Eighteen healthy, adult female New Zealand white rabbits weighing 3,000 to 3,500 g underwent bilateral ACL reconstruction using an extensor digitorum longus tendon autograft harvested from the same surgical site. The physical conditions of all knees were carefully monitored during the experimental period. All experimental animal procedures were performed in strict accordance with our Institutional Animal Care and Use Committee. All knees were distributed into three groups according to use the biologic agent; control group (n=12) without any biologic agent, fibrin group (n=12) and MSCs group (n=12) with stem cell and fibrin.

Bone marrow aspiration was carried out from the iliac crest of the rabbits in the control group as previously described (25, 26). Bone marrow (1 ml) was collected and put in complete Dulbecco’s Modified Eagle Medium (DMEM-lg, Gibco, USA) containing 10% fetal bovine serum and cephazolin l g/ml. The sample was washed twice with DMEM-lg and centrifuged at 2,000 rpm for 6 min. The supernatant was removed and the precipitating component was suspended in 10 ml of stem cell growth medium. The stem cells were counted and plated (1.0 to 1.5×10⁷ cells) in a 100 mm dish. Cultures were incubated at 37°C in an atmosphere 5% CO2. After three days, non-adherent cells were removed when the medium was replaced with fresh medium. Once the colonies of MSCs reached 70~80% confluence, adherent cells were obtained from the flask with 0.05% trypsin. Homogenous MSCs were obtained after 2 weeks of culture. Passage two MSCs (P-2) were used.

Each animal was then anesthetized with ketamine (35 mg/kg) and positioned in supinen. Both knees of each rabbit were shaved, scrubbed with betadine and covered with a sterile drape exposing only the unilateral knee. After midline incision, a medial para-patellar approach was performed and then the native ACL and extensor digitorum longi (EDL) tendon were identified (28) and the length was measured (Fig. 1A). The EDL tendon was harvested to a length of 20 mm or more and was kept immersed in sterile saline until implantation (Fig. 1B). After excision of the native ACL at its femoral and tibial origins, a 2 mm-diameter bone tunnel was made at the footprints of the native ACL on the tibia and lateral femoral condyle (Fig. 1C, D). The reproducible and accurate tunnel positioning in relation to the native ACL footprint was necessary to achieve loading conditions similar to those of native knee joint. The free end of each EDL graft secured with 3-0 vicryl. In the fibrin group, 2 mm of fibrin glue (TISSEEL Kit, IMMUNO AG Vienna, Austria) was injected to the end of graft and MSCs embedded in fibrin glue by mixing was injected in the MSCs group. A total of 1×10⁶ bone marrow stem cell with 1 ml fibrin glue were injected in the end of graft and bone tunnel in MSCs groups. In fibrin group, a total of 2 ml fibrin glue was injected in graft and tunnel. The prepared graft was passed through the tibial and femoral tunnels (Fig. 1E). The gap between the graft and the tunnel was filled with each biologic agent. A total of 1×10⁷ stem cells with 1 ml fibrin glue were injected to the end of graft and filled into the tunnel in MSCs groups. After extra-osseous post-screws were inserted near the tunnel orifices, the end of each implanted graft was anchored to screws using sutures (Fig. 1F). The incision was closed in layers with the arthrotomy and subcutaneous layers. For initial stability, the postoperative knee was maintained in extension by an elastic bandage. Then the contralateral knee was operated on using the same method. All animals were euthanized at 12 weeks postoperatively for histologic analysis and for the evaluation of tunnel widening using micro-CT analysis.

The knee joint was obtained by amputating at femoral and tibial shaft after euthanasia. CT imaging was performed using a Quantum GX μCT imaging system (PerkinElmer, Waltham, MA, USA). The X-ray source was set to levels of 90 kV and 88 mA with a field of view of 72 mm with a slice thickness of 0.144 mm. The scanning time was 14 min in a 360° rotation. A matrix size of 512×512×512 was then used to reconstruct using 3D Viewer, existing software within the Quantum GX. Following scanning, image segmentation was performed by Analyze software (AnalyzeDirect, Overland Park, KS, USA). To calculate the tunnel widening, the oblique-coronal and oblique-sagittal plane were obtained parallel to the tunnel (Fig. 2). The tunnel diameter was measured on 2 mm proximal and distal from the femoral and tibial articular surface. The tunnel widening was defined as difference between measured diameter on CT scanning and drilled hole diameter (2 mm).

Then, Raw CT values were converted to Hounsfield Unit (HU) using intensity values of dry air was −1,000 HU and that of a water was 0 HU. To define each 1.5 mm×1.5 mm squared specific regions of interest (ROI) within the rabbit knees, the oblique plane images of the bone tunnel 2 mm proximal and distal from the femoral and tibial articular surface were analyzed using Quantum GX software. HUs were measured in each ROI in the tunnel and around tunnel in cancellous bone. All radiographic measurements were performed by two blinded observers not participating in surgery.

After euthanasia, each knee was carefully stripped from the surrounding soft tissues except from the autograft itself (Fig. 3). For histopathologic examination of the knee joints, each sample was fixed in 10% buffered formalin. The specimens were decalcified in a decalcifying agent (Calci-Clear Rapid, National Diagnostics, GA) for 12 hours, cut along the grafted tendon in the oblique-sagittal direction, and embedded in paraffin. The paraffin block were sectioned at a 3 μm thickness using a microtome (Leica Microsystems, Inc, Wetzlar, Germany) and each section was stained with Hematoxylin and Eosin (H & E). The stained slides of the knee joint sections were observed under a light microscope and were read by a pathologist.

Statistical evaluation was made using two ways repeated measures Kruskal-Wallis test to detect significant differences in the CT density scales and tunnel widening. When significant differences were found, post hoc comparisons with the Mann-Whitney test were made. The study results were evaluated with SPSS version 20.0 for Windows (SPSS Inc, Chicago, Illinois). Difference were considered statistically significant at p<0.05.

The average combined length of the tunnels was 3.8 cm. There were no statistically significant differences among the three groups in terms of tunnel lengths (p>0.05).

Comparing the three groups, there are significant difference of femoral and tibial tunnel widening on the oblique-sagittal view (p=0.001, 0.018, respectively). The mean extent of widening of the femoral and tibial tunnel in oblique-sagittal image was significantly greater in the control than in the fibrin and MSCs groups (p<0.05) (Table 1). No significant differences were evident in terms of the mean tunnel widening on oblique-coronal views among the three groups (p>0.05).

HUs in the femoral tunnel (control : 95.06±38.6, fibrin : 103.69±21.8, MSCs : 105.75±13.2) and in the tibial tunnel (control : 98.43±57.8, fibrin : 102.77±32.4, MSCs : 101.93±57.9) were similar to tendon structures in previous study (29). In addition, HUs around the femoral tunnel (control : 424.88±70.0, fibrin : 407.54±29.5, MSCs : 387.30±50.1) and around the tibial tunnel (control : 405.85±55.2, fibrin : 401.68±31.4, MSCs : 421.08±42.7) were similar to prior values obtained from the cancellous bone of humans (29). HUs in and around the bone tunnel were not significantly different between the three groups (p>0.05) (Table 2).

Gross observation revealed intact intra-articular grafts in all cases. For evaluation of tendon-bone integration and graft architecture, the intra-osseous portion of each graft was obtained and analyzed. H & E staining 12 weeks after surgery revealed preservation of smooth bone to graft bonding and good architecture were observed in the control group. However, there were fewer fibrocytes (hypocellularity) and less compact collagen fiber compared with those of other groups (Fig. 4A, D). Focal areas of disorganized architecture and degenerative alterations were evident in the control group. In the fibrin group, although disorganized architecture or partial tear of the graft was observed, there are good cellularity and more compact collagen fiber compared with the control group (Fig. 4B, E). In the MSCs group, excellent restoration of the bone to graft bonding to the well-formed tunnel was observed. The good architecture with hypercellularity and compact collagen deposition was shown in compared with other groups (Fig. 4C, F).