Daping Hospital and Research Institute of Surgery of the Third Military Medical University, Chongqing, China
Center of Bone Metabolism and Repair, State Key Laboratory of Trauma, Burns, and Combined Injury, Trauma Center, Daping Hospital and Research Institute of Surgery of the Third Military Medical University, Chongqing 400042, China
Fibroblast growth factor (FGF) family members are involved in the regulation of articular cartilage homeostasis. The aim of this study was to investigate the function of FGF receptor 1 (FGFR-1) in the development of osteoarthritis (OA) and its underlying mechanisms.
FGFR-1 was deleted from the articular chondrocytes of adult mice in a cartilage-specific and tamoxifen-inducible manner. Two OA models (aging-associated spontaneous OA, and destabilization-induced OA), as well as an antigen-induced arthritis (AIA) model, were established and tested in Fgfr1-deficient and wild-type (WT) mice. Alterations in cartilage structure and the loss of proteoglycan were assessed in the knee joints of mice of either genotype, using these 3 arthritis models. Primary chondrocytes were isolated and the expression of key regulatory molecules was assessed quantitatively. In addition, the effect of an FGFR-1 inhibitor on human articular chondrocytes was examined.
The gross morphologic features of Fgfr1-deficient mice were comparable with those of WT mice at both the postnatal and adult stages. The articular cartilage of 12-month-old Fgfr1-deficient mice displayed greater aggrecan staining compared to 12-month-old WT mice. Fgfr1 deficiency conferred resistance to the proteoglycan loss induced by AIA and attenuated the development of cartilage destruction after surgically induced destabilization of the knee joint. The chondroprotective effect of FGFR-1 inhibition was largely associated with decreased expression of matrix metalloproteinase 13 (MMP-13) and up-regulation of FGFR-3 in mouse and human articular chondrocytes.
Disruption of FGFR-1 in adult mouse articular chondrocytes inhibits the progression of cartilage degeneration. Down-regulation of MMP-13 expression and up-regulation of FGFR-3 levels may contribute to the phenotypic changes observed in Fgfr1-deficient mice.
Osteoarthritis (OA) is one of the most prevalent chronic joint diseases and is characterized by progressive cartilage destruction and insufficient extracellular matrix synthesis. Thus far, none of the strategies employed to prevent and treat OA have been effective. The final option for advanced OA is the surgical approach of total joint replacement (1). Studies have shown that changes in growth factor signalings and their downstream target genes may be involved in the development of OA (2–5). However, the pathogenic mechanisms of OA are still largely unknown. Better understanding of the molecular events in OA development would provide important information that could facilitate the identification of novel therapeutic targets for the prevention and treatment of OA.
Chondrocytes are the only cell type in adult human articular cartilage in which the main responsibility is the synthesis and degradation of extracellular matrix (6). Numerous molecules and pathways, such as hypoxia-inducible factor 2α, discoidin domain receptor 2, and hedgehog signaling, in the articular chondrocytes have been shown to be involved in cartilage metabolism (7–11). Recently, the role of fibroblast growth factor (FGF) family members in the regulation of cartilage homeostasis has received specific attention (12). FGF-18 is a well-established anabolic growth factor that contributes to the metabolism of articular cartilage (13, 14). Although several lines of evidence support the tight association between FGF-2 and articular cartilage metabolism, the role of FGF-2 in cartilage homeostasis is controversial (12, 15).
Some studies of human articular chondrocytes have demonstrated that FGF-2 stimulates the production of matrix metalloproteinase 13 (MMP-13), which is the main collagenase responsible for collagen degradation (16). Results of other studies have suggested that FGF-2 functions as a chondroprotective factor in cartilage homeostasis (17–19). It has been reported that FGF-2 can suppress interleukin-1 (IL-1)–induced catabolic effects on human cartilage (18). In mice with Fgf2 deficiency, spontaneous OA, as well as instability-induced OA, is accelerated (19). In response to tissue injury or mechanical compression, FGF-2 can be released from the extracellular matrix to activate intracellular ERK signaling and regulate expression of chondrocyte-specific genes, suggesting that FGF-2 plays a homeostatic role in articular chondrocytes (20).
The involvement of FGF ligands in the maintenance of articular cartilage suggests that FGF receptors (FGFRs) may also play an important role in articular cartilage homeostasis. Notably, FGFR-1 and FGFR-3 are highly expressed in human articular chondrocytes and have been implicated in cartilage metabolism (21). Valverde-Franco et al showed that Fgfr3-knockout (KO) mice exhibited abnormal cartilage metabolism and early signs of OA, suggesting that FGFR-3 signaling may have a chondroprotective role during the development of OA (22). It is hypothesized that activation of FGFR-1 may exert antianabolic and procatabolic effects on adult human articular cartilage (12). However, since conditional deletion of Fgfr1 is known to have a lethal effect on mouse embryos (23), it is impossible to investigate the in vivo function of FGFR-1 in articular cartilage homeostasis using conventional Fgfr1-KO mice.
In this study, we investigated the role of FGFR-1 in articular cartilage function postnatally, using mice with tamoxifen-inducible and cartilage-specific conditional knockout (cKO) of the Fgfr1 gene (hereinafter referred to as Fgfr1 cKO mice). We evaluated the effects of FGFR-1 on the function of articular chondrocytes and the degeneration of articular cartilage in mice using 2 models of OA, as well as an antigen-induced arthritis (AIA) model. Our study demonstrated that Fgfr1 deficiency in mice attenuates articular cartilage degeneration in all 3 arthritis models. In addition, we found that blockade of FGFR-1 signaling could antagonize the IL-1β–induced up-regulation of MMP-13 and enhance the expression of FGFR-3 in human articular chondrocytes.
MATERIALS AND METHODS
Fgfr1fl/fl mice and Col2a1-CreERT2 mice were generated in the laboratories of Dr. Deng and Dr. Di Chen, respectively (23, 24). Both Col2a1-CreERT2 and Fgfr1fl/fl mice were backcrossed and maintained on a C57BL/6J background. Col2a1-CreERT2;Fgfr1fl/fl mice with Fgfr1 deficiency were obtained from offspring of the breeding of Fgfr1fl/fl mice with Col2a1-CreERT2–transgenic mice. The efficiency of Cre recombination in the cartilage of these adult mice was evaluated in a manner as previously reported (25, 26). Fgfr1 cKO mice and their Cre-negative wild-type (WT) control littermates were intraperitoneally injected, at age 8 weeks, with tamoxifen at a dose of 1 mg/10 gm body weight, administered daily for 5 days. Genotyping for Fgfr1 and the Cre transgene was carried out using polymerase chain reaction (PCR) (primer sequences available from the corresponding author upon request).
Surgical model of OA in mice.
Before surgical induction of OA, mice were intraperitoneally injected, at age 8 weeks, with tamoxifen, administered daily for 5 days. Destabilization of the medial meniscus (DMM) surgery was performed on the right knee joints of 10-week-old male mice, according to previously described procedures (27). DMM surgery will lead to the development of progressive cartilage degeneration. As a control, sham surgery was performed with medial capsulotomy only. Pathologic changes in the medial tibial plateau and medial femoral condyle of the knee joint were scored at 4 weeks or 8 weeks postoperatively. All procedures were approved by the Institutional Animal Care and Use Committee of Daping Hospital (Chongqing, China).
Mouse model of AIA.
Induction of AIA in mice was performed according to previously described methods (28). Briefly, to induce inflammatory arthritis, 10-week-old male mice were immunized by intradermal injection at the base of the tail with a 100-μl emulsion containing 100 μg of methylated bovine serum albumin (mBSA; in 0.9% saline) and an equal volume of Freund's complete adjuvant (catalog no. F5881; Sigma). Ten days after immunization, mBSA (10 μl of 20 μg/ml mBSA in 0.9% sterile saline) or an equal volume of vehicle was injected intraarticularly into the right or left knee joints. All knee joints were harvested on day 17 after immunization and processed for histologic assessment.
Microscopic scoring of mouse cartilage degeneration was performed in accordance with the recommendations of the Osteoarthritis Research Society International (OARSI) (29). Briefly, after the mice were killed, the knee joints were dissected free of skin and muscle. The joints were then fixed in 4% paraformaldehyde, decalcified in 20% formic acid, and embedded in paraffin. Either sagittal or frontal serial sections were obtained across the entire knee joint. Specifically, three 5-μm sections were placed on each slide, with each joint fully harvested in ∼10-slide intervals. Some intervening sections were used for immunostaining. Sections of the joints at every 50 μm (∼10 slides per joint) were stained with Safranin O–fast green to assess cartilage destruction.
Histologic grading of cartilage degeneration was performed using the OARSI-recommended subjective scoring system (on a scale of 0–6) (29). In addition, cartilage aggrecan depletion was scored (on a scale of 0–3) as a complement measure of cartilage degeneration, as previously described (30). All sections from WT and cKO mouse joints in each of the 3 arthritis models were stained on the same day, and all slides in each shelf were stained simutaneously to avoid staining variations between slides. The intensity of Safranin O staining in the growth plate of the femora and tibiae was used as an internal control between batches.
The severity of cartilage destruction was expressed as a summed score (sum of the 4 highest scores in all slides) and as a maximal score for the medial femora and medial tibiae separately within each section. We summed the 4 highest scores from 4 sections that showed the most significant histologic changes in all slides from each joint, to demonstrate both the severity and the extent of cartilage damage.
Tissue sections were deparaffinized using xylene and deprived of endogenous peroxidase activity with 3% H2O2, and then blocked with normal goat serum for 30 minutes. Sections were incubated overnight with rabbit anti–FGFR-1 polyclonal antibody (1:100 dilution; Abcam), rabbit anti–FGFR-3 polyclonal antibody (1:100 dilution; Santa Cruz Biotechnology), rabbit anti–MMP-13 polyclonal antibody (1:50 dilution; ProteinTech), and rabbit anti–type X collagen polyclonal antibody (1:200 dilution; Calbiochem-Merck). After rinsing the sections with phosphate buffered saline, a horseradish peroxidase–conjugated secondary antibody was applied, followed by staining with diaminobenzidene.
Isolation and culture of chondrocytes.
To obtain mouse primary chondrocytes, we harvested the knee joints from 5-day-old Col2a1-CreERT2;Fgfr1fl/fl mice and digested the joints with 0.1% collagenase (Gibco). Chondrocytes were maintained in 6-well plates at 1 million cells per well. On culture day 3, the cells were treated with 4-hydroxytamoxifen (4OH-TM) (1 μM) or vehicle for 48 hours. Alternatively, chondrocytes were stimulated with recombinant human IL-1β (1 ng/ml; R&D Systems) for 24 hours after incubation with 4OH-TM (i.e., 4OH-TM was still present during the culture period after the addition of IL-1β).
To obtain human articular chondrocytes, surgically removed human cartilage tissue was harvested from patients with OA who were undergoing total knee joint replacement or from patients who were hospitalized at Daping Hospital (Chongqing, China) because of a traffic accident. Collection of these samples was approved by our Institutional Review Board and Ethics Committee, and samples were obtained following the patients' provision of written consent. Articular chondrocytes were digested from the cartilage according to previously described methods (18). Isolated chondrocytes were seeded onto 12-well plates at 1 million cells per well and cultured in Dulbecco's modified Eagle's medium/F-12 containing 10% fetal calf serum for 3 days. Prior to treatment with IL-1β, the chondrocytes were preincubated with the FGFR-1 inhibitor PD166866 (100 nM; Sigma) for 2 hours and then incubated with IL-1β (1 ng/ml) and PD166866 for 24 hours.
Overexpression of Fgfr1 in ATDC5 cells.
ATDC5 prechondrocytes were maintained in 6-well plates at 1 million cells per well, and pcDNA3.1 or pcDNA3.1-Fgfr1 plasmids were transfected using Lipofectamine 2000 (Invitrogen), in accordance with the manufacturer's instructions. Cultures were harvested for extraction of RNA from the cells after 48 hours of incubation.
RNA was extracted from the chondrocytes using TRIzol reagent (Invitrogen). Total RNA was reverse-transcribed (RT) to complementary DNA (cDNA) using a PrimerScript RT reagent kit (Takara). For real-time PCR, cDNA was amplified with a Two-Step QuantiTect SYBR Green RT-PCR kit (Takara) in an Mx3000P PCR machine (Stratagene). Relative gene expression was calculated with MxPro software, which was provided by the manufacturer (Stratagene). All samples were measured in triplicate and normalized to an internal control, cyclophillin A. The annealing temperature was 57°C, and primer sequences were designed using the Primer Premier program, version 5.0.
Proteins were lysed from the articular chondrocytes using a radioimmunoprecipitation assay buffer containing protease inhibitors (Roche). Proteins of the same quality were separated on sodium dodecyl sulfate–polyacrylamide electrophoresis gels and transferred to PVDF membranes (Millipore). The membranes were blocked with 5% milk resolved in Tris buffered saline–Tween buffer, and probed by Western blotting with diluted antibodies. The primary antibodies used were anti–MMP-13 (Protein Group), with β-actin as a loading control (Sigma).
Results are expressed as the mean ± SD. All statistical comparisons between groups were conducted using Student's unpaired t-test, utilizing the SPSS program (version 13.0). P values less than 0.05 were considered statistically significant.
Histologic characterization of the joints of Fgfr1 cKO mice.
We first confirmed the efficiency of Cre-mediated recombination in adult mice by lacZ staining of the articular cartilage. Consistent with previous results (26), we observed positive staining for lacZ in the articular chondrocytes of Col2a1-CreERT2;Rosa26R mice at 5 months of age (results not shown), which indicated that generation of Col2a1-CreERT2 mice could lead to efficient Cre recombination in mouse articular chondrocytes at the adult stage. In addition, no lacZ-positive chondrocytes were observed in Col2a1-CreERT2;Rosa26R mice that had not received treatment with tamoxifen (results available from the corresponding author upon request).
To examine the role of FGFR-1 in articular cartilage homeostasis, Fgfr1 cKO mice and their WT littermates received, at age 8 weeks, treatment with tamoxifen, administered intraperitoneally for 5 days. Thereafter, spontaneous changes in the histologic features of the articular cartilage were analyzed in Fgfr1 cKO and WT male mice at ages 3, 6, and 12 months. No gross abnormalities in the articular cartilage and no changes in the morphologic features of the joints were observed in either the Fgfr1 cKO or WT mice. Staining of whole knee joint sections with Safranin O revealed no obvious differences in cartilage thickness or alterations in histologic features between Fgfr1 cKO and WT mice at ages 3 and 6 months (Figure 1A).
In WT mice at age 12 months, the intensity of Safranin O staining of the knee joints appeared to be greatly decreased in most cartilage regions (Figure 1A, arrow) compared to that in WT mice at age 6 months, but the integrity of tidemark was still well preserved and the intensity of Safranin O staining in the growth plate was comparable (Figure 1A, open arrowhead) (additional results available from the corresponding author upon request). In contrast, in the knee joints of Fgfr1 cKO mice, there were no apparent changes in proteoglycan intensity at age 12 months compared to that at age 6 months (Figure 1A and results available from the corresponding author). Compared to WT littermate control mice, the intensity of Safranin O staining of cartilage was greater in Fgfr1 cKO mice at age 12 months (Figure 1A, arrow, and results available from the corresponding author).
We also scored alterations in the proteoglycan content in the medial femoral condyle and medial tibial plateau across the whole joint, and the 5 highest scores from an individual joint were summed. We observed that the summed scores of proteoglycan loss in both the tibia and the femur were significantly decreased in Fgfr1 cKO mice compared to WT mice at age 12 months (Figure 1B). These results indicate that progression of spontaneous, aging-related proteoglycan loss is attenuated by Fgfr1 deficiency.
Resistance to AIA after cartilage-specific knockout of Fgfr1.
To determine whether FGFR-1 contributes to the development and progression of chemically induced arthritis, we created an AIA model in Fgfr1 cKO and WT mice and assessed the knee joints for histologic alterations. As expected, we observed that administration of mBSA led to the loss of sulfated proteoglycan in the articular cartilage of WT and Fgfr1 cKO mice. However, the extent of proteoglycan loss in the Fgfr1 cKO mice was less than that in WT control mice (Figures 2A and B). In the saline-injected knee joints, there were no apparent gross changes in either histologic features or proteoglycan content in the articular cartilage of either Fgfr1 cKO or WT mice (Figures 2A and B). These results indicate that mice with cartilage-specific knockout of Fgfr1 are resistant to the proteoglycan loss induced by AIA.
To quantify the extent of cartilage degeneration, Safranin O–stained sections of mBSA-treated knee joints were evaluated. Results of Safranin O staining of the medial femoral condyle and medial tibial plateau revealed that proteoglycan loss (as assessed using both the summed and maximal scores) was reduced in Fgfr1 cKO mice compared to WT mice (Figure 2C). These findings support the hypothesis that suppression of FGFR-1 protects against inflammation-induced proteoglycan depletion.
Delayed progression of articular cartilage destruction following DMM surgery in Fgfr1 cKO mice.
To determine the effect of Fgfr1 deficiency on the development of OA induced by injury, we performed DMM surgery in the right knee joints of Fgfr1 cKO and WT mice. It has been reported that female mice in this model are resistant to surgery-induced cartilage degeneration (31), and therefore we used only male mice in these experiments. DMM surgery in the right knee joints resulted in a progressive degeneration of cartilage, with little or no synovitis. Moreover, we observed that the expression of FGFR-1 was progressively up-regulated in WT mice following DMM surgery (results available from the corresponding author upon request). In contrast, the knee joints of sham-operated mice of either genotype (Fgfr1 deficient or WT) were normal or showed only mild changes at 4 or 8 weeks after the procedure (Figures 3A and 4A).
Histologic analysis of the knee joints showed that cartilage degeneration developed both at 4 and at 8 weeks after DMM surgery (Figures 3A and 4A). However, following DMM surgery, the responses related to instability-induced OA development were significantly different between the Fgfr1 cKO mice and their WT littermates. At 4 weeks after DMM surgery, the joints of WT mice exhibited a large depletion of proteoglycan in the superficial layer and some erosion that extended to the calcified cartilage (Figure 3A). In contrast, the knee joints of Fgfr1 cKO mice displayed only mild aggrecan loss within 4 weeks after DMM surgery (Figure 3A).
To quantify the extent and severity of the cartilage damage, we compared the summed and maximal structural damage scores in the knee joints of Fgfr1 cKO and WT mice at 4 weeks following surgery. We found that the summed and maximal scores for cartilage damage in the femur and tibia were decreased significantly in Fgfr1 cKO mice compared to WT mice at 4 weeks following DMM surgery (Figure 3B).
Within 8 weeks after DMM surgery, the knee joints of WT mice showed loss of uncalcified cartilage in some areas and exposed subchondral bone, whereas the knee joints of Fgfr1 cKO mice exhibited significantly lower levels of damage (Figure 4A). In the Fgfr1 cKO mice, progression of cartilage erosion was evident, especially in the medial tibial plateau, but the summed and maximal cartilage damage scores were still significantly lower in the Fgfr1 cKO mice compared to WT mice (Figure 4B). These results demonstrate that the development of OA induced by destabilization is significantly attenuated in mice with Fgfr1 deficiency.
Down-regulated expression of MMP-13 and increased expression of FGFR-3 in mice with Fgfr1 deficiency.
To explore the mechanisms underlying the delayed progression of joint degeneration in Fgfr1 cKO mice, we analyzed chondrocyte gene expression in Fgfr1-deficient and WT mice. The primary chondrocytes were isolated from Col2a1-CreERT2;Fgfr1fl/fl mice and treated with 4OH-TM on culture day 3 to drive Cre-mediated recombination. We performed real-time PCR to quantify the expression of Fgfr1, aggrecan, Fgfr3, Mmp13, Mmp3, and Adamts5 in WT and Fgfr1-deficient mouse primary chondrocytes.
When Fgfr1 was deleted from the articular chondrocytes, expression of aggrecan and Fgfr3 was increased significantly, while Mmp13 levels were significantly decreased (Figure 5A). The expression levels of other critical proteinase genes, such as Mmp3 and Adamts5, were not significantly changed in Fgfr1-deficient mouse chondrocytes compared to WT control mouse chondrocytes (Figure 5A).
Results of Western blotting showed that targeted deletion of the Fgfr1 gene significantly abolished the IL-1β–mediated up-regulation of MMP-13 expression (Figure 5B). We also found that overexpression of Fgfr1 in mouse prechondrogenic ATDC5 cells led to up-regulation of Mmp13 (Figure 5C). These results demonstrate that the deletion of the Fgfr1 gene in mouse primary chondrocytes will lead to down-regulation of the catabolic enzyme MMP-13 and accumulation of proteoglycan, indicating that suppression of FGFR-1 may inhibit catabolic action and promote anabolic metabolism in primary chondrocytes.
We next performed immunohistochemical staining of the knee joints of mice at age 12 months. Results of immunostaining showed that the expression levels of FGFR-1, type X collagen, and MMP-13 were significantly decreased in the knee joints of Fgfr1 cKO mice compared to WT mice (Figures 6A–C). These results indicate that inhibition of FGFR-1 signaling may delay the process of matrix degradation and articular condrocyte hypertrophy.
FGFR-1 and FGFR-3 are considered to be the predominant FGFRs expressed in human articular chondrocytes (21), and these 2 receptors may play opposite functions in cartilage homeostasis. We thus determined the expression of FGFR-3 in the knee joints of WT and Fgfr1 cKO mice that were subjected to either DMM surgery or sham operation.
Consistent with the results of messenger RNA (mRNA) quantification (Figure 5A), the expression of FGFR-3 in the sham-operated knee joints was increased in Fgfr1-deficient mice compared to WT mice (Figure 6D). At 4 weeks after DMM surgery, the expression of FGFR-3 decreased in both the WT and Fgfr1 cKO mice compared to that in the respective sham-operated control mice (Figure 6D). In the DMM groups, FGFR-3 expression was notably higher in the knee joints of Fgfr1-deficient mice compared to WT mice (Figure 6D). Our results suggest that FGFR-3 is down-regulated in the articular cartilage when OA is induced by DMM surgery. Deletion of the Fgfr1 gene will lead to a relatively higher expression of FGFR-3, even under conditions of DMM-induced joint instability.
With regard to the expression of FGFR-1 in articular cartilage chondrocytes, we compared the Fgfr1 mRNA levels in the chondrocytes between patients with OA and healthy control subjects, and found that Fgfr1 mRNA levels were significantly increased in OA patients compared to healthy controls (results available from the corresponding author upon request). When we treated the OA human articular chondrocytes with the FGFR-1 selective inhibitor PD166866, we found that the chondrocytes expressed significantly lower levels of MMP13 and ColX and higher levels of aggrecan. The MMP13 mRNA levels were elevated by 6.9-fold after treatment of the articular chondrocytes with IL-1β for 24 hours. However, treatment of the chondrocytes with PD166866 remarkably reversed the IL-1β–induced up-regulation of MMP13 (detailed data available from the corresponding author upon request). These results suggest that suppression of FGFR-1–mediated signaling could antagonize the up-regulation of MMP-13 induced by catabolic factors such as IL-1β.
OA is one of the most prevalent aging-related joint diseases and is the leading cause of disability in the elderly. Currently, there is no effective way to prevent and treat cartilage destruction in OA, which is a reflection of the insufficient understanding of the molecular mechanisms of the initiation and progression of OA. In our study, we provide evidence, for the first time, to indicate that genetic inhibition of Fgfr1 in mature adult mouse articular cartilage attenuated the development of OA.
Understanding of the physiologic role of FGFR-1 in the maintenance of articular cartilage homeostasis in vivo has been elusive, due to the lethal effects of conventional Fgfr1 knockout on mouse embryos (32). Adult mice with conditional knockout of the Fgfr1 gene in chondrocytes (Col2-Cre;Fgfr1fl/fl mice) will remain alive, but these mice undergo significant skeletal changes, including an increased height of the hypertrophic chondrocyte zone and significantly disturbed bone homeostasis (33), which may cause intrinsic changes in the articular cartilage and affect the susceptibility of articular cartilage to OA at the adult stage.
To determine the specific function of FGFR-1 in articular chondrocytes, we specifically disrupted Fgfr1 gene expression in articular cartilage of mice at an adult stage, in a tamoxifen-inducible and chondrocyte-specific manner, by crossing Fgfr1-floxed mice with Col2-CreERT2–transgenic mice. Fgfr1 cKO mice and their Cre-negative WT littermates were administered tamoxifen at the age of 8 weeks. Mice with the Fgfr1 deletion exhibited no gross abnormalities of body size, body weight, skeletal structure, or histologic features when compared to WT mice. Thus, these Fgfr1-deficient mice with grossly normal skeletons, joint structures, and morphologic features were suitable for study of OA at an adult stage.
To investigate the biologic effect of FGFR-1 on cartilage degeneration, 3 different arthritis models, including aging-associated spontaneous OA, surgery-induced OA, and AIA, were used. Histologic examination of the knee joints of mice in these 3 models revealed that deletion of the Fgfr1 gene resulted in a substantial protection against loss of aggrecan and against articular cartilage structural damage in the knee joint, demonstrating that down-regulation of FGFR-1 in articular cartilage has a chondroprotective effect, thereby slowing the development of OA. These observations are consistent with those from a recent in vitro study showing that FGFR-1 mainly transmits an FGF-2–mediated catabolic signal in human articular chondrocytes (21).
The primary characteristic of OA is an imbalance between anabolic effects on extracellular matrix synthesis and catabolic effects on matrix degradation. These dual processes of reduced matrix synthesis and/or increased production of degradative proteinases will lead to generation of articular cartilage matrix that is unable to transmit normal mechanical stress. The proteolytic functions of the ADAMTS and MMP family members play important roles in the development of OA. Adamts5, a key member in the ADAMTS family, is responsible for aggrecan degradation in mouse models of OA (34, 35), a feature that constitutes an early event in the development of OA (36). Furthermore, MMP-13 is considered the most active collagenase for type II collagen degradation, and greatly contributes to OA development (37). Genetically modified mice with constitutionally active MMP-13 expression showed development of OA, and Mmp13 deficiency protected mice against OA cartilage damage (38, 39).
FGF-2 has been shown to stimulate production of MMP-13 by activating the NF-κB and MAPK pathways in human articular chondrocytes (40). Moreover, IL-1β could stimulate the synthesis and secretion of multiple degradative enzymes (such as MMP-13) in cartilage, which contributes to OA development (41). In the present study, we found that deletion of the Fgfr1 gene decreased MMP-13 expression and significantly reversed the IL-1β–induced up-regulation of MMP-13 in primary chondrocytes. Consistent with this finding, we also found that inhibition of FGFR-1 with the selective inhibitor PD166866 antagonized the IL-1β–induced up-regulation of MMP-13 in human articular chondrocytes. In contrast, we did not detect any significant changes in Adamts5 expression in Fgfr1-deficient mouse chondrocytes. In addition, we observed that disruption of the Fgfr1 gene decreased the expression of type X collagen, enhanced the expression of aggrecan, and increased the accumulation of proteoglycan in articular cartilage.
These changes may be at least partly responsible for the decelerated progression of cartilage degeneration in Fgfr1-deficient mice. Our results suggest that Fgfr1 deficiency suppresses the expression of MMP-13 in articular chondrocytes, indicating that FGFR-1, as a critical regulator of MMP-13 expression, may be used as a therapeutic target for OA treatment.
The FGF family plays a critical role in cartilage development. In growth plate cartilage, FGFR-3 is mainly expressed in the proliferating and prehypertrophic chondrocytes. FGFR-3 suppresses the proliferation and differentiation of chondrocytes, while FGFR-1 functions as a potent mitogenic stimulator (42, 43). However, few studies have explored the physiologic role of FGFR signaling in vivo in adult articular chondrocytes. It has been reported that FGFR-1 and FGFR-2 are the predominant receptors in healthy mature articular cartilage in mice (19). In contrast, the major FGFRs expressed in human articular chondrocytes are considered to be FGFR-1 and FGFR-3 (21).
During OA development, FGFR-3 expression is down-regulated in articular chondrocytes from patients with OA (21), and early development of OA in Fgfr3-KO mice could be attributed to an increase in MMP-13 expression (22). We also found that expression of FGFR-3 was decreased in the knee joints of mice with OA induced by DMM surgery. These results suggest that down-regulation of FGFR-3 is at least partially responsible for the development of OA, and that FGFR-3 plays a potentially chondroprotective role in the articular cartilage.
FGFR-3–transmitted signaling in articular chondrocytes may have an effect on cartilage metabolism that is opposite to that of FGFR-1. In this study, we found that the absence of FGFR-1 in mouse articular cartilage not only could antagonize up-regulation of MMP-13 but also could lead to increased expression of FGFR-3 in articular chondrocytes. These results indicate that deletion of the Fgfr1 gene may, in addition to antagonizing the catabolic action of certain proteinases such as MMP-13 on cartilage, also delay the progression of OA by elevating the expression of FGFR-3. It is conceivable that knockdown of Fgfr3 in articular chondrocytes would attenuate the chondroprotective phenotype in Fgfr1-deficient mice. Studies involving chondrocyte-specific Fgfr1/Fgfr3 double-KO mice would provide further solid evidence to support the opposing roles of FGFR-1 and FGFR-3 in the development and progression of OA; these studies are now under way.
Overexpression of FGFR-3 could inhibit the Indian hedgehog (IHH) signaling pathway in the growth plate, and inhibition of IHH signaling is able to attenuate the severity of OA (11, 44). Whether IHH is involved in the chondroprotective effects of FGFR-3 requires further investigation. In addition, it has been reported that FGFR-3 is down-regulated by FGF-2 via the FGFR-1/ERK/MAPK pathway in human articular chondrocytes (21). Whether deletion of FGFR-1 in chondrocytes leads to up-regulation of FGFR-3 via down-regulation of the ERK/MAPK pathway needs further study.
In summary, these results show, for the first time, that conditional deletion of the Fgfr1 gene in mature mouse articular chondrocytes protects the knee joint cartilage against destruction, probably by suppressing the production of MMP-13 and enhancing the expression of FGFR-3. Our study provides valuable insights into the potential roles of FGFR-1 in articular cartilage metabolism. Therefore, these findings suggest that targeting of FGFR-1 could be a potential therapeutic strategy in patients with OA.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. L. Chen had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Weng, D. Chen, L. Chen.
Acquisition of data. Weng, Yi, Huang, Luo, Wen, Du, D. Chen, L. Chen.
Analysis and interpretation of data. Weng, Q. Chen, Deng, D. Chen, L. Chen.