Graft-tunnel motion may impair early graft incorporation and may lead to osteoclast-mediated bone resorption, contributing to tunnel widening. Early, aggressive postoperative rehabilitation may have detrimental effects on graft-to-bone healing.

A scar tissue interface forms rather than a normal ligament insertion site following attachment of a tendon graft to bone. The specific cell types that initiate the process of tendon-to-bone healing are unknown. We hypothesized that inflammatory cell accumulation following tendon-to-bone repair results in this scar interface. We used a rodent model to examine the temporal and spatial pattern of accumulation of hematopoietic lineage cells in the early phase of tendon-to-bone healing. Thirty-six Lewis rats underwent anterior cruciate ligament (ACL) reconstruction in the left knee using a flexor digitorum longus tendon graft. Six animals were sacrificed at 4, 7, 11, 14, 21, and 28 days after surgery. Serial sections were analyzed for proliferating cells (PCNA), recruited macrophages (ED1 ), resident macrophages (ED2), neutrophils, T-lymphocytes (CD3), mast cells, immature progenitor cells/pericytes (expressing the NG2 cell-surface chondroitin sulfate proteoglycan), and newly-formed blood vessels (Factor VIII). Neutrbphils, EDlt and ED2+ macrophages accumulated sequentially in the healing tendon graft, with progressive cell ingrowth from the interface towards the inner tendon. Neutrophils and EDl+ cells were seen in the tendon-bone interface at 4 days after surgery, while ED2+ macrophages were not identified until 11 days. These cells progressively repopulated the tendon graft. NG2-positive progenitor cells were found along the edge of the bone tunnel in the interface, but these cells did not invade the tendon. Occasional T-lymphocytes and mast cells were seen in the tendon-bone interface. There was no proliferation of intrinsic tendon cells, indicating that the tendon does not directly contribute to healing. We hypothesize that cytokines produced by infiltrating macrophages are likely to contribute to the formation of a fibrous scar tissue interface rather than a normal insertion site.