Cartilage repair using bone morphogenetic protein 4 and muscle-derived stem cells

Authors


Abstract

Objective

Muscle-derived stem cells (MDSCs) isolated from mouse skeletal muscle exhibit long-time proliferation, high self-renewal, and multipotent differentiation. This study was undertaken to investigate the ability of MDSCs that were retrovirally transduced to express bone morphogenetic protein 4 (BMP-4) to differentiate into chondrocytes in vitro and in vivo and enhance articular cartilage repair.

Methods

Using monolayer and micromass pellet culture systems, we evaluated the in vitro chondrogenic differentiation of LacZ- and BMP-4–transduced MDSCs with or without transforming growth factor β1 (TGFβ1) stimulation. We used a nude rat model of a full-thickness articular cartilage defect to assess the duration of LacZ transgene expression and evaluate the ability of transplanted cells to acquire a chondrocytic phenotype. We evaluated cartilage repair macroscopically and histologically 4, 8, 12, and 24 weeks after surgery, and performed histologic grading of the repaired tissues.

Results

BMP-4–expressing MDSCs acquired a chondrocytic phenotype in vitro more effectively than did MDSCs expressing only LacZ; the addition of TGFβ1 did not alter chondrogenic differentiation of the BMP-4–transduced MDSCs. LacZ expression within the repaired tissue continued for up to 12 weeks. Four weeks after surgery, we detected donor cells that coexpressed β-galactosidase and type II collagen. Histologic scoring of the defect sites 24 weeks after transplantation revealed significantly better cartilage repair in animals that received BMP-4–transduced MDSCs than in those that received MDSCs expressing only LacZ.

Conclusion

Local delivery of BMP-4 by genetically engineered MDSCs enhanced chondrogenesis and significantly improved articular cartilage repair in rats.

Damage to articular cartilage frequently results from injury or disease. Because of its limited intrinsic healing capacity, articular cartilage cannot fully regenerate; thus, articular cartilage injuries eventually lead to secondary degenerative disease of the involved joint (1, 2). Researchers have designed clinical methods and proposed various experimental approaches by which to achieve better repair of injured articular cartilage, including abrasion arthroplasty (3, 4), microfracture (5), and transplantation of chondrocytes (6–9), perichondrium (10), meniscal allografts (11), periosteum (12), or osteochondral grafts (13). However, there is no known treatment that enables full restoration of injured articular cartilage to its original phenotype.

Tissue engineering based on cell and gene therapy is one of the most promising new approaches by which to repair various tissues, including articular cartilage. This process involves the use of various cell types that can repair articular cartilage by acting as chondroprogenitor cells, as gene delivery vehicles that produce a therapeutic protein at the site of interest, or as both. The variety of cells available for use in cartilage tissue engineering ranges from undifferentiated pluripotent stem cells to well-differentiated chondrocytes (6–9, 14, 15). Chondrocytes are a natural and logical choice for articular cartilage repair applications. However, a limited donor-site capacity to provide a large quantity of chondrocytes is a major impediment to autologous chondrocyte transplantation. Therefore, there are ongoing efforts to identify other cell populations that contain chondroprogenitor cells and are easily isolated and expandable.

In light of its availability and the relative ease of muscle cell isolation, skeletal muscle is an attractive source of cells for use in cartilage tissue engineering applications. Adachi et al recently reported comparable healing of cartilage defects treated with collagen gel containing either muscle-derived cells (MDCs) or chondrocytes (16), which suggests that skeletal muscle may contain cells that can aid in cartilage repair. Several studies have provided evidence of the existence of pluripotent stem cells in skeletal muscle that can differentiate into various lineages, including myogenic, hematopoietic, osteogenic, endothelial, and neuronal (17–20). However, the differentiation of muscle-derived stem cells (MDSCs) into the chondrogenic lineage has not yet been demonstrated.

The administration of growth factors that can enhance cartilage healing is another important facet of cartilage repair. Several growth factors, including transforming growth factor β (TGFβ), bone morphogenetic proteins (BMPs), insulin-like growth factor 1, and basic fibroblast growth factor, can improve chondrocyte proliferation and extracellular matrix (ECM) synthesis in vitro and in vivo (21–28). Studies performed in vitro and in the developing skeleton have led to the identification of BMP-4 as a promising candidate for the promotion of chondrogenesis (29, 30). Steinert et al observed particularly high expression of chondrogenic markers in populations of BMP-4–transfected mesenchymal progenitor cells (31).

Although the results of some studies suggest that direct injection of therapeutic proteins can promote cartilage healing, the relatively short half-lives of these proteins in vivo often necessitate high dosages or repeated injections. It also can be difficult to deliver these growth factors to the injury site. Moreover, the possibility exists that unregulated high dosages or repeated injections of growth factors could negatively influence both injured and normal structures and lead to adverse effects, such as hypertrophy or osteophyte formation (32). Gene therapy, a process whereby donor cells are genetically engineered to deliver a therapeutic protein to the site of injury and thereby promote tissue repair, provides a promising alternative to the direct injection of therapeutic proteins.

We designed this study to determine if MDSCs could serve as a source of progenitor cells for cartilage repair and as gene delivery vehicles for BMP-4. We posited that the impressive self-renewal and long-term proliferation ability of MDSCs (17) would enable enhanced and prolonged gene expression within cartilage defects, which would promote cartilage repair. Our primary aim was to evaluate the ability of BMP-4–expressing MDSCs to differentiate into chondrocytes in vitro and repair articular cartilage defects in vivo.

We first cultured cells in a monolayer or in a 3-dimensional pellet culture system either in normal differentiation medium or chondrogenic medium (CM) with or without TGFβ1 supplementation. We then transplanted the transduced MDSCs into cartilage defects created in athymic rats and investigated the distribution of the LacZ transgene in the repaired tissues and the expression of chondrogenic differentiation markers by the donor cells. Finally, we evaluated the regenerative ability of the donor cells and determined the role of ex vivo BMP-4 gene therapy in the healing of articular cartilage.

MATERIALS AND METHODS

Isolation of primary MDSCs.

MDSCs were isolated from the hind limb skeletal muscle of 3-week-old female C57BL/10ScSn mdx/mdx mice (The Jackson Laboratory, Bar Harbor, ME) via a modified preplate technique that has been described previously (20). The characterization of the marker profile and behavior of the MDSCs compared with that of other populations of myogenic cells isolated from skeletal muscle using the preplate technique (including MDCs) is described elsewhere (17, 18, 20).

Retroviral transduction of MDSCs.

Retroviral vectors expressing BMP-4 or LacZ (CLBMP-4 or CLLacZ) were generated as described previously (33, 34). MDSCs were transduced separately with retroviral vectors at a multiplicity of infection of 5 in the presence of Polybrene (8 μg/ml). The transduced cells were expanded for 2 weeks before being used in animal experiments. MDSCs transduced with the CLLacZ retroviral vector were termed “MDSC.” Some of the LacZ-transduced cells then were transduced with the retroviral vector CLBMP-4; this process resulted in coexpression of LacZ and BMP-4 by the MDSCs. These cells were termed “MDSC-B4.” A previously described BMP-4 bioassay (33) was used to estimate the level of BMP-4 secreted by the transduced cells.

Monolayer cell culture and immunostaining for type II collagen (CII).

MDSC and MDSC-B4 were plated at a density of 2,500 cells/chamber on Falcon culture slides (Becton Dickinson, Franklin Lakes, NJ) and cultured in 1 of 3 types of media: 1) normal differentiation medium (Dulbecco's modified Eagle's medium [DMEM] supplemented with 1% fetal bovine serum, 1% horse serum, 0.5% chicken embryo extract, and 2% penicillin/streptomycin), 2) CM (high-glucose DMEM supplemented with dexamethasone, sodium pyruvate, ascorbate-2-phosphate, proline, insulin–transferrin–selenium+Premix, L-glutamine, and 1% penicillin/streptomycin) (Cambrex Bioscience, Walkersville, MD), or 3) CM (see description above) supplemented with TGFβ1 (10 ng/ml).

After 4 days in culture, cells were fixed with cold acetone for 20 minutes, blocked for 60 minutes with 10% sheep serum containing 0.5% Triton X-100, and incubated with primary antibody (rabbit anti-rat CII, 1:300; Chemicon, Temecula, CA) in blocking solution for 3 hours at room temperature in a humid chamber. The secondary antibody (Cy3-conjugated sheet anti-rabbit antibody, 1:300; Sigma, St. Louis, MO), was applied for 1 hour at room temperature. Nuclei were revealed with 4′,6-diamidino-2-phenylindole. Four chambers of the culture slide were used for each experimental group, and an E-800 microscope (Nikon, Melville, NY) was used with Northern Eclipse software (Empix Imaging Institute, Cheektowaga, NY) to calculate the number of CII-positive colonies in 5 fields per chamber.

Pellet cell culture and histochemical staining for glycosaminoglycans.

Cell pellets were made with MDSC and MDSC-B4, as previously described (35). The cells first were trypsinized and counted. Next, 2 × 105 cells in 0.5 ml of CM were centrifuged at 500g in 15-ml polypropylene conical tubes. The pellets were incubated at 37°C in 5% CO2, and the medium was changed every 2–3 days. Pellets were harvested after 14 days, fixed overnight in 10% buffered formalin, dehydrated, and embedded in paraffin. After sectioning, the 5-μm–thick pellet sections were deparaffinized, placed in 3% acetic acid for 3 minutes, and transferred to Alcian blue solution for 30 minutes. The slides were rinsed with running tap water for 1 minute and counterstained with nuclear fast red.

Repair of osteochondral defects.

The policies and procedures of our animal laboratory are in accordance with those published by the US Department of Health and Human Services. The research techniques used in these experiments were approved by the Animal Research and Care Committees at Children's Hospital of Pittsburgh and the University of Pittsburgh (protocol #12/02). Thirty-six 12-week-old athymic rats were used in this study. The animals were anesthetized by exposure to 3% isoflurane and O2 gas (1.5 liter/minute) delivered through an inhalation mask. A medial parapatellar skin incision was made, and the knee joint was exposed via lateral dislocation of the patella. A 1.5-mm biopsy punch (Robbins Instruments, Chatham, NJ) was used to create a full-thickness articular cartilage defect (1.5 × 1. 5 × 2.0 mm) in the trochlear groove of each femur. Cells were mixed with fibrin glue (Tisseel VH; Baxter Healthcare, Glendale, CA) before transplantation. The animals were divided into 3 treatment groups. The defects in the rats in group 1 were treated with 500,000 MDSC embedded in fibrin glue, defects in the rats in group 2 were treated with 500,000 MDSC-B4 embedded in fibrin glue, and defects in rats in group 3 (controls) were treated with acellular fibrin glue. After surgery, the rats were allowed to move freely within their cages.

Macroscopic evaluation and LacZ transgene expression at the repair site.

Rats were euthanized 4, 8, 12, and 24 weeks after surgery. At each time point, 6 defects from each treatment group were examined macroscopically and 2 samples of the repaired tissues were sharply excised en bloc. The tissues were flash frozen and cryostat sectioned. The sections were stained with β-galactosidase (β-gal) and analyzed to determine the average percentage of LacZ-positive cells.

Evaluation of chondrogenic differentiation of transplanted cells in vivo.

Sections obtained from the defect areas at the 4-week time point were stained with β-gal to assess LacZ expression and with Alcian blue to assess proteoglycan expression. Also, immunohistochemical staining was performed to evaluate the differentiation of transplanted cells into chondrocytes by colocalizing β-gal and CII. Staining for CII was performed using the same protocol described for the in vitro monolayer culture above. The dilution of the primary antibody (rabbit anti-rat CII) and secondary antibody (Cy3-conjugated sheet anti-rabbit antibody) was 1:100. To stain for β-gal, tissue sections were blocked with 2% horse serum for 1 hour and incubated overnight at room temperature in a humid chamber with the primary antibody (monoclonal mouse anti–β-gal, 1:200; Sigma) in 1% horse serum. The sections were then incubated with streptavidin 488 for 1 hour (1:1,000; Sigma) and mounted on slides.

Histologic evaluation of cartilage repair.

After macroscopic examination, 4 distal femora per group per time point were dissected and fixed with 10% buffered formalin for 48– 72 hours. They then were decalcified with decalcifying solution (Decalcifier II; Surgipath, Richmond, IL) for 24 hours and embedded in paraffin. Sagittal sections (5 μm thick) were obtained from the center of each defect and were stained with Safranin O–fast green. The histologic grading scale described by O'Driscoll et al (36) was used to evaluate the quality of the repaired tissue. Evaluation was performed by 2 of the authors (RK and SK), who were blinded as to which treatment the rats received.

Statistical analysis.

All data are expressed as the mean ± SD. Differences were analyzed by one-way analysis of variance. P values less than 0.05 were considered significant.

RESULTS

BMP-4 protein production by transduced cells.

The mean ± SD amount of BMP-4 protein secreted by MDSC-B4 1 week after transduction was 115 ± 20 ng/106 cells/24 hours; secretion at this level continued for at least 4 weeks in vitro. The secreted BMP-4 stimulated alkaline phosphatase activity in C2C12 cells (data not shown).

Chondrogenic differentiation of cells in monolayer and micromass pellet cultures.

Neither MDSC nor MDSC-B4 expressed CII after being cultured for 4 days in monolayer in normal differentiation medium (data not shown). When cultured in CM, MDSC formed very few CII-positive colonies (Figure 1a). The addition of TGFβ1 to the CM containing MDSC increased the number of positive colonies formed by the cells (Figure 1b). MDSC-B4 cultured in CM with or without TGFβ1 acquired a chondrocyte-like morphology and formed colonies surrounded by CII-positive matrix (Figures 1c and d).

Figure 1.

Type II collagen (CII) immunostaining in monolayer cell culture. After 4 days, only a few CII-positive colonies were detected in muscle-derived stem cell cultures (MDSCs) expressing LacZ (MDSC) grown in chondrogenic medium (CM) (a), but their number increased when transforming growth factor β1 (TGFβ1) (10 ng/ml) was added to the medium (b). Many CII-positive colonies were observed in MDSCs coexpressing LacZ and bone morphogenetic protein 4 (MDSC-B4), grown in CM with or without TGFβ1 (c and d). (Original magnification × 600.) There were significantly more CII-positive colonies in MDSC-B4 cultures grown in CM than in MDSC cultures grown in the same medium (e). Values in e are the mean and SD of 5 random fields from 4 cell culture chambers. ∗ = P = 0.0076.

Quantitative analysis (n = 4) demonstrated that monolayer cultures of MDSC-B4 contained significantly more CII-positive colonies than did MDSC cultures grown in CM (P = 0.0076) (Figure 1e). The addition of TGFβ1 to the CM containing MDSC-B4 did not lead to an increased number of CII-positive colonies relative to the number of colonies formed by the same cells cultured in CM without TGFβ1. Indeed, we observed a slight decrease in the number of CII-positive colonies when MDSC-B4 were cultivated in CM containing TGFβ1. However, there was no statistically significant difference in terms of the number of CII-positive colonies formed by MDSC-B4 or MDSC cultivated in CM containing TGFβ1.

MDSC cultured in CM did not form compact pellets, nor did the cultures (with or without TGFβ1) contain cells morphologically similar to chondrocytes (Figures 2a and b). However, culturing MDSC in CM that contained TGFβ1 resulted in the formation of Alcian blue–positive ECM (Figure 2b). In contrast, MDSC-B4 cultured in CM with or without TGFβ1 formed very compact pellets that contained hypertrophic chondrocyte-like cells surrounded by Alcian blue–positive matrix (Figures 2c and d).

Figure 2.

Chondrogenic differentiation of cells grown in pellet culture. MDSC cultured in CM for 14 days with or without TGFβ1 formed loose pellets (see insets) that did not contain chondrocyte-like cells (a and b); however, the addition of TGFβ1 to the medium resulted in a matrix that was Alcian blue positive (b). MDSC-B4 cultured in CM either with or without TGFβ1 formed compact pellets that contained chondrocyte-like cells (arrows) surrounded by an Alcian blue–positive matrix (c and d). (Original magnification × 600 in main panels; × 200 in insets.) See Figure 1 for definitions.

Expression of the LacZ transgene and chondrogenic differentiation markers in vivo.

Repaired tissue in the defects treated with MDSC contained many LacZ-positive cells 4 and 8 weeks after transplantation (Figures 3a and b). Some LacZ-positive cells remained detectable 12 weeks after transplantation (Figure 3c). However, we were unable to detect LacZ-positive cells in the defect area 24 weeks after transplantation (results not shown). Quantitative analysis of transgene expression revealed 37 ± 12.07% LacZ-positive cells (mean ± SD) at 4 weeks after transplantation, 28.03 ± 13.8% LacZ-positive cells at 8 weeks, and 8.89 ± 1.34% LacZ-positive cells at 12 weeks (Figure 3d).

Figure 3.

LacZ transgene expression after MDSC transplantation. LacZ-positive cells were observed in the repaired tissue 4 weeks (a), 8 weeks (b), and 12 weeks (c) after surgery, but not at the 24-week time point (original magnification × 400). Quantitative analysis of transgene expression showed a gradual decrease in the percentage of LacZ-positive cells within the reparative tissue (d). Values in d are the mean and SD in 5 random fields from the sagittal sections of articular cartilage in 2 treated defects. See Figure 1 for definitions.

The repaired tissue harvested 4 weeks after transplantation of MDSC-B4 contained many round hypertrophic chondrocytes displaying LacZ-positive (blue) nuclei (Figure 4a); furthermore, Alcian blue staining revealed a proteoglycan-rich matrix (Figure 4b). Immunohistochemical staining to colocalize cells expressing β-gal (donor cell marker) and CII (chondrogenic differentiation marker) within the repaired defects confirmed that some donor cells expressed CII and hence adopted a chondrocyte-like phenotype in vivo (Figure 4c).

Figure 4.

Chondrogenic differentiation of MDSC-B4 in vivo. LacZ-positive (a) and Alcian blue–positive (b) hypertrophic chondrocyte-like cells were present within the repaired cartilage defects 4 weeks after transplantation. Immunohistochemistry revealed colocalization of β-galactosidase (donor cell marker [green]) and type II collagen (chondrogenic differentiation marker [red]) in the repaired tissue (c). (Original magnification × 600.) See Figure 1 for definitions.

Macroscopic examination of the transplant site.

Gross examination of the cartilage defects 8 weeks after surgery revealed glossy white, well-integrated, repaired tissue in the MDSC-B4 treatment group. Treated regions in the MDSC group and the control group appeared patchy and only moderately well integrated with the surrounding normal cartilage (Figure 5). Twelve weeks after transplantation, the original defects in the MDSC-B4 group contained glossy white repaired tissue that appeared to be well integrated with the surrounding normal cartilage. At the same time point, however, regenerated tissue in the MDSC group and the control group appeared patchier, and the margin between the regenerated tissue and the normal cartilage was easily distinguishable.

Figure 5.

Gross appearance of the cartilage defect healing 8, 12, and 24 weeks after surgery. Only the repaired cartilage defects in the MDSC-B4 treatment group exhibited white, glossy, smooth surfaces throughout the 24-week healing period. See Figure 1 for definitions.

At the 24-week time point, defects in the MDSC-B4 treatment group were the only ones that still contained smooth repaired tissue that was well integrated with the normal cartilage. By this time point, the surface of the regenerated cartilage in the MDSC and control groups had become rough and the margin between the regenerated tissue and original cartilage was very clear. There also were signs of osteoarthritic changes, such as osteophytes, in the control group (Figure 5).

Histologic evaluation of repaired cartilage.

Histologic observations 8 weeks after osteochondral transplantation revealed that the defects in 2 of the 4 MDSC recipients were filled with repaired tissue that appeared to be well integrated with the surrounding articular cartilage. Most of the cells in the repaired tissue displayed chondrocytic morphology and were surrounded by matrix that was stained weakly by Safranin O (Figure 6a). The defects in the MDSC-B4 recipients had a smooth surface, and the repaired tissue in all 4 specimens was well integrated with the surrounding articular cartilage. Round chondrocyte-like cells were present within the matrix, which was stained more strongly by Safranin O (Figure 6d). Although the defect in 1 of the 4 specimens in the control group contained repaired tissue, the integration of the repaired tissue with the surrounding articular cartilage was very poor. Unlike repaired tissue in the MDSC and MDSC-B4 treatment groups, most of the tissue in the control group was fibrotic, and the matrix was not stained metachromatically by Safranin O (Figure 6g).

Figure 6.

Histologic evaluation and grading of articular cartilage repair 8, 12, and 24 weeks after transplantation. Eight weeks after surgery, the regenerated tissues in the MDSC and MDSC-B4 treatment groups had smooth surfaces, were completely integrated with the surrounding normal cartilage, and contained round chondrocytes surrounded by Safranin O–positive extracellular matrix (a and d). At the same time point, repaired tissue in the control treatment group was poorly integrated with the surrounding cartilage and contained fibrotic tissue (g). Twelve weeks after transplantation, the repaired tissue in the MDSC treatment group was not completely integrated with the adjacent cartilage and contained fewer chondrocyte-like cells, which were surrounded by Safranin O–negative matrix (b). The repaired tissue in the MDSC-B4 treatment group at this time was well integrated with the normal cartilage and contained chondrocyte-like cells and Safranin O–positive matrix (e). In the control treatment group 12 weeks after transplantation, the defects had clear margins and were partially filled with fibrotic tissue, which is indicative of early degenerative changes (h). At the 24-week time point, the defects in the MDSC and control groups exhibited incomplete repair and irregular surfaces with degenerative changes and contained fibrocartilage-like tissue (c and i). At the same time point, however, defects in the MDSC-B4 treatment group were filled with regenerated tissue that displayed nearly normal hyaline cartilage morphology, and the matrix was intensely metachromatically stained (f). (Original magnification × 40 in the upper panels; × 400 in the lower panels.) The histologic grading scores 8, 12, and 24 weeks after transplantation, respectively, in the MDSC-B4 treatment group, in the MDSC treatment group, and in the control group are shown (j). Values in j are the mean and SD of 4 knees per treatment group. ∗ = P < 0.05. See Figure 1 for definitions.

Twelve weeks after transplantation, the repaired tissue in the MDSC-treated defects had become thinner. In 1 of the 4 specimens, the repaired tissue had not completely bonded with the adjacent cartilage, whereas bonding had occurred in the remaining 3 specimens. The repaired tissue contained fewer chondrocyte-like cells, and those cells were surrounded by Safranin O–negative matrix (Figure 6b). In the MDSC-B4 treatment group, the repaired tissue also appeared thinner at 12 weeks than it had at 8 weeks, but the margins in 2 of the 4 specimens were completely bonded with the adjacent cartilage. Most of the cells in the repaired tissue displayed chondrocytic morphology and were surrounded by matrix that was stained at a moderate level by Safranin O (Figure 6e). In the control group, the repaired tissue became thinner, contained mostly fibroblast-like cells, and did not stain positive for Safranin O. Furthermore, evidence of early degenerative changes was apparent in the adjacent articular cartilage (Figure 6h).

Twenty-four weeks after transplantation, the repaired tissue in the MDSC-treated group appeared even thinner. The margins of the repaired tissue had not bonded with the adjacent cartilage in any of the 4 specimens. The repaired tissue contained fibrocartilage and fibrous tissue and stained negative for Safranin O (Figure 6c). In the MDSC-B4 treatment group, the repaired tissue had a smooth surface and the margins were completely integrated with the surrounding cartilage in 2 of the 4 specimens. Morphologically, most of the cells in the repaired tissue resembled normal hyaline cartilage, and the matrix was stained strongly by Safranin O (Figure 6f). In the control group, the repaired tissue was very thin, and the margins had not bonded with the adjacent cartilage in any of the 4 specimens. We observed no chondrocyte-like cells within the repaired tissue, the matrix stained negative for Safranin O, and the adjacent tissue displayed evidence of degenerative changes (Figure 6i).

Histologic grading of the repaired cartilage.

We used a histologic grading scale (36) to semiquantitatively evaluate the quality of the tissue repair. Regenerated tissue from the MDSC treatment group received significantly higher histologic scores than did tissue from the control group 8 and 12 weeks after transplantation (P = 0.036 and P = 0.02, respectively) (Figure 6j). However, there was no significant difference between the MDSC treatment group and the control group at the 24-week time point. The score in the MDSC-B4 treatment group was significantly higher than that in the control group 8, 12, and 24 weeks after transplantation (P < 0.0001, P = 0.0004, and P = 0.0001, respectively). Although there was no significant difference between the scores assigned to regenerated cartilage in the MDSC and MDSC-B4 treatment groups 8 and 12 weeks after transplantation, the score in the MDSC-B4 group was significantly higher than that in the MDSC group 24 weeks after transplantation (P = 0.019).

DISCUSSION

BMP4-expressing MDSCs used in ex vivo gene therapy produce bone by endochondral ossification, a process by which cartilage and bone form within the injected tissue (34, 37). We know of no other published studies designed to investigate the potential of these cells in promoting articular cartilage repair. The secretion of chondrogenic, antiarthritic, or chondroprotective proteins by locally transplanted cells could result in faster repair and restoration of normal hyaline cartilage within lesions. Here we evaluated the ability of BMP-4–expressing MDSCs to undergo chondrogenic differentiation in vitro and in vivo. We also assessed the feasibility of using MDSC-based ex vivo gene therapy to enhance the healing of full-thickness articular cartilage defects in vivo.

As evidenced by the results of this study, environmental stimuli, such as culture medium conditions and circulating growth factors, are essential to the differentiation of stem cells toward a specific lineage. MDSC and MDSC-B4 cultured in monolayer in normal differentiation medium did not undergo chondrogenic differentiation. When cells were cultured in CM they began to express CII, an indicator of chondrogenesis. We detected significantly more CII-positive colonies in the MDSC-B4 cultures than in the MDSC cultures maintained in CM for 4 days. This finding demonstrates the importance of BMP-4 secretion by the transduced cells and indicates that BMP-4 influences the differentiation of multipotent MDSCs toward the chondrogenic lineage.

We also investigated the effect of TGFβ1 on MDSC and MDSC-B4 cultures in vitro. Research has shown that TGFβ1 has a significant effect on chondrogenesis by mesenchymal stem cells derived from bone marrow (35, 38–43) and periosteum (44, 45). The addition of TGFβ1 to the MDSC cultures in CM increased the number of CII-positive colonies. However, the addition of TGFβ1 to the MDSC-B4 cultures in CM slightly decreased the number of CII-positive colonies. This finding suggests that exposure to TGFβ1 (10 ng/ml) for 4 days did not enhance the chondrogenic differentiation exhibited by MDSC-B4.

Studies involving embryonic stem cells and mesenchymal progenitor cells also have demonstrated the beneficial effect of BMP-4 on chondrogenic differentiation in vitro (30, 31, 46). Other research groups have used the culturing of cells in a micromass pellet culture system as an assay for chondrogenesis, and this approach appears to be optimal for assessing the chondrogenic differentiation of mesenchymal stem cells (35, 41–43). Here we used the pellet culture system and found that BMP-4 plays an important role in promoting the chondrogenic differentiation of MDSCs. This finding confirms the results obtained from the monolayer cell culture. BMP-4 promoted aggregation of the cells, as evidenced by the size and structure of the pellets, and led to their adoption of a chondrocyte-like morphology. The MDSC formed smaller, less condensed pellets that did not contain chondrocyte-like cells, whereas the MDSC-B4 formed bigger, more condensed pellets that contained cells that displayed hypertrophic chondrocyte morphology. This study shows that, under appropriate culture conditions, BMP-4–expressing MDSCs can acquire a chondrocytic phenotype in vitro.

To validate these in vitro results, we attempted to use genetically engineered MDSCs to repair articular cartilage in vivo. We began by demonstrating that transduction of the MDSCs with the LacZ marker gene enabled detection of the transplanted cells within the treated defects for up to 12 weeks after transplantation. Adachi et al (16) previously reported that MDCs and chondrocytes used for osteochondral transplantation expressed LacZ in the repaired tissues for only up to 4 weeks after transplantation. During the cartilage healing process, 4 weeks of protein expression may be insufficient to generate mature repaired tissue. Our results, compared with those generated by MDC or chondrocyte transplantation (16, 47, 48), indicate that the better survival of transplanted MDSCs led to significantly enhanced and prolonged LacZ transgene expression within the cartilage lesions.

LacZ-positive and Alcian blue–positive hypertrophic chondrocytes were present within the repaired tissues 4 weeks after transplantation of MDSC-B4. Furthermore, we were able to colocalize β-gal with CII, a chondrogenic marker, by immunohistochemistry. These findings indicate that the transplanted cells acquired a chondrocyte-like phenotype and participated in the repair of articular cartilage.

Macroscopic examination of the knee joints before preparation for histologic examination was predictive of the histologic results. Regenerated cartilage harvested from the MDSC- and MDSC-B4–treated defects received significantly higher histologic grading scores than did repaired tissue harvested from the control defects 8 and 12 weeks after transplantation. Furthermore, the repaired cartilage obtained from the MDSC-B4 treatment group received significantly higher scores than that obtained from the control or MDSC treatment groups 24 weeks after transplantation. This finding indicates that continuous endogenous BMP-4 supplied by MDSCs genetically modified to express BMP-4 over an extended period of time can enhance articular cartilage healing. Although we observed no ossification of the reparative tissue during the 24-week healing period in this study, we cannot exclude the possible occurrence of ossification at a later time.

Scaffold optimization is an essential step to ensure successful cartilage tissue engineering. Many researchers have chosen to focus on the study of cell-seeded matrices in an effort to engineer synthetic cartilage in vitro for implantation in vivo. The use of cells suspended in fibrin glue instead of solid grafts enables the suspended cells to settle even in narrow clefts, and fast-forming clots are able to prevent any initial displacement (31). Research has demonstrated that fibrin glue has no adverse effects on cell viability and is a suitable matrix for applications designed to promote chondrogenesis (49). Furthermore, some findings suggest that fibrin and fibronectin, the primary components of fibrin glue, could play a role in cell–matrix and matrix–matrix interactions (50). Our study confirmed that a fibrin glue matrix allows MDSCs to proliferate and acquire a chondrocyte-like phenotype after osteochondral transplantation.

This study demonstrates the feasibility of using ex vivo BMP-4 gene therapy involving retroviral transduction of MDSCs to repair articular cartilage. Skeletal muscle is a promising source of progenitor cells that can undergo chondrogenic differentiation in vitro and in vivo under proper stimulation. The MDSCs used here served as good carriers of a therapeutic gene and enabled the delivery of appropriate amounts of BMP-4 protein to the injury site. This technique resulted in improved repair of articular cartilage that persisted for 24 weeks after transplantation. Using other populations of MDSCs isolated from normal mice and retrovirally transduced to express BMP-4, we have been able to produce similar results in vitro and in vivo (data not shown). These findings suggest that BMP-4 gene therapy based on retrovirally transduced MDSCs is a potential strategy by which to improve articular cartilage healing.

Acknowledgements

The authors wish to thank Ryan Sauder for editorial assistance and Marcelle Huard and Jing Zhou for technical assistance.

Ancillary