Soft tissue repair and reconstruction is a rapidly advancing field within orthopaedic surgery. Tissue engineering for tendon and ligament restoration reconstruction and regeneration represents a potential breakthrough to improve clinical outcomes following soft tissue injuries. Two blossoming fields within soft tissue regeneration are the use of mesenchymal stromal cells and the design of biosynthetic scaffolds. Stromal cells provide a potential reservoir of cells to repopulate a soft tissue defect with appropriate, functional, differentiated cells. Scaffolds provide interim tissue integrity and a backdrop against which the body's healing apparatus can work effectively.
While stromal cells and scaffolds are each useful in their own right, their potential synergistic effect is much greater. Tissue engineering, which combines stromal cells with biosynthetic scaffold materials, represents an exceptionally powerful tool with many potential applications in the augmentation of soft tissue repair. In order to harness this potential synergy, the optimal stromal cell should first be selected. The ideal stromal cell for soft tissue reconstruction should be one that is readily available and able to survive within the environment of a supporting scaffold, undergo terminal differentiation, and produce the desired proteins to facilitate repair or replacement of the injured tissues. Thus, for tendon repair, the ideal cells employed should readily differentiate into fibroblasts, more specifically, tenocytes, which produce abundant collagen.
Tenocytes are derived from embryonic mesenchyme. Mesenchymal stromal cells (MSCs) therefore provide a pool for tenocyte regeneration. MSCs are derived from many sources including bone marrow stroma,1 skeletal muscle,2 synovium,3 and adipose tissue.4 Previous work has shown that MSCs can be stimulated to differentiate into fibroblasts under the right culture conditions.5 A question arises as to whether a given subpopulation of MSCs, derived from the aforementioned tissues, is comparatively more adept at tenocyte differentiation. The concept of comparative MSC differentiation predispositions was explored in the work of Yoshimura et al.6 This study evaluated the productivity of rat MSCs derived from periosteum, synovium, adipose tissue, skeletal muscle, and bone marrow, varying culture conditions to drive differentiation toward chondrogenesis, osteogenesis, and adipogenesis. Their work indicated that synovial derived MSCs were more adept at chondrogenesis, while bone marrow derived MSCs were more efficient at osteogenesis. Elucidating a difference in MSC abilities to give rise to tenocytes would provide valuable information in deciding which MSC population to utilize in the augmentation of tendon repair.
MSC culture conditions will most often define the cells' differentiation pathway. Most MSC culture has been focused on pushing differentiation toward chondrogenesis and osteogenesis; however, a study by Lee et al.5 showed that MSCs exhibit fibroblastic differentiation in the presence of connective tissue growth factor. In another study, Wolfman et al.7 described ectopic tendon formation in rats using growth and differentiation factors-5, -6, and -7 (GDF 5, 6, and 7), indicating that these growth factors also play a role in tenocytic differentiation.
The purpose of this study was to investigate the comparative ability of bone marrow derived stromal cells (BMSCs) and skeletal muscle derived stromal cells (SMSCs) to differentiate into tenocytes; and to determine if tenogenesis in these cell types can be preferentially augmented by the addition of GDF-5 to the culture media.
MATERIALS AND METHODS
Seven, 1-year-old, female mongrel dogs, weighing between 19.4 and 22.5 kg (average 20.5 kg) were used for all portions of this study excluding immunohistochemistry (IHC), for which six separate dogs of the same demographic were utilized at a later date. The study was approved by our Institutional Animal Care and Use Committee (IACUC).
Bone Marrow Harvest and Culture
The animals were sedated using an intravenous mixture of 13 mg of ketamine and 6 mg of diazepam. Following adequate sedation the hind limbs were shaved and then prepared and draped using sterile technique. A bone marrow aspiration needle (Angiotech, Gainesville, FL) was then introduced into the tibial marrow cavity bilaterally. Between 6 and 8 ml of marrow aspirate were collected from each tibia into syringes containing 2 ml of Heparin. The samples were centrifuged at 1,500 rpm for 5 min. The supernatant was discarded and the remaining cell pellet was resuspended in MEM with 10% FBS and 1% antibiotic mixture (Penicillin, Streptomycin, and Amphotericin B). The cells and media were then plated and incubated at 37°C and 5% CO2. Cells were washed and fed every 3 days and passaged when 80–100% confluent.
Skeletal Muscle Harvest and Culture
Dogs were euthanized following bone marrow aspiration for reasons related to other IACUC approved studies, using an intravenous infusion of Pentobarbital. One dog was used for bone marrow harvest only and not euthanized, as it was participating in a different ongoing study. This animal was subsequently euthanized but did not contribute additional tissues to this study.
Immediately following euthanasia, two samples of skeletal muscle (average wt.: 0.43 g) were harvested from one hind limb of each animal. One of the two specimens was immediately frozen at −80°C, for subsequent RNA extraction. The remaining sample was minced and immersed in 10 ml of a 0.2% collagenase II solution for between 4 and 5 h, at room temperature. The digested tissue was then gently pressed through a sterile 60 µm filter using a sterile 12 cm3 syringe plunger. The resulting filtrate was centrifuged at 1,500 rpm for 5 min and the supernatant was descanted. The cell pellet was resuspended in 1× PBS and re-centrifuged at 1,500 rpm for 5 min. Following this wash, the cells were resuspended in MEM with 10% FBS and 1% antibiotic mixture (Penicillin, Streptomycin, and Amphotericin B). The cells and media were then plated and incubated at 37°C and 5% CO2. Cells were washed and fed every 3 days and passaged when 80–100% confluent.
RNA Extraction From Skeletal Muscle Tissue
Tissue specimens were thawed, weighed and then minced at room temperature. Samples were homogenized inside a Mikro-Dismembrator (B. Braun Biotech International, Melsungen, Germany) at 1,500 rpm for 15 min in the presence of 1 ml of Trizol reagent (Invitrogen, Carlsbad, CA). Homogenized samples were centrifuged at 12,000 rpm for 2 min. The supernatant was retained and protein precipitation was performed by adding 200 µl of chloroform (Sigma–Aldrich, St. Louis, MO), vortexing, and incubating on ice for 5 min. The samples were then centrifuged at 12,000 rpm for 10 min. RNA precipitation was performed on the supernatant by adding 500 µl of 2-propanol (Sigma–Aldrich), manually mixing, and incubating on ice for 30 min. The samples were centrifuged again for 10 min at 12,000 rpm and the supernatant was discarded. The RNA precipitate was washed in 1 ml of cold ethanol by gentle resuspension of the pellet and centrifugation at 12,000 rpm for 1 min. The ethanol was then decanted and the RNA was kept frozen at −80°C. RNA pellets were thawed and purified using an RNeasy Mini spin column (Qiagen, Valencia, CA). Residual DNA was digested using DNase (Qiagen).
Osteogenic and Adipogenic Differentiation of Stromal Cell Cells
Following four passages, cell samples were induced toward osteogenic and adipogenic differentiation using a Human Mesenchymal Stem Cell Functional Identification Kit (R&D Systems Inc., Minneapolis, MN). Cells were fixed and stained with Alizarin Red and Oil Red O, respectively, following differentiation for a qualitative analysis.
Tenogenic Differentiation of Stromal Cells
Following four passages, cells were placed on 60 mm plates at 2 × 105 cells per dish. Duplicate dishes were constructed for each dog, tissue type, and time point. The cells were cultured in MEM with 10% FBS and 1% antibiotic solution. Once confluent, ascorbic acid 2-phosphate (Sigma–Aldrich) was added to the media at a concentration of 50 µg/ml. Duplicate samples also received GDF-5 (MBL, Woburn, MA) at a concentration of 100 ng/ml upon induction and with subsequent media changes during the entire 2-week culture period.
Reverse Transcriptase-Polymerase Chain Reaction for Measurement of Relative Expression of Tenocyte Markers and Matrix Proteins
RNA samples were quantified using a NanoDrop™ 2000 (ThermoFischer Scientific, Waltham, MA). cDNA were synthesized from total RNA using a random primer and the Trancriptor First Strand cDNA Synthesis Kit (Roche, Mannheim, Germany). Quantitative real time PCR (RT-PCR) was then performed to measure and compare relative expression levels of CD44, tenomodulin (TNMD), collagen type I (COL1), and collagen type III (COL3) using the LightCycler® 480 Real-Time PCR system and LightCycler® FastStart DNA Master SYBR Green kit (Roche). HPRT was used as an internal control gene for RT-PCR. Relative expression levels were compared using the method. The RT-PCR primers used in these reactions are listed in Table 1.
Table 1. Primers Used in RT-PCR Reactions
Mesenchymal stromal cells from bone marrow and skeletal muscle were harvested from two and four additional dogs, respectively. The cells were expanded in culture over four passages, and induced toward tenogenic differentiation as described. During tenogenic differentiation, cells were cultured on sterile glass cover slips at the concentration of 2 × 104 cells per well of a 6-well plate. After being washed with PBS, cover slips were fixed in 4% paraformaldehyde and prepared for immunohistochemistry and Hematoxylin and Eosin staining. For immunofluorescence, the cover slips were blocked with 5% Bovine Serum Albumin or 5% donkey serum in PBS for 1 h at room temperature. Cover slips were then incubated overnight at 4°C with either rabbit polyclonal anti-collagen I (1:200, Abcam, Cambridge, MA,) or goat polyclonal anti-tenomodulin (1:200, Santa Cruz Biotechnology, CA) in PBS containing 0.3% Triton X-100 (PBST). Following another PBST wash, the slides were incubated with Alexa Fluor 488 goat-anti-rabbit secondary antibody or Alexa Fluor 488 donkey-anti-goat secondary antibody (1:200, Molecular Probes, Eugene, OR) in PBST for 2 h at room temperature. Cell nuclei were counter stained using 4′,6-diamidino-2-phenylindole (DAPI) (Vector Laboratories, Inc., CA). The cover slips were observed with a confocal microscope (LSM510; Zeiss, Germany). Fresh canine patellar tendon was used as a positive control for both tenomodulin and collagen I staining (Fig. 1).
The effects of stem cell origin and GDF-5 media augmentation on the outcome of relative expression were evaluated using linear mixed models in which day, cell type (BMSC vs. SMSC), and growth factor augmentation (yes vs. no) were modeled as fixed effects, and dog identification was incorporated as a random effect. Separate analyses were performed for day 7 outcomes and day 14 outcomes for CD44, tenomodulin, collagen type I, and collagen type III. All statistical analyses were performed with SAS version 9.1 (SAS Institute Inc., Cary, NC).
cDNA was synthesized from the skeletal muscle tissue of six dogs. Following three passages in culture, cDNA was synthesized from SMSCs and BMSCs of five dogs each. cDNA was synthesized just prior to induction of tenogenesis, and at 7 and 14 days following induction, of SMSCs from five dogs and BMSCs from four dogs. Two sets of cDNA were synthesized post-induction at each time point, as there were two separate culture conditions that were being observed. Insufficient RNA extraction yields in some samples precluded our ability to create cDNA from both cell types in each dog.
Following three passages in culture, SMSCs had a mean expression of myosin that was 56,843 times less than that of the normal skeletal muscle tissue harvested at animal sacrifice (p < 0.001). BMSCs had a mean expression of myosin that was 114,731 times less than that of the normal skeletal muscle tissue (p < 0.001). The difference in the relative expression of myosin between skeletal muscle derived stem cells and bone marrow derived stem cells at this time point was not significant.
Further analysis following three passages revealed that SMSCs had a mean expression of CD44 that was 48.28 times greater than that of the normal skeletal muscle tissue harvested at animal sacrifice (p = 0.002). BMSCs had a mean expression of CD44 that was 20.74 times greater than that of skeletal muscle tissue (p = 0.007). The difference in the relative expression of CD44 between SMSCs and BMSCs at this time point was not statistically significant.
Following induction toward osteogenesis, both BMSC and SMSC cultures stained positive with Alizarin Red (Fig. 2a and b), indicating calcified osteoid matrix formation. Following induction toward adipogenesis, both BMSC and SMSC cultures stained positive with Oil Red O (Fig. 3a and b), indicating lipid production. In both assays, SMSC replication was qualitatively more robust than BMSC replication and led to a more pronounced staining pattern.
Seven days following induction of tenogenesis, CD44 expression in BMSCs and SMSCs was 2.39 and 0.083 times that of pre-induction levels, respectively. This did not represent a significant difference for either cell types (p = 0.06 and 0.13). At 14 days, BMSCs and SMSCs expression was 1.47 and 2.1 times that of pre-induction levels, respectively. Again these were not significantly different from pre-induction expression levels (Fig. 4a).
At 7 and 14 days post-induction, the mean relative expression of tenomodulin observed in SMSCs was 0.26 and 26.31 times that of pre-induction expression, respectively. This represented a significant increase at 14 days (p = 0.02). The mean relative expression of tenomodulin observed in BMSCs at these time points was 0.76 and 1.39 times that of pre-induction expression. Neither of these values represented a significant change from pre-induction expression (Fig. 4b).
SMSC expression of collagen type I, measured at 7 and 14 days following induction, was 0.07 and 3.01 times that of pre-induction expression, respectively. The expression level observed at day 14 was significantly increased from pre-induction (p < 0.001). BMSCs showed an increase in collagen type I expression at both time points, reaching 3.30 and 6.81 times that of pre-induction expression levels; however only the increase seen at 14 days was significant (p = 0.002; Fig. 4c).
Collagen type III expression in SMSCs on day 7 and 14 was 0.06 and 2.15 times that of pre-induction expression, respectively. The expression noted at day 14 represented a significant increase (p = 0.02). BMSCs showed an increase in collagen type III expression at both time points, reaching 1.99 and 5.51 times that of pre-induction expression levels; however, only the increase on day 14 was significant (p < 0.001; Fig. 4d).
The effect of time on gene expression was also assessed separately with respect to growth factor exposure (GDF-5 and no GDF-5) irrespective of cell type (Fig. 5a–d). At day 14 both the GDF-5 exposed cells and the GDF-5 naïve cells showed significantly increased production of collagen types I and III. No other significant changes in relative expression were noted in this analysis.
Immunohistochemistry performed on SMSCs and BMSCs demonstrated the presence of collagen type I prior to tenocytic induction. There was no apparent increase in collagen I staining intensity over time following induction of either cell type (Figs. 6 and 7). There was, however, a change in cell morphology noted in the BMSCs, which assumed a more spindle shaped appearance (Fig. 6). The SMSCs increased in cell number and demonstrated polarity at day 14. Furthermore, the SMSCs began to line up in a streaming orientation (Fig. 7).
Prior to induction both SMSCs and BMSCs did not demonstrate any positive staining for tenomodulin (Figs. 8 and 9). The BMSCs began to show some positive staining at day 7, which subsequently increased at day 14. There was a slight increase in staining intensity noted in GDF-5 treated BMSCs when compared to non-treated cells; however, it was not possible to determine if this was significant. The arrangement of BMSCs was again noted to be devoid of any noticeable architecture (Fig. 8). Tenomodulin staining of SMSCs was not positive at day 7, but was positive at day 14. The SMSCs also exhibited an increase in cell number over time and a streaming architecture with respect to cellular arrangement and orientation with this staining preparation as well (Fig. 9).
Hematoxylin and eosin staining were performed and observed at 10× and 20× magnification for each cell culture at each time point (Figs. 10–13). Again noted was the haphazard array of BMSCs at day 14 compared to the more organized, streaming orientation displayed by the SMSCs.
Bone marrow derived stromal cells have been used to augment tendon repair in animal models, with improvement of biomechanical properties.8–10 Recent studies have also shown that stromal cells from different tissues of origin have divergent capabilities with regard to transformation into various terminally differentiated states.6 The results of Yoshimura et al.6 revealed that stromal cells were more capable of differentiation into cells that they were intimately associated with anatomically. Synovial stromal cells outperformed bone marrow stromal cells in chondrogenesis, while the converse was true of osteogenesis.
Skeletal muscle has recently been shown to be a reservoir of mesenchymal stromal cells capable of multipotent differentiation.11, 12 Due to their anatomic continuity with tendons, we hypothesized that muscle derived stromal cells would have a greater potential for tenocyte differentiation than bone marrow derived stromal cells. We found that both types of stromal cells responded to tenogenic induction with increased matrix gene expression; however, only skeletal muscle derived stromal cells increased tenomodulin expression significantly. Furthermore, the histology of skeletal muscle derived stromal cells demonstrated a streaming architecture more consistent with the cellular organization observed in tendon.
This study successfully isolated stromal cells from bone marrow using historically proven methods of serial expansion.13, 14 These cells possessed the stromal cell characteristics of multipotency and self-renewal in culture, and also showed expression of the stromal cell biomarker CD44.1, 15 Muscle derived stromal cells were also successfully isolated through similar methods and demonstrated a loss of myosin expression and an increase of CD44 expression when compared to skeletal muscle tissue expression levels. These cells also exhibited multipotency and self-renewal in culture. The multipotent nature of our isolated cells help to distinguish them from satellite cells, another quiescent self-renewing cellular constituent of skeletal muscle. With these attributes confirmed in both stromal cell populations, we turned our attention toward generating tenocytes.
Tenocyte differentiation of BMSCs has been previously reported in Rhesus monkeys through direct gene transfer of bone morphogenic protein-12, a homologue of GDF-7.16 Following gene transfer, BMSCs were found to have increased expression of scleraxis, CD44, and collagen type I. The authors of this study claimed that CD44 expression in conjunction with an absence of collagen type III expression indicated that these cells were tenocytes and not fibroblasts, based on previous cell culture work by Schulze-Tnazil et al.17 and Ohyama et al.18 Following tenocytic induction, our cells also showed CD44 expression; however, they expressed both collagen I and III. This may either indicate a non-uniform differentiation within our cultures, or may more closely mirror tenocyte expression patterns in vivo, where type III collagen is found as a matrix constituent.19 Tenocyte differentiation of our stromal cells was however more strongly demonstrated in our SMSCs with respect to their significant increase in tenomodulin expression.
Tenomodulin is a type II transmembrane glycoprotein mapped to the Xq22 genome locus in humans.20 It serves an anti-angiogenic role in tendons, and allows them to remain hypovascular structures throughout development.20 It is expressed during embryonic tendon development, and in vitro is only detectable in tenocytes.21 Tenomodulin is up-regulated by scleraxis and serves a late marker in the differentiation pathway of tenocytes, as high levels of tenomodulin are detectable in elongated tenocytes in mature tendons and may play a role in collagen fiber orientation.21 Our results demonstrated that the mean relative tenomodulin expression did not increase above pre-induction levels until the second week of tenogenic culture, supporting the proteins role as a late marker of tenocyte differentiation. Furthermore, the higher mean relative expression of tenomodulin observed in SMSCs in conjunction with the organization they demonstrated in culture also support tenomodulin's possible role in collagen fiber orientation.
The limitations of this study include its small sample size and the high variance in RT-PCR results. This combination led to increases in standard error within our data sets, making it difficult to uncover significance in some of our results. Tissue was procured from animals following euthanasia and immediately processed; added ischemia time was negligible and we did not feel that the pentobarbital used to euthanize the animals had an effect on our samples, but this may be viewed as an uncontrolled factor limiting our study as well. Another limitation of this study was that the GDF-5 used was a human recombinant protein. It is conceivable that the level of cross reactivity between this growth factor and the canine stromal cells was not high enough to show a difference in its ability to stimulate tenocytic differentiation, where a canine recombinant protein might have been able to do so. Furthermore, only one concentration of GDF-5 was utilized (100 ng/ml). This concentration was selected based on unpublished data from our laboratory, which showed that this concentration of GDF-5 imparted a collagen gel cell patch had the highest modulus of elasticity when compared to other concentrations. Moreover, in a study by Park et al.22 investigating various concentrations of GDF-5 on tenocyte gene expression induced in adipose derived MSCs, 100 ng/ml was found to be the optimal concentration. Our study did not incorporate testing for negative markers of tenocyte expression following differentiation of our stromal cells; however, the purpose of the study was to show that bone marrow and skeletal muscle derived stromal cells have the potential to augment a tendon repair via differentiation into a cells capable of tendon matrix production, which was demonstrated. Finally, results in cell culture might not be reproducible in tissue culture or in vivo.
Mayo Foundation and NIAMS (AR44391) funded this study. None of the authors have any financial disclosure.