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
Experimental Rheumatology, Radboud University Medical Center, Nijmegen, The NetherlandsOrthopaedics Research Lab, Radboud University Medical Center, Nijmegen, The Netherlands
Address correspondence and reprint requests to: P.M. van der Kraan, Experimental Rheumatology, Radboud University Medical Center, Geert Grooteplein 28, 6525 GA Nijmegen, The Netherlands. Tel: 31-24-3616568; Fax: 31-24-3540403.
Ageing is the main risk factor for osteoarthritis (OA). We investigated if expression of transforming growth factor β (TGFβ)-family components, a family which is crucial for the maintenance of healthy articular cartilage, is altered during ageing in cartilage. Moreover, we investigated the functional significance of selected age-related changes.
Design
Age-related changes in expression of TGFβ-family members were analysed by quantitative PCR in healthy articular cartilage obtained from 42 cows (age: ¾–10 years). To obtain functional insight of selected changes, cartilage explants were stimulated with TGFβ1 or bone morphogenetic protein (BMP) 9, and TGFβ1 and BMP response genes were measured.
Results
Age-related cartilage thinning and loss of collagen type 2a1 expression (∼256-fold) was observed, validating our data set for studying ageing in cartilage. Expression of the TGFβ-family type I receptors; bAlk2, bAlk3, bAlk4 and bAlk5 dropped significantly with advancing age, whereas bAlk1 expression did not. Of the type II receptors, expression of bBmpr2 decreased significantly. Type III receptor expression was unaffected by ageing. Expression of the ligands bTgfb1 and bGdf5 also decreased with age. In explants, an age-related decrease in TGFβ1-response was observed for the pSmad3-dependent gene bSerpine1 (P = 0.016). In contrast, ageing did not affect BMP9 signalling, an Alk1 ligand, as measured by expression of the pSmad1/5 dependent gene bId1.
Conclusions
Ageing negatively affects both the TGFβ-ALK5 and BMP-BMPR signalling routes, and aged chondrocytes display a lowered pSmad3-dependent response to TGFβ1. Because pSmad3 signalling is essential for cartilage homeostasis, we propose that this change contributes to OA development.
. Eventually, OA leads to pain and disability, greatly affecting quality of life of patients. Many risk factors have been identified for OA including: obesity, female gender, and occupational risks, but its main risk factor is (old) age
. To better understand why ageing is a risk factor for OA, research has centred on age–related changes in cartilage, as this is the main tissue affected by this pathology.
During ageing, many changes accumulate in cartilage. On a macroscopic level, progressive thinning of the cartilage occurs, most prominently at load-bearing areas
. On a more detailed level, in the extracellular matrix (ECM), structure, size and expression of the essential structural proteoglycan aggrecan change during ageing
Crosslinking by advanced glycation end products increases the stiffness of the collagen network in human articular cartilage: a possible mechanism through which age is a risk factor for osteoarthritis.
. Together, these changes adversely affect the mechanical properties of the cartilage matrix, as they lead to a loss of water retaining capacity and matrix stiffening respectively. On a cellular level, up to 40% of chondrocytes are lost during ageing
. Possibly, this loss occurs because ageing chondrocytes produce less of and/or respond less to essential anabolic growth factors like insulin-like growth factor 1 (IGF1), fibroblast growth factor 2 (FGF2), transforming growth factor β (TGFβ) and bone morphogenetic protein 7 (BMP7)
Both TGFβ and BMP7 belong to the TGFβ-family. Canonical signalling by members of this family induces receptor-regulated Smad (R-Smad) activation via phosphorylation, which is crucial for chondrocyte homeostasis
TGF-beta signaling in chondrocyte terminal differentiation and osteoarthritis: modulation and integration of signaling pathways through receptor-Smads.
. Phosphorylated Smad3 (pSmad3) is chondroprotective and inhibits many processes deleterious for cartilage, including matrix metalloprotease 13 (MMP13) production and cellular hypertrophy
. Phosphorylated Smad1/5 (pSmad1/5) is associated with ECM production, e.g., glycosaminoglycans and collagen type II expression, but, in contrast to pSmad3, also with MMP13 production and deleterious chondrocyte hypertrophy
TGF-beta signaling in chondrocyte terminal differentiation and osteoarthritis: modulation and integration of signaling pathways through receptor-Smads.
. However, in these experiments in mice, ageing and OA were concomitant. Because OA chondrocytes are profoundly different from ageing chondrocytes, e.g., in anabolic activity
, and both processes possibly interfere, we wanted to further study cartilage ageing in a system in which we could observe age-related changes independently of an OA process. To achieve this, we investigated ageing in non-OA cartilage of the bovine metacarpophalangeal (MCP) joint. Bovine material can be obtained in a wide age range, and examination of the cartilage allows for inclusion of only healthy joints with an intact cartilage surface.
In this study we investigated the impact of ageing on expression of receptors and ligands of the TGFβ-family. We were predominantly interested in expression of TGFβ and bone morphogenetic protein (BMP) receptors, as especially little is known of the latter in the context of ageing. We show that in old cartilage the Smad2/3-inducing receptors bAcvr1b and bTgrbr1 are profoundly lower expressed than in young cartilage, just as the BMP-receptors bBmpr1 and bBmpr2. Furthermore, we show that reduction of Tgfbr1 expression is reflected in a reduced response to TGFβ1. Overall, we conclude that changes in TGFβ-family signalling occur before onset of OA symptoms but could predispose cartilage to this disease.
Materials and methods
Sample collection
Bovine MCP joints were obtained from a local abattoir within 3 h post mortem. Joints were opened and the cartilage surface was examined for signs of OA. Only joints showing an intact cartilage surface were included in this study. Subsequently, a ∼0.5 cm2 full thickness cartilage slice was obtained from the middle of the medial condyle adjacent to the intertrochlear notch of the metacarpal bone. This cartilage sample was shortly rinsed in saline and flash frozen in liquid nitrogen for later use. In total 42 female cows were included with ages ranging from 9 months old up to 10 years old (see Supplementary Fig. 1).
Detection of gene expression
A ∼0.5 cm2 cartilage slice was homogenized using a Mikro-dismembrator (B.Braun, Germany) for 1 min at 1.5 × 103 RPM. Subsequently, RNA was isolated with a Fibrous tissue kit (Qiagen) according to manufacturers protocol. Next, 8 μl of sample was treated with 1 μl DNAse (Life Technologies, USA) for 10 min at room temperature to remove DNA contamination. Hereafter, DNAse was inactivated at 65°C with 1 μl 25 mM EDTA (Life Technologies, USA) for 10 min. To perform reverse transcriptase (RT) reaction; 1.9 μl ultra pure water, 2.4 μl 10× DNAse buffer, 2.0 μl 0.1 M DTT, 0.8 μl 25 mM dNTP, 0.4 μg oligo dT primer, 1 μl 200 U/μl M-MLV RT (all Life Technologies, USA) and 0.5 μl 40 U/μl RNAsin (Promega, the Netherlands) was added, and samples were incubated for 5 min at 25°C, 60 min at 39°C, and 5 min at 95°C using a thermocycler. The obtained cDNA was diluted 10× in ultra pure water, and gene expression was measured using 0.25 μM of validated cDNA-specific primers (see Table 1) (Biolegio, the Netherlands) in a quantitative real time polymerase chain reaction (qPCR) using SYBR green master mix (Applied Biosystems, USA). The following protocol was used: after 10 min at 95°C, 40 cycles of 15 s 95°C and 1 min 60°C each were run. A melting curve was made to verify gene specific amplification. For calculations of the −ΔCt, the average of the following four reference genes was used: bovine (b) glyceraldehyde 3-phosphate dehydrogenase (bGapdh), ribosomal protein S14 (bRps14), ribosomal protein L22 (bRpl22) and beta glucuronidase (bGusb). These reference genes correlate highly in our samples, showing that ageing does not affect their interrelation (see Supplementary Fig. 2).
Table 1Template, efficiency and sequence of the primers used in this study.
Samples were fixed overnight in phosphate buffered formalin (10%). Four animals were used per age-group. Hereafter, the samples were dehydrated in a tissue processing apparatus (Pathos, Milestone Medical Inc.) and embedded in paraffin. Six μm thick sections were cut and deparaffinised. Subsequently, IHC was performed: first, sections were treated with citrate buffer (0.1 M sodium citrate and 0.1 M citric acid) for 2 h at room temperature for antigen unmasking. Next, endogenous peroxidase was inactivated by incubation with 1% hydrogen peroxidase in methanol for 30 min. Then, sections were blocked with 5% normal rabbit serum. Afterwards, sections were incubated overnight at 4°C with goat polyclonal TGFβ RI Antibody (V-22) 1:200 (v/v) (sc-398-G Santa Cruz Biotechnology, USA) or goat polyclonal ALK-1 Antibody (D-20) 1:200 (v/v) (sc-19546 Santa Cruz Biotechnology, USA)
. Hereafter, sections were washed with PBS (pH 7.0) and incubated for 30 min with a biotin-labelled rabbit anti goat antibody (Dako, Glostrup, Denmark). Next streptavidin-linked horseradish peroxidase was added (Vector Laboratories, Baiklin Game, California, USA), and staining was visualised using dimethylaminoazobenzene.
To quantify IHC, a 0–3 scoring was used (½ values allowed) by a scorer using blinded sections (as recommended by
). A value of three represented strongest staining in contrast to a value of 0, which represented the weakest staining (see Supplementary Fig. 3). Four animals were used per age group, and per animal four sections were scored for each of the three cartilage zones: surface, middle and deep.
Ex vivo stimulation of explants
MCP joints were obtained from 22 animals aged 0.5 up to 11 years old and opened under aseptic conditions. With the use of a biopsy punch, ∼7 mm2 explants were made. Subsequently, these explants were incubated for 24 h in standard cell culture conditions in DMEM/F12 1:1 (Gibco, USA) containing penicillin/streptomycin and pyruvate but no FCS. After 24 h, explants were stimulated with 1 ng/ml rhTGFβ1 or 5 ng/ml rhBMP9 (both R&D systems, USA) for 24 h. Per condition, four pooled explants were used, except for the unstimulated control for which two times four pooled explants were used. After stimulation, mRNA was isolated using the aforementioned protocol.
Statistics
Gene expression of individual animals is plotted with a solid line depicting the best fit regression analysis and 95% confidence interval (CI) as dotted line. For every analysis, data was checked for normality using the Shapiro–Wilk test. A Kruskal–Wallis test with Dunn's multiple comparison post hoc test was used to analyse the scoring of IHC. All statistics were conducted using GraphPad Prism v 5.0.
Results
Bovine cartilage as a valid model for age-related changes in cartilage
To validate our bovine model and chosen age range (¾–10 years old) as a model for ageing cartilage, we first characterized well known markers of ageing cartilage in our data set. Macroscopically, with advancing age, thinning of the cartilage was observed and cartilage colour changed from purple towards yellow. Microscopically, aged cartilage showed decreased cellularity compared to young cartilage (for quantification of cell number and thickness see
), and a clear tidemark could be detected in old cartilage using a hematoxilin staining [Fig. 1(A)]. Closer examination showed that this tidemark actually consisted out of multiple tidemarks [Fig. 1B]. Furthermore, in young cartilage the osteochondral junction was less well defined, as shown by the close interaction of bone marrow and blood vessels with the cartilage, and absence of a tidemark. Moreover, the superficial zone of old cartilage looks irregular when looking at the lamina splendens as visualized by Safranin O/Fast green staining [Fig. 1A]. On gene expression, a very profound ∼250-fold drop in bovine collagen type 2 alpha 1 (bCol2a1) expression was observed (P < 0.0001, R2 = 0.63) when comparing young vs old [Fig. 1C]. In contrast, aggrecan (bAcan) expression did not drop extensively (∼2-fold) (P < 0.01, R2 = 0.16), and expression of matrix gla protein (bMgp) was significantly (∼8 fold) higher in old cartilage vs young cartilage, showing that not of every gene expression decreases with ageing. Because these observations on macroscopic, microscopic and gene expression level reflect well known markers of ageing cartilage
, we concluded that our data set is a valid model for studying ageing processes in articular cartilage.
Fig. 1Age-related changes in bovine cartilage. (A) Safranin O/Fast green and Hematoxilin/Eosin staining of articular cartilage obtained from a 2 years old (left) or 10 years old cow (right). Pictures are representative of three animals each. 100× Magnification, scale bar = 100 μm. Thickness and cellularity are reduced and a clear tidemark is visible (arrow) in 10 year old cartilage. Note that the surface is intact. (B) 800× Magnification, scale bar = 12.5 μm. Multiple tidemarks (stars) in healthy 10 year old cartilage. No tidemark is visible between non-calcified and calcified cartilage in 2 year old cartilage. (C) Relative gene expression of bovine collagen type 2 alpha 1 (bCol2a1), aggrecan (bAcan) and matrix Gla protein (bMgp) in ageing cartilage as determined by qPCR. Regression analysis (solid line) with 95% CI (dotted line) depicted. Expression of bCol2a1 decreases whereas expression of bMgp increases during ageing.
. Therefore, we first measured gene expression of the four Smad1/5 phosphorylating receptors: bAcvrl1 (=ALK1), bAcvr1 (=ALK2), bBmpr1 (=ALK3) and bBmpr1b (=ALK6) [Fig. 2A]. bAcvrl1 expression was unaffected by ageing. bAcvr1 expression did drop ∼3-fold but, in view of the relatively high p value (P = 0.02, R2 = 0.13), not very convincingly. In contrast bBmpr1 expression clearly diminished ∼4-fold during ageing (P = 0.001, R2 = 0.24). bBmpr1b expression was not reliably detected, despite the use of multiple primer sets. Subsequently, we measured the three Smad2/3 phoshorylating receptors: bAcvr1b (=ALK4), bTgfbr1 (=ALK5) and bAcvr1c (=ALK7) [Fig. 2B]. Of these three receptors, expression of bAcvr1b and bTgfbr1 was readily detected. In contrast, we were unable to reliably detect bAcvr1c expression, again despite the use of multiple primer sets. Expression of both bAcvr1b and bTgfbr1 decreased ∼8-fold and ∼4-fold respectively (P = 0.0009, R2 = 0.30 for bAcvr1b and P = 0.0004 R2 = 0.29 for bTgfbr1) with advancing age. Next we validated part of this gene expression data by immunohistochemically staining the corresponding proteins [Fig. 2C] and quantifying this [Fig. 2D]. bAcvrl1 and bTgfbr1 were selected to relate this study to our previous results obtained in mice
. Staining for bTgfbr1 clearly showed abundant positive cells in young animals, but with advancing age Tgfbr1 staining decreased in the superficial, middle and deep zone of the cartilage. This was in contrast with the bAcvrl1 staining, which was only reduced in the deep zone. Taken together, our data show a profound decrease in expression of the Smad2/3 phosphorylating receptors bAcvr1b and bTgfbr1 with advancing age, whereas of the Smad1/5 phosphorylating receptors bBmpr1 expression was clearly diminished but bAcvrl1 expression was not.
Fig. 2Age-related drop in TGFβ-family receptor expression in healthy cartilage. (A) Relative gene expression of the Smad1/5 phosphorylating receptors: activin A receptor type IL (bAcvrl1), activin A receptor type I (bAcvr1) and bone morphogenetic protein receptor type IA (bBmpr1a), in ageing cartilage as determined by qPCR. Regression analysis (solid line) with 95% CI (dotted line) depicted. Expression of bBmpr1a clearly decreases with advancing age. (B) Relative gene expression of the Smad2/3 phosphorylating receptors: transforming growth factor beta receptor I (bTgfbr1) and activin A receptor type IB (bAcvr1b) in ageing cartilage as determined by qPCR. Regression analysis (solid line) with 95% CI (dotted line) depicted. Expression of bTgfbr1 and bAcvr1b clearly decreases with age. (C) IHC of bAcvrl1 (left) and bTgfbr1 (right) in 1 year and 5 years old animals in the three zones of articular cartilage. Pictures are representative of four animals each. 1000× Magnification, scale bar = 10 μm. During ageing, bTgfbr1 expression decreases below the detection threshold and unstained cells start to appear (arrows). (D) Scoring of Acvrl1 and Tgfbr1 staining in the superficial, middle and deep zone of 1 and 5 year old bovine cartilage using a 0–3 scale (see Supplementary Fig. 3). Four animals were scored in quadruple for each group for each zone. Error bars represent 95% CI. Statistics were calculated using a Kruskal–Wallis test; * = P ≤ 0.05, ** = P ≤ 0.01, *** = P ≤ 0.001.
Expression of TGFβ-family type II and type III receptors in ageing cartilage
Next to receptor type I expression, the presence of type II receptors is necessary for TGFβ-family signalling and therefore we measured expression of these receptors
. Furthermore, because several ligands of the TGFβ-family can use multiple type I receptors (e.g., TGFβ1), we measured expression of the type III receptors Endoglin (bEng) and Betaglycan (bTgfrbr3) because these receptors can direct receptor choice of such a ligand [Fig. 3]. With advancing age, gene expression of the type II receptor bBmpr2 clearly decreased ∼6-fold (P < 0.0001, R2 = 0.38) [Fig. 3]. In contrast, bTgfbr2 expression was not significantly decreased with age (P = 0.0517, R2 = 0.10) [Fig. 3]. Expression of both bEndoglin and bTgfbr3 did not change during ageing.
Fig. 3Age-related changes in TGFβ-family type II and type III receptors in healthy cartilage. Relative gene expression of bone morphogenetic protein receptor type II (bBmpr2), transforming growth factor beta receptor II (bTgfbr2), endoglin (bEng), and transforming growth factor beta receptor III (bTgfbr3) in ageing cartilage as determined by qPCR. Regression analysis (solid line) with 95% CI (dotted line) depicted. Only expression of bBmpr2 clearly decreases with advancing age.
Decreased expression of TGFβ1 and GDF5 in ageing cartilage
In addition to the changes we observed on receptor expression, we also looked into the expression of ligands. As TGFβ is crucial for healthy articular cartilage, we investigated expression of the three TGFβ isotypes: TGFβ1, TGFβ2 and TGFβ3. Of these three TGFβ isoforms, only expression of bTgfb1 was significantly ∼3-fold (P = 0.0002, R2 = 0.32) less expressed in old cartilage compared to young cartilage [Fig. 4]. In contrast, expression of both Tgfb2 and Tgfb3 was not lowered with advancing age. Next, we measured bGdf5 expression, because reduced expression of this growth factor has been identified as a risk factor for OA
. bGdf5 expression significantly (P = 0.0090, R2 = 0.10) decreased ∼2-fold with advancing age. Additionally, we measured bBmp2 and bBmp7 expression in view of their importance for cartilage biology. bBmp2 expression was not affected by age, whereas bBmp7 expression was remarkably low (undetected) in most samples, This made it difficult to draw a clear conclusion regarding bBmp7 expression. Taken together our data shows that expression of at least two important growth factors for cartilage homeostasis decreases with advancing age.
Fig. 4Age-related drop in TGFβ-family ligands. Relative gene expression of transforming growth factor beta 1 (bTgfb1), transforming growth factor beta 2 (bTgfb2), transforming growth factor beta 3 (bTgfb3), growth differentiation factor 5 (bGdf5), bone morphogenetic protein 2 (bBmp2) and bone morphogenetic protein 7 (bBmp7) in ageing cartilage as determined by qPCR. Regression analysis (solid line) with 95% CI (dotted line) depicted. Expression of both bTgfb1 and bGdf5 clearly decreases with advancing age.
Functional implications of altered receptor expression in ageing cartilage
In view of the changes we observed in receptor expression, we wanted to investigate the functional implications of these changes. We focused on the TGFβ receptors Acvrl1 (ALK1) and Tgfbr1 (ALK5) because of TGFβ's importance in cartilage biology. First, we analysed our data set for clues of decreased TGFβ signalling. Previously, we have identified expression of bSmad7 as an indicator of TGFβ signalling in bovine explants
. When analysing bSmad7, we observed a clear ∼4-fold reduction in expression with advancing age (P = 0.0089, R2 = 0.17) [Fig. 5A]. To prove that aged cartilage indeed is less responsive to TGFβ, we stimulated cartilage explants of 22 animals (aged 0.5–11 years old) ex vivo with TGFβ1 for 24 h and measured bSmad7 expression. However, TGFβ-induced bSmad7 expression was not significantly different between young and old animals [Fig. 5C]. A possible reason for this is that TGFβ1 can induce bSmad7 expression via both bTgfbr1 and bAcvrl1
and bAcvrl1 expression was unaffected by ageing. Therefore, we subsequently measured bSerpine1 expression, a gene which is induced by TGFβ in a specifically pSmad3-dependent way
, and thus a good indicator of only bTgfbr1 signalling. Although bSerpine1 expression did not change with age [Fig. 5B], old cartilage clearly responded less to TGFβ1 than young cartilage, because age-dependent bSerpine1 expression in response to TGFβ1 is characterized by a negative regression coefficient (P = 0.0157) [Fig. 5D]. This negative coefficient reflects a ∼3-fold lower response to TGFβ1 of 10 year old cartilage compared to 6 months old cartilage. Finally, we wanted to investigate if Acvrl1 functionality did not decrease with age. Therefore we stimulated explants with BMP9, a high affinity Acvrl1 ligand
. BMP9 stimulation induced bId1 expression to a similar extent (∼4-fold) in both young and old animals showing that Acvrl1 (ALK1) function was not diminished [Fig. 5(E)]. Taken together, these results indicate that, in ageing chondrocytes, loss of bTgfbr1 expression indeed has functional consequences for cellular response to TGFβ1.
Fig. 5Age-related reduction in Smad3-dependent TGFβ1 responsiveness in cartilage. (A) Relative gene expression of SMAD family member 7 (bSmad7) in ageing cartilage as determined by qPCR. Regression analysis (solid line) with 95% CI (dotted line) depicted. (B) Relative gene expression of serpin peptidase inhibitor, clade E member 1 (bSerpine1) in ageing cartilage as determined by qPCR. Regression analysis (solid line) with 95% CI (dotted line) depicted. (C) Relative gene expression of bSmad7 in ex vivo explants in response to 24 h TGFβ stimulation (1 ng/ml) corrected for unstimulated explants (ΔΔCt). N = 22 animals. Regression analysis (solid line) with 95% CI (dotted line) depicted. (D) Relative gene expression of bSerpine1 in ex vivo explants in response to 24 h TGFβ1 stimulation (1 ng/ml) corrected for unstimulated explants (ΔΔCt). N = 22 animals. Regression analysis (solid line) with 95% CI (dotted line) depicted. (E) Relative gene expression of bId1 in ex vivo explants in response to 24 h TGFβ (1 ng/ml) or BMP9 (5 ng/ml) stimulation corrected for unstimulated explants (ΔΔCt). N = 8 animals. Regression analysis (solid line) depicted.
Ageing affects cartilage on multiple levels, including chondrocyte homeostasis. In this paper we investigated the impact of ageing on the expression of TGFβ-family signalling components in healthy articular cartilage. With the use of bovine cartilage we were able to separate ageing from OA, two processes that are often concomitant and interfering. We report that ageing decreases Smad2/3 phosphorylating receptors and that this is reflected in a lowered Smad3-dependent TGFβ1 response.
In chondrocyte biology, ageing and (early) OA can have diametrically opposed effects. For example, ageing negatively affects production of essential matrix components like glycosaminoglycans and collagen type 2a1, whereas in early OA expression of both matrix components is markedly upregulated
. Another example is that while cell division is as good as absent in (mature) ageing cartilage, proliferation and clonal expansion of chondrocytes are common during OA
Apoptotic cell death is not a widespread phenomenon in normal aging and osteoarthritis human articular knee cartilage: a study of proliferation, programmed cell death (apoptosis), and viability of chondrocytes in normal and osteoarthritic human knee cartilage.
. These opposite effects of ageing and OA on chondrocyte biology are most likely a reflection of antagonistic regulation of gene expression. Therefore, it is difficult to study the effects of ageing on chondrocyte gene expression in joints with concurrent OA, and ideally both processes should be studied separately. In mice, we found this separation difficult to do, as e.g., in our hands the widely used C57BL/6 strain spontaneously developed OA as young as 8 months old, leaving little room to study ageing separately. However, in the bovine MCP joint we were able to study ageing without concomitant OA.
Ageing affects cartilage on a macroscopic, microscopic and cellular level. We were able to identify many previously defined features of ageing cartilage in our bovine data set, including progressive thinning of the cartilage, loss of superficial zone organization, tidemark duplications and profound loss of Col2a1 expression
. Therefore using bovine cartilage to study ageing is a valuable tool to replace healthy human cartilage, which is often very difficult to obtain reliably.
Phosphorylated Smad3 (pSmad3) is an essential transcription factor for the maintenance of healthy cartilage. Mice lacking Smad3 show rapid and profound degeneration of their articular cartilage, a process in which chondrocyte hypertrophy plays a major role
TGF-beta signaling in chondrocyte terminal differentiation and osteoarthritis: modulation and integration of signaling pathways through receptor-Smads.
. We show that, compared to young chondrocytes, aged chondrocytes express less of the Smad2/3 phosphorylating receptors bAcvr1b and bTgfbr1. Furthermore, this decreased bTgfbr1 expression is coupled to diminished pSmad3-dependent TGFβ1 signalling (bSerpine1 expression). The direct dependency of TGFβ-induced Serpine1 expression on pSmad3 has been well established
, but a dependency of Serpine1 on pSmad2 has not clearly been established. Therefore, and because we did not measure pSmad2 levels directly, we cannot conclude from our experiments if pSmad2 signalling was also diminished. However, we do think this is likely
. Overall, taken the importance of pSmad3 in cartilage biology into account, our observations suggest that ageing chondrocytes are more prone to become hypertrophic, and give a possible reason for age as the main risk factor of OA.
A diminished response to TGFβ1 with advancing age was suggested by our previous studies showing age-related loss of TGFβ type I and type II receptors in vivo in murine cartilage
. Here we actually confirm that a reduced response is indeed the case. Previously, the effects of TGFβ on chondrocytes of different ages have been studied using chondrocytes cultured in monolayer, but this has yielded variable results on a common parameter as proteoglycan production
Age-related changes in the response of human articular cartilage to IL-1alpha and transforming growth factor-beta (TGF-beta): chondrocytes exhibit a diminished sensitivity to TGF-beta.
Age-related changes in the response of human articular cartilage to IL-1alpha and transforming growth factor-beta (TGF-beta): chondrocytes exhibit a diminished sensitivity to TGF-beta.
, highlighting the difficulty of studying cellular response of aged cells in vitro. Because the outcome of TGFβ signalling is very dependent on cellular context, e.g., which phase of the cell cycle cells are in or ECM composition
, we think that using explants with cells in their native environment is a more suitable approach to study TGFβ signalling in aged cells compared to the use of matrix deprived monolayer chondrocytes. Of note, we have previously also shown that compared to young cartilage, aged bovine cartilage has a diminished capacity to induce Smad2/3 signalling in response to mechanical stimulation
. This indicates that the receptor changes we observe also have functional consequences when cells are not directly stimulated with an exogenous ligand.
Although we chose to focus on age-related changes in TGFβ signalling, BMP signalling is also relevant for cartilage homeostasis, for example in regulation of proteoglycan content. Locally produced BMP2 and BMP7 have been shown to be important for maintenance of cartilage GAG content
Elevated extracellular matrix production and degradation upon bone morphogenetic protein-2 (BMP-2) stimulation point toward a role for BMP-2 in cartilage repair and remodeling.
. Both BMP2 and BMP7 use Bmpr2 as type II receptor and share Bmpr1 as type I receptor. Remarkably, we observed profound downregulation of both receptors ((∼4-fold) in bBmpr1 expression, ∼6-fold bBmpr2) with advancing age, and are the first to report so. Possibly this decrease results in lower BMP2 and BMP7 signalling in elder cartilage, resulting in less proteoglycan maintenance. However, although we did observe age-related downregulation of BMP receptors, we only observed a ∼2 fold drop in aggrecan expression. This indicates that the relative importance of BMP signalling in aggrecan production is maybe limited. Additionally, Bmpr1 is also the receptor used by GDF5. Notably we also observed that GDF5 expression itself decreases with ageing. Reduced GDF5 expression is associated with OA, because a SNP lowering expression of this growth factor was the first genetic risk factor to be identified for OA
. In view of this, it is relevant that we did not observe an age-related decrease in Acvrl1 expression or function. Previously we have suggested that a shift in balance between Acvrl1 and Tgfbr1 receptors could be a cause of OA, and that Acvrl1 is involved in chondrocyte Mmp13 expression
. In our samples we also observed this shift, because Acvrl1 expression remained stable over time whereas Tgfbr1 expression diminished. Therefore we can confirm this shift as a robust age-related effect in cartilage that occurs across species. In contrast to previous papers, which studied cartilage ageing in mice, we did not observe a loss of Tgfb2, Tgfb3, or Tgfbr2 expression
. Possibly, this difference is due to a species-specific effect, or interference of an OA-related effect like inflammation. Additionally, it could imply an additional level of regulation of these factors on protein level, as we did not study this.
Our study does have limitations. First of all, we did not use human material and therefore species specific observations cannot be excluded. However, we observed many changes documented for human cartilage in bovine cartilage
, indicating that similar processes happen in both species. Secondly, in all cases we predominantly analysed mRNA expression and did not study the translation of this signal into protein. Nonetheless, with the use of mRNA we were able to detect many age-related changes in TGFβ-family signalling that have been confirmed on protein level in other species, like the change in Acvrl1/Tgfbr1 ratio
. Finally, we macroscopically selected healthy cartilage. This doesn't fully exclude microscopic changes, thus making it possible that very early OA samples entered our sample pool.
In conclusion, ageing bovine cartilage displays many characteristics of ageing human cartilage, making it a valuable tool to supplement rare human material. The separation of ageing and OA allowed us to study ageing independently of OA, which revealed that expression of TGFβ and BMP receptors decreases with advancing age. The decrease in expression of the TGFβ-receptor Tgfbr1 resulted in a lowered phosphorylated Smad3-dependent response of older cartilage to TGFβ1. Due to the importance of phosphorylated Smad3 in cartilage biology, reduction of Smad3 activation during ageing could explain the close relation between ageing and OA development, highlighting ageing as the main risk factor for the development of this disease.
Author contributions
Conception and design: Arjan van Caam, Esmeralda Blaney Davidson, Peter van der Kraan.
Collection and assembly of data: Arjan van Caam, Wojciech Madej, Eva Thijssen.
Analysis and interpretation of data: Arjan van Caam, Wojciech Madej, Esmeralda Blaney Davidson, Amaya Garcia de Vinuesa, Ellen van Geffen, Marie-José Goumans, Peter ten Dijke, Peter van der Kraan.
Drafting of the manuscript: Arjan van Caam, Wojciech Madej, Esmeralda Blaney Davidson, Peter van der Kraan.
Critical revision: Arjan van Caam, Wojciech Madej, Esmeralda Blaney Davidson, Amaya Garcia de Vinuesa, Eva Thijssen, Marie-José Goumans, Peter ten Dijke, Peter van der Kraan.
Final approval of the article: Arjan van Caam, Wojciech Madej, Esmeralda Blaney Davidson, Amaya Garcia de Vinuesa, Eva Thijssen, Marie-José Goumans, Peter ten Dijke, Peter van der Kraan.
Funding
This study was supported by a grant from: The Netherlands Organisation for Scientific Research (NWO, ZonMW) for AvC, AGdV, MJG, PtD, and the Dutch Arthritis Association (Reumafonds) for EBD, ET and PvdK.
Conflict of interest
The authors have no conflict of interest.
Appendix A. Supplementary data
The following is the supplementary data related to this article:
Crosslinking by advanced glycation end products increases the stiffness of the collagen network in human articular cartilage: a possible mechanism through which age is a risk factor for osteoarthritis.
TGF-beta signaling in chondrocyte terminal differentiation and osteoarthritis: modulation and integration of signaling pathways through receptor-Smads.
Apoptotic cell death is not a widespread phenomenon in normal aging and osteoarthritis human articular knee cartilage: a study of proliferation, programmed cell death (apoptosis), and viability of chondrocytes in normal and osteoarthritic human knee cartilage.
Age-related changes in the response of human articular cartilage to IL-1alpha and transforming growth factor-beta (TGF-beta): chondrocytes exhibit a diminished sensitivity to TGF-beta.
Elevated extracellular matrix production and degradation upon bone morphogenetic protein-2 (BMP-2) stimulation point toward a role for BMP-2 in cartilage repair and remodeling.
01041-4&id=gr1.jpg){kind=link}
01041-4&id=gr2.jpg){kind=link}
01041-4&id=gr3.jpg){kind=link}
01041-4&id=gr4.jpg){kind=link}
01041-4&id=gr5.jpg){kind=link}