC administration, this study only focused on comparison between systemic and local injection of MSCs into a rat oral implant model.

Sealing and defense at the PIE–implant interface are very important because dental implants in the oral mucosa are at high risk of inflammation. However, sealing between the PIE and implant is much weaker than that between the junctional epithelium (JE) and teeth [3], possibly owing to an inferiority of adhesion structures at the PIE-implant interface [15]. We therefore aimed to assess the influence of MSCs during implant treatment. Our previous report showed a positive effect of systemically injected MSCs for the improvement of peri-implant tissue sealing and acceleration of tissue healing [11].

In the systemic group, a strong HRP reaction was seen only in the coronal portion of the PIE on the implant surface (Fig. 2a). In the control and local groups, HRP reaction was not only found in the coronal PIE region on the surface of the PIE but also in the connective tissue. Furthermore, in the middle and apical PIE regions of these latter groups, the deep layers of PIE cells exhibited the strongest HRP reaction. This result meant that the PIE with these groups had only a weak epithelial sealing, and had been penetration of the external factors to the surrounding tissue of implant.

The systemic group exhibited a significant improvement in blocking HRP penetration (Fig. 2b) compared with both the local and control groups, which were comparable.

In the systemic group, immunohistochemical staining of Ln-332 showed a positive reaction along the whole implant-PIE interface at 4 weeks (Fig. 3a). In the local group, the Ln-332 deposition pattern in the PIE was comparable to that of the control group. In the oral mucosa around both local and control group implants, Ln-332-positive staining was apparent at the apical portion of the implant-PIE interface, but the upper portion of the interface did not exhibit Ln-332 detection. Only in the control group was the PIE-connective tissue interface intensely stained at the end of the PIE. Absence of Ln-332 staining was noted in the buccal mucosa underlying the OSE or OE in all groups.