Repair of articular cartilage defect by autologous transplantation of basic fibroblast growth factor gene–transduced chondrocytes with adeno-associated virus vector

Abstract

Objective

To examine the effects of basic fibroblast growth factor (bFGF) gene–transduced chondrocytes on the repair of articular cartilage defects.

Methods

LacZ gene or bFGF gene was transduced into primary isolated rabbit chondrocytes with the use of a recombinant adeno-associated virus (AAV) vector. These gene-transduced chondrocytes were embedded in collagen gel and transplanted into a full-thickness defect in the articular cartilage of the patellar groove of a rabbit. The efficiency of gene transduction was assessed according to the percentage of LacZ-positive cells among the total number of living cells. The concentration of bFGF in the culture supernatant was measured by enzyme-linked immunosorbent assay to confirm the production by bFGF gene–transduced chondrocytes. At 4, 8, and 12 weeks after transplantation, cartilage repair was evaluated histologically and graded semiquantitatively using a histologic scoring system ranging from 0 (complete regeneration) to 14 (no regeneration) points.

Results

LacZ gene expression by chondrocytes was maintained until 8 weeks in >85% of the in vitro population. LacZ-positive cells were found at the transplant sites for at least 4 weeks after surgery. The mean concentration of bFGF was significantly increased in bFGF gene–transduced cells compared with control cells (P < 0.01). Semiquantitative histologic scoring indicated that the total score was significantly lower in the bFGF-transduced group than in the control group throughout the observation period.

Conclusion

These results demonstrated that gene transfer to chondrocytes by an ex vivo method was established with the AAV vector, and transplantation of bFGF gene–transduced chondrocytes had a clear beneficial effect on the repair of rabbit articular cartilage defects.

Damage of articular cartilage leads to joint dysfunction associated with pain or limited range of motion and usually progresses to degeneration of the articular surface, resulting in osteoarthritis. It is well recognized that articular cartilage is a highly differentiated tissue with a limited capacity for self-repair. Current therapy for osteoarthritis consists of short-acting antiinflammatory drugs, intraarticular injection of steroids or other agents, such as hyaluronic acid, and surgical intervention. However, these treatments may not relieve joint pain completely. Therefore, cartilage repair seems to be essential for the prevention of a catastrophic outcome in a joint. Several studies describing the successful repair of osteochondral defects by the transplantation of cultured chondrocytes have been reported (1). However, a major problem with cartilage repair by autologous chondrocyte transplantation is that a large quantity of chondrocytes from normal articular cartilage is required, whereas donor sites have a limited capacity to provide chondrocytes.

Many studies have demonstrated that basic fibroblast growth factor (bFGF) is one of the most potent of the various growth factors for cartilage repair (2–6). In order to establish an efficient approach for the treatment of cartilage defects, it may be advantageous to maintain a certain level of growth factor locally for a long time.

A new therapeutic approach to cartilage repair, gene therapy, has been described, in which genes are transduced into chondrocytes with the use of naked DNA or viral vectors (7–10). However, problems with this method are related to the ability to obtain high-efficiency transduction, to maintain long-term expression of the therapeutic gene, and safety. Recently, the adeno-associated virus (AAV) has been recognized as a tool for transducing a gene into target cells (11–14). Gene therapy with AAV has several advantages, including the lack of virulence of the wild-type virus, the safety, since there is no replication activity alone, the ability to transduce to nondividing cells, the integration into the host genome, and the long-term expression of the transduced gene.

In this study, we attempted to use this new delivery vector to repair cartilage defects, in an ex vivo method. The purpose of this study was to evaluate the utility of the AAV vector for ex vivo gene delivery to chondrocytes and to investigate the repair of an articular cartilage defect by transplantation of bFGF gene–transduced chondrocytes.

MATERIALS AND METHODS

AAV vector production.

Two AAV constructs were prepared for this study: AAV-LacZ contained the bacterial β-galactosidase (LacZ) gene and AAV-bFGF contained the bFGF gene, which harbors a nuclear localization signal under the regulation of the cytomegalovirus immediate early promoter. The AAV subtype 2 vector plasmid used in this study, pLacZ, was derived from the vector plasmid pW1, which contains the LacZ gene, as previously described (15). Recombinant bFGF gene (GenBank accession no. X07285) was obtained from Takeda Chemical Industries (Osaka, Japan). A fragment containing bFGF complementary DNA was amplified by polymerase chain reaction using the following primer pairs with the Eco RI or the Xho I site: 5′-ATGAATTCATGGCTGCCGGCAGCATCACTTCGCTT-3′ and 5′-ATCTCGAGAGAGTCAGCTCTTAGCAGAC-3′. The fragment was subcloned between the Eco RI and Xho I sites of the pLacZ AAV vector plasmid to replace the LacZ gene (pbFGF). An AAV helper plasmid containing subtype 2 AAV rep and cap genes, which are required for replication and capsid formation of AAV vectors, pIM45 was used. A plasmid containing the E2A, E4, and VA genes of the adenovirus genome, pladeno-1, was used in place of helper adenovirus for AAV vector production.

Subconfluent human fetal kidney cells (293 cells) were cotransfected by the calcium phosphate coprecipitation method with pbFGF, pIM45, and pladeno-1 to produce the AAV-inducing bFGF gene (AAV-bFGF). After 48 hours, the cells were harvested and lysed in Tris HCl buffer (10 mM Tris HCl, 150 mM NaCl, pH 8.0) through 3 cycles of freezing and thawing. One round of sucrose precipitation and 2 rounds of CsCl density-gradient ultracentrifugation were performed to isolate AAV-bFGF from the lysates. The vector titer was determined by quantitative DNA dot-blot hybridization of the DNase I–resistant fraction.

Isolation of chondrocytes.

Thirty-nine 10-week-old Japanese white rabbits (Oriental Yeast Company, Tokyo, Japan), weighing an average of 1.8 kg, were used in this study. They were divided into 3 groups: 9 for the LacZ-transduced group, 12 for the bFGF-transduced group, and 18 for the control group. Under intravenous anesthesia with pentobarbital sodium (Somnopentyl; Schering-Plough, Union, NJ), articular cartilage tissues (4 × 4–mm slices) were harvested from the patellar groove of the right knee, washed 3 times in phosphate buffered saline (PBS), and cut into small pieces. The pieces were treated with 0.05% trypsin and 0.001M EDTA (Gibco BRL, Gaithersburg, MD) for 30 minutes at 37°C and digested sequentially with 0.25% collagenase (type II collagenase; Worthington, Lakewood, NJ) for 3 hours at 37°C. The isolated chondrocytes were washed 3 times with PBS.

The mean number of cells collected from each rabbit was 2.4 × 10⁵ (SD 0.5 × 10⁵). These cells were divided into 4–6 culture wells and cultured in 24-well flat-bottomed plates (Falcon, Lincoln Park, NJ) at a concentration of 5 × 10⁴ cells/well in 0.5 ml of Dulbecco's modified Eagle's medium (DMEM; Sigma, St. Louis, MO) supplemented with 10% fetal calf serum (FCS) and antibiotics (100 units/ml of penicillin G, 0.1 mg/ml of streptomycin; Gibco BRL) (DMEM–FCS), at 37°C in an atmosphere of 5% CO₂ in air.

Gene transduction into chondrocytes.

Chondrocytes were cultured for 3 days, removed from the growth medium, and washed once with serum-free medium. To the culture wells for the transduced group was added 500 μl of serum-free DMEM containing AAV-LacZ or AAV-bFGF to enable quantification of transgene expression at the optimal number of viral particles (10⁵ particles per cell) determined from the LacZ gene group experiments. To culture wells for the control group was added 500 μl of serum-free DMEM containing Tris buffer alone. After incubation for 1 hour at 37°C, 500 μl of DMEM–FCS was added to each culture well for both culture groups. One sample from each rabbit was used for autologous transplantation. The remaining samples were used for in vitro experiments.

Twenty samples from the LacZ-transduced group and 20 from the control group were used for the experiment to determine the efficacy of gene transduction in vitro. Culture medium was exchanged twice a week after gene transduction up to the time of analysis. At 3, 7, 14, 28, and 56 days after transduction, LacZ expression was assessed using the X-Gal staining technique (16), as follows. Cells were washed 3 times with PBS and fixed with 0.5% glutaraldehyde for 10 minutes, followed by 2 rinses in PBS containing 1 mmole/liter of MgCl₂. The cells were finally incubated with X-Gal substrate (1 mg/ml of X-Gal, 1 mmole/liter of MgCl₂, 5 mmoles/liter of K₄Fe[CN]₆/K₃Fe[CN]₆ in PBS) for 12 hours at 37°C. Efficiency of gene transduction was calculated as the average percentage of X-Gal–positive cells per total number of living cells in 3 randomly selected fields viewed with an optical microscope.

Measurement of bFGF concentration in culture medium.

Samples from 12 rabbits in the bFGF-transduced group and from 12 rabbits in the control group were used for determinations of accumulated bFGF production in the culture supernatant. The culture medium was not changed at each sampling of either group. At 3, 7, and 14 days after transduction, culture supernatants were collected from every 4 bFGF-transduced or control group culture wells, respectively, and after centrifugation, were stored at −80°C until analyzed. The bFGF concentration in the culture supernatants was measured by enzyme-linked immunosorbent assay (ELISA) using a bFGF-specific ELISA kit (Quantikine; R&D Systems, Minneapolis, MN) according to the manufacturer's instructions.

Autologous transplantation of gene-transduced chondrocytes into an articular cartilage defect.

Chondrocytes from the LacZ-transduced, bFGF-transduced, and control groups were cultured for 1 week after gene transduction, collected from the culture wells by trypsinization, and then centrifuged. The supernatant was removed, and chondrocytes were embedded in a 0.2% solution of type I collagen (Cellgen; Koken, Tokyo, Japan) at a density of 1 × 10⁶/ml. For autologous transplantation, chondrocytes were suspended in the collagen gel by incubation at 37°C for 1 hour.

Rabbits were anesthetized with pentobarbital sodium, and the left hind leg of each rabbit was sterilized for surgery. A 3-cm medial parapatellar incision was made over the knee, and the patella was dislocated laterally. A full-thickness defect in the articular cartilage (5 mm in diameter; 3 mm deep) was made in the patellar groove using a hand drill. The collagen gel containing ∼7.5 × 10⁴ autologous chondrocytes was transplanted into the full-thickness defect. A periosteal flap of ∼5 × 5 mm was harvested from the anteromedial surface of the tibia and sutured to the peripheral rim of the artificial defect with 5-0 nylon thread. The cambium layer of the periosteal flap was faced toward the joint space. The rabbits were allowed to move freely immediately after surgery.

Among the 39 rabbits, LacZ-transduced chondrocytes were transplanted into 9, bFGF-transduced chondrocytes into 12, and chondrocytes without gene transduction into the remaining 18.

Evaluation of LacZ expression at the site of transplantation.

Rabbits from the LacZ-transduced (n = 3) and control (n = 2) groups were killed at 1, 2, and 4 weeks after transplantation. The specimens were harvested from the patellar groove, embedded in TissueTek OCT compound (Sakura Finetek USA, Torrance, CA), and immediately frozen in nitrogen liquid. The frozen specimens were sectioned into 20-μm slices with a cryotome (Coldtome CM-502; Sakura Seiki, Tokyo, Japan) and double-stained with X-Gal and hematoxylin and eosin (H&E).

Histologic evaluation of repair cartilage.

Rabbits from the bFGF-transduced (n = 4) and control (n = 4) groups were killed at 4, 8, and 12 weeks after transplantation. The distal part of the femur was resected en bloc, fixed with 10% buffered formalin, and decalcified with a 0.5M EDTA solution. Sagittal sections were prepared and stained with H&E, toluidine blue, or Safranin O–fast green. The histologic features of each specimen were evaluated semiquantitatively using the histologic scoring system described by Wakitani et al (17). This system consists of 5 categories (cell morphology, matrix staining, surface regularity, cartilage thickness, and integration of donor with host) scored on a 0–14-point scale, where 0 = complete regeneration and 14 = no regeneration.

Statistical analysis.

Data are expressed as the mean ± SD. The statistical significance of differences was calculated with the use of StatView software (version J-5.0; Abacus Concepts, Berkeley, CA). One-way analysis of variance and the Mann-Whitney U test were used for analyzing statistical significance. P values less than 0.05 were considered significant.

RESULTS

In vitro experiment.

Efficiency of gene transduction of chondrocytes.

The efficiency of gene transduction was determined for chondrocytes transfected with AAV-LacZ at 7 days after transduction. The mean ± SD percentage of LacZ-positive cells among the total number of living cells was 43.7 ± 8.8%, 62.4 ± 5.1%, 97.7 ± 0.6%, and 98.2 ± 1.5% at a vector dose of 10³, 10⁴, 10⁵, and 10⁶ particles/cell, respectively (Figure 1). The percentage of successfully transduced chondrocytes increased in a vector dose–dependent manner. A vector dose of >10⁶ particles/cell did not improve the transduction rate. The optimal dose of virus that was required to achieve transduction of ∼100% of the chondrocytes was determined to be 10⁵ particles/cell.

Figure 1. Vector dose–dependent LacZ expression in cultured chondrocytes. On day 7 after adeno-associated virus–LacZ transduction into chondrocytes, LacZ expression was assessed by X-Gal staining. The percentages of LacZ-positive cells among the total number of living cells were 43.7%, 62.4%, 97.7%, and 98.2% at doses of 10³, 10⁴, 10⁵, and 10⁶ particles/cell, respectively.

LacZ gene expression was highly maintained until 56 days after gene transduction. The mean ± SD percentage of LacZ-positive cells was 69.4 ± 15.1%, 97.7 ± 0.6%, 97.2 ± 1.8%, 95.8 ± 2.9%, and 85.8 ± 6.2% at 3, 7, 14, 28, and 56 days after transduction, respectively (Figure 2). The greatest population of LacZ-expressing cells was 97.7% at 7 days (Figure 3A), and more than 85% of the population was maintained up to 56 days. Cells without LacZ transduction in the control group failed to reveal LacZ expression at any sampling point (Figure 3B). There was no microscopic evidence of cell death or cytopathologic changes in the transduced cells as determined by optical microscopy.

Figure 2. Time-dependent expression of LacZ in cultured chondrocytes. The percentages of LacZ-positive cells were 69.4%, 97.7%, 97.2%, 95.8%, and 85.8% at 3, 7, 14, 28, and 56 days after transduction, respectively.

Figure 3. Photomicrographs of transduced chondrocytes stained with X-Gal. A, LacZ group chondrocytes were stained with X-Gal on day 7 after adeno-associated virus–LacZ transduction. LacZ-positive cells are stained blue. B, Control group chondrocytes showed no LacZ-positive cells. (Original magnification × 100.)

Expression of bFGF gene by transduced chondrocytes.

Production of bFGF was detected in both bFGF-transduced cells and control cells. The mean ± SD bFGF concentration in culture supernatants from the bFGF-transduced cells was 88.2 ± 9.8 ng/ml, 130.9 ± 28.8 ng/ml, and 240.6 ± 22.5 ng/ml at 3, 7, and 14 days after transduction, respectively (Figure 4). In control cells, the bFGF concentration was 35.0 ± 15.8 ng/ml, 44.5 ± 16.4 ng/ml, and 62.3 ± 25.8 ng/ml at 3, 7, and 14 days after transduction, respectively. The bFGF concentration was significantly greater in bFGF-transduced cells than in the control cells on all sampling days (P < 0.01).

Figure 4. Concentration of basic fibroblast growth factor (bFGF) in culture supernatants of bFGF-transduced chondrocytes. Culture supernatants of control and bFGF-transduced cells were collected on days 3, 7, and 14 after transduction, and the bFGF concentration was determined by enzyme-linked immunosorbent assay. Transduction of the bFGF gene significantly elevated the secretion of bFGF (∗ = P < 0.01).

The mean number of chondrocytes in the bFGF-transduced group was 14.1 × 10⁴ (SD 1.1 × 10⁴) and 34.8 × 10⁴ (SD 6.2 × 10⁴) at 7 and 14 days after transduction, respectively. In the control group, the numbers were 6.4 × 10⁴ (SD 1.5 × 10⁴) and 17.0 × 10⁴ (SD 3.2 × 10⁴) at 7 and 14 days after transduction, respectively. The mean number of chondrocytes in the bFGF-transduced group was significantly higher than that in the control group during the period of culture (P < 0.01).

In vivo experiment.

LacZ expression at the transplant site.

Throughout the observation period, X-Gal staining of LacZ-transduced cells showed cells with blue nuclei distributed across the entire regenerated cartilage under the layer covered by the transplanted periosteal flap. The controls, which received transplants without LacZ-transduced chondrocytes, did not show LacZ-positive cells at any week of sampling. No adverse effects related to the virus were observed in this ex vivo gene transfer experiment.

Macroscopic findings at the site of transplantation of the bFGF-transduced chondrocytes.

Macroscopic observation of the transplant site showed regeneration of the articular cartilage defect in both the bFGF-transduced and control groups. At 12 weeks, the margin between the regenerated tissue and the original cartilage was not distinguishable in both groups. The surface of the regenerated cartilage closely resembled normal cartilage in the bFGF-transduced group, but that in the control group could still be distinguished from surrounding normal cartilage. No sign of osteoarthrosis, such as erosion of cartilage or osteophyte formation, was seen in any of the knees during the observation period.

Histologic findings of regenerated cartilage following transplantation of bFGF-transduced chondrocytes.

At 4 weeks after transplantation, in the tissues obtained from the bFGF-transduced group, the deep part of the regenerated tissue was composed of round chondrocytes, with an extracellular matrix that stained weakly with Safranin O (Figure 5A). There was no integration of the edges of the regenerated tissue with the adjacent normal cartilage or reconstitution of the osteochondral junction in any specimen. In the control group, the extracellular matrix was faintly stained with Safranin O (Figure 5B).

Figure 5. Photomicrographs of sagittal sections of articular cartilage defects in rabbits after transplantation with basic fibroblast growth factor (bFGF)–transduced chondrocytes (A, C, E, and G) and control chondrocytes (B, D, F, and H). Photomicrographs of the junction area are also shown (G and F). In the bFGF-transduced group, at 4 weeks (A), the deep layer is composed of round chondrocytes, but the matrix is weakly stained. At 8 weeks (C), the matrix is more distinctly stained, but the superficial region is weakly stained. At 12 weeks (E and G), the matrix is intensely metachromatically stained, and there is reconstitution of the osteochondral junction in most specimens. In the control group, at 4 weeks (B) and 8 weeks (D), the matrix is faintly stained. At 12 weeks (F and H), the deep layer of the matrix stained well, but staining is reduced in the superficial layers. Solid arrows indicate regenerated cartilage; open arrows indicate surrounding normal cartilage. Sections were stained with Safranin O–fast green. Bars in A and F = 1 mm; bars in G and H = 100 μm.

At 8 weeks, the bFGF-transduced group showed an extracellular matrix that was more distinctly stained with Safranin O in the deep part, while the superficial part was weakly stained (Figure 5C). Although both edges were integrated with the adjacent normal cartilage, reconstitution of the osteochondral junction was not seen in any specimen. The tissues from the control group at 8 weeks were essentially the same as those at 4 weeks (Figure 5D).

In the bFGF-transduced group at 12 weeks, the intensity and thickness of the extracellular matrix that was metachromatically stained with Safranin O were increased as compared with the findings at 4 and 8 weeks, and the microstructure of the regenerated tissue resembled the surrounding normal cartilage (Figures 5E and G). There was reconstitution of the osteochondral junction in most of the specimens. In the control group, the deep layer of the regenerated cartilage matrix stained well with Safranin O, but staining was reduced in the superficial layers (Figures 5F and H). There was no reconstitution of the osteochondral junction in any of the control specimens.

The Wakitani score was 6.5 ± 0.6, 4.3 ± 1.0, and 2.8 ± 1.0 points (mean ± SD) at 4, 8, and 12 weeks after transplantation, respectively, in the bFGF-transduced group. In the control group, the score was 8.8 ± 1.0, 7.0 ± 1.4, and 5.5 ± 1.7 points at 4, 8, and 12 weeks after transplantation, respectively (Table 1). The scores in both groups gradually decreased throughout the experimental period. However, the score in the bFGF-transduced group became significantly lower than that in the control group with the passage of time postoperatively (P < 0.01).

Table 1. Histologic scores of regenerated cartilage at 4, 8, and 12 weeks*

Weeks after transplantation	Control group	bFGF-transduced group
* The histologic features were scored on a 0–14-point scale as described by Wakitani et al (17), where 0 represents complete regeneration and 14 represents no regeneration. Values are the mean ± SD. bFGF = basic fibroblast growth factor. † P < 0.01 versus controls.
4	8.8 ± 1.0	6.5 ± 0.6†
8	7.0 ± 1.4	4.3 ± 1.0†
12	5.5 ± 1.7	2.8 ± 1.0†

DISCUSSION

Autologous chondrocyte transplantation has been successfully applied in recent years to the treatment of focal cartilage defects in a series of patients (18). Autologous chondrocytes for grafting are harvested from non–weight-bearing areas, cultured in vitro, reinjected into the cartilage defect, and the area is covered with a periosteal flap that is sutured in place. However, this method may be limited to small local cartilage defects, since the number of cells collected from donor sites to be used for cultivation is still limited.

Augmentation of the procedure with bFGF for help in repairing the cartilage has been reported to be efficacious (4–6, 19). Weisser et al (4) reported that among several different growth factors used to treat transplanted chondrocytes, positive effects on cartilage repair were observed only with the bFGF-treated chondrocyte implants (4). Previous studies showed that exogenous bFGF induces the proliferation of chondrocytes, the maturation of cartilage, and the differentiation of mesenchymal cells, and it stimulates the synthesis of cartilaginous matrix (5, 6). Otsuka et al (19) reported that continuous administration of bFGF using an osmotic pump had a clearly beneficial effect on repair of cartilage defects. However, bFGF alone did not lead to complete structural restitution of hyaline cartilage to repair the full-thickness defects of articular cartilage. Prolonged local expression of bFGF by transduction of the genetic code of bFGF into chondrocytes would be an efficient treatment for articular cartilage defects.

For the repair of cartilage defects with gene therapy, it may be necessary to obtain high-efficiency transduction and continuous local expression of the therapeutic gene. Several studies have demonstrated that the AAV vector has the ability to highly efficiently transduce a gene into cells, to integrate into the host genome, and to express the transduced gene for a long time (20–22). There are studies of the utility of the AAV vector for joint disease that demonstrated high-efficiency gene delivery to the synovium in vivo (23) or gene transduction to cultured chondrocytes in vitro (24). Delivering genes directly to the surface of the abnormal articular cartilage in order to accelerate cartilage repair could result in a long-term treatment. The advantage of ex vivo gene delivery would be direct delivery of the therapeutic gene to the abnormal articular cartilage and the ability to limit the area of gene expression to the cartilage defect alone. In our previous study, ex vivo gene transfer to periosteum-derived cells using an AAV vector induced LacZ expression for 4 weeks in vivo (25).

In this study, high-efficiency LacZ gene transduction into chondrocytes was obtained long-term in vitro, and LacZ gene expression in vivo was sustained without any adverse effects. These findings suggest that gene transfer to an articular cartilage defect by use of the ex vivo method was established with the AAV vector. Cartilage repair was slightly inferior to that described in previous reports, even though we used 10-week-old rabbits for the experiment. One of the reasons was thought to be differentiation to fibrous chondrocytes during the 1-week culture before transplantation. The cell number and the bFGF secretion were significantly increased in bFGF-transduced chondrocytes compared with the control chondrocytes in vitro. Furthermore, the histologic appearance of the transplant site in the bFGF-transduced group was fully repaired compared with that in the control group. The repair at a comparatively early stage was apparently different between bFGF-transduced and null chondrocytes even at 12 weeks. Continuous bFGF secretion by gene transfer seemed to be an effective way to promote cartilage repair.

These results demonstrate that repair of full-thickness defects in rabbit articular cartilage can be enhanced by transplantation of bFGF gene–transduced chondrocytes. This method seems to be one of the best techniques for achieving repair of articular cartilage defects.

Acknowledgements

We thank Avigen, Inc. (Alameda, CA) for supplying the plasmid for the production of the AAV vector.