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
Freie Universität Berlin, Institute of Chemistry and Biochemistry, Berlin, GermanyCharité-Universitätsmedizin Berlin, Department of Orthopaedic, Trauma and Reconstructive Surgery, Campus Benjamin Franklin, Berlin, Germany
Freie Universität Berlin, Institute of Chemistry and Biochemistry, Berlin, GermanyBerlin-Brandenburg Center for Regenerative Therapies, Charité-Universitätsmedizin Berlin, Berlin, Germany
Address correspondence and reprint requests to: G. Schulze-Tanzil, Institute of Anatomy, Paracelsus Medical University, Salzburg and Nuremberg, Prof. Ernst Nathan Str. 1, 90419 Nuremberg, Germany. Tel: 49-(0)911-398-6772; Fax: 49-(0)911-398-6774.
Affiliations
Charité-Universitätsmedizin Berlin, Department of Orthopaedic, Trauma and Reconstructive Surgery, Campus Benjamin Franklin, Berlin, GermanyInstitute of Anatomy, Paracelsus Medical University, Salzburg and Nuremberg, Nuremberg, Germany
Anterior cruciate ligament (ACL) degeneration leads to knee instability and favors osteoarthritis (OA) progression. During ageing the growth factor sensitivity of ligaments changes but nothing is known about BMP2-signalling and -sensitivity in degenerated ACLs.
This study addressed the question whether a dysregulated BMP2 signalling might contribute to age- and OA-dependent ACL degeneration.
Method
ACL samples from patients with/without OA of different ages (<60 and ≥60 years, males, females) were graded histopathologically (n = 45). After stimulation of cultured ACL fibroblasts with 5 nM BMP2 for different time points, phosphorylation of SMAD1/5/8 and gene expression of crucial BMP2 signalling proteins, ligamentogenic and chondrogenic transcription factors, scleraxis (SCX) and SOX9, were analyzed.
Results
ACL samples displayed different grades of degeneration, often associated with synovitis and calcium deposits. Degeneration correlated significantly with synovitis. ACL fibroblasts expressed BMP type I receptors ALK3 and ALK6 and the BMP type II receptor BMPRII. Donors could be divided into “responders” and “non responders” since their BMP2 mediated SMAD1/5/8 phosphorylation level differed.
Basal ID1 expression was lower in cells derived from OA compared with non-OA patients and BMP2 led to an ID1 induction in both. Irrespective of BMP2 stimulation, the donor age significantly influenced the expression profile of BMP6 and SCX but not BMP signalling. The BMP2-mediated SMAD6 expression differed between OA and healthy ACL fibroblasts.
Conclusion
Our data indicate that the expression level of BMP2/SMAD target genes such as ID1 and SMAD6 was reduced in ACL fibroblasts derived from OA compared with non OA patients.
. Particularly the anterior cruciate ligament (ACL) as an intra-articular ligament covered by a synovial membrane and therefore, prone to synovitis might be strongly affected by OA. Patients with cartilage defects and OA in the knees can suffer from different grades of ACL and posterior cruciate ligament (PCL) degeneration
Twenty-year outcomes of a longitudinal prospective evaluation of isolated endoscopic anterior cruciate ligament reconstruction with patellar tendon autografts.
, which could explain chondroid metaplasia. Chondro-osteogenic bone morphogenetic proteins (BMPs) such as BMP2 and -4 have recently been implicated in histopathological changes in tendons
Higher BMP/Smad sensitivity of tendon-derived stem cells (TDSCs) isolated from the collagenase-induced tendon injury model: possible mechanism for their altered fate in vitro.
Higher BMP/Smad sensitivity of tendon-derived stem cells (TDSCs) isolated from the collagenase-induced tendon injury model: possible mechanism for their altered fate in vitro.
Calcification in calcium pyrophosphate dihydrate (CPPD) crystalline deposits in the knee: anatomic, radiographic, MR imaging, and histologic study in cadavers.
. However, it remains unclear whether a dysbalanced BMP2 signalling might be involved in calcium deposition in ACLs. It should be considered, whether the sensitivity of ACL fibroblasts towards extracellular BMP stimulations is altered in OA or ageing facilitating ACL degeneration. For example, in porcine ACLs an age dependent expression of growth factor receptors was described
. Significant differences in the expression profile could be detected for the transforming growth factor receptor 1 (TGFBR1), fibroblast growth factor receptor (FGFR), vascular endothelial growth factor receptor 2 (VEGFR2) and to a lesser extent for TGFBR3 and the platelet derived growth factor receptor (PDGFR)
. In humans with osteoarthritic knee joints, the articular cartilage expression of BMP receptors IA (BMPRIA, also called ALK3) correlated with the grade of OA degeneration. This correlation was higher for ALK3 than for its primary ligand, BMP2. ALK3 was therefore suggested as an indicator of cartilage degeneration
. BMP2 activates intracellular signalling by binding to heterooligomeric serine/threonine kinase receptor complexes composed of two type I BMP receptors (ALK3, BMPRIB, also called ALK6 or ACTRIA) and two type II BMP receptors (BMPRII, ACTRIIA or ACTRIIB), which can be presented as preformed complexes on the cell surface (reviewed in
). After ligand binding autophosphorylated BMP type II receptors phosphorylate type I receptors at the GS domain, a repeat of about thirty serine and glycine residues, where two receptor bound SMADs1/5/8 (R-SMADs) are activated by phosphorylation. Conformational changes lead to release from the receptor complex to bind the common mediator SMAD4 (Co-SMADs). The trimeric SMAD complex translocates to the nucleus to regulate BMP specific gene transcription.
We wanted to characterize SMAD dependent immediate downstream signalling events in the ACL fibroblasts after stimulation with BMP2 and measure expression of BMP antagonists and chondrogenic and ligamentogenic transcription factors in ACL fibroblasts derived from OA and non-OA patients of different ages. The results should provide information whether dysbalanced BMP signalling contributes to ACL degeneration, leading to functional insufficiency and accelerated OA progression.
Materials and methods
Sample acquisition
Forty-five ACL tissue samples of patients of different ages (<60 and ≥60 years, 12 males, 33 females, mean age: 63.6, age range: 23–81) were harvested during joint replacement surgery in accordance with the institutional ethical committee of the Charité-Universitätsmedizin Berlin, Campus Benjamin Franklin. Most of the patients revealed OA of the knee (n = 41). Some patients underwent ACL or total knee reconstruction due to traumatic damage (n = 4). An exclusion criterion was previous ACL replacement surgery to be sure to harvest the original ACL. For Western blot and qPCR analysis a subset of samples (n = 20) was recruited. n = 10 derived from old and n = 10 from younger donors. n = 16 samples derived from OA, n = 4 from non-OA patients.
Histological staining procedures and histological scoring
Tissue samples of ACL midsubstance including the synovial sheath were fixed in 4% paraformaldehyde (PFA) solution for 24–48 h and embedded in paraffin. Paraffin longitudinal sections of the samples underwent Hematoxylin–Eosin (HE) staining, alcian blue staining (AB) to assess the sulphated glycosaminoglycan (sGAG) content and van Kossa staining to detect calcium deposits following standard procedures.
Stained sections were graded by histology based on HE and AB (n = 45) using the degeneration score published by Hasegawa et al.
. Van Kossa stainings (n = 45) were evaluated using a self-designed scoring system (Supplemental Table I). Samples, which received score numbers between three and seven were positive for calcium deposits. As far as available (n = 20), the synovial membrane covering the ACL was checked for synovitis using the scoring system of Berger et al.
. Histopathology of pseudoanonymous samples was performed by an experienced anatomist, who was blinded for age, diagnosis (OA, non OA) of the donors and results of other analyses.
ACL fibroblast isolation and culturing
Surrounding connective tissue was removed and pure ACL tissue was cut into 1–2 mm2 slices and incubated in T25 culture flasks with growth medium (Ham's F-12/Dulbecco's Modified Eagle's [DMEM] Medium 1:1) containing 10% fetal calf serum (FCS), 50 IU/mL streptomycin, 50 IU/mL penicillin, 0.5 μg/mL partricin, 1% essential amino acids (all from Biochrom, Berlin, Germany) and 25 μg/mL ascorbic acid (Sigma–Aldrich) at 37°C and 5% CO2. Emigrating ACL fibroblasts were expanded in T75 culture flasks until passage 4–6 and seeded in 3.5 cm diameter petri dishes (Sarstedt [Cellplus], Nümbrecht, Germany) at 10,000–10,714 cells/cm2 for protein and qPCR analysis. ACL fibroblasts were gently rinsed three times and serum starved for 3 h using growth medium containing 0.5% FCS, before treated with 5 nM BMP2 (produced from Escherichia coli and kindly provided by Prof. Dr. Walter Sebald, University of Würzburg, Germany)
Osteoinductive capacity and heat stability of recombinant human bone morphogenetic protein-2 produced by Escherichia coli and dimerized by biochemical processing.
for 5, 10, 15, 30 or 60 min for Western blot and 60, 120 and 240 min for qPCR analyses. Controls were incubated with a similar volume (30.9 or 61.8 μL) of PBS for each time point.
Gene expression analysis
Total RNA was isolated from the petri dishes with unfixed cells using the RNeasy-Mini-Kit (Qiagen GmbH, Hilden, Germany) combined with RNAse (Qiagen) treatment according to the manufacturer's instructions and immediately stored at −80°C. RNA purity and quantity was measured using Nanodrop ND1000 (Peqlab Biotechnologie GmbH, Erlangen, Germany). Equal amounts of total RNA (500 ng in a volume of 20 μL) were reverse transcribed using the QuantiTect reverse transcription Kit (Qiagen) and immediately stored at −20°C. Nine microlitres containing 5.5 ng cDNA were amplified by qPCR in a 20-μL reaction mixture plated in triplicate using specific validated primer pairs (Supplemental Table II, Applied Biosystems, Foster City, USA). Contamination was excluded by a UNG digestion step. For all used primers, amplification efficiencies and mean normalized expression (MNE) ratios
ACL fibroblasts were rinsed twice with ice-cold PBS and then incubated with 0.3 mL lysis buffer (1 tablet of complete Mini Protease Inhibitor Cocktail [Roche Diagnostics, Basel, Switzerland], 100 μL 200 mM 1,4-dithio-DL-threitol (DTT) [Carl-Roth], 100 μL 0.1 M ethylene glycol tetraacetic acid (EGTA) [Carl Roth], 250 μL 1 M 4-(2-hydroxyethyl)-1 piperazin-ethanesulfonic acid (HEPES) [Biochrom], 100 μL 0.5 M magnesium chloride (MgCl2) [Sigma–Aldrich], 100 μL 10% Triton x-100 [Sigma–Aldrich], ultrapure water ad 10 mL) for 5 min. The suspension with cell fragments was centrifuged at 17,000× g and 4°C for 30 min. The supernatants containing total cell proteins was mixed with 6× Laemmli buffer (1 M Tris HCl pH 6.8, 4.2% sodium dodecyl sulphate [SDS], 2.1% β-mercaptoethanol (Sigma–Aldrich), bromphenolblue (Serva, Heidelberg, Germany), 7.5% glycerine (Sigma–Aldrich) to reach a target volume of 100 μL and denaturized by incubation at 95°C for 10 min. SDS polyacrylamide gel electrophoresis (SDS-PAGE) was performed for protein separation with 12.5% separating and 5% collecting gels at 80 V for 30 min and 120 V for 60–90 min. Then, proteins were transferred onto an activated PVDF membrane (0.45 μm pore size [Merck Millipore, Billerica, USA]) using 120 V for 90 min. The membranes were blocked (1:9 Roti®-Block [Carl Roth] in distilled water) at room temperature (RT) for 2 h before incubated with the specific primary antibodies overnight at 4°C (rabbit anti-phospho(P)-SMAD1/5/8 [#9511, cell signalling], rabbit pan-SMAD1 [D59D7, #6944, Cell Signalling Technology, Danvers, USA]) or for 1 h at RT (β-actin, A5441, Sigma–Aldrich). Then, the membranes were washed with washing buffer (0.5 mL Tween-20 [Sigma–Aldrich], PBS ad 1000 mL). The enzyme-coupled secondary antibodies (goat anti-mouse and goat anti-rabbit [P0447 and P0448] DAKO Cytomation, Hamburg, Germany) were added for 2 h at RT. After washing, the chemiluminescence reaction was initiated using horseradish peroxidase substrate peroxide solution (Merck Millipore) and luminol reagent (Merck Millipore). The arising chemiluminescence was detected using high performance chemiluminescence films (GE Healthcare Limited, Chalfont St. Giles, UK). Densitometric evaluation of the Western blot analyses was performed using Bio 1D software (Vilber-Lourmat Deutschland GmbH, Eberhardzell, Germany). Donors, who revealed a 1.5-fold or higher β-actin matched P-SMAD1/5/8 density compared with the respective controls were arbitrary classified as BMP2 “responders”.
Statistical analysis
All values were expressed as mean with 95% confidence interval. A Rout test with an abrasiveness of 1% was performed to identify outliers. Statistical significance was set at a P value of ≤0.05. Most data, for which a Gaussian distribution could be proven (D'Agostino and Pearson omnibus normality tests) was analyzed using one-sample-t-test and one-way analysis of variance (ANOVA). Non-parametric data was monitored using the Wilcoxon signed rank test. For ordinally scaled data from histopathological scoring differences among groups were assessed by ANOVA with continuous variables and Kruskal Wallis test. Kruskal Wallis is equivalent to Mann–Whitney test for non-parametric one way ANOVA. Spearman's non parametric correlation coefficient p was determined comparing two ordinal variables using SPSS22 (IBM Germany, Ehningen, Germany). All other calculations were undertaken using GraphPad Prism 6 (GraphPad software inc, San Diego, USA).
Results
Histopathological changes of the ACLs
The majority (75.6%, n = 34 of 45) of ACL samples showed a severe degeneration characterized by ECM fiber bundle deterioration and chondroid metaplasia, detectable by increased or inhomogeneous AB staining [Fig. 1(C2) ]. 64.4% (n = 29 of 45) revealed calcium deposits of different sizes with randomized localization within the ACLs and also in the covering synovial membrane [Supplemental Table IIIa and b, Fig. 1(A2, C2, D2), arrows]. 65% (n = 13 of 20) of the available samples exhibited symptoms of synovitis characterized by thickening of the intimal cell layers, increased cellularity in the subintimal stroma and accumulation of inflammatory cells [Fig. 1(B3)].
Fig. 1Representative histological images of a nearly unchanged (A1–D1) and heavily degenerated (A2, A3, C2, D2 and D3) samples and surrounding synovial sheaths. Samples were stained with HE (A1–B3), AB (C1,3) or van Kossa (D1,3). Altered cell morphology and distribution, disintegration of the extracellular matrix, inflammation of the synovial layer and increased sGAGs (blue) and calcium deposition (black) were detectable. Arrows indicate calcium deposits. Scale bars: 200 μm.
A higher ACL degeneration was also associated with a significantly higher likelihood for the presence of synovitis [Fig. 2(A)]. However, the degeneration of the ACL did not show any age- or gender-dependency in the samples investigated in this study. The assessment of ACL degeneration and calcium deposition, revealed no significant correlation [Fig. 2(B)].
Fig. 2Correlations between histopathological score systems. (A) Synovitis and degeneration scores revealed a significant correlation indicating that a high grade of degeneration increases the development of a synovitis in the knee. (B) Calcium deposition measured by the van Kossa score and ACL degeneration did not show a significant correlation. Spearman's non parametric correlation coefficient is shown. Degeneration and van Kossa score: n = 45. Synovitis score: n = 20.
A distinct trend, which did not reach the significance level towards a higher rate of calcium deposition in ACLs of the older patients, but no gender-dependency could be observed (not shown).
Expression of BMP receptors and responsiveness towards BMP2 in human ACL fibroblasts
We found the expression of the three BMP receptors ALK3, ALK6 and BMPRII
in unstimulated, serum starved ACL fibroblasts with no differential expression with respect to age and degeneration status of the tissue (Fig. 3).
Fig. 3The BMP2 responsive BMP receptors ALK3, ALK6 and BMPRII are expressed in human ACL cells. Serum starved and unstimulated ACL cells were analyzed for the gene expression of ALK3 (A), ALK6 (B) and BMPRII (C) by qPCR. Mean values of the MNE ± 95% confidence intervals (n = 20) are shown. Samples were normalized to the reference gene HPRT.
During the time course of BMP2 stimulation, ACL fibroblasts of the different donors showed different responses in the canonical BMP signalling pathway (Fig. 4 and Supplemental Table IV). Most of the donors (80%, n = 16 of 20) were identified to be responsive towards BMP2 stimulation in terms of activation of the canonical signalling pathway via phosphorylation of SMAD1/5/8. This group was classified as “responders”. In four cases we observed no or only slightly higher phosphorylation level of SMAD1/5/8 in response to BMP2 compared to unstimulated controls, which did not excess the threshold of 1.5-fold. This group (20%, n = 4 of 20) was classified as “non-responders” [Fig. 4(A) and (B)]. It was noticeable that in three out of the four non-responders we detected a strong phosphorylation of R-SMADs1/5/8 also in the unstimulated controls (Supplemental Table IV), which could be reasoned by either high secretion of BMPs by the ligamentocytes or ligand independent activity of the canonical BMP signalling pathway. The P-SMAD1/5/8 responsiveness of the ACL fibroblasts was not related to age and degeneration level of the ACL, but some correlations between the responsiveness and the gene expression levels of particular targets could be detected as described below.
Fig. 4Western blot and densitometric evaluation of ACL cells of two representative donors indicating either BMP2 responsiveness or BMP2 non-responsiveness with respect to SMAD1/5/8 phosphorylation. (A) Quantification of western blot analysis depicting phosphoprotein levels normalized to β-actin. Serum starved ACL fibroblasts were stimulated for 5, 10, 15, 30 and 60 min with 5 nM BMP2. The non-responsive donor (black bars) did not respond to BMP2 stimulations but showed higher basal P-SMAD1/5/8 level. The responsive donor (gray bars) showed high activation of SMAD1/5/8 compared to unstimulated controls. (B) Corresponding western blots for P-SMAD1/5/8 and β-actin of one non-responder (non-resp.) and one responder (resp.), which were used for quantifications are depicted.
The BMP2 target gene ID1 was significantly upregulated compared to unstimulated control samples already 1 h after stimulation with BMP2. This significant increase could be observed until the latest measured time point 4 h after BMP2 stimulation [Fig. 5(A) and (B)]. In unstimulated controls of donors with OA we found lower ID1 gene expression levels compared with unstimulated ligamentocytes of non-OA donors, with a significant difference at the 1 and 4 h time point [Fig. 5(A)]. Moreover, in “non-responders” with high basal P-SMAD1/5/8 protein amounts the increased activation of the canonical BMP pathway was reflected by significantly increased ID1 gene expression in unstimulated controls [Fig. 5(B)].
Fig. 5Differential ID1 gene expression in human ACL cells of patients with OA or high basal P-SMAD1/5/8 level. Serum starved ACL fibroblasts were stimulated for 1, 2 and 4 h with 5 nM BMP2 under serum-starved conditions. BMP2 stimulation of ACL fibroblasts resulted in a significant increase in the early target gene ID1 at each time point during the stimulation (indicated with a and b for significant BMP2 induced expression increase in all groups compared to unstimulated controls). (A) The OA group (n = 16) showed significantly less basal ID1 expression in the samples which remained unstimulated for 1 and 4 h compared to the cells derived from non-OA samples (n = 4) (B). In ACL cells showing high basal activation levels of SMAD1/5/8 by phosphorylation (n = 3), the ID1 gene expression was significantly increased in the cells which remained unstimulated for 1 and 2 h. Mean values of the MNE ± 95% confidence intervals (n = 20) are shown. Samples were normalized to the reference gene HPRT. a: P ≤ 0.05.
Chondrogenesis of periodontal ligament stem cells by transforming growth factor-beta3 and bone morphogenetic protein-6 in a normal healthy impacted third molar.
, we investigated gene expression of these BMPs after stimulation with BMP2. Interestingly, the gene expression of BMP6 did not differ between unstimulated and BMP2 stimulated cells at all taken time points [Fig. 6(A) and (B)], whereas BMP4 showed a slight downregulation as response to BMP2 stimulation [Fig. 6(C)]. Furthermore, cells of older donors and “non-responder” revealed a significant higher BMP6 expression compared to cells from younger donors or “responders” [Fig. 6(A) and (B)], indicating a differentially regulated expression pattern of BMP ligands during ageing and in response to a differently activated canonical BMP signalling cascade.
Fig. 6Relative gene expression of BMP6 and BMP4 in human ACL fibroblasts in response to BMP2 treatment. Gene expression of BMP6 in serum starved ACL cells was not altered by stimulation with 5 nM BMP2 over the whole time kinetics of 1, 2 or 4 h. (A) Gene expression of BMP6 was significantly higher (except for the samples that remained unstimulated for 4 h) in ACLs of older patients (n = 10) compared to younger individuals (n = 10). (B) Gene expression of BMP6 was increased in unstimulated and stimulated ACLs of non-responders (n = 4) compared to responders (n = 16) on the level of phosphorylation of SMAD1/5/8. (C) Gene expression of BMP4 was slightly decreased by BMP2 stimulation in all patients (n = 20). Mean values of the MNE ± 95% confidence intervals (n = 20) are shown. Samples were normalized to the reference gene HPRT. a: P ≤ 0.05; b: P ≤ 0.01.
Since we identified BMPs to be differentially expressed under different conditions in ACL fibroblasts we wanted to know if inhibitory components of the BMP pathway were also altered in their gene expression. SMAD6 is known to inhibit BMP signalling by competing with R-SMADs for receptor binding
. In general, BMP2 treatment increased the expression of SMAD6 after 2 and 4 h of stimulation compared to control samples [Fig. 7(A) and (B)]. In line with the decreased expression of ID1 under unstimulated conditions in donors with OA [Fig. 5(A)], we observed a decreased level of SMAD6 in this group compared to healthy individuals [Fig. 7(A)]. Moreover, the downregulated level in the OA situation was also detectable in BMP2 stimulated samples with significant changes in the 4 h stimulated samples. Comparing the two groups differing in their basal P-SMAD1/5/8 protein amounts, SMAD6 gene expression was higher in the group with high basal P-SMAD1/5/8 protein content [Fig. 7(B)]. Increased activation of the canonical BMP pathway in this particular group resulted in both increased ID1 and SMAD6 expression especially in unstimulated samples [Fig. 5, Fig. 7]. Noggin (NOG) is a secreted antagonist of several BMP ligands, e.g., BMP2 and BMP4 which prevents the binding of the BMP dimer to the oligomeric BMP receptor complex. BMP6 is known to be NOG insensitive (reviewed in
, we wanted to know if this mechanism is similar in ACL fibroblasts. For all measured time points we detected a clear but not significant upregulation of NOG in BMP2 stimulated samples. There was no differential expression related to age, BMP2 responsiveness status or tissue degeneration [Fig. 7(C)].
Fig. 7Relative gene expression of inhibitory factors of the BMP pathway. Serum starved ACL fibroblasts were stimulated for 1, 2 and 4 h with 5 nM BMP2 under serum-free conditions. (A) BMP2 stimulation seemed to increase the expression of the inhibitory SMAD protein SMAD6 in OA-derived fibroblasts and P-SMAD1/5/8 responders. The expression of SMAD6 was reduced in the OA group (n = 16) with significant difference in the 4 h BMP2 stimulation time point. (B) In ACL samples with high basal SMAD1/5/8 phosphorylation status (n = 3) the gene expression levels of SMAD6 were significantly increased in the 1 and 2 h stimulation controls. (C) Gene expression of the extracellularly acting BMP2 antagonist noggin was clearly and early induced by BMP2 stimulations but showed no differential expression in different groups (n = 20). Mean values of the MNE ± 95% confidence intervals are shown. Samples were normalized to the reference gene HPRT. a: P ≤ 0.05; b: P ≤ 0.01.
Gene expression of tenogenic and chondrogenic transcription factors in response to BMP2
The tenogenic transcription factor SCX was upregulated by BMP2 treatment at all investigated time points with significant upregulation in the 4 h stimulated samples [Fig. 8(A)]. Cells derived from older donors showed a higher SCX gene expression independent from stimulation with BMP2. However, the master chondrogenic transcription factor SOX9 remained unaffected by age [Fig. 8(B)]. Cells from “non-responders” revealed generally a higher SOX9 gene expression. This difference was significant at 4 h of BMP2 stimulation. Moreover, the SOX9 gene expression correlated significantly with the van Kossa score indicating calcification of the ACL [Fig. 8(C)].
Fig. 8Gene expression of tenogenic, but not chondrogenic markers are upregulated by BMP2 stimulation. Serum starved ACL fibroblasts were stimulated for 1, 2 and 4 h with 5 nM BMP2 under serum-free conditions. (A) BMP2 stimulation upregulated the expression of SCX, a marker for tendons and ligaments in younger (n = 10) and older (n = 10) patients and showed a general higher gene expression in older patients. (B) The chondrogenic transcription factor SOX9 is not regulated by BMP2 stimulations but shows higher expression in the non-responder group (n = 4) compared to responders (n = 16). (C) Since SOX9 has essential roles also in the endochondral bone formation the expression of SOX9 in our unstimulated 1 h samples correlated significantly with the appearance of calcium deposits in the ACL tissues quantified by scoring of van Kossa stainings. Mean values of the MNE ± 95% confidence intervals are shown. Samples were normalized to the reference gene HPRT. a: P ≤ 0.05.
. In accordance, in most of our investigated samples, BMP2 evoked a clear response in the ACL fibroblasts characterized by a time dependent SMAD1/5/8 phosphorylation, but also some non-responders could be distinguished. Apart from tightly regulated SMAD1/5/8 activation, many regulatory factors are involved in BMP signalling such as ligand and receptor availability. Three different receptor complexes are known to mediate BMP2 signalling: ALK2 (also designated as Activin A receptor type I: ACVRI), -3 (BMPRIA), -6 (BMPRIB) and the BMPRII, ACVRIIa and ACVRIIb. BMP/TGFβ receptors form heterotetramers. Within the receptor complex a type I and a type II receptor chain interacts with the dimeric ligand
. We found an expression of ALK3, -6 and BMPRII, but no significant correlation between ALK3, -6, BMPRII expression and SMAD1/5/8 responder or non-responder groups. However, there was no direct correlation between the severity of histopathological degenerative changes and the BMP2 responsiveness as well as with the age of the cell donors. So far, there were no published data available reporting the expression level of these receptor components in tendons and ligaments. Ongoing studies should also involve ALK1 and ALK5 expression analysis and study MAPkinase (ERK1/2) activation, since both receptor chains are involved in fibrosis of other tissues
. Nevertheless, the receptor expression analysis remained restricted to RNA level as a limitation of the present study. The composition of the ECM is also known to influence BMP ligand availability to the receptors
Histopathological changes in the human posterior cruciate ligament during aging and osteoarthritis: correlations with anterior cruciate ligament and cartilage changes.
. Matrix metalloproteinases (MMPs) which are responsible for ECM remodeling under physiological and pathological conditions such as OA bind and degrade BMP2
. Therefore, calcium deposits could result from dysbalanced BMP2 signalling in ACLs. Nevertheless, calcium deposits might not definitely reflect an ossification process, since they did not contain any cells. In 64.4% of the samples calcium deposits could be detected. These deposits were not detected clinically by X-ray. Compared to the report of Hasegawa et al.
who described 9.4% calcium-positive samples, it is a rather high number of positive samples. Although the donor mean age was very similar in both studies, this could be explained by the peculiarities of the donor collectives: Hasegawa et al., used samples from the tissue bank
, whereas the samples of the present study derived mostly from OA patients. The appearance and distribution of calcium deposits was very similar to that observed by Berleman et al.
, but in our study there was only a very slight trend to more deposits in female (data not shown). Others suggested the osteogenic differentiation of tendon-derived stem cells to be involved in the same feature in tenopathic tendon samples
, but due to only few controls included in our study we could not support this observation. Upstream signalling can also influence BMP2 signalling. Regulation e.g., by inhibitors such as NOG or SMAD6 but also miRNA and methylation of the receptor chains interferes with BMP2 signal transduction
. We found an induction of NOG and SMAD6 suggesting a negative feedback loop, which is in agreement with the report of Gazzerro et al. analyzing cultured osteoblasts
. In addition, ID1 was induced, but it remains unclear whether it is a direct or secondary effect of BMP2, since diverse BMP family members such as TGFβ induce ID1
as well as probably other growth factors. Interestingly, unstimulated ligamentocytes of OA patients revealed a lower ID1 gene expression than those of non-OA controls. BMP6 is a BMP which has like BMP2 also angiogenic properties
. Its role in tendon and ligaments is so far unknown. Looking at the whole donor collective BMP6 remained mainly unaffected by BMP2 stimulation of the ACL fibroblasts, whereas the gene expression of the osteogenic BMP4
was not significantly inhibited by the stimulatory regime. Interestingly, BMP6 gene expression was significantly higher in age compared with younger donors of the ACL cells, irrespectively whether cells were stimulated or not. BMP6 mediated osteogenic differentiation of osteoblasts and chondrogenic/osteogenic differentiation of periodontal ligament-derived stem cells
Chondrogenesis of periodontal ligament stem cells by transforming growth factor-beta3 and bone morphogenetic protein-6 in a normal healthy impacted third molar.
. It was not only higher in cells, which derived from old donors but also in those, that did not respond to the BMP2 with SMAD1/5/8 phosphorylation. This might suggest an induction of chondroid ACL metaplasia and might serve as a putative marker for the age of ligamentocytes.
Taking into account that this observation is only based on gene expression data BMP2 might not further induce the osteogenic pathway in ACL fibroblast. A balanced interplay between SOX9 and SCX determines the differentiation of early tendon/ligament cell progenitors towards the chondrogenic or tenogenic lineage
and might be also important to maintain mature tendon homeostasis. Lordaz-Diez et al. reported a down-regulation of SOX9 and aggrecan and up-regulation of SCX, and tenomodulin through the SMAD pathway during limb bud tenogenesis. TGFβ signalling induced SOX9 expression thereby mediating ectopic chondrogenesis
. However, we found a significant correlation between calcium deposition in tissue samples and SOX9 expression in the ACL fibroblasts isolated from the same tissue. This observation suggests the involvement of other signalling pathways in the induction of SOX9 and thereby contributing to ligament degeneration. Interestingly, ACL fibroblasts, which did not respond to BMP2 revealed a higher SOX9 expression.
This fact suggests also that BMP2 might not be directly responsible for the phenomena of chondroid metaplasia. Kumagai et al. demonstrated that degenerated ACLs revealed a higher Sox9 and a lower SCX protein synthesis
. We analyzed also the gene expression of SCX, but could not find any significant correlation between SOX9 and SCX gene expression. Interestingly, BMP2 induced significantly the SCX gene expression in the cultured ACL fibroblasts. SCX is known to play a role in mediating BMP4 gene expression during early tendon differentiation
. Since a non-significant inhibition of BMP4 gene expression could be observed at the same time points when SCX was induced by BMP2 a negative feed-back loop can be suggested in ACL fibroblasts by the data of the present study.
Some limitations have to be accounted for this study: the number of controls included in this study was rather low due to the difficulty to receive unchanged human ACL samples. Only a subset of samples (n = 20) could be used for gene expression analysis based on cultured primary human ACL fibroblasts. The cell isolation and expansion procedure might affect BMP2 signalling. Furthermore, the BMP2 signalling analysis was mainly restricted to gene expression data since the protein expression analysis remained limited to SMAD1/5/8 phosphorylation and histopathology.
In summary, comparing ACL fibroblasts of OA and non-OA patients we found differences in the expression level of the BMP2/SMAD target genes ID1 and SMAD6 suggesting that OA might modify BMP2 signalling. However, no direct influence of the donor age on the investigated BMP2/SMAD target genes could be detected.
Author contributions
KR: performed the densitometric and SPSS analyses, supervised the qPCR and Western blot experiments, provided experimental design and wrote the manuscript. CM: conducted cell culturing, Western blots and histological stainings. MU, ACK and KS: performed the qPCR experiments/analyses, cell culturing. BK: performed the statistical and qPCR data analysis. PK: contributed project design and critical discussion of data to the study. MJ and SA: provided the samples and clinical data analysis. GST: contributed study design, supervision, performed histopathology, wrote and revised the manuscript. All authors reviewed the manuscript critically and approved the final version and have no conflicts of interest.
Conflict of interest
All authors declare that there is no conflict of interest.
Acknowledgment
This study was funded by the Focus area DynAge at FU Berlin, through the DFG-excellence initiative (to GST and PK) and by DFG grant (SCHU1979/9-1 to GST). We acknowledge the technical support of Mr Tarik-Alp Sargut and Ms Karolin Holzwardt.
Appendix A. Supplementary data
The following are the supplementary data related to this article:
Twenty-year outcomes of a longitudinal prospective evaluation of isolated endoscopic anterior cruciate ligament reconstruction with patellar tendon autografts.
Higher BMP/Smad sensitivity of tendon-derived stem cells (TDSCs) isolated from the collagenase-induced tendon injury model: possible mechanism for their altered fate in vitro.
Calcification in calcium pyrophosphate dihydrate (CPPD) crystalline deposits in the knee: anatomic, radiographic, MR imaging, and histologic study in cadavers.
Osteoinductive capacity and heat stability of recombinant human bone morphogenetic protein-2 produced by Escherichia coli and dimerized by biochemical processing.
Chondrogenesis of periodontal ligament stem cells by transforming growth factor-beta3 and bone morphogenetic protein-6 in a normal healthy impacted third molar.
Histopathological changes in the human posterior cruciate ligament during aging and osteoarthritis: correlations with anterior cruciate ligament and cartilage changes.
30099-1&id=gr1.jpg){kind=link}
30099-1&id=gr2.jpg){kind=link}
30099-1&id=gr3.jpg){kind=link}
30099-1&id=gr4.jpg){kind=link}
30099-1&id=gr5.jpg){kind=link}
30099-1&id=gr6.jpg){kind=link}
30099-1&id=gr7.jpg){kind=link}
30099-1&id=gr8.jpg){kind=link}