If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Unloading results in rapid loss of TGFβ signaling in articular cartilage: role of loading-induced TGFβ signaling in maintenance of articular chondrocyte phenotype?
Address correspondence and reprint requests to: P.M. van der Kraan, Experimental Rheumatology, Radboudumc, PO Box 9101, 6500 HB Nijmegen, The Netherlands. Tel: 31-24-3616568; Fax: 31-24-354-0403.
Recently it was shown that loading of articular cartilage explants activates TGFβ signaling. Here we investigated if in vivo chondrocytes express permanently high TGFβ signaling, and the consequence of the loss of compressive loading-mediated TGFβ signaling on chondrocyte function and phenotype.
Method
Bovine articular cartilage explants were collected within 10 min post mortem and stained immediately and after 30, 60 and 360 min for phosphorylated-Smad2, indicating active TGFβ signaling. Explants were unloaded for 48 h and subsequently repeatedly loaded with a compressive load of 3 MPa. In addition, explants were cultured unloaded for 2 weeks and the effect of loading or exogenous TGFβ on proteoglycan level and chondrocyte phenotype (Col10a1 mRNA expression) was analyzed.
Results
Unloading of articular cartilage results in rapid loss of TGFβ signaling while subsequent compressive loading swiftly restored this. Loading and exogenous TGFβ enhanced expression of TGFβ1 and ALK5. Unloading of explants for 2 weeks resulted in proteoglycan loss and increased Col10a1 expression. Both loading and exogenous TGFβ inhibited elevated Col10a1 expression but not proteoglycan loss.
Conclusion
Our data might imply that in vivo regular physiological loading of articular cartilage leads to enduring TGFβ signaling and TGFβ-induced gene expression. We propose a hypothetical model in which loading activates a self-perpetuating system that prevents hypertrophic differentiation of chondrocytes and is crucial for cartilage homeostasis.
Loading of articular cartilage is absolutely essential for its maintenance. Reduced joint loading leads to cartilage atrophy and degeneration, both in humans and animal models
. TGFβ signaling has been shown to be essential for the preservation of articular cartilage. Loss of the intracellular TGFβ signaling molecule Smad3 or functional loss of the TGFβ type II receptor in chondrocytes results in loss of articular cartilage in mice
Expression of a truncated, kinase-defective TGF-beta type II receptor in mouse skeletal tissue promotes terminal chondrocyte differentiation and osteoarthritis.
. In humans, an inactivating mutation of Smad3 results in early onset of OA while a Smad3 polymorphism is associated with the total burden of radiographic OA
. Based on these observations we hypothesized that under normal in vivo conditions articular cartilage is always subject to active TGFβ signaling and that loss of in vivo loading will result in loss of signaling. Remarkably, this has never been investigated yet. In this study it was investigated if TGFβ signaling was lost when articular cartilage was taken out of its natural environment and if the changes in TGFβ signaling could be reversed by in vitro compressive loading. Moreover, we studied the effect of prolonged unloading in vitro on proteoglycan content and chondrocyte phenotype, as measured by expression of Col10a1, and if induced changes could be prevented by loading or exogenously added TGFβ. Our data indicate that unloading results in rapid loss of TGFβ signaling and this might lead to changes in chondrocyte phenotype. Based on our observations a self-regulatory loading-driven model is proposed that keeps articular cartilage healthy, connecting compressive loading to cartilage homeostasis via TGFβ.
Material and methods
For all performed experiments full cartilage thickness explants were harvested from bovine metacarpophalangeal joints (MCP) of skeletally mature cows (age- 4–5 years old) obtained from the local abattoir. 0.7 ± 0.12 mm thick explants were isolated with a 4 mm Ø biopsy punch (Kai-medical, Japan). All explants (if cultured) were cultured in standard culture conditions (37°C, 5% CO2 and 95% humidity) in DMEM/F-12 medium (Gibco®, UK) containing Antibiotic-Antimycotic (contains 10,000 units/mL of penicillin, 10,000 μg/mL of streptomycin, and 25 μg/mL of Fungizone®) (Gibco®, USA) unless stated differently. No serum was added to the medium unless stated differently.
Effect of unloading on TGFβ signaling in articular cartilage [Fig. 1(A)]
Bovine articular cartilage explants were harvested from the MCP joint of skeletally mature cows (age 3–5 years old). Joints were processed within 10 min post mortem (joint loading stopped).
Fig. 1Schematic representation of experiments conducted in this study. (A) The effect of unloading on TGF β signaling in articular cartilage was investigated by measuring Smad2P and Smad2/3P responsive genes on indicated time points. (B) Effect of repeated physiological mechanical compression on TGFβ signaling in articular cartilage was investigated by measuring Smad2P and Smad2/3P responsive genes on indicated time points before and 2 h and 48 h after compression for multiple compressions. (C) Effect of loading on GAG content and Col10a1 expression was measured by culturing cartilage explants for 2 weeks ex vivo. Explants were compressed every 48 h the first week and hereafter either cultured for 1 week in serum free medium for Col10a1 or compressed again for GAG measurement.
For immunohistochemical (IHC) analysis, explants (4 mm Ø) were fixed in 4% phosphate buffered formalin (pH 7.0) directly after isolation or first cultured for 30 min, 2 h, 6 h or 24 h at standard culture conditions in DMEM/F-12 medium (Gibco®, UK). No serum was added to the medium.
For gene expression analysis, explants were isolated within 3 h post mortem. One group of explants was flash frozen in liquid nitrogen immediately after opening. The remaining groups were placed in medium with or without the ALK4/5/7 kinase blocker SB-505124 (Sigma–Aldrich, St. Louis, MO, USA)
(5 μM) or vehicle control (0.5 μl/ml Dimethyl sulfoxide (DMSO)) for 24 or 48 h. This experiment was repeated in seven animals.
Effect of repeated physiological mechanical compression on TGFβ signaling in articular cartilage [Fig. 1(B)]
Five groups were used in this experiment [see Fig. 1(B)]. Explants were harvested within 3 h post mortem. After 48 h of equilibration, the first group of explants was frozen, whereas the other groups of explants were subjected to 3 MPa dynamic mechanical compression for 30 min with a frequency of 1 Hz
[Fig. 1(B)]. At 2 h after the first compression, a second group of explants was frozen. The remaining groups of explants were again cultured for 48 h after which the third group of explants was frozen. At the same day the last two groups of explants were subjected to mechanical compression. Two hours after the second compression, the fourth group of explants was frozen and the fifth group was further cultured for 48 h and then frozen.
The same experimental set up was repeated in the presence of SB-505124 (5 μM) or DMSO. The specimens were pre-incubated with SB-505124 (or DMSO) for 1 h prior to the compression to ensure penetration of the agent
. SB-505124 or DMSO was also present in the medium during and after dynamic mechanical compression. These experiments were repeated four times.
To immunohistochemically investigate the induction of pSmad2 by mechanical compression after 48 h of equilibration, explants were stimulated with 3 MPa for 30 min with 1 Hz. Then explants were fixed in 4% phosphate buffered formalin (pH 7.0) at 1 h after the compression. For the staining details see section: IHC Analysis.
Effect of loading on glycosaminoglycan (GAG) content and Col10a1 expression [Fig. 1(C)]
The first group of explants was isolated and flash frozen immediately after joint opening. After an equilibrium period of 48 h the medium of four other groups was changed for DMEM/F-12 containing 10 % Fetal Bovine Serum, 20 ng/ml of rhIGF-1 (PeproTech, NJ, USA) or 10 ng/ml rhTGFβ1 (Biolegend, CA, USA) or combination of 20 ng/ml of rhIGF-1 + 10 ng/ml rhTGFβ1 and refreshed every 72 h. An additional group of explants was subjected to mechanical compression every 48 h for 14 days. At day 14, explants from all groups were flash frozen and GAG content was measured using Dimethylmethylene Blue (DMB).
To analyze if a lack of mechanical load on articular cartilage explants results in induction of Col10a1 a first group of explants was isolated and immediately frozen. Four other groups were cultured in unloaded condition for 14 days in DMEM/F-12 medium supplemented with 10 % FBS or 1 ng/ml rhTGFβ1 or 10 ng/ml rhTGFβ1 or 50 ng/ml of Activin A (R&D Systems, MN, USA). To investigate if mechanical compression is able to inhibit non-loading induced induction of Col10a1 expression an extra group of explants was subjected to mechanical compression three times during the first week of the experiment (every 72 h). During the second week of the experiment, only medium was changed every 72 h. This experiment was conducted 6 times.
Dynamic mechanical compression of articular cartilage explants
To compress cartilage, a BOSE® ElectroForce® BioDynamicTM bioreactor (5160 BioDynamic System) equipped with a 50 lbf load-cell was used (BOSE Bose Corp. ElectroForce Systems Group, MN, USA). First, a preset compression force of 5 N (0.3 MPa) was applied to guarantee contact between plates and specimen. Subsequently, explants were subjected to 3 MPa, force controlled, unconfined, dynamic mechanical compression using a 1 Hz sine wave and desired pressure for 30 min (1800 cycles). Unloaded controls were also placed in the bioreactor incubator but in a separate well.
Gene expression analysis
Samples were homogenized using a micro dismembrator (B. Braun Biotech International, Melsungen, Germany). Total RNA was isolated using RNeasy Fibrous tissue kits (Qiagen Inc., Valenzia, CA, USA) according to manufacturers protocol. Isolated RNA was transcribed into cDNA using M-MLV reverse transcriptase and single step RT-PCR: 5 min at 25°C, 60 min at 39°C, and 5 min at 95°C. Gene expression was measured using 0.5 μM of validated primers (see Table I) (Biolegio, the Netherlands) in a quantitative real time polymerase chain reaction (qPCR) using SYBR green (Applied Biosystems, Darmstadt, Germany). A melting curve was made to verify gene specific amplification. Two reference genes were used: glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and ribosomal protein S14 (RPS14).
Table ITemplate, efficiency and sequence of the primers used in this study
Samples were fixed overnight in phosphate buffered formalin, dehydrated and embedded in paraffin. Six μm thick sections were cut and mounted on Superfrost™ Plus Microscope Slides (Thermo Scientific, Waltham, USA). After deparaffinization, citrate buffer (0.1 M sodium citrate and 0.1 M citric acid) was used for 2 h at RT for antigen unmasking. Hydrogen peroxide 1% v/v in methanol was used for 30 min to block endogenous peroxidase. Afterwards, sections were incubated overnight at 4°C with specific primary antibodies against c-terminally phosphorylated SMAD2P (rabbit pAb anti Phospho-Smad2 (Ser465/467) (1:100) (Cell Signalling Technology, Danvers, Massachusetts, USA). Biotin-labelled secondary antibodies were used (Dako, Glostrup, Denmark). Together with a biotin–streptavidin detection system used according to the manufacturers' protocol (Vector Laboratories, Baiklin Game, California, USA). Staining was visualized using dimethylaminoazobenzene (DAB) reagent
Cartilage explants were weighed and digested overnight at 60°C using papain (1 mg/ml papain, 0.1 M sodium acetate, 10 mM l-cysteine hydrochloride and 50 mM ethylenediaminetetraacetic acid sodium salt, pH 6.0). After digestion, samples were centrifuged for 15 min at 15,000 RPM and supernatant was diluted 20 times in ultra pure water. 200 μl of the DMB solution was added to 40 μl of diluted digest, and absorbance at λ = 595 nm was measured immediately using a 96-well plate reader (Biorad, CA, USA).
Statistical analysis
All quantitative data analysis were expressed as a Tukey box blot with mean showed as “+” and outliers showed as “•”. All datasets were checked for normality using the Shapiro–Wilk's test and then for equality of variances by Levene's test.
Linear mixed models with Bonferroni multiple comparison post tests were used to estimate the effect of time and treatment (+DMSO or + SB-505124) on gene expression levels. One way ANOVA with Fisher's LSD post-test was used to estimate the effect of compression or the effect of lack of the compression on gene expression (LSD does not correct for multiple comparisons, however we compared only the effect of the compression on induction of the gene expression or the effect of lack of the compression on gene drop, no multiple comparisons were required). The same approach was used to estimate the effect of compression or lack of the compression and treatment (+DMSO or +SB-505124) on Smad7 expression levels. One way ANOVA with Tukey's post-test was used to estimate the effect of treatments on GAG content. The same approach was used to estimate the effect of treatments on bCol10a1 gene expression. Unpaired one tailed t-test was used to estimate the effect of time on Pai1 expression levels. One way ANOVA with Tukey's post-test was used to estimate the effect of addition of TGFβ1 or Activin A on Smad7, Pai1 and Tgfb1 expression levels. The same approach was used to estimate the effect of dynamic mechanical compression on the expression levels of Alk1.
All the analyses were performed with the statistical software packages: SPSS 20.0 (SPSS, Chicago, USA).
Results
Unloading results in loss of TGFβ signaling
To investigate if unloading results in loss of TGFβ signaling, articular cartilage was obtained within 10 min post mortem from the MCP joint of mature cows and fixed immediately or after in vitro incubation. At the earliest time points, the majority of chondrocytes clearly stained positive for active TGFβ signaling (phosphorylated-Smad2, Smad2P) throughout all zones of the articular cartilage [Fig. 2(A)]. However, already after 2 h of unloaded culture, cells in middle zone of the cartilage had lost staining which was even more pronounced after 6 h. At 6 and 24 h only very few cells stained positive for Smad2P. To ensure that this loss in Smad2P staining was not a cutting artifact we left intact MCP joints unopened for 6 h or overnight and thereafter isolated the cartilage. After 6 h, cartilage had highly reduced Smad2P staining. Moreover, cartilage stored overnight showed significantly reduced Pai1 (a marker for active TGFβ signaling) expression when compared to fresh tissue (Supplementary Fig. 2).
Fig. 2Loss of active Smad2/3 signaling in unloaded cartilage. (A) Phosphorylated Smad2 levels in adult cartilage after unloading, as detected by IHC. Representative individual shown. (B) Relative gene expression of the TGFβ-responsive genes: Alk5, Smad7 and Pai1 24 h and 48 h after unloading compared to fresh samples. Tukey box plot, + = mean, N = 7, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001. (C) The effect of the ALK4/5/7 blocker SB-505124 on relative gene expression of bAlk5, bSmad7 and bPai1 24 h and 48 after unloading, compared to fresh samples. DMSO was used as vehicle control. Tukey box plot, + = mean, N = 7, N.S. = not significant, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001. (D) Relative gene expression of Aggrecan or Smad3 24 h and 48 h after unloading. Tukey box plot, + = mean, N = 7.
To determine whether this rapid drop in Smad2/3 signaling was reflected in TGFβ signaling-dependent gene expression, expression of the known ALK5/Smad3 responsive genes Smad7
was assessed. Expression of all three genes was significantly reduced after 24 h and further lowered after 48 h [Fig. 2(B)]. Addition of the ALK4/5/7 blocker SB-505124 which blocks TGFβ signaling did not result in a more severe loss of gene expression, indicating that no residual TGFβ signaling was present in the unloaded, cultured cartilage and that absence of loading has similar effects as actively blocking Smad2/3 signaling [Fig. 2(C)]. Notably, a decrease in gene expression is not a general phenomenon in unloaded cartilage, as the expression levels of for example Acan (aggrecan) and Smad3 were maintained [Fig. 2(D)].
Reloading repeatedly induces TGFβ signaling
Next, we evaluated if compressive reloading could restore Smad2P signaling and TGFβ-dependent gene expression and if this was a repeatable process. Therefore explants were loaded 48 h after isolation and again 48 h after the first in vitro loading. Compressive loading rapidly induced Smad2P staining in cartilage explants [Fig. 3(A)].
Fig. 3Repeatable restoration of active Smad2/3 signaling by dynamic mechanical compression. (A) Induction of Smad2 phosphorylation by dynamic mechanical compression (3 MPa) of adult articular cartilage, as detected by IHC. (B) Relative gene expression of the TGFβ-responsive genes: Alk5, Smad7 and Pai in response to repeated dynamic mechanical compression and unloading. Tukey box plot, + = mean, N = 4, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001. n.d. = not detected. (C) Relative gene expression of Smad7 in response to repeated dynamic mechanical compression and unloading in the presence of the ALK4/5/7 inhibitor; SB-505124. DMSO was used as vehicle control. Tukey box plot, N = 4, + = mean, N.S. = not significant, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001. (D) Scheme reflecting the effects of dynamic compression on TGFβ-signaling in the absence or presence of SB-505124. (E) Relative gene expression of Alk1 in (repeatedly) dynamically compressed (3 MPa) cartilage compared to unloaded controls 48 h after compression. Tukey box plot, + = mean, N = 8, ***P ≤ 0.001.
Two hours after the first 30 min of loading, gene expression of Alk5, Smad7 and Pai1 was significantly induced, indicating that loading restores TGFβ signaling [Fig. 3(B)]. Strikingly, 48 h after the first in vitro loading gene expression had dropped again to unloaded levels. Two hours after a second loading for 30 min, again gene expression of Smad7 and Pai1 was strongly elevated. We could confirm our earlier observations
that the loading-induced expression of Smad7 can be fully blocked by the ALK4/5/7 inhibitor SB-505124 [Fig. 3(C)], indicating that compression-induced Smad7 expression indeed runs via active Smad2/3P.
Notably, SB-505124 did not inhibit all compression-induced gene expression. For example, a loading-induced ∼4-fold increase in Bmp2 expression (Supplementary Fig. 3) was unaffected. Because Bmp2 expression was also not responsive to exogenously added TGFβ, regulation of this gene is most likely TGFβ-independent. This observation thus shows that SB-505124 does not affect compression-induced gene expression that is induced independently of TGFβ. In Fig. 2(D) a schematic representation of the effects of loading and unloading is depicted showing the repeated mechano-responsiveness of TGFβ signaling in cartilage. Finally, we were also able to confirm our earlier observations that loading reduces ALK1 expression [Fig. 2(E)]
As a possible source for the observed induction of Smad2/3P we investigated two ligands capable of inducing these Smads: TGFβ and Activin-A. Incubation of bovine explants with exogenously added TGFβ resulted in comparable up-regulation of gene expression as induced by compressive loading, in contrast to Activin-A, which did not induce expression of either Alk5, Smad7 or Pai1 although it was bioactive (Supplementary Fig. 4).
Subsequently, the potential physiological relevance of the loading-induced TGFβ signaling was investigated. We postulated that TGFβ signaling either sustains the proteoglycan (aggrecan) content of cartilage or blocks hypertrophic differentiation of chondrocytes or both. Culturing bovine explants for 2 weeks resulted in a significant loss of glycosaminoglycans (GAGs) (nearly 60%) from the extracellular matrix (ECM). Compressive loading was totally ineffective to prevent this loss [Fig. 4(A)]. In contrast, this GAG loss could be prevented by addition of 10% fetal calf serum or 20 ng/ml Insulin-like Growth Factor-1 (IGF1) to the medium. However, addition of 10 ng/ml TGFβ was completely ineffective and even lowered IGF1 effects on GAG content [Fig. 4(B)]. We conclude that it is unlikely that TGFβ plays a direct role in maintenance of GAG content in articular cartilage.
Fig. 4Dynamic mechanical compression protects against chondrocyte hypertrophy. (A) The sulfated GAG content (w/w) of adult articular cartilage, cultured for 2 weeks ex vivo either with or without 10% FCS, 10 ng/ml TGFβ1 or subjected to dynamic mechanical compression (3 MPa). Tukey box plot, + = mean, N = 8, ***P ≤ 0.001. (B) GAG content (w/w) of articular cartilage cultured for 2 weeks ex vivo in serum free medium, or in the presence of either: 10% FCS, 20 ng/ml IGF1, 10 ng/ml TGFβ1, or a combination of IGF1 and TGFβ1. Tukey box plot, N = 9, *P ≤ 0.05, ***P ≤ 0.001. (C) Expression of the hypertrophy marker Col10a1 after 2 weeks ex vivo culture in serum free medium or in the presence of 10% FCS, 1 or 10 ng/ml TGFβ1, or 50 ng/ml Activin A. Tukey box plot, + = mean, N = 4 (for Activin A, N = 2), ***P ≤ 0.001. (D) The effect of repeated dynamic mechanical compression (3 MPa) on Col10a1 expression after 2 weeks ex vivo culture. Tukey box plot, + = mean, N = 6. **P ≤ 0.01, ***P ≤ 0.001.
Apart from GAG loss, culturing of bovine explants in the absence of loading resulted in strongly increased expression of Col10a1, an accepted marker for early hypertrophic differentiation of chondrocytes
. The increase in Col10a1 expression was not affected by addition of 10% fetal calf serum, underling that serum factors are not able to inhibit hypertrophic differentiation of articular chondrocytes [Fig. 4(C)]. In contrast, addition of 1 ng/ml TGFβ fully blocked induction of Col10a1 gene expression. Of note, addition of 10 ng/ml activin did not affect the increase in Col10a1 expression. When we investigated whether compressive loading inhibited the up-regulation of Col10a1 expression, loading for 30 min at time point 48, 96 and 144 h after isolation significantly prevented the up-regulation of Col10a1 in bovine explants measured after 2 weeks [Fig. 4(D)]. Unfortunately, in this experimental setting, a 14 day culture period, we were unable to include the inhibitor SB-505124 because addition of this compound for such a long period resulted in significantly decreased cell viability in the cartilage explants (Supplementary Fig. 5), making us unable to show the importance of Smad2/3P in this process.
Discussion
This study is the first to demonstrate that removal of articular cartilage from in its in vivo situation results in rapid loss of TGFβ signaling and that subsequent compressive loading can repeatedly restore this signaling. This suggests that the absence of loading will result in the loss of TGFβ signaling in articular cartilage. The consequence of this is reduced expression of TGFβ1 and the TGFβ type 1 receptor ALK5, together with increased expression of ALK1
. Moreover, prolonged unloading leads to proteoglycan loss and change in chondrocyte phenotype, as determined by increased Col10a1 expression. However, compressive loading or addition of exogenous TGFβ prevent the increase in Col10a1 expression but not of proteoglycan loss.
Our data demonstrate the repeated mechano-responsiveness of TGFβ signaling in cartilage, which shows rapid activation upon compression and inactivation upon unloading, we propose the following hypothetical model for this loss and activation of TGFβ signaling by compressive mechanical loading: Articular cartilage contains high amounts of TGFβ (up to 300 ng/g
), but inactive and bound to the latency-associated peptide (LAP) and ECM. LAP forms a so-called straitjacket that keeps the mature form of TGFβ1 associated with LAP, but unfolding of LAP by mechanical force (40 pN) can release active TGFβ
, inactivating it again. Although no tools are available yet to investigate this mechanism in situ in intact articular cartilage, we propose that such a mechanism explains the repeated mechanosensitivity of TGFβ in cartilage.
We hypothesize that this loading-released TGFβ will bind to its receptors, but will also rapidly bind the ECM becoming unavailable again
, but in an inactive form that will be bound to the ECM. Moreover, expression of ALK5 will be up-regulated whereas expression of ALK1 will be down-regulated, favoring TGFβ-dependent Smad2/3 signaling and decreasing Smad1/5/8 signaling (Fig. 5). Chondrocyte terminal differentiation is stimulated by Smad1/5/8 activation and inhibited the Smad2/3
TGF-beta signaling in chondrocyte terminal differentiation and osteoarthritis: modulation and integration of signaling pathways through receptor-Smads.
. This is in line with our current observation that regular loading appears to prevent early hypertrophic differentiation of chondrocytes.
Fig. 5Proposed hypothetical model for compression-mediated protection of articular cartilage integrity. Compression of articular cartilage leads to release of active TGFβ from the ECM. This active TGFβ signals via ALK5, resulting in phosphorylation of Smad2/3. Subsequently, phosphorylated Smad2/3 inhibits hypertrophic differentiation of articular chondrocytes, as characterized by Col10a1. Additionally, phosphorylated Smad2/3 activates a positive feedback loop by not only inducing expression of ALK5 and lowering expression of ALK1, but also by inducing the expression of inactive TGFβ1, which after production will bind to the ECM, returning the system to its original state.
Other factors than TGFβ will undoubtedly play a role in the mechanical regulation of articular cartilage and several mechanosensitive actors have been identified in chondrocytes
. Our current study shows that loading and exogenous TGFβ can block hypertrophic differentiation of chondrocytes, as measured by Col10a1 mRNA expression, but is not able to block proteoglycan loss in vitro. In vivo, proteoglycan synthesis will be maintained by systemic levels of IGF-I and BMP9 and by (load-induced) factors such as BMP2
. The observation that the biological consequence of the loading/TGFβ driven process is not the direct maintenance of proteoglycan content might appear to be in contrast with the study of Morales et al.. However, in that study cartilage of 6 months old calves was used, where growth still takes place, while cartilage from skeletally mature cows was used in our study
. Both TGFβ and compressive loading inhibited the up-regulation of the early hypertrophy marker Col10a1, suggesting that loading-induced TGFβ signaling blocks hypertrophic differentiation of chondrocytes in articular cartilage. Unfortunately we were not able to use SB-505124 in our long term cultures but in our short term cultures we could show that loading-induced TGFβ signaling can be blocked by this inhibitor, in line with our previous work
. Importantly, our results seem to indicate that loading and cartilage homeostasis are interconnected via TGFβ signaling.
Our study has a number of limitations. We used expression of Col10a1 as a marker for changes in chondrocyte phenotype in the direction of hypertrophy. Because we had to perform our cultures without fetal calf serum, to prevent continuous presence of TGFβ, we were not able to perform our in vitro cartilage cultures endlessly. This made it impossible to demonstrate the induction of late hypertrophic markers, such as MMP13. However, although we only demonstrated elevated mRNA expression in the time span studied, we still think that our results indicate a phenotypic shift towards hypertrophy since this is supported by other studies that show that loss of TGFβ signaling results in chondrocyte hypertrophy
Expression of a truncated, kinase-defective TGF-beta type II receptor in mouse skeletal tissue promotes terminal chondrocyte differentiation and osteoarthritis.
Another limitation is our loading regime. We use simple compressive loading as a simplified model for the mechanical forces acting in vivo on articular cartilage. The loading protocol we used results in a permanent deformation during the 30 min loading cycle of approximately 10%
. The force we used is in a physiological range (3 Mpa) but the loading itself will be quite different from in vivo loading. In addition, there are regions in articular cartilage which are considered non-load bearing in vivo. In our concept it should be expected that these areas deteriorate. However, it can be argued if truly non-load bearing articular cartilage exists in joints. Furthermore, these locations might experience high shear stress, which has also been shown to be able to activate TGFβ
. Moreover, our data indicate that short physiological compressive loading once every day will be sufficient to maintain TGFβ-induced gene expression. Infrequent compressive loading might be experienced by this so-called “non-loaded” cartilage but still be sufficient to maintain homeostasis.
Finally, a considerable limitation is the lack of absolute proof that the observed processes run via TGFβ. SB-505124 gives an indication that an ALK4/5/7 ligand is important, but we were limited in its use due to its toxic long term effects. Unfortunately, no tools are currently available to investigate our proposed hypothesis more deeply in situ. Ideally we would knock out the TGFβ type II receptor TGFBR2
Deletion of the transforming growth factor β receptor type II gene in articular chondrocytes leads to a progressive osteoarthritis-like phenotype in mice.
to show the importance of TGFβ in our proposed model, but we have been unable to target chondrocytes in situ with the currently available tools to manipulate gene expression.
Absence of loading results in cartilage loss. Chondrocyte hypertrophy is associated with cartilage degradation and proteoglycan loss
Increased type II collagen degradation and very early focal cartilage degeneration is associated with upregulation of chondrocyte differentiation related genes in early human articular cartilage lesions.
. Our finding of early hypertrophic differentiation of chondrocytes in the absence of loading-induced TGFβ signaling might provide an explanation for the loss of articular cartilage that is observed after long term cartilage unloading. Increased numbers of hypertrophic chondrocytes and expression of matrix degrading enzymes have been described in articular cartilage of rats after immobilization
, could be involved in the age-dependency of OA. If the hypothetical model we propose is valid, age-related loss of this loading-induced mechanism will make articular cartilage more prone to hypertrophic changes of articular chondrocytes and OA development. Furthermore, this suggests that unraveling the exact molecular mechanism underlying this system could provide tools to interfere with OA development and or progression.
Acknowledgment
Dr Gerjon Hannink is kindly acknowledged for his assistance with the statistical evaluation. Reumafonds (LLP-7) and ZonMW (40-00812-98-090200) are greatly acknowledged for their financial support. None of the authors have any support or other benefits from commercial sources or other conflicts of interest regarding the work reported in the manuscript, or any other competing financial interests.
Appendix A. Supplementary data
The following are the supplementary data related to this article:
Supplementary Figure 1: Counterstaining of bovine cartilage explants. Hematoxylin staining of bovine cartilage explants as used in Fig. 1.
Supplementary Figure 2: Active TGFβ-signaling is lost in articular cartilage of intact MCP joints. (A) Phosphorylated Smad2 (left) and Hematoxylin (right) staining of cartilage explants isolated 6 h after unloading of the MCP joint. The MCP joint was left intact for the whole duration of this experiment. Arrows depict cells without pSmad2 staining. (B) Gene expression of Pai1 relative to the average expression of the reference genes: Gapdh and Rps14, in cartilage explants isolated rapidly after unloading or after an overnight (O/N) period. Tukey box plot, + = mean, N = 8, **P ≤ 0.01.
Supplementary Figure 3. Compression induces Bmp2 expression but this is not affected by SB-505124. Relative gene expression of Bmp2 in dynamically compressed cartilage compared to unloaded controls 2 h after compression in the presence of SB-505124 (gray) or vehicle/DMSO (white). Additionally, explants were stimulated with 5 ng/ml TGFβ for 2 h but this did not significantly induce Bmp2 expression. Tukey box plot, + = mean, N = 5, ***P ≤ 0.001, n.s. = not significant.
Supplementary Figure 4: rhActivin A does not induce Smad7, Pai1 or Tgfb1 expression in bovine cartilage explants. (A) Relative gene expression of Smad7, Pai1 and Tgfb1 in cartilage explants 24 h after stimulation with rhTGFβ1 or rhActivin A compared to unstimulated samples. Tukey box plot, + = mean, N = 4, ***P ≤ 0.001. (B) Relative gene expression of Smad7 and Pai1 5 h after stimulation of primary bovine chondrocytes with various doses of rhActivin A showing that the used rhActivin A is bioactive and compatible with bovine cells. Experimental duplo shown.
Supplementary Figure 5: SB-505124 negatively affects chondrocyte viability in long term explants culture. Relative viability of cartilage explants after 2 weeks ex vivo culture in the presence of DMSO or 5 μM SB-505124 as measured by XTT assay according to manufacturers protocol (Roche Diagnostics GmbH, Germany). Tukey box plot, + = mean, N = 5, *P ≤ 0.05.
Conception and design: W Madej, A van Caam, P Buma, P van der Kraan, collection and assembly of data: W Madej, A van Caam, analysis and interpretation of data: W Madej, A van Caam, E Blaney Davidson, P Buma, P van der Kraan, drafting of the manuscript: W Madej, A van Caam, Pieter Buma, Peter van der Kraan.
Competing interests statement
There are no conflicts of interests for any of the authors.
References
Palmoski M.
Perricone E.
Brandt K.D.
Development and reversal of a proteoglycan aggregation defect in normal canine knee cartilage after immobilization.
Expression of a truncated, kinase-defective TGF-beta type II receptor in mouse skeletal tissue promotes terminal chondrocyte differentiation and osteoarthritis.
TGF-beta signaling in chondrocyte terminal differentiation and osteoarthritis: modulation and integration of signaling pathways through receptor-Smads.
Deletion of the transforming growth factor β receptor type II gene in articular chondrocytes leads to a progressive osteoarthritis-like phenotype in mice.
Increased type II collagen degradation and very early focal cartilage degeneration is associated with upregulation of chondrocyte differentiation related genes in early human articular cartilage lesions.
30105-4&id=gr1.jpg){kind=link}
30105-4&id=gr2.jpg){kind=link}
30105-4&id=gr3.jpg){kind=link}
30105-4&id=gr4.jpg){kind=link}
30105-4&id=gr5.jpg){kind=link}