The Effect of Purified Human Bone Marrow-Derived Mesenchymal Stem Cells on Rotator Cuff Tendon Healing in an Athymic Rat

Purpose: To evaluate the ability of purified human bone marrow-derived mesenchymal stem cells (MSCs) to augment healing of an acute small- to medium-sized rotator cuff repair in a small-animal model, evaluating the structure and composition of the healing tendon-bone interface with histologic and biomechanical analyses. Methods: Fifty-two athymic rats underwent unilateral detachment and transosseous repair of the supraspinatus tendon augmented with either fibrin glue (control group) or fibrin glue with 106 human MSCs (experimental group) applied at the repair site. Flow cytometry verified the stem cell phenotype of the cells as CD73(+), CD90(+), CD105(-), CD14(-), CD34(+), and CD45(-). Rats were killed at 2 and 4 weeks, with 10 from each group used for biomechanical testing and 3 for histologic analysis. Results: Safranin O staining identified increased fibrocartilage formation at the repair site at 2 weeks in the human MSC group (18.6% +/- 2.9% vs 9.1% +/- 1.6%, P = .026). Picrosirius staining identified decreased energy (36.88 +/- 4.99 J vs 54.97 +/- 8.33 J, P = .04) and increased coherence in the human MSC group (26.96% +/- 15.32% vs 14.53% +/- 4.10%, P = .05), indicating improved collagen orientation. Biomechanical testing showed a significant increase in failure load (11.5 +/- 2.4 N vs 8.5 +/- 2.4 N, P = .002) and stiffness (7.1 +/- 1.2 N/mm vs 5.7 +/- 2.1 N/mm, P < .001) in the experimental group compared with the control group at 2 weeks. These effects dissipated by 4 weeks, with no significant differences in fibrocartilage formation (35% +/- 5.0% vs 26.6% +/- 0.6%, P = .172) or biomechanical load to failure (24.6 +/- 7.1 N vs 21.5 +/- 4.1 N, P = .361) or stiffness (13.5 +/- 3.1 N/mm vs 16.1 +/- 5.6 N/mm, P = .384). All failures occurred at the bone-tendon interface. Conclusions: Rotator cuff repair augmentation with purified human MSCs improved early histologic appearance and biomechanical strength of the repair at 2 weeks, although the effects dissipated by 4 weeks with no significant differences between groups.