OEC adhesion assays were conducted according to previously published methods [16, 20]. Non-adherent or weakly attached cells were removed by shaking (3 × 5 min at 75 rpm) using a rotary shaker (NX-20, Nissin, Tokyo, Japan). Adherent cells were then counted and calculated as a percentage of the initial count, which was used to define adhesive strength of the cells.

Scratch assays were performed on Ti plates to model wounding using various numbers of MSCs in the upper Transwell chamber. The techniques were conducted as described previously [16, 20]. Briefly, confluent OEC monolayers were scratched with a cell scraper and cultured for 48 h. OECs at the edge of the wound were visualized immunofluorescently using antibodies against actin to observe cell migration, and then the migrating cells were counted on the wound area.

Data are presented as means ± standard deviation (SD). One-way analysis of variance (ANOVA) with Fisher’s least significant difference test was performed. Significance was established when p < 0.05. Experiments were performed with triplicate samples and were repeated three or four times to verify reproducibility.