Osteoarthritis (OA) is the most common degenerative joint disorder affecting about 10% of the global population [1
]. OA is characterized by progressive degradation of articular cartilage [2
] leading to pain, stiffness, and disabilities. There is no proven treatment for OA at present, therapies target late stage pain or inflammation, but in most cases prosthetic joint replacement remains as final therapy option [3
]. OA pathogenesis involves the interaction of cartilage-surrounding joint tissues including synovial tissue, bone, ligaments, meniscus, and the synovial fluid [4
]. Since articular cartilage is avascular, aneural, and alymphatic chondrocytes exhibit a very limited self-regenerative capacity [7
]. However, during the past decades, the existence of highly regenerative mesenchymal stem cells (MSCs) was confirmed in cartilage-surrounding tissues including synovium, bone marrow, and synovial fluid [8
]. These cells might migrate to damaged cartilage areas and differentiate to chondrocytes [9
There is strong evidence that synovial adipose tissue-derived MSCs (sASC) contribute to the repair of cartilage injuries [10
]. However, the chondrogenic capacity of MSCs detected in OA cartilage is insufficient [10
] and the reasons for this are not completely understood. Recent studies demonstrated that peripheral sympathetic nerve fibers are involved in chondrogenic extracellular matrix (ECM) deposition during endochondral ossification [12
]. Furthermore, sympathetic nerve fibers expressing tyrosine hydroxylase (TH), which is the key enzyme of catecholamine biosynthesis, have been detected in healthy and OA joint tissues [15
]. TH-positive nerve fibers release high concentrations of norepinephrine (NE) into the synovial fluid as shown in previous studies [18
NE is one major catecholaminergic neurotransmitter of the sympathetic nervous system and binds to specific adrenergic receptors (AR) depending on its concentration [19
]. All α and β adrenergic receptors subtypes (α1A, α1B, α1D, α2A, α2B, α2C, β1, β2, β3) belong to the G protein-coupled receptor family. Depending on the activation of alpha subunits of ARs (Gs, Gi, or Gq), different intracellular signaling pathways become activated [21
]. At low concentrations (≤10−7
M), NE mainly acts via α-ARs with subsequent protein kinase a (PKA) and cAMP inhibition (Gαi signaling). In contrast, NE at high concentrations (≥10−7
M) preferentially acts via β-ARs leading to PKA and cAMP increase (Gαi signaling). Furthermore, β-arrestin regulates AR-mediated signaling by binding to phosphorylated ARs which in turn leads to the alternative ERK1/2 pathway [22
]. Previous studies investigated the effects of NE on articular cartilage and described that NE induced catabolic effects in chondrocytes by inhibiting type II collagen synthesis via β2-AR or accelerated cartilage ECM degradation via α2A-AR suggesting that mainly the α2A- and β2-AR subtypes might play a role in disturbed cartilage regeneration during OA development [25
]. Recently, we demonstrated that bone marrow-derived MSCs (BMSCs) obtained from trauma patients express α- and β-ARs [18
] and that NE inhibits chondrogenic differentiation with concomitant induction of hypertrophy by the β2-AR under normoxic conditions (20% O2
]. However, no studies using sASC regarding NE sensitivity and chondrogenic potential under NE influence have been performed until now. Furthermore, most existing studies on the regeneration potential of MSCs have been performed under normoxia, although cartilage microenvironment contains only about 2% O2
representing the “physioxic condition” [28
]. Therefore, the aim of the present study was to address the effect of NE on sASC proliferation and chondrogenic differentiation under physioxic conditions.
Cartilage injuries as well as OA-associated cartilage degeneration still represent a huge orthopedic challenge [30
]. Joint resident MSCs have been shown to be involved in cartilage regeneration processes [31
], however, these MSCs exhibit an insufficient chondrogenic potential, although their number is increased in OA articular cartilage tissue [32
]. The reasons for this contradictory phenomenon are not completely understood. Recent studies demonstrated that TH-positive nerve fibers are present in trauma and OA synovium [16
]. Moreover, a considerable amount of the sympathetic neurotransmitter NE was detected in synovial fluid samples of trauma and OA patients [18
], which might influence the regenerative capacity of the stem cells migrating from the synovial tissue to damaged cartilage areas [9
]. The present study is the first demonstrating noradrenergic effects on sASC chondrogenesis and accordingly on articular cartilage homeostasis under physioxic conditions.
The initial step of this study was to confirm the molecular MSC characteristics of the isolated sASCs. Analysis of multiple independent samples over months indicated that our isolation and cell culture method resulted in homogenous sASC populations expressing the MSC-specific surface marker panel [34
]. Next, the chondrogenic differentiation capacity of sASCs derived from late OA patients was analyzed. We demonstrated in this study for the first time that these sASCs condensated to typical chondrogenic pellets and synthesized considerable amounts of cartilage-specific ECM over 21 days under physioxia.
Another important prerequisite for the present study investigating the effects of NE on sASCs function was the analysis of the complete AR profile of these cells. We could show for the first time that sASCs expressed α1B, α2A-, α2B-, α2C-, and β2-AR. Kotova et al. also detected α1B, α2A-, and β2-AR, but not α2B- and α2C on abdominal subcutaneous adipose tissue-derived MSCs [35
]. Interestingly, in the same tissue but in another study α1A, α1B-, α2A-, α2B-, and β1-AR were strongly but β2-AR only weakly expressed at mRNA and protein level [36
]. One possible reason for this apparent discrepancy might be the tissue-specific expression of different AR subtypes. In addition, the pathophysiological situation caused by a catabolic microenvironment with increased NE concentrations might influence the expression level of ARs [36
]. However, it was not possible to compare our sASCs derived from OA patients with ASCs from healthy or knee trauma donors, first, because healthy synovial samples were not available and second, because in most cases neither healthy nor knee trauma synovial tissue samples are surrounded or overgrown by adipose tissue in a similar manner to the OA synovium. In order to consider possible autocrine effects, the expression of TH, the key enzyme of catecholamine biosynthesis, was analyzed in sACSs. However, TH expression was not detectable, neither in monolayer nor in chondrogenic sASC cultures. Thus, autocrine effects can be excluded. This result is in line with our previous study demonstrating that BMSCs derived from knee trauma patients are TH-negative [18
]. The expression of ARs during chondrogenesis was also an important point in this study. At the mRNA level α2A-, α2B-, and β2-AR were strongly expressed both at the beginning and at the end of differentiation enabling NE to unfold its effects continuously over 21 days, while α1A-, α1B-, α2C-, and β1-AR expression was lower suggesting that chondrogenesis itself does not influence AR expression. A similar AR expression profile in OA articular chondrocytes was recently described by Speichert et al. and by Lorenz et al. [26
] indicating that the AR profile of cells from the same embryonal mesodermal origin might be conserved [38
]. After detecting the major ARs, we demonstrated that these receptors activate the downstream intracellular ERK1/2 signaling pathway, without influencing the alternative PKA pathway, which is in line with previous studies investigating murine chondrocytes under NE influence [25
Even though sASCs express different ARs, NE had no effect on the proliferation and also their viability was not influenced indicating that even highly-concentrated NE is not toxic to sASCs. One possible explanation for proliferative non-responsiveness of sASCs to NE might be the immediate vicinity to a catabolic and inflammatory microenvironment in the OA-affected joint. A similar non-responsiveness to NE was also observed by Lorenz et al. in monolayer OA chondrocyte proliferation experiments [26
The relationship between NE concentration and chondrogenesis was investigated as a next step. These experiments revealed a clear inhibition of chondrogenesis when NE was added in high concentrations suggesting the involvement of β-ARs. We observed that pellets treated with NE were smaller in size. For this reduction of pellet size different processes can be responsible: First, cell apoptosis could result in smaller pellets. However, we demonstrated that cell viability was not affected by any treatment as indicated by an unchanged LDH activity. Furthermore, the cell number was constant in all pellets reflected by equal dsDNA content. Second, the pellet size might be reduced as a result of suppressed ECM synthesis and indeed, we evidenced significantly lower sGAG and type II collagen concentrations in pellets treated with NE. This finding is in line with earlier studies demonstrating decreased sGAG and type II collagen synthesis during BMSC or chondrocyte differentiation under NE influence [18
]. The third possible explanation of reduced pellet size might be the acceleration of hypertrophy or the increase of matrix-degrading enzyme expression. However, in contrast to NE-treated BMSC pellets, hypertrophic differentiation characterized by increased COL10A1, RUNX2, and MMP13 expression was not observed in the present study [18
]. One reason for suppressed hypertrophy might be the incubation of the pellets under physioxia as recently described by others [29
In order to confirm the involvement of the β2-AR, the only β-AR expressed by chondrogenic pellets, specific AR antagonists were tested in combination with NE. Based on previous studies [18
], we expected that the β2-AR antagonist propranolol might reverse the NE-mediated inhibition of sASC chondrogenesis. However, propranolol was only partly able to neutralize NE effects, while the α2-AR antagonist yohimbine significantly abrogated the inhibitory effect of NE on the pellet volume and type II collagen protein. The only existing study analyzing alpha-adrenergic signaling in chondrocytes was performed by Jiao et. al. identifying α2a-AR as responsible AR for promoting degenerative remodeling in the temporomandibular joint by induction of catabolic activities in chondrocytes [25
]. Typically, α2a-AR used to be targeted by lower NE concentrations such as 10−8
], but in the present study the effects of high NE concentrations were neutralized by α2a-AR and not or only partly by β2-AR. The reason for these mixed α2-/β2-AR-mediated effects might be the instability of NE under cell culture conditions, as demonstrated. After one or two days in culture medium, less than 40% of the initial NE is still available, thus, the 10−7
M initial NE concentration might decrease to 10−8
M. Another possible explanation would be the switch of β2-AR from Gαs to Gαi signaling as described previously by us [41
]. The fact that the α1-ARs were not involved in any effect was not surprising, since no studies describing α1-AR-induced influences on chondrogenesis exist.
In conclusion, this study demonstrated that sASCs obtained from OA patients exhibit a strongly decreased chondrogenic capacity in the presence of NE mediated by α2a-dependent ERK1/2 phosphorylation and reversed by the specific α2-AR antagonist yohimbine. Thus, NE might suppress sASC-dependent regeneration of articular cartilage and contribute to the manifestation of OA. Therefore, the inhibition of α2a-adrenergic signaling pathways represents a promising approach for the development of novel OA strategies.
4. Materials and Methods
Adipose synovial tissue was obtained from patients with OA during knee joint replacement surgery. The experimental cohort included 32 patients (characteristics of patients in Table 1
). Patients were informed about the purpose of the study and gave written consent. A non-selective beta blocker medication, targeting not only β1- but also the β2-AR, was a criterion for exclusion in this study. The project was approved by the Ethics Committee of the University of Regensburg (Ethikkommission der Medizinischen Fakultät der Universität Regensburg, vote number/project ID 13-101-0135, approved: 26 August 2015) and of the Ethics Committee Goethe University Frankfurt am Main (Ethik-Kommission des Fachbereichs Medizin Universitätsklinikum der Goethe-Universität, vote number/project ID 148-17B, approved: 10 May2017). All experiments were performed in accordance with relevant guidelines and regulations.
4.2. Isolation and FACS Characterization of Human ASCs
Human sASCs were isolated as described previously [42
]. The cells were seeded in 75 cm2
tissue culture flasks and cultivated in Dulbecco’s modified Eagle medium (DMEM/F12; Gibco Invitrogen, Thermo Fisher Scientific, Darmstadt, Germany) containing 1% penicillin/streptomycin (Gibco Invitrogen) and 10% MSC qualified FBS (Gibco Invitrogen) at 37 °C in a humidified atmosphere containing 2% O2
and 5% CO2
. The MSC characteristics of isolated sASCs were investigated by FACS analysis according to the suggestions of “The Mesenchymal and Tissue Stem Cell Committee of the International Society for Cellular Therapy” [34
4.3. In Vitro Proliferation of sASC
Human synovial sASCs were seeded at a cell number of 2 × 105 in a 75 cm2 flask. In addition to untreated control, cells were treated for 7 days with NE at different concentrations (10−9–10−6 M, Sigma Aldrich, Munich, Germany) at 37 °C in a humidified atmosphere containing 2% O2 and 5% CO2. NE was freshly added at day 0, 3, and 6. After seven days, total viable and dead cell number was determined.
4.4. Determination of Cell Viability
To determine possible toxic effects of treatments lactate dehydrogenase release (LDH Cytotoxicity Detection Kit; TaKara MK401, Shiga, Japan) was measured in supernatants of monolayer cell cultures at day 7. In addition, LDH in supernatants during chondrogenesis at day 1, 7, 14, and 21 was analyzed. Cells lysed with 1% Triton X-100 were taken as positive control and medium without cells as negative control.
4.5. Chondrogenic Differentiation of sASCs
In vitro chondrogenesis was performed as described earlier [18
] by using serum-free high glucose DMEM containing 1% P/S, 100 nM dexamethasone, 200 µM ascorbate-2-phosphate, 10 ng/mL TGF-β3, 10 ng/mL BMP-6 and ITS+3 premix (Sigma). Pellets (200.000 cells per pellet) were formed by centrifugation (491× g
) in 96-well plates with conical bottom (Nunc/Fisher Scientific, Schwerte, Germany) and were cultivated for 21 days at 37 °C in a humidified atmosphere containing and 2% O2
and 5% CO2
. Pellets were treated with NE (10−9
M, Sigma). In addition, pellets were treated with specific α1-AR antagonist doxazosin (10−7
M, Tocris Bioscience, Bristol, UK), specific α2-AR antagonist yohimbine (10−6
M, Tocris Bioscience), and specific β2-AR antagonist propranolol (10−6
M, Tocris Bioscience, Bristol, UK), alone or in combination with NE (10−6
M). The differentiation medium with freshly diluted supplements was changed every two days.
4.6. Norepinephrine Quantification
The stability of NE at the cell culture conditions used was examined by adding a concentration of 10−6
M NE to the culture medium at time point zero and high-pressure liquid chromatography of medium samples after 24 h as previously described by us [44
4.7. Western Blot Analysis
In order to examine whether monolayer chondrocytes respond to NE, the two major AR-dependent signaling pathways—the phosphorylation of PKA and ERK1/2—were investigated. The sASCs were treated with NE (10−9
M) and/or with specific AR antagonists as described previously [25
]. ASCs were lysed and pellets were mechanically homogenized using Polytron PT-1200 (Kinematica, Thermo Fisher Scientific, Darmstadt, Germany) homogenizer and proteins were isolated using PhosphoSafe™ Extraction Reagent (Merck Millipore, Darmstadt, Germany). Samples were loaded onto 10% SDS-PAGE and electro-transferred to a polyvinylidene difluoride (PVDF) membrane. Membranes were blocked with 5% bovine serum albumin for 1 h at room temperature before incubation with primary antibodies detecting total ERK1/2 (#9107, mouse, Cell Signaling Technology, Frankfurt am Main, Germany), phosphorylated ERK1/2 (#4370; rabbit, Cell Signaling Technology), total PKA (ab32514, rabbit, Abcam, Cambridge, UK), phosphorylated PKA (ab32390, rabbit, Abcam) and GAPDH (MA5-15738, mouse, Thermo Fisher Scientific, Darmstadt, Germany) at 4 °C overnight. The membranes were incubated with a HRP-conjugated secondary antibody (swine anti-rabbit P039901-2, rabbit anti-mouse P026002-2, both from DAKO, Agilent Technologies, Hamburg.Germany) for 1 h at room temperature. The target protein was detected using the enhanced chemiluminescence (ECL, Thermo Fisher Scientific, Darmstadt, Germany) reagents, with GAPDH as endogenous control. Densitometric values of the detected bands were quantified using the ImageJ Software (https://imagej.nih.gov/ij/download.html
4.8. RNA Isolation, Endpoint and Real-Time Quantitative PCR
Monolayer sASCs were lysed and pellets were homogenized mechanically (Polytron PT-1200, Kinematica). RNA isolation was performed using the NucleoSpin RNA kit (Machrey Nagel, Düren, Germany) according to the manufacturer’s instructions. cDNA synthesis was carried out using qScript cDNA Supermix (Quanta Biosciences, Beverly, MA, USA). Gene expression of α- (α1A, α1B, α1D, α2A, α2B, α2C), β-AR (β1, β2, β3) subtypes and TH was determined by qPCR using Taq PCR Master Mix kit (Qiagen, Hilden, Germany). The PCR products were run on a 1.8% (wt
) agarose gel, stained with GelRed Nucleic Acid Gel Stain (Biotium, Fremont, CA, USA). Average score of AR and TH gene expression in sASCs from three different OA patients was calculated. In order to analyze the effects of NE and specific antagonists on chondrogenic differentiation of sASCs, real-time quantitative PCR was performed using Quanta PerfeCta SYBR Green FastMix (Quanta Biosciences) in a qTOWER3
real time PCR Thermocycler (Analaytik, Jena, Jena, Germany). Gene expression of chondrogenic markers (SOX9, COL2A1), of fibrous cartilage markers (COL1A1), and hypertrophic markers (RUNX2,COL10A1, and MMP13) was quantified. GAPDH served as endogenous control. Relative gene expression was determined using qPCR software (Analytic, Jena). All primers were synthesized by Thermo Fisher Scientific (Supplementary Table S1
4.9. Macroscopic and Histological Investigations
Macroscopic images of day 7, 14, and 21 pellets were taken using a standard binocular with Polaroid PDMC-3 camera. Surface areas of spherical pellets were analyzed (ImageJ software) and the pellets volume was calculated mathematically using the average radius. Then, pellets were fixed after day 7, 14, and 21 in 4% paraformaldehyde and infiltrated with increasing concentrations (10, 20, and 30%) of sucrose, each concentration for one day. Pellets were embedded in Tissue-Tek (Sakura, Alphen aan den Rijn, Netherlands) and sectioned at 8 µm thickness using a cryotom ( Cryostar NX70, Thermo Fisher Scientific, Darmstadt, Germany). Cryosections were stained with dimethylmethylene blue (DMMB, Sigma Aldrich, Munich) to visualize the sulfated glycosaminoglycans (sGAGs).
After 5 min rehydration in 1× PBS, cryosections to be stained for adrenergic receptors were demasked using citrate buffer (10 mM sodium citrate, 0.05% tween 20, pH 6) for 20 min at 95 °C. Sections for type II collagen staining were digested with 1 mg/mL pepsin in 1× McIlvaine buffer (pH 3.6) for 12–15 min at 37 °C. Then endogenous alkaline phosphatase (Bloxall, Vector Labs, Linaris, Dossenheim, Germany, 10 min, room temparature) peroxidase (0.3% H2O2 10 min, room temperature) were blocked. Non-specific binding sites were blocked using secondary antibody-specific sera (VECTASTAIN® ABC-AP Staining KIT, Vecor Labs, Linaris, Dossenheim, Germany or HRP-AEC Kit, Linaris, Dossenheim, Germany) for 45 min at room temperature. Sections were then incubated with the primary rabbit antibodies directed against α2a AR (1:200; ab85570), β2-AR (1:200; ab213651), and type II collagen (1:200, ab34712) at 4 °C overnight. Specific staining was visualized using horseradish peroxidase labelled secondary antibodies and ALP or peroxidase substrate solutions (VECTASTAIN® ABC-AP Staining KIT, Vecor Labs, Linaris, Dossenheim, Germany or HRP-AEC Kit, Linaris, Dossenheim, Germany).
4.11. Biochemical Analysis of sGAGs and Type II Collagen Protein
For the quantification of sGAGs and type II collagen levels as well as double-stranded DNA (dsDNA), pellets were mechanically homogenized using a Polytron PT-1200 (Kinematica) homogenizer and digested as described previously [18
]. Concentration of dsDNA was determined using the Quant-iT PicoGreen assay kit (Invitrogen). The sGAG content of digested pellets was measured using a colorimetric assay based on dimethylmethylene blue (DMMB, Sigma) [18
]. Type II collagen protein levels were quantified by ELISA (Chondrex, AMS Biotechnology (Europe) Ltd, Abingdon OX14 4SE, UK). sGAG and type II collagen levels were normalized to the dsDNA content.
4.12. Statistical Analysis
Statistical analysis was performed using SigmaPlot software (SigmaPlot V.13, Systat Software, Erkrath, Germany). All experiments were carried out with cells of 4–10 patients. Comparisons between groups were performed using ANOVA on ranks or Wilcoxon/Mann–Whitney-Test followed by Bonferroni or Dunn’s correction. p values less than 0.05 were considered significant.