Synovial mesenchymal stem cells accelerate early remodeling of tendon-bone healing
- First Online:
- Cite this article as:
- Ju, Y., Muneta, T., Yoshimura, H. et al. Cell Tissue Res (2008) 332: 469. doi:10.1007/s00441-008-0610-z
- 419 Views
Tendon-bone healing is important for the successful reconstruction of the anterior cruciate ligament by using the hamstring tendon. Mesenchymal stem cells (MSCs) have attracted much interest because of their self-renewing potential and multipotentiality for possible clinical use. We previously reported that MSCs derived from synovium had a higher proliferation and differentiation potential than the other MSCs that we examined. The purpose of this study was to investigate the effect and mechanism of the implantation of the synovial MSCs on tendon-bone healing in rats. Half of the Achilles’ tendon grafts of rats were inserted into a bone tunnel from the tibial plateau to the tibial tuberosity with a suture-post fixation. The bone tunnel was filled with MSCs labeled with fluorescent marker DiI or without MSCs as the control. The tendon-bone interface was analyzed histologically, and collagen fibers were quantified. At 1 week, the tendon-bone interface was filled with abundant DiI-positive cells, and the proportion of collagen fiber area was significantly higher in the MSC group than in the control group. By 2 weeks, the proportion of oblique collagen fibers, which appeared to be Sharpey’s fibers, was significantly higher in the MSC group than in the control group. At 4 weeks, the interface tissue disappeared, and the implanted tendon appeared to attach to the bone directly in both groups. DiI-labeled cells could no longer be observed. Implantation of synovial MSCs into bone tunnel thus accelerated early remodeling of tendon-bone healing, as shown histologically.
KeywordsMesenchymal stem cellsTendon-bone healingAnterior cruciate ligament reconstructionCell therapySynoviumRat (Sprague Dawley)
Anterior cruciate ligament (ACL) reconstruction by using autologous hamstring tendon grafts and bone-patellar tendon-bone grafts (BPTB) has recently become a popular reconstruction method. One of the advantages of ACL reconstruction with hamstring tendon grafts rather than with BPTB is the reduced harvest-site morbidity such as slow recovery in quadriceps muscle strength and anterior knee pain (Sachs et al. 1989; Kartus et al. 2001; Goldblatt et al. 2005). However, one of the disadvantages of ACL reconstruction with hamstring tendon grafts is the relatively slow tendon-bone healing. Tendon pullout from bone tunnels occurs up to 12 weeks after ACL reconstruction with hamstring tendon graft in a dog model (Tomita et al. 2001). The mechanical stability at the tendon-bone interface in an early period is mainly dependent on the method of fixation. Therefore, acceleration of tendon-bone healing could allow patients to return to sporting activity earlier and more safely.
Previous studies have attempted to enhance tendon-bone healing by using biological materials. Rodeo et al. (1999) have reported that recombinant human bone morphogenetic protein-2 (rhBMP-2) accelerates the healing process histologically and biomechanically in a dog model. Martinek et al. (2002) have demonstrated that BMP-2 gene transfer improves the integration of semitendinosus tendon grafts in bone tunnels after reconstruction of the ACL in rabbits. Moreover, Anderson et al. (2001) have shown that a growth factor mixture derived from bone augments the healing of a tendon graft in a rabbit model of ACL reconstruction.
Mesenchymal stem cells (MSCs) have attracted much interest because of their self-renewing potential and multipotentiality for possible clinical use (Prockop 1997). We have previously reported that MSCs derived from the synovium have a higher proliferation and differentiation potential than the other mesenchymal tissue-derived MSCs that we have examined from both human (Sakaguchi et al. 2005; Shirasawa et al. 2006; Mochizuki et al. 2006) and rat (Yoshimura et al. 2007) sources.
Previous studies have demonstrated that bone-marrow-derived cells can be used to enhance the tendon-bone healing. Ouyang et al. (2004) have reported that bone-marrow-derived cells improve tendon-bone healing histologically in a rabbit model. Lim et al. (2004) have shown that implantation of bone-marrow-derived MSCs results in improved tendon-bone healing by an existing intervening zone of cartilage, and that they perform significantly better in biomechanical testing than controls in rabbit ACL reconstruction with hamstring tendon. However, the function and fate of implanted MSCs remain unknown because of the lack of cell labeling and tracing. If MSCs could be traced after implantation, this could provide useful information for understanding the fate and mechanism of implanted MSCs.
We have hypothesized that the implantation of synovial MSCs might enhance tendon-bone healing. The purpose of this study has been to investigate the effect and mechanism of the implantation of synovial MSCs on tendon-bone healing in rats.
Materials and methods
Nineteen 12-week-old mature Sprague-Dawley rats weighing approximately 400 g were used. Animal care was in strict accordance with the guidelines of the animal committee of our institute. One rat was used for the harvesting of the MSCs that were later implanted in this study and evaluated in the differentiation assay in terms of whether the cells used in this study did indeed have multipotentiality, one of the characteristics of MSCs. The remaining 18 rats were divided into two groups: the MSC group and the control group. In each group, we killed three rats at 1, 2, and 4 weeks, respectively, for histological examination.
Preparation of MSCs
Synovial membrane of bilateral knee joints was excised from one rat. The tissues were minced, digested for 3 h at 37°C with type II collagenase (0.2%; Sigma, Lakewood, N.J.), and passed through a 40-μm filter (Becton Dickinson, Franklin Lakes, N.J.) to yield a single-cell suspension. The cell number was counted with a hematocytometer. Nucleated cells from the synovium were plated at 104 cells/60 cm2-dish and cultured for 7 days as Passage 0. Dishes were trypsinized and harvested to count cell number cultured for 14 days as Passage 1. Some of the Passage 1 cells were used in the differentiation assay, but most of the Passage 1 cells were stored at −80°C until the experiment began. All of the cells were cultured in complete medium (αMEM; Invitrogen, Carlsbad, Calif.), 20% fetal bovine serum, lot-selected for rapid growth of human MSCs (Invitrogen), 100 units/ml penicillin, 100 μg/ml streptomycin, and 2 mM L-glutamine (Invitrogen) and incubated at 37°C with 5% CO2. At 24 h after initial plating, the cells were washed twice with phosphate-buffered saline to remove nonadherent cells.
Stored cells were thawed, incubated at 37°C with 5% CO2 for 1 week, and used after cell labeling. For cell labeling, we used CM-DiI (1,1'-dioctadecyl 3,3,3',3'-tetramethylindocarbocyanine pechlorate) cell-labeling solution (Molecular Probes, Eugene, Ore.), a lipophilic membrane-bound fluorescent dye. The cells were incubated in the working solution for 15 min at 37°C for fluorescent microscopic observation.
One-hundred cells at Passage 1 were plated and cultured for 7 days in 60-cm2 dishes. The dishes were then stained with 0.5% crystal violet in 4% paraformaldehyde for 5 min.
For chondrocyte differentiation, a micromass culture system was used. Approximately 8×105 cells were placed in a 15-ml polypropylene tube (Falcon, Bedford, Mass.) and pelleted into micromasses by centrifugation at 450g for 10 min. The pellets were cultured for 21 days in chondrogenic media, which contained 500 ng/ml BMP-2 (Yamanouchi Pharmaceutical, Tokyo, Japan), in addition to high-glucose Dulbecco’s modified Eagle’s medium (Invitrogen) supplemented with 10 ng/ml transforming growth factor-β3, 10-7 M dexamethasone, 50 μg/ml ascorbate-2-phosphate, 40 μg/ml proline, 100 μg/ml pyruvate, and 50 mg/ml ITS+TMPremix (Becton Dickinson, Franklin Lakes, N.J.). For microscopy, the pellets were embedded in paraffin, cut into 5-μm-thick sections, and stained with 1% toluidine blue (Yoshimura et al. 2007; Sekiya et al. 2001).
Passage 1 cells were plated at 100 cells per 60-cm2 dish and cultured in complete medium for 7 days, and then the medium was switched to adipogenic medium consisting of complete medium supplemented with 0.5 μM dexamethasone (Sigma), 0.5 mM isobutylmethylxanthine (Sigma), and 50 μM indomethacin (Sigma). After 4 days, the adipogenic cultures were fixed in 4% paraformaldehyde for at least 1 h and stained with fresh Oil red-O solution for 2 h. The Oil red-O solution was prepared by mixing three parts stock solution (0.5% in isopropanol; Sigma) with two parts water and filtering it through a 0.2-μm filter. After being stained, the dishes were washed three times, and the Oil red-O positive colonies were observed.
Passage 1 cells were plated at the same densities as indicated for the adipogenesis assay and cultured in complete medium for 7 days. The medium was then switched to calcification medium in the presence of 100 nM dexamethasone, 10 mM β-glycerophosphate, and 50 μM ascorbic acid (Sigma), and the cells were incubated for 21 days. These dishes were stained with fresh Alizarin red solution.
Rats were divided into two groups: the MSC group (n=9) and the control group (n=9). In the MSC group, a total of 1×107 MSCs immobilized in 0.2 ml atelocollagen gel (3% type-I collagen; Koken, Tokyo, Japan) were injected into the bone tunnel. In the control group, 0.2 ml atelocollagen without cells was injected (Fig. 1c). Injection was performed after the graft passage and from both of the tunnel orifices. The needle was placed around the graft so that the injectant spread circumferentially through the entire portion of the graft. All animals were kept in a standardized cage and allowed to be active without any fixation device.
The tendon-tibia complex was fixed in a 4% paraformaldehyde solution immediately after being harvested from each limb. The specimens were decalcified and cast in paraffin blocks, and the proximal tibiae were sectioned sagittally, parallel to the longitudinal axis of the bone tunnel. Then, sections (5 μm thick) were stained with Masson trichrome to identify collagen fibers. In the MSC group, the serial sections were also observed with fluorescence microscopy for DiI to trace the implanted cells.
Quantification of collagen fibers
The areas of the entire visible tendon-bone interface and collagen fibers within the interface were measured, respectively, in histological sections stained with Masson trichrome by Scion Image analysis software for Windows (Scion Corporation, Frederick, Md.), which is based on the popular NIH image software originally designed for the Macintosh platform. The collagen fibers area was divided by the tendon-bone interface area to obtain the proportion of the collagen fibers to the entire tendon-bone interface.
For quantification of the oblique collagen fibers, the sections were imposed onto Scion Image, and the entire tendon-bone interface area was measured. Then, tissues other than oblique collagen fibers were erased manually, and the oblique collagens fiber area was measured. The oblique collagen fibers area was divided by the entire tendon-bone interface area to obtain the proportion of the oblique collagen fibers to the entire tendon-bone interface.
Three sections were examined per specimen, and the quantification of collagen fibers was analyzed (blind) by two examiners who were not informed of the group assignment.
To assess differences, Mann-Whitney U tests were used. A value of P<0.05 was considered significant.
Characteristics of synovial cells as MSCs
We have examined the effect and fate of MSCs on the healing of the tendon-bone interface in rats. Our results indicate that the application of synovial MSCs has the potential to accelerate the early remodeling of tendon-bone healing histologically by producing more collagen fibers at 1 week and forming more oblique collagen fibers connecting the bone to tendon resembling Sharpey’s fibers at 2 weeks. Moreover, MSCs do not interfere with tendon-bone healing at 4 weeks.
Several surgical models have previously been performed in animals to investigate tendon-bone healing. Some investigators have mimicked ACL reconstruction in animals. Rodeo et al. (1999) have transplanted the proximal site of the long digital extensor tendon into an oblique drill hole in the proximal tibia. Yagishita et al. (2005) have reported that, when the medial half of the patella tendon of Japanese white rabbits is transplanted into a drill hole in the tibial tuberosity, good graft isometry is obtained. We have used the same drill hole and graft pathway as that in the study of Yagishita et al. (2005), and so we believe that the appropriate tension has been applied to the grafted tendon.
Previous reports have attempted to enhance tendon healing in the bone tunnel by using biological materials. Rodeo et al. (1999) have reported that rhBMP-2 accelerates the healing process histologically and biomechanically in a dog model. They have demonstrated that the presence of rhBMP results in increased bone ingrowth into the interface zone because of rhBMP-induced differentiation of the mononuclear cells into osteoblastic cells. Anderson et al. (2001) have reported that a growth factor mixture derived from bone augments the healing of a tendon graft in a rabbit ACL reconstruction model. Martinek et al. (2002) have demonstrated that BMP-2 gene transfer improves the integration of semitendinosus tendon grafts in bone tunnels after reconstruction of the ACL in rabbits. The effect of BMP-2 and bone-derived proteins is thought to induce bone growth into the space between the tendon graft and the bony wall; this will enhance maximum failure load of the tendon-bone junction.
Bone-marrow-derived cells have been used to enhance tendon-bone healing. Ouyang et al. (2004) report that bone marrow stromal cells improve the healing of tendon to bone histologically in a rabbit model; the histology of the experimental group shows more perpendicular collagen fibers and the proliferation of cartilage-like tissue within the bone tunnels. Lim et al. (2004) have revealed that bone-marrow-derived MSCs promote healing by an intervening zone of cartilage, and that they perform significantly better than controls in biomechanical testing in rabbit ACL reconstruction by using hamstring tendon. However, our form of bone-tendon healing enhancement is different from that of previous studies. Whereas our studies have shown the acceleration of the remodeling of the tendon-bone interface by the production of collagen fibers and formation of Sharpey’s fibers in the early period, we have been unable to observe fibrocartilage formation as observed in previous studies. The difference between our results and those of the other studies is attributed to the cell carrier or the source of the mesenchymal cells, because we have used atelocollagen gel as a cell carrier and synovial MSCs, whereas previous studies involved the use of a fibrin sealant as a cell carrier and bone-marrow-derived MSCs.
Despite the use of allogenic cells in this study, we have not observed features of immune reactions, such as synovial inflammation macroscopically. Immune tolerance against MSCs is still controversial (Tse et al. 2003; Tolar et al. 2006).
Before this report, the fate of implanted MSCs during tendon-bone healing was still unknown. Although several tendon-bone healing studies have been carried out with MSCs, implanted MSCs have not been tracked in previous investigations. The use of DiI, a membrane-bound fluorescent dye, is one way to track the fate of implanted cells (Mothe and Tator 2005). Previous studies have shown that DiI typically exhibits low cell toxicity and does not compromise cell viability or differentiation potential (Crawford and Braunwald 1991; Ponticiello et al. 2000). DiI is weakly fluorescent in aqueous solutions but is highly fluorescent and photostable when incorporated into lipid membrane. Because of this, even if the dye were to leak out of a dead cell, it would not emit significant fluorescence in an aqueous environment such as the tendon-bone interface.
We have demonstrated that collagen fibers in the tendon-bone interface occur in significant amounts in the MSC group in comparison to the control group at 1 week. Most fibroblasts in the interface are DiI-positive in the MSC group, indicating that the implanted MSCs directly differentiate into fibroblasts with the potential of producing collagen fibers, although MSCs are indistinguishable from the fibroblasts in appearance. MSCs are known to secrete some trophic mediators (Caplan and Dennis 2006). In our study, implanted MSCs might have secreted some cytokines to promote collagen fiber synthesis.
At 2 weeks, although the collagen fiber area in the tendon-bone interface becomes similar in both groups, the proportion of oblique collagen fiber area is significantly higher in the MSC group than in the control group. We speculate that implanted MSCs promote bone-tendon healing and act mainly as a stimulus for the collagen maturation necessary to form the Sharpey’s fibers in the environment of the tendon-bone interface at this period.
By 4 weeks, DiI-labeled cells can no longer be observed. We propose three mechanisms for this. First, implanted cells could be totally replaced with host cells because of rapid cell turnover. Second, DiI labeling could disappear, although implanted cells still exist in this model; DiI fluorescence might decrease concomitantly with the repetition of cell doubling. Third, implanted cells might disappear because of apoptosis after maturation of the tendon-bone interface; the apoptotic cells might release transforming growth factor-β (Jakowlew 2006), promoting tendon-bone healing (Yamazaki et al. 2005).
MSCs retain considerable plasticitiy and may differentiate into a variety of cell types in vitro and in vivo, including adipocytes (MacKenzie and Flake 2002), cardiac myocytes (Lovell and Mathur 2002), chondrocytes (Caterson et al. 2002), endothelial cells (Al-Khaldi et al. 2003), fibroblasts (Direkze et al. 2004), myofibroblasts (Direkze et el. 2003), osteoblasts (Petite et al. 2000), pericytes (Shi and Gronthos 2003), skeletal myocytes (Camargo et al. 2004), tenocytes, and thymic stroma (Liechty et al. 2000). Direkze et al. (2004) have reported that implanted bone marrow contributes to myofibroblast and fibroblast populations in tumor stroma in a mouse model of pancreatic insulinoma. Our results indicate that implanted MSCs might differentiate into fibroblasts in the bone-tendon interface and promote the production of collagen fibers, thereby accelerating the early remodeling of tendon-bone healing.
Several limitations should be noted in this study. First, a significant limitation is that we have not performed biomechanical testing for the tendon-bone complex to determine whether the MSC-treated group actually enhances early tendon-bone healing biomechanically in the rat model. We consider our model not to be suitable for performing reproducible and reliable mechanical testing because of the size of the animals. Previous studies have reported that the maximum pull-out load progressively increases as collagen fibers grow, and as the interface healing matures histologically in the dog model (Rodeo et al. 1999). We can only speculate from this previous study that our results might enhance the pull-out load in the early period of tendon-bone healing. Examination of the maximum pull-out load of the tendon-bone interface in large animals is the objective of our next study.
Second, half of the Achilles’ tendon has been used as a graft in this model. We have employed the Achilles’ tendon as a graft because the extensor hallucis longus or hamstring tendon is too thin to be used as a graft in the rat model. We have selected the same side of the Achilles’ tendon because the graft is usually harvested from the same side of the limb during ACL reconstruction. Although some degree of limping has been seen in our rats postoperatively, we have observed that they can extend their ankle joints. Thus, we consider that appropriate tension is applied to the grafted tendon.
Third, we have used an extra-articular model to avoid variables of graft tensioning or positioning. The complex biological environment of a tendon graft in an intra-articular application is not reproduced by this model. Further study is thus needed to evaluate the effect of MSCs in an intra-articular model.
Care must be taken when applying our results to humans. The tendon-bone healing process occurs at a faster rate in rats than in human. Further experimental investigations in large animals will thus be needed to determine whether our results are applicable to humans. If the treatment of MSCs produces a clinical effect, even in only the early phase after ligament surgery, it could improve rehabilitation and produce a better outcome for ligament reconstruction.
In conclusion, implanted synovial MSCs improve early remodeling of the tendon-bone healing at 1 and 2 weeks histologically. However, the effect is not observed at 4 weeks.
We thank Kenichi Shinomiya, MD, PhD, for continuous support and Miyoko Ojima for expert help with histology. Recombinant human BMP2 was kindly provided by Astellas Pharma.