During osteoarthritis (OA) chondrocytes show deviant behavior resembling terminal differentiation of growth-plate chondrocytes, characterized by elevated MMP-13 expression. The latter is also a hallmark for OA. TGF-β is generally thought to be a protective factor for cartilage, but it has also displayed deleterious effects in some studies. Recently, it was shown that besides signaling via the ALK5 (activin-like kinase 5) receptor, TGF-β can also signal via ALK1, thereby activating Smad1/5/8 instead of Smad2/3. The Smad1/5/8 route can induce chondrocyte terminal differentiation. Murine chondrocytes stimulated with TGF-β activated the ALK5 receptor/Smad2/3 route as well as the ALK1/Smad1/5/8 route. In cartilage of mouse models for aging and OA, ALK5 expression decreased much more than ALK1. Thus, the ALK1/ALK5 ratio increased, which was associated with changes in the respective downstream markers: an increased Id-1 (inhibitor of DNA binding-1)/PAI-1 (plasminogen activator inhibitor-1) ratio. Transfection of chondrocytes with adenovirus overexpressing constitutive active ALK1 increased MMP-13 expression, while small interfering RNA against ALK1 decreased MMP-13 expression to nondetectable levels. Adenovirus overexpressing constitutive active ALK5 transfection increased aggrecan expression, whereas small interfering RNA against ALK5 resulted in increased MMP-13 expression. Moreover, in human OA cartilage ALK1 was highly correlated with MMP-13 expression, whereas ALK5 correlated with aggrecan and collagen type II expression, important for healthy cartilage. Collectively, we show an age-related shift in ALK1/ALK5 ratio in murine cartilage and a strong correlation between ALK1 and MMP-13 expression in human cartilage. A change in balance between ALK5 and ALK1 receptors in chondrocytes caused changes in MMP-13 expression, thereby causing an OA-like phenotype. Our data suggest that dominant ALK1 signaling results in deviant chondrocyte behavior, thereby contributing to age-related cartilage destruction and OA.

Osteoarthritis (OA)3 is the most common joint disease. Patients with OA suffer from cartilage degeneration resulting in loss of joint function. The major risk factor for OA is age. Cartilage cells, chondrocytes, maintain the homeostasis of normal articular cartilage. However, in OA, chondrocytes have a distorted phenotype and produce MMP-13. This enzyme is the most potent type II collagen-degrading protein. Thus, in OA, the chondrocytes themselves induce destruction of articular cartilage. The articular chondrocytes in OA cartilage behave analogous to hypertrophic growth plate chondrocytes that have undergone terminal differentiation and start to produce MMP-13 in high amounts (1, 2). Until now it is not clear why chondrocytes in OA articular cartilage behave this way.

Chondrocyte terminal differentiation is repressed by the Smad2/3 pathway (3, 4). This pathway is activated upon binding of TGF-β to its type II receptor, which then recruits the type I receptor ALK5 (activin-like kinase 5). This complex in turn phosphorylates the intracellular effectors Smad2 and Smad3 (5, 6, 7, 8, 9). TGF-β signaling has been found to be important for maintenance and protection of healthy cartilage (4).

However, besides being a potent protective factor for cartilage, TGF-β also has the ability to induce OA under certain conditions (10, 11, 12, 13, 14). Until now this has been a puzzling, unexplained TGF-β property.

It has recently been shown that TGF-β also has the ability to signal via the alternative receptor ALK1 instead of the well-characterized ALK5 route. Activation of ALK1 has been found to result in activation of the Smad1/5/8 pathway (15), a pathway that is known to induce chondrocyte terminal differentiation (16). The balance between signaling via either ALK5 or ALK1 can determine the response of cells to TGF-β stimulation, which can be totally opposite (17, 18). For instance, in endothelial cells, ALK5 inhibits migration, whereas ALK1 stimulates migration and proliferation (19).

Our group has previously demonstrated that with age ALK5 expression and Smad2 phosphorylation strongly decrease in murine cartilage (20). Additionally, signaling via ALK5 drastically decreased with development of OA (21).

We hypothesize that with this decrease in ALK5 signaling, a shift toward a dominant TGF-β signaling via ALK1 occurs, thereby inducing a chondrocyte phenotype resembling hypertrophic chondrocytes expressing elevated MMP-13 levels. This would explain the contradictory results that have been found until now. Therefore, we have studied ALK1 and ALK5 expression in articular cartilage in aging mice, experimental models of OA, and human OA. Moreover, we have examined the functional consequences of dominant ALK1 or ALK5 signaling in chondrocytes, revealing that a shift in balance toward ALK1 induces an OA phenotypic change, thereby providing a new view on TGF-β signaling during aging and OA.

Chondrocytes were obtained from mouse cartilage by digestion with collagenase. They were immortalized by transduction with SV40 large T Ag and subsequently cloned by minimal dilution as previously described (22). All chondrocyte cell lines had both ALK5 and ALK1 expression. The chondrocyte cell line (H1) was cultured in 105 cells per well in 24-well plates in DMEM/HAM’s F12 (1:1) with 10% FCS and penicillin and streptomycin.

The H1 chondrocytes were stimulated with TGF-β (1 ng/ml) under serum-free conditions to evaluate functional TGF-β signaling via ALK5 or ALK1. This experiment was repeated in bovine primary chondrocytes to ensure that the chondrocyte cell line responded similarly to primary cells. The primary bovine chondrocytes were isolated from bovine metacarpophalangeal joint as previously described (23).

Additionally, H1 chondrocytes were transduced with either adenovirus overexpressing constitutive active ALK1 (Ad-caALK1) or Ad-caALK5 or Ad-LacZ control (multiplicity of infection of 50) for 2 h and cultured for 48 h thereafter to evaluate whether active ALK1 resulted in Smad1/5/8P and active ALK5 resulted in Smad2P or to investigate expression patterns on the mRNA level.

For blocking experiments, short hairpin RNA (shRNA) GFP-tagged plasmids containing shRNA for ALK1 or ALK5 or a negative control (SABiosciences) were used. H1 chondrocytes were plated in 75-cm2 flasks (106 cells/flask), and 5.8 μg of shRNA was transfected per flask using Lipofectamine 2000 reagent (Invitrogen). After 2 days the GFP-positive cells were separated from the nonpositive cells using FACS sorting (Epics Elite; Beckman Coulter).

Quantitative PCR was performed as previously described on an ABI Prism 7000 sequence detection system (Applied Biosystems) (24). Efficiencies (E) for all primer sets were determined (Table I) using a standard curve of five serial cDNA dilutions in water in duplicate. Primers were accepted if the deviation from the slope of the standard curve was <0.3 compared with the slope of the GAPDH standard curve and if the melting curve showed only one product.

Table I.

Primers used for quantitative PCR analysis

GeneR2EaForward Primer (5′→3′)Reverse Primer (5′→3′)
Mouse     
 GAPDH 0.997 2.05 GGCAAATTCAACGGCACA GTTAGTGGGGTCTCGCTCCTG 
 ALK1 0.999 1.90 ACCATCGTGAATGGCATCGT GGTCATTGGGCACCACATC 
 ALK5 0.999 2.02 AGCGGTCTTGCCCATCTTC AGCAATGGCTGGCTTTCCT 
 Collagen II 0.992 2.15 TTCCACTTCAGCTATGGCGA GACGTTAGCGGTGTTGGGAG 
 Aggrecan 0.992 2.15 TCTACCCCAACCAAACCGG AGGCATGGTGCTTTGACAGTG 
 MMP-13 0.992 1.93 ACCTTGTGTTTGCAGAGCACTAACTT CTTCAGGATTCCCGCAAGAGT 
Human     
 GAPDH 0.999 1.92 ATCTTCTTTTGCGTCGCCAG TTCCCCATGGTGTCTGAGC 
 ALK1 0.997 2.05 GACTCAAGAGCCGCAATGTG GGTCGGCGATGCAACAC 
 ALK5 0.993 2.05 CGACGGCGTTACAGTGTTTCT CCCATCTGTCACACAAGTAAAATTG 
 Collagen II 0.995 2.00 CACGTACACTGCCCTGAAGGA CGATAACAGTCTTGCCCCACTT 
 Aggrecan 0.997 2.01 GCCTGCGCTCCAATGACT ATGGAACACGATGCCTTTCAC 
 MMP-13 0.998 1.95 ATTAAGGAGCATGGCGACTTCT CCCAGGAGGAAAAGCATGAG 
GeneR2EaForward Primer (5′→3′)Reverse Primer (5′→3′)
Mouse     
 GAPDH 0.997 2.05 GGCAAATTCAACGGCACA GTTAGTGGGGTCTCGCTCCTG 
 ALK1 0.999 1.90 ACCATCGTGAATGGCATCGT GGTCATTGGGCACCACATC 
 ALK5 0.999 2.02 AGCGGTCTTGCCCATCTTC AGCAATGGCTGGCTTTCCT 
 Collagen II 0.992 2.15 TTCCACTTCAGCTATGGCGA GACGTTAGCGGTGTTGGGAG 
 Aggrecan 0.992 2.15 TCTACCCCAACCAAACCGG AGGCATGGTGCTTTGACAGTG 
 MMP-13 0.992 1.93 ACCTTGTGTTTGCAGAGCACTAACTT CTTCAGGATTCCCGCAAGAGT 
Human     
 GAPDH 0.999 1.92 ATCTTCTTTTGCGTCGCCAG TTCCCCATGGTGTCTGAGC 
 ALK1 0.997 2.05 GACTCAAGAGCCGCAATGTG GGTCGGCGATGCAACAC 
 ALK5 0.993 2.05 CGACGGCGTTACAGTGTTTCT CCCATCTGTCACACAAGTAAAATTG 
 Collagen II 0.995 2.00 CACGTACACTGCCCTGAAGGA CGATAACAGTCTTGCCCCACTT 
 Aggrecan 0.997 2.01 GCCTGCGCTCCAATGACT ATGGAACACGATGCCTTTCAC 
 MMP-13 0.998 1.95 ATTAAGGAGCATGGCGACTTCT CCCAGGAGGAAAAGCATGAG 
a

Efficiency.

For Western blot analyses each sample was measured for protein concentration, after which 20 μg of each protein sample was loaded on a SDS 7.5% polyacrylamide gel. Proteins were transferred onto a nitrocellulose membrane (Hybond P; Amersham Pharmacia Biotech) using the iBlot system (Invitrogen). The membrane was incubated with Abs against Smad1/5/8P (1/1000) and Smad2P (1/1000) (Cell Signaling Technology) overnight at 4°C. The secondary HRP-conjugated Ab was goat-anti-rabbit (1/1500) (Dako), and the membrane was developed using the ECL Plus Western blotting detection system (GE Healthcare).

Male C57BL/6N mice aged 1 year (n = 7) and 2 years (n = 10) were used for comparison of young vs old mice.

C57BL/6N mice aged 10 wk were used to induce OA by transaction of the medial meniscus (n = 10) (destabilization of the medial meniscus (DMM) model). In this surgical mouse model the anterior attachment of the medial meniscus to the tibia is transected, as previously described by Glasson et al. (25). Mice were sacrificed 8 wk after surgery.

STR/ort mice aged 3 mo (n = 10), 6 mo (n = 20), 9 mo (n = 6), and 1 year (n = 10) were used as a spontaneous model of OA. The histological lesions seen in this model resemble those seen in humans (26, 27).

Knee joints of sacrificed mice were isolated for histology. OA was confirmed by histological evaluation. Animals were considered to have hyaline cartilage degeneration, the signature pathologic feature of OA, if they had cartilage matrix degradation and/or regeneration/repair, chondrocyte death, chondrocyte replication, chondrocyte proliferation, cartilage lesions, bone lesion soft tissue laxity, or a combination of lesions (28).

Animals were kept in filter top cages with woodchip bedding under standard pathogen-free conditions. They were fed a standard diet with tap water ad libitum. The Animal Ethics Committee of the Radboud University Nijmegen approved all animal procedures.

Knee joints were fixed in phosphate-buffered formalin for 7 days. They were dehydrated using an automated tissue-processing apparatus (Tissue-Tek VIP; Sakura) and embedded in paraffin. Tissue sections of 7 μm were prepared.

Immunohistochemistry was performed as previously described (21). Specific primary Abs against ALK1 (1/100), ALK5 (1/100), Id-1 (inhibitor of DNA binding-1) (1/1000), and PAI-1 (plasminogen activator inhibitor-1) (1/100) were incubated overnight at 4°C (purchased from Santa Cruz Biotechnology). As a negative control, the primary Ab was replaced with goat or rabbit IgGs. A biotin-streptavidin detection system was used according to the manufacturer’s protocol (Vector Laboratories). Bound complexes were visualized via reaction with 3′,3′-diaminobenzidine (Sigma-Aldrich) and H2O2 resulting in a brown precipitate. Sections were counterstained with heamatoxylin and mounted with Permount.

For the different Ags, the number of positive articular chondrocytes in the tibia was determined by a blinded observer using the Qwin image analysis system (Leica Imaging Systems) and a Leica DC 300F digital camera as previously described (20). Briefly, a line with a defined width that corresponds to the width of the noncalcified cartilage was drawn across the cartilage in the image. The computer software defined the positive particles according to preset values of immunopositivity. For each knee joint, at least three tissue sections were measured and thereafter averaged per joint. The number of positive cells was corrected for the average number of chondrocytes in healthy joints determined in hematoxylin-stained sections.

Samples of human cartilage were obtained from patients undergoing knee or hip replacement surgery. The cartilage was instantly frozen in liquid nitrogen and crushed using a dismembrator. The crushed material was dissolved in RLT buffer supplied with the RNeasy Mini kit (Qiagen). Samples were treated with proteinase K, and subsequently RNA was isolated further with the RNeasy Mini kit according to the manufacturer’s protocol, after which an RT-PCR was performed followed by a quantitative PCR.

The Medical Ethics Committee of Radboud University Nijmegen Medical Centre approved the study protocol.

Results were analyzed with Student’s t test and considered significant if the p value was <0.05. Correlations were tested with Pearson’s correlation and considered significant if the p value was <0.05.

Alternative signaling of TGF-β via ALK1 has been shown in endothelial cells, but it was not known whether chondrocytes have the ability to signal via both type I TGF-β receptors as well. Therefore, expression of both type I TGF-β receptors, ALK1 and ALK5, on chondrocytes was evaluated by quantitative RT-PCR. This showed that both ALK1 and ALK5 were clearly expressed in murine cartilage, murine chondrocyte cell lines, bovine cartilage, and human cartilage (data not shown). Stimulation of H1 chondrocytes with TGF-β led to both Smad1/5/8 and Smad2/3 phosphorylation within 15 min, which lasted for at least 5 h (Fig. 1,A). The enhanced phosphorylation levels vanished after 24 h (Fig. 1,A). To determine if, similar to the chondrocyte cell line, primary chondrocytes were also able to signal via both routes in response to TGF-β, the experiment was repeated with primary bovine chondrocytes. These chondrocytes also responded to TGF-β by phosphorylation of Smad1/5/8 and Smad2. The bovine primary chondrocytes had a slower response on Smad1/5/8 phosphorylation (Fig. 1 A). These data show that TGF-β is not only able to signal via the Smad2/3 pathway, but also via the Smad1/5/8 route in chondrocytes.

FIGURE 1.

TGF-β signals through ALK1 and ALK5 in chondrocytes. A, Chondrocytes, both the murine H1 cell lines as the primary bovine chondrocytes, are able to signal via both ALK1 and ALK5 upon TGF-β stimulation, thereby displaying phosphorylation of Smad1/5/8 and Smad2, respectively. B, Chondrocytes (H1 cell line) were transduced with Ad-caALK1 or Ad-caALK5 after which Western blot was performed for Smad1/5/8P and Smad2P. Adenoviral overexpression of constitutive ALK5 leads to Smad2 phosphorylation, while constitutive ALK1 results in Smad1/5/8 phosphorylation.

FIGURE 1.

TGF-β signals through ALK1 and ALK5 in chondrocytes. A, Chondrocytes, both the murine H1 cell lines as the primary bovine chondrocytes, are able to signal via both ALK1 and ALK5 upon TGF-β stimulation, thereby displaying phosphorylation of Smad1/5/8 and Smad2, respectively. B, Chondrocytes (H1 cell line) were transduced with Ad-caALK1 or Ad-caALK5 after which Western blot was performed for Smad1/5/8P and Smad2P. Adenoviral overexpression of constitutive ALK5 leads to Smad2 phosphorylation, while constitutive ALK1 results in Smad1/5/8 phosphorylation.

Close modal

To assess whether activation of the ALK1 pathway leads to the expected Smad1/5/8 phosphorylation and ALK5 to Smad2/3 phosphorylation, we transduced the H1 chondrocytes with Ad-caALK1, Ad-caALK5, or Ad-LacZ as a control. The ΔΔCT values of transfected samples vs nontransfected controls was 10.99 for ALK1 and 11.3 for ALK5, and thus transfection led to a comparable increase in mRNA expression levels. It was clear that transfection with Ad-ALK1 specifically led to phosphorylation of Smad1/5/8P, whereas ALK5 resulted in Smad2 phosphorylation (Fig. 1 B).

Our group has previously demonstrated that aging was accompanied by a decrease in signaling via ALK5 in murine models for OA (20, 21) To investigate potential changes in ALK1 expression with age, we stained sections of knee joints of mice aged 1 year or 2 years immunohistochemically for ALK5 and ALK1. The number of both ALK1 and ALK5 immunopositive cells decreased with age (Fig. 2, A and B). For ALK1, the number of cells staining positive was reduced, with 15% in the medial tibial cartilage and 17% in the lateral tibial cartilage, but this was not significant. However, the number of ALK5-positive cells decreased, with even 80% and 70%, respectively. Thus, the drop in ALK5 expression is much more pronounced than the drop in ALK1 expression with age, hence altering the receptor balance with age (Fig. 2 C).

FIGURE 2.

The ALK1/ALK5 ratio increases with age. Murine knee joints of mice aged 1 or 2 years were isolated and prepared for paraffin sections in order to perform immunohistochemistry for ALK1 and ALK5. The number of cells staining positive in the tibia were measured with a computerized imaging system and corrected for the total number of cells in H&E-stained sections. The number of positive cells for ALK5 is strongly reduced with age (A), whereas the reduction in ALK1 staining is not significant (B). This results in a shift in the ALK1/ALK5 ratio with age favoring the ALK1 side (C). Representative sections are displayed in D. ∗, p < 0.05; ∗∗, p < 0.005; ∗∗∗, p < 0.001.

FIGURE 2.

The ALK1/ALK5 ratio increases with age. Murine knee joints of mice aged 1 or 2 years were isolated and prepared for paraffin sections in order to perform immunohistochemistry for ALK1 and ALK5. The number of cells staining positive in the tibia were measured with a computerized imaging system and corrected for the total number of cells in H&E-stained sections. The number of positive cells for ALK5 is strongly reduced with age (A), whereas the reduction in ALK1 staining is not significant (B). This results in a shift in the ALK1/ALK5 ratio with age favoring the ALK1 side (C). Representative sections are displayed in D. ∗, p < 0.05; ∗∗, p < 0.005; ∗∗∗, p < 0.001.

Close modal

In the DMM model, the ligament attaching the medial meniscus to the tibia is transected. This resulted in joint instability and OA development. As expected, the initial changes could be seen in the cartilage on the medial side of the tibia. The number of ALK1- and ALK5-positive cells was measured in the tibial cartilage both on the medial and the lateral side of the joint. No change in ALK5 and ALK1 expression was found on the lateral side of the joint, but a drastic decrease in both ALK1- and ALK5-positive cells in the medial tibial plateau was observed (Fig. 3). The number of ALK5-positive cells had decreased 89% compared with a 68% decrease of cells positive for ALK1 (Fig. 3). This emphasizes that in experimental OA the number of cells expressing ALK5 declines more rapidly than the number of cells expressing ALK1, thus increasing the ALK1/ALK5 ratio.

FIGURE 3.

The ALK1/ALK5 ratio increases with OA in the DMM model. Expression of ALK1 and ALK5 was determined in the knee joints of mice in which the medial meniscus was destabilized (DMM model) and compared with sham-operated controls. Immunohistochemistry was performed for ALK1 and ALK5 using specific Abs. The numbers of cells in the tibia staining positive were measured with a computerized imaging system and corrected for the total number of cells in H&E-stained sections. The number of positive cells for ALK5 is strongly reduced in the medial tibial cartilage surface in the DMM model (A). ALK1 expression is also reduced at the same site, but less than ALK5 (B). The ALK1/ALK5 ratio increases in the affected sites and thus in OA (C). Representative sections are displayed in D. ∗, p < 0.05; ∗∗, p < 0.005; ∗∗∗, p < 0.001.

FIGURE 3.

The ALK1/ALK5 ratio increases with OA in the DMM model. Expression of ALK1 and ALK5 was determined in the knee joints of mice in which the medial meniscus was destabilized (DMM model) and compared with sham-operated controls. Immunohistochemistry was performed for ALK1 and ALK5 using specific Abs. The numbers of cells in the tibia staining positive were measured with a computerized imaging system and corrected for the total number of cells in H&E-stained sections. The number of positive cells for ALK5 is strongly reduced in the medial tibial cartilage surface in the DMM model (A). ALK1 expression is also reduced at the same site, but less than ALK5 (B). The ALK1/ALK5 ratio increases in the affected sites and thus in OA (C). Representative sections are displayed in D. ∗, p < 0.05; ∗∗, p < 0.005; ∗∗∗, p < 0.001.

Close modal

STR/ort mice spontaneously develop OA at an early age. In STR/ort mice 2–3 mo old, initial changes can be seen on the medial tibia. Eighty percent of male mice have OA lesions by 6 mo of age. To assess the type I TGF-β receptor expression during spontaneous OA, changes in the numbers of cells staining positive for ALK1 and ALK5 in time were analyzed. This revealed a decrease in the number of positive cells for both ALK1 and ALK5 with age, and thus with progression of OA. The ALK1/ALK5 ratio clearly increased with progression of OA (Fig. 4). Where the number of cells staining positive for ALK5 was equal to ALK1 in the lateral tibia at 3 mo, the number of cells staining positive for ALK5 in the medial tibia was only 1% compared with 26% of the cells staining positive for ALK1. Strikingly, STR/ort mice develop OA first on the medial plateau. This is also the plateau that shows the first signs of decrease in ALK5 expression. These data again demonstrate that with progression of OA, the ALK1/ALK5 ratio increases (Fig. 4 C).

FIGURE 4.

The ALK1/ALK5 ratio increases with OA progression in the STR/ort model. Expression of ALK1 and ALK5 was determined in the knee joints of STR/ort mice that spontaneously develop OA. Immunohistochemistry was performed for ALK1 and ALK5 using specific Abs. The number of cells in the tibia staining positive were measured with a computerized imaging system and corrected for the total number of cells in H&E-stained sections. In the STR/ort model ALK5 is already extremely low in the medial side of the tibia and decreases with time as well in the lateral tibia (3 and 6 mo significantly different from all other time points). A, ALK1 staining is still present at 3 mo of age in the medial tibial cartilage. Staining reduces during aging but persisted compared with ALK5. On the lateral side ALK1 expression decreases until 9 mo of age (B). The ALK1/ALK5 ratio increases with age and OA progression (C). Representative sections are displayed in D. ∗, p < 0.05; ∗∗, p < 0.005; ∗∗∗, p < 0.001.

FIGURE 4.

The ALK1/ALK5 ratio increases with OA progression in the STR/ort model. Expression of ALK1 and ALK5 was determined in the knee joints of STR/ort mice that spontaneously develop OA. Immunohistochemistry was performed for ALK1 and ALK5 using specific Abs. The number of cells in the tibia staining positive were measured with a computerized imaging system and corrected for the total number of cells in H&E-stained sections. In the STR/ort model ALK5 is already extremely low in the medial side of the tibia and decreases with time as well in the lateral tibia (3 and 6 mo significantly different from all other time points). A, ALK1 staining is still present at 3 mo of age in the medial tibial cartilage. Staining reduces during aging but persisted compared with ALK5. On the lateral side ALK1 expression decreases until 9 mo of age (B). The ALK1/ALK5 ratio increases with age and OA progression (C). Representative sections are displayed in D. ∗, p < 0.05; ∗∗, p < 0.005; ∗∗∗, p < 0.001.

Close modal

Changes in receptor balance do not necessarily imply downstream alterations. Therefore, we investigated the changes in the corresponding specific downstream markers PAI-1 (downstream ALK5) (19) and Id-1 (downstream ALK1) (29) in a model for primary OA using STR/ort mice.

The Id-1/PAI-1 ratio was already very high at 3 mo of age in the medial tibial cartilage, where OA is first found. On the lateral side of the tibia this ratio was much lower at an age of 3 mo. However, both the medial and the tibial side of the tibia showed a pronounced increase in the Id-1/PAI-1 ratio during aging from 3 mo to 1 year. Thus, not only did we find altered receptor expression during OA progression, but also alterations in the downstream markers of the corresponding signaling pathways (Fig. 5).

FIGURE 5.

The Id-1/PAI-1 ratio increases with OA progression in the STR/ort model. Knee joints of STR/ort mice aged 3, 6, or 9 mo or 1 year were isolated for histology, and immunohistochemistry was performed for PAI-1 and Id-1. The numbers of cells staining positive in the tibia were measured with a computerized imaging system and corrected for the total number of cells in H&E-stained sections. A, The number of PAI-1-positive cells in tibial cartilage is drastically reduced within 9 mo. B, The number of Id-1-positive cells also decreases, but the decrease is less pronounced than for PAI-1, and the Id-1-positive cells still persist in the lateral tibial cartilage. C, The Id-1/PAI-1 ratio increases in time, with most marked increase in the lateral tibia. ∗, p < 0.05; ∗∗, p < 0.005; ∗∗∗, p < 0.001.

FIGURE 5.

The Id-1/PAI-1 ratio increases with OA progression in the STR/ort model. Knee joints of STR/ort mice aged 3, 6, or 9 mo or 1 year were isolated for histology, and immunohistochemistry was performed for PAI-1 and Id-1. The numbers of cells staining positive in the tibia were measured with a computerized imaging system and corrected for the total number of cells in H&E-stained sections. A, The number of PAI-1-positive cells in tibial cartilage is drastically reduced within 9 mo. B, The number of Id-1-positive cells also decreases, but the decrease is less pronounced than for PAI-1, and the Id-1-positive cells still persist in the lateral tibial cartilage. C, The Id-1/PAI-1 ratio increases in time, with most marked increase in the lateral tibia. ∗, p < 0.05; ∗∗, p < 0.005; ∗∗∗, p < 0.001.

Close modal

We found an increase in ALK1/ALK5 ratio with OA progression. This would suggest a correlation between ALK1 and deleterious effects in OA cartilage. To investigate this, cartilage samples of 20 OA patients undergoing joint replacement surgery were obtained for RNA isolation. The mRNA expression levels of ALK5, ALK1, essential matrix molecules collagen type II and aggrecan, and the most important degrading enzyme of collagen type II and marker of OA MMP-13 were measured by quantitative PCR. ALK1 expression significantly correlated with expression of MMP-13 in these samples (Fig. 6). In contrast, ALK5 expression showed no correlation with MMP-13 expression but was significantly correlated with mRNA levels of the cartilage matrix constituents aggrecan and type II collagen. This indicates that ALK1 expression is related to MMP-13 expression, while ALK5 expression is related to the expression of cartilage matrix molecules (Fig. 6).

FIGURE 6.

ALK1 is significantly correlated with MMP-13 expression, whereas ALK5 is significantly correlated with cartilage matrix components in human cartilage. Human cartilage was obtained from patients undergoing knee or hip replacement surgery. RNA was isolated from the cartilage and a quantitative PCR was performed for ALK1, ALK5, aggrecan, collagen type II, and MMP-13. ALK1 expression is significantly correlated with MMP-13 expression (A and D) while ALK5 expression is significantly related to type II collagen and aggrecan expression (A–C). ∗, p < 0.05; ∗∗, p < 0.005; ∗∗∗, p < 0.001.

FIGURE 6.

ALK1 is significantly correlated with MMP-13 expression, whereas ALK5 is significantly correlated with cartilage matrix components in human cartilage. Human cartilage was obtained from patients undergoing knee or hip replacement surgery. RNA was isolated from the cartilage and a quantitative PCR was performed for ALK1, ALK5, aggrecan, collagen type II, and MMP-13. ALK1 expression is significantly correlated with MMP-13 expression (A and D) while ALK5 expression is significantly related to type II collagen and aggrecan expression (A–C). ∗, p < 0.05; ∗∗, p < 0.005; ∗∗∗, p < 0.001.

Close modal

We showed that there was a correlation between ALK1 and MMP-13 while ALK5 correlated with matrix molecule expression in human OA cartilage. However, this merely suggests an involvement of ALK1 in MMP-13 up-regulation. To investigate whether ALK1 is a potential cause of OA-like changes, chondrocytes were transduced with Ad-caALK5 and Ad-caALK1. Overexpression of caALK5 resulted in a significantly increased expression of aggrecan mRNA, indicating an anabolic response (Fig. 7 A). Overexpression of caALK1 also showed a trend toward up-regulation of collagen type II and aggrecan, but this was not significant. However, overexpression of caALK1 resulted in a strikingly different pattern on MMP-13 expression. ALK1 signaling induced a significant increase in MMP-13 mRNA levels. Thus, caALK1 overexpression resulted in elevated expression of the chondrocyte terminal differentiation marker and marker for OA, MMP-13, while caALK5 induces elevated aggrecan mRNA levels and had no significant effect on MMP-13 mRNA levels.

FIGURE 7.

Overexpression of ALK1 or blocking of ALK5 both lead to up-regulation of MMP-13 expression. A, Chondrocytes (H1 cell line) were transduced with Ad-caALK1 or Ad-caALK5. After 2 days RNA was isolated for quantitative PCR. Collagen type II, aggrecan, and MMP-13 expression was measured. The experiment was performed four times. Ad-caALK1 expression significantly up-regulates MMP-13 expression, whereas Ad-caALK5 significantly up-regulates aggrecan expression. Ad-caALK1 also showed a trend toward up-regulation of collagen type II and aggrecan, but this was not significant. B, Chondrocytes (H1 cell line) were transfected with plasmids overexpressing ALK1 or ALK5 shRNA or a nonspecific shRNA. RNA was isolated after 2 days and quantitative PCR was performed for ALK1, ALK5, aggrecan, collagen type II, and MMP-13. Knockdown of ALK5 resulted in increased MMP-13 expression, whereas ALK1 knockdown resulted in nondetectable levels of MMP-13 mRNA. ∗, p < 0.05; ∗∗, p < 0.005; ∗∗∗, p < 0.001.

FIGURE 7.

Overexpression of ALK1 or blocking of ALK5 both lead to up-regulation of MMP-13 expression. A, Chondrocytes (H1 cell line) were transduced with Ad-caALK1 or Ad-caALK5. After 2 days RNA was isolated for quantitative PCR. Collagen type II, aggrecan, and MMP-13 expression was measured. The experiment was performed four times. Ad-caALK1 expression significantly up-regulates MMP-13 expression, whereas Ad-caALK5 significantly up-regulates aggrecan expression. Ad-caALK1 also showed a trend toward up-regulation of collagen type II and aggrecan, but this was not significant. B, Chondrocytes (H1 cell line) were transfected with plasmids overexpressing ALK1 or ALK5 shRNA or a nonspecific shRNA. RNA was isolated after 2 days and quantitative PCR was performed for ALK1, ALK5, aggrecan, collagen type II, and MMP-13. Knockdown of ALK5 resulted in increased MMP-13 expression, whereas ALK1 knockdown resulted in nondetectable levels of MMP-13 mRNA. ∗, p < 0.05; ∗∗, p < 0.005; ∗∗∗, p < 0.001.

Close modal

As we found significant changes in the ALK1/ALK5 ratio in vivo and a clear association between higher ALK1 and MMP-13, the most crucial question that remained was whether this reduction in ALK5 and thus a shift in ALK1/ALK5 ratio that we found with age could actually induce elevation of MMP-13. Therefore, chondrocytes were transfected with a plasmid containing shRNA against ALK5. This resulted in reduced levels of ALK5 expression (reduction of 1.8 cycles) and induced increased levels of MMP-13 mRNA compared with the negative plasmid control (5.7 cycles). Transfection of chondrocytes with shRNA for ALK1 resulted in nondetectable levels ALK1 and nondetectable levels of MMP-13 and thus a reduction in MMP-13. No changes were found in the expression of aggrecan or collagen type II. Although Ad-caALK5 significantly increased aggrecan expression, the difference between Ad-caALK5- and Ad-caALK1-induced aggrecan expression was not significant. Thus, it is not expected that inhibition of only one receptor will result in significant changes in aggrecan expression. These findings show that the change in balance between ALK1 and ALK5 that was found with age and even more so during OA will result in elevated MMP-13 levels in chondrocytes downstream, which is a hallmark for OA (Fig. 7 B).

Our findings introduce a new view on TGF-β signaling function with age and OA. Moreover, it provides an explanation for the enigmatic findings in the past and proposes a new mechanism that contributes to understanding OA development.

Growth factors such as TGF-β are able to stimulate chondrocytes in the cartilage of young animals to renew their extracellular matrix (4, 30, 31, 32, 33, 34, 35). This has always been viewed as the main function of TGF-β in cartilage. However, during aging and under OA conditions, we have shown that cartilage shows drastically reduced TGF-β signaling via ALK5, as demonstrated by diminished expression of ALK5 and reduced Smad2 phosphorylation (21). In aging mice that do not show cartilage degradation yet, there is already a decrease in the expression of the ALK5 receptor and in Smad2 phosphorylation in cartilage (20). These data indicate that decreased TGF-β signaling through the canonical ALK5 pathway could be a prerequisite to OA development. In endothelial cells, Goumans et al. suggested that the balance between ALK5 and ALK1 determines the outcome of the response to TGF-β (19). Given the reduced ALK5 expression during OA, this led to the hypothesis that a skewed balance between ALK1 and ALK5 during OA favored the ALK1 side leading to an OA phenotype.

From the studies in endothelial cells it has become clear that TGF-β has the ability to signal via ALK1 besides ALK5, thereby phosphorylating Smad1/5/8 (19, 36). Our data show that chondrocytes, like endothelial cells, also express both ALK5 and ALK1 and that they have the ability to respond to TGF-β using both routes. We show that with age and in OA cartilage, the reduction in ALK5 is far more drastic than the mild reduction in ALK1-positive cells, thus leading toward a skewed balance when compared with normal cartilage. The change in the ratio of Id-1/PAI-1 expression was in line with a shift from a dominant ALK5 to a more pronounced ALK1 signaling. The initial PAI-1 signaling was relatively high, similar to our previous findings of Smad2P expression in young mice (20). Thus, with age, not only is the anabolic function of TGF-β signaling reduced, but the terminal-differentiation potentiating function that we found in vitro becomes more dominant, thereby potentially making aged cartilage prone to develop OA.

ALK1 stimulates the Smad1/5/8 route, which is known to stimulate terminal differentiation and MMP-13 expression in growth plate chondrocytes. It has been frequently hypothesized that OA chondrocytes undergo differentiation toward a hypertrophic-like state. Chondrocytes from OA cartilage, even in unaffected areas, already express markers of hypertrophy when compared with normal cartilage (1, 2, 37). In human OA cartilage, ALK5 correlated with high aggrecan expression as well as high levels of collagen type II expression, both crucial constituents of the cartilage matrix. In contrast, samples with a high ALK1 expression also had a high MMP-13 expression, which is a hallmark of chondrocyte hypertrophy and OA.

As in endothelial cells, also altered signaling in chondrocytes leads to a change in outcome: rather than an anabolic response in a predominant ALK5 signaling situation, a shift in receptor balance favoring ALK1 signaling, either by overexpression of ALK1 or by blocking ALK5, led to elevation of MMP-13. This not only confirms our human cartilage data, but moreover adds a function to the correlation that was found: changing the receptor balance between ALK1 and ALK5 favoring the ALK1 side has a functional consequence and is able to induce an OA-like phenotype.

Therefore, we postulate that in human OA cartilage, ALK1 signaling apparently stimulates type II collagen degradation via MMP-13 while ALK5 promotes anabolic pathways in chondrocytes, thereby stimulating the synthesis of cartilage matrix molecules. Thus, with age and OA, ALK1 pushes the chondrocytes toward a hypertrophic-like state, expressing MMP-13.

We propose the following hypothesis for OA development and progression. During aging the ALK1/ALK5 ratio increases and Smad1/5/8 signaling (ALK1) becomes dominant relative to Smad2/3 signaling (ALK5). This results in stimulation of articular chondrocytes to differentiate to a chondrocyte with a hypertrophic-like phenotype, expressing high levels of MMP-13. High MMP-13 expression results in degradation of the cartilage collagen network, making the cartilage matrix vulnerable to normal loading. This is the developmental stage of cartilage degeneration in age-related primary OA. When cartilage lesions have developed, a subpopulation of chondrocytes will show high MMP-13 expression, but another subpopulation of chondrocytes in the affected cartilage will show an attempted repair reaction. As a consequence, OA cartilage will be a mixture of cells with a hypertrophic-like phenotype expressing high ALK1 and MMP-13 levels and a population of chondrocytes with high ALK5 expression and high expression of matrix molecules such as aggrecan and type II collagen. The ratio of both populations will be dissimilar in different samples of OA cartilage as a result of OA stage and location of cartilage sampling (distance to severe OA lesions).

A central role for TGF-β in OA becomes apparent. Initially TGF-β, signaling via ALK5 in chondrocytes, is a protective factor for articular cartilage, blocking terminal differentiation and MMP-13 synthesis and stimulating the synthesis of matrix molecules. During aging the role of TGF-β changes. Due to variation in receptor expression (ratio ALK1/ALK5), TGF-β becomes a factor stimulating terminal differentiation and MMP-13 expression via ALK1 signaling. A role for TGF-β in OA joints is confirmed by the role of TGF-β in the formation of osteophytes and synovial fibrosis, as we have demonstrated in previous studies (31, 32, 35, 38). TGF-β is protective for cartilage at a young age but is highly involved in OA pathology, inducing cartilage degradation, osteophyte formation, and synovial fibrosis in older individuals. This concept explains our enigmatic findings in earlier studies with regard to TGF-β, with TGF-β being both a potent protective factor for cartilage and a factor that has the ability to induce OA under certain conditions (10, 11, 12, 13, 14). Whether TGF-β induction of osteophytes and synovial fibrosis requires ALK5 or ALK1 as a preferred route remains to be investigated. However, a role for ALK1 in fibrosis has been suggested (39). Osteophytes undergo chondrogenic differentiation similar to what is observed in the growth plate. It can be anticipated that the cells forming the osteophytes require ALK1 signaling at a certain stage. This would make blocking ALK1 during OA even more appealing.

In conclusion, we show that there are changes in ALK1 and ALK5 expression with age and OA favoring dominant ALK1 signaling. Shifting the receptor balance between ALK1 and ALK5 in chondrocytes alters the eventual outcome downstream; in case of aging and OA, this will lead to increased MMP-13 expression. This model provides both an explanation for the development of OA during aging, the shift in chondrocyte phenotype and MMP-13 expression during OA, and provides an explanation for contradictory findings in TGF-β function in cartilage in the past. Our findings not only have implications for the view on OA development, but they also change the way that TGF-β might be used as a reparative factor for cartilage during OA, as the receptor balance within the tissue is crucial for the subsequent downstream response.

We thank the Department of Orthopedics, Radboud University Nijmegen Medical Centre Nijmegen, for the collection of human tissue samples.

The authors have no financial conflicts of interest.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This research was supported by the Dutch Arthritis Association “National Reumafonds”.

3

Abbreviations used in this paper: OA, osteoarthritis; Ad-caALK, adenovirus overexpressing constitutive active ALK; ALK, activin-like kinase; DMM, destabilization of the medial meniscus; Id-1, inhibitor of DNA binding-1; PAI-1, plasminogen activator inhibitor-1; shRNA, short hairpin RNA.

1
Kamekura, S., Y. Kawasaki, K. Hoshi, T. Shimoaka, H. Chikuda, Z. Maruyama, T. Komori, S. Sato, S. Takeda, G. Karsenty, et al
2006
. Contribution of runt-related transcription factor 2 to the pathogenesis of osteoarthritis in mice after induction of knee joint instability.
Arthritis Rheum.
54
:
2462
-2470.
2
Kawaguchi, H..
2008
. Endochondral ossification signals in cartilage degradation during osteoarthritis progression in experimental mouse models.
Mol. Cells
25
:
1
-6.
3
Ferguson, C. M., E. M. Schwarz, P. R. Reynolds, J. E. Puzas, R. N. Rosier, R. J. O'Keefe.
2000
. Smad2 and 3 mediate transforming growth factor-β1-induced inhibition of chondrocyte maturation.
Endocrinology
141
:
4728
-4735.
4
Yang, X., L. Chen, X. L. Xu, C. L. Li, C. F. Huang, C. X. Deng.
2001
. TGF-β/Smad3 signals repress chondrocyte hypertrophic differentiation and are required for maintaining articular cartilage.
J. Cell Biol.
153
:
35
-46.
5
Nakao, A., T. Imamura, S. Souchelnitskiy, M. Kawabata, A. Ishisaki, E. Oeda, K. Tamaki, J. Hanai, C. H. Heldin, K. Miyazono, P. ten Dijke.
1997
. TGFβ receptor mediated signalling through Smad2, Smad3 and Smad4.
EMBO J.
16
:
5353
-5362.
6
Lin, H. Y., A. Moustakas.
1996
. TGF-β receptors: structure and function.
Cell. Mol. Biol.
40
:
337
-349.
7
Moustakas, A., S. Souchelnytskyi, C. H. Heldin.
2001
. Smad regulation in TGF-β signal transduction.
J. Cell Sci.
114
:
4359
-4369.
8
Inman, G. J., F. J. Nicolas, C. S. Hill.
2002
. Nucleocytoplasmic shuttling of Smads 2, 3, and 4 permits sensing of TGF-β receptor activity.
Mol. Cell
10
:
283
-294.
9
Roberts, A. B..
1999
. TGF-β signaling from receptors to the nucleus.
Microbes Infect.
1
:
1265
-1273.
10
Takahashi, N., K. Rieneck, P. M. van der Kraan, H. M. van Beuningen, E. L. Vitters, K. Bendtzen, W. B. van den Berg.
2005
. Elucidation of IL-1/TGF-β interactions in mouse chondrocyte cell line by genome-wide gene expression.
Osteoarthritis Cartilage
13
:
426
-438.
11
van Beuningen, H. M., H. L. Glansbeek, P. M. van der Kraan, W. B. van den Berg.
2000
. Osteoarthritis-like changes in the murine knee joint resulting from intra-articular transforming growth factor-β injections.
Osteoarthritis Cartilage
8
:
25
-33.
12
Glansbeek, H. L., H. M. van Beuningen, E. L. Vitters, P. M. van der Kraan, W. B. van den Berg.
1998
. Stimulation of articular cartilage repair in established arthritis by local administration of transforming growth factor-β into murine knee joints.
Lab. Invest.
78
:
133
-142.
13
van Beuningen, H. M., P. M. van der Kraan, O. J. Arntz, W. B. van den Berg.
1994
. In vivo protection against interleukin-1-induced articular cartilage damage by transforming growth factor-β1: age-related differences.
Ann. Rheum. Dis.
53
:
593
-600.
14
van Beuningen, H. M., P. M. van der Kraan, O. J. Arntz, W. B. van den Berg.
1993
. Protection from interleukin 1 induced destruction of articular cartilage by transforming growth factor β: studies in anatomically intact cartilage in vitro and in vivo.
Ann. Rheum. Dis.
52
:
185
-191.
15
Goumans, M. J., C. Mummery.
2000
. Functional analysis of the TGFβ receptor/Smad pathway through gene ablation in mice.
Int. J. Dev. Biol.
44
:
253
-265.
16
Ito, H., H. Akiyama, C. Shigeno, T. Nakamura.
1999
. Noggin and bone morphogenetic protein-4 coordinately regulate the progression of chondrogenic differentiation in mouse clonal EC cells, ATDC5.
Biochem. Biophys. Res. Commun.
260
:
240
-244.
17
Wu, X., J. Ma, J. D. Han, N. Wang, Y. G. Chen.
2006
. Distinct regulation of gene expression in human endothelial cells by TGF-β and its receptors.
Microvasc. Res.
71
:
12
-19.
18
Seki, T., K. H. Hong, S. P. Oh.
2006
. Nonoverlapping expression patterns of ALK1 and ALK5 reveal distinct roles of each receptor in vascular development.
Lab. Invest.
86
:
116
-129.
19
Goumans, M. J., G. Valdimarsdottir, S. Itoh, A. Rosendahl, P. Sideras, P. ten Dijke.
2002
. Balancing the activation state of the endothelium via two distinct TGF-β type I receptors.
EMBO J.
21
:
1743
-1753.
20
Blaney Davidson, E. N., A. Scharstuhl, E. L. Vitters, P. M. van der Kraan, W. B. van den Berg.
2005
. Reduced transforming growth factor-β signaling in cartilage of old mice: role in impaired repair capacity.
Arthritis Res. Ther.
7
:
R1338
-R1347.
21
Blaney Davidson, E. N., E. L. Vitters, P. M. van der Kraan, W. B. van den Berg.
2006
. Expression of TGF-β and the TGF-β signaling molecule SMAD-2P in spontaneous and instability-induced osteoarthritis: role in cartilage degradation, chondrogenesis and osteophyte formation.
Ann. Rheum. Dis.
65
:
1414
-1421.
22
van Beuningen, H. M., R. Stoop, P. Buma, N. Takahashi, P. M. van der Kraan, W. B. van den Berg.
2002
. Phenotypic differences in murine chondrocyte cell lines derived from mature articular cartilage.
Osteoarthritis Cartilage
10
:
977
-986.
23
Glansbeek, H. L., P. M. van der Kraan, E. L. Vitters, W. B. van den Berg.
1993
. Correlation of the size of type II transforming growth factor beta (TGF-β) receptor with TGF-β responses of isolated bovine articular chondrocytes.
Ann. Rheum. Dis.
52
:
812
-816.
24
Blaney Davidson, E. N., E. L. Vitters, F. M. Mooren, N. Oliver, W. B. Berg, P. M. van der Kraan.
2006
. Connective tissue growth factor/CCN2 overexpression in mouse synovial lining results in transient fibrosis and cartilage damage.
Arthritis Rheum.
54
:
1653
-1661.
25
Glasson, S. S., T. J. Blanchet, E. A. Morris.
2007
. The surgical destabilization of the medial meniscus (DMM) model of osteoarthritis in the 129/SvEv mouse.
Osteoarthritis Cartilage
15
:
1061
-1069.
26
Das Gupta, E. P., T. J. Lyons, J. A. Hoyland, D. M. Lawton, A. J. Freemont.
1993
. New histological observations in spontaneously developing osteoarthritis in the STR/ORT mouse questioning its acceptability as a model of human osteoarthritis.
Int. J. Exp. Pathol.
74
:
627
-634.
27
Mason, R. M., M. G. Chambers, J. Flannelly, J. D. Gaffen, J. Dudhia, M. T. Bayliss.
2001
. The STR/ort mouse and its use as a model of osteoarthritis.
Osteoarthritis Cartilage
9
:
85
-91.
28
Pritzker, K. P., S. Gay, S. A. Jimenez, K. Ostergaard, J. P. Pelletier, P. A. Revell, D. Salter, W. B. van den Berg.
2006
. Osteoarthritis cartilage histopathology: grading and staging.
Osteoarthritis Cartilage
14
:
13
-29.
29
Lebrin, F., M. J. Goumans, L. Jonker, R. L. Carvalho, G. Valdimarsdottir, M. Thorikay, C. Mummery, H. M. Arthur, P. ten Dijke.
2004
. Endoglin promotes endothelial cell proliferation and TGF-β/ALK1 signal transduction.
EMBO J.
23
:
4018
-4028.
30
Blaney Davidson, E. N., E. L. Vitters, W. B. van den Berg, P. M. van der Kraan.
2006
. TGF β-induced cartilage repair is maintained but fibrosis is blocked in the presence of Smad7.
Arthritis Res. Ther.
8
:
R65
31
van Beuningen, H. M., P. M. van der Kraan, O. J. Arntz, W. B. van den Berg.
1994
. Transforming growth factor-β1 stimulates articular chondrocyte proteoglycan synthesis and induces osteophyte formation in the murine knee joint.
Lab. Invest.
71
:
279
-290.
32
van Beuningen, H. M., H. L. Glansbeek, P. M. van der Kraan, W. B. van den Berg.
1998
. Differential effects of local application of BMP-2 or TGF-β1 on both articular cartilage composition and osteophyte formation.
Osteoarthritis Cartilage
6
:
306
-317.
33
Frenkel, S. R., P. B. Saadeh, B. J. Mehrara, G. S. Chin, D. S. Steinbrech, B. Brent, G. K. Gittes, M. T. Longaker.
2000
. Transforming growth factor β superfamily members: role in cartilage modeling.
Plast. Reconstr. Surg.
105
:
980
-990.
34
Grimoud, E., D. Heymann, F. Redini.
2002
. Recent advances in TGF-β effects on chondrocyte metabolism potential potential therapeutic roles of TGF-β in cartilage disorders.
Cytokine Growth Factor Rev.
13
:
241
-257.
35
Scharstuhl, A., H. L. Glansbeek, H. M. van Beuningen, E. L. Vitters, P. M. van der Kraan, W. B. van den Berg.
2002
. Inhibition of endogenous TGF-β during experimental osteoarthritis prevents osteophyte formation and impairs cartilage repair.
J. Immunol.
169
:
507
-514.
36
Oh, S. P., T. Seki, K. A. Goss, T. Imamura, Y. Yi, P. K. Donahoe, L. Li, K. Miyazono, P. ten Dijke, S. Kim, E. Li.
2000
. Activin receptor-like kinase 1 modulates transforming growth factor-β1 signaling in the regulation of angiogenesis.
Proc. Natl. Acad. Sci. USA
97
:
2626
-2631.
37
Yang, K. G., D. B. Saris, R. E. Geuze, M. H. van Rijen, Y. J. van der Helm, A. J. Verbout, L. B. Creemers, W. J. Dhert.
2006
. Altered in vitro chondrogenic properties of chondrocytes harvested from unaffected cartilage in osteoarthritic joints.
Osteoarthritis Cartilage
14
:
561
-570.
38
Scharstuhl, A., E. L. Vitters, P. M. van der Kraan, W. B. van den Berg.
2003
. Reduction of osteophyte formation and synovial thickening by adenoviral overexpression of transforming growth factor β/bone morphogenetic protein inhibitors during experimental osteoarthritis.
Arthritis Rheum.
48
:
3442
-3451.
39
Pannu, J., S. Nakerakanti, E. Smith, P. ten Dijke, M. Trojanowska.
2007
. Transforming growth factor-β receptor type I-dependent fibrogenic gene program is mediated via activation of Smad1 and ERK1/2 pathways.
J. Biol. Chem.
282
:
10405
-10413.