Objectives To investigate the effects of interleukin 1β (IL1β) treatment on the Notch1/Hes1 pathway in chondrocytes in vitro.
Methods Mouse articular chondrocytes in primary culture were challenged with IL1β, alone or combined with Notch1 and IL1β pathway inhibitors. Notch1 and Hes1 expressions were investigated by immunocytochemistry, western blot and real-time quantitative (q)PCR. IL1β-responsive genes were assessed by real-time qPCR and a specific siRNA against Hes1 was used to identify Hes1 target genes.
Results Notch1 labelling remained nuclear and stable in intensity irrespective of treatment, suggesting a steady state activation of this pathway in our model. IL1β transiently increased Hes1 mRNA (2.5-fold) and protein expression in treated versus naive chondrocytes. Hes1 mRNA level then decreased below control and its cyclic pattern of expression was lost. This was associated with nuclear translocation of the cytoplasmic Hes1 protein. IL1β induced increase in Hes1 mRNA was transcriptional, occurred through nuclear factor (NF)κB activation and appeared to be associated with downregulation by its own protein. Hes1 induction was insensitive to the γ-secretase inhibitor N-(N-(3,5-difluorophenacetyl)-l-alanyl)-S-phenylglycine t-butyl ester (DAPT), which suggested its independence from novel Notch1 activation. Hes1 expression was efficiently silenced by a specific siRNA. This experiment revealed that Hes1 did not mediate IL1β-induced downregulation of Sox9, type II collagen and aggrecan transcription but mediated IL1β induction of matrix metalloproteinase (MMP)13 and ADAM metallopeptidase with thrombospondin type 1 motif, 5 (ADAMTS5). The Hes1-related repressor Hey1 was expressed at a very low level and was not inducible by IL1β.
Conclusion Hes1 is a novel IL1β target gene in chondrocytes which influences a discrete subset of genes linked to cartilage matrix remodelling and/or degradation.
Statistics from Altmetric.com
In physiological conditions, articular chondrocytes remain in a quiescent state while interactions with the extracellular matrix (ECM) contribute to phenotype stability.1 Growth plate chondrocytes follow a maturation process characterised by the expression of specific genes.2 Some of these genes, including matrix metalloproteinases (MMPs) and aggrecanases (ADAM metallopeptidase with thrombospondin type 1 motif, 5 (ADAMTS5)), are expressed in physiological remodelling and pathological ECM degradation.
We recently reported that MMP13 expression is influenced by the Notch pathway in murine chondrocytes induced to dedifferentiate in vitro.3 The Notch pathway is an array of cascades regulating cell fate determination during development (for review see Kopan and Ilagan4). Four Notch receptors and several counter receptors constitute the general Notch pathway,5 which is also found in chondrocytes.6 Notch receptors are activated by extracellular proteolytic cleavages by transarterial chemoembolisation (TACE) and γ-secretase. The cleaved notch intracellular domain (NICD) translocates into the nucleus and activates the hairy enhancer of split (HES) family of transcriptional factors (Hes, Hey).6 7 Hes1 is considered as a master gene in the Notch signalling cascade.8 In certain cases, Hes1 and Hey1 cooperate through heterodimerisation as shown for Col2a1 transcriptional repression in mesenchymal cells.9 The target genes of these regulators remain poorly known and the Notch cascade appears differentially involved in chondrogenesis, depending on cell populations, that is, mesenchymal cell lines,10 cultured articular chondrocytes11 or hypertrophic chondrocytes.12 Genetic defects of Notch cascade members have been linked to different diseases,13 suggesting non-redundant functions.
In cartilage, Notch1 has been proposed to be involved in the osteoarthritis (OA) degradative process14 but the role played by Hes/Hey remains elusive. OA is characterised by loss of articular cartilage matrix, as a result of proteolytic enzymes action on proteoglycans and type II collagen, associated with chondrocyte dedifferentiation.15 16 However, changes in chondrocytic phenotype, including catabolic, anabolic, fibroblastic and hypertrophic changes, are also observed, besides OA, in different cartilage pathogenic statuses and/or localisations.16 17
Proinflammatory cytokines such as interleukin 1β (IL1β) increase OA degradative processes.18 IL1β upregulates MMPs 1, 3, 13 and ADAMTS5, and perturbates the chondrocytic phenotype by reducing expression of aggrecan and type II collagen.19 Here, we report that IL1β induces Hes1 expression in chondrocytes in vitro. Hes1 then induces MMP13 and ADAMTS5 expression independently of Notch activation.
Reagents were purchased from Sigma Aldrich (St Quentin, France), unless stated otherwise. Inhibitors of γ-secretase (N-(N-(3,5-difluorophenacetyl)-l-alanyl)-S-phenylglycine t-butyl ester (DAPT)), nuclear factor (NF)κB (BAY 1170-82), extracellular-regulated kinases (ERK1/2) (PD 98059), p38MEK (SB 203580) and c-Jun N-terminal kinase II (JNK II) were obtained from Calbiochem (France Biochem, Meudon, France).
Sample collection and cell culture
Experiments followed protocols approved by the French/European ethics committee. Cell culture medium and fetal calf serum (FCS) were from Invitrogen (Cergy-Pontoise, France). Collagenase D and complete protease inhibitor were from Roche Diagnostics (Meylan, France). Methods for immature murine articular chondrocyte (IMAC) collection and their phenotyping have been previously described.These cells display a typical chondrocytic phenotype: robust type II collagen and proteoglycan expression associated with a low level of type I collagen expression. IMACs respond to cytokine stimulation by strongly expressing cyclo-oxygenase 2, MMPs and producing nitric oxide (NO) and prostaglandin E2 (PGE2).20 21 Isolated femoral heads and tibial plateaus from Swiss mice litters, 5 to 6 days old (Janvier, France), underwent two 45-min incubations with collagenase D (3 mg/ml in Dulbecco's modified Eagle medium (DMEM), Gibco-BRL) at 37°C under 5% CO2. Soft tissues were detached and cartilage pieces were incubated overnight with collagenase D solution (0.5 mg/ml) at 37°C. Chondrocytes were resuspended in culture medium (DMEM; 1 g/litre glucose) supplemented with 5 % FCS and antibiotics, and seeded in six-well plates at 1.5×105 cells/well. Hes1 and Hey1 expression was analysed under different conditions: 5% FCS, no FCS and no FCS plus IL1β (2.5 ng/ml). Cells were deprived of serum 24 h before the onset of IL1β experiments. The 2.5 ng ml IL1β concentration was determined in a preliminary experiment as yielding the maximal linear response for Hes1 mRNA induction. Human cartilage samples were collected following orthopaedic surgery.
Western blot analysis
Cells were challenged with IL1β for 3 h, in parallel with untreated control cells, washed twice with ice-cold Ca/Mg-free phosphate-buffered saline (PBS) balanced salt solution and scraped off the flask in cold lysis buffer containing complete protease inhibitors (Roche). The pellet was collected in lysis buffer containing dithiothreitol (DTT) (10mM) and protease inhibitors, and the suspension was incubated for 30 min on ice on a rocking platform set at 150 rpm. The suspension was centrifuged at 14 000 g for 10 min, and the supernatant was stored at −80°C until use. Proteins were size separated by use of NuPage 4% to 12% Bis-Tris gel (Invitrogen). The molecular weight marker used was SeeBlue, code LC5925 (Invitrogen). The gel was electroblotted on to nitrocellulose in transfer buffer in a XCell II Blot Module (Invitrogen). Equal transfer of proteins was confirmed by staining the nitrocellulose membranes with Ponceau Red. The rabbit polyclonal antibodies against Notch1 (C-20R), Hes1 (H-140) and Hey1 (HRT1, H-120) were from Santa Cruz (Tebu, Le Perray, France). These antibodies recognised mouse and human proteins. Each antibody was applied to membranes overnight at a 1:200 dilution at 4°C. A secondary peroxidase-conjugated goat anti-rabbit IgG antibody (Dako P0160, Glostrup, Denmark) (1:20 000) was added and incubated for 1 h at room temperature. Blots were revealed by enhanced chemiluminescence as instructed by the manufacturer (Amersham Biosciences, France). Semiquantitative scanning densitometry involved the National Institutes of Health (NIH) image software and results were standardised by use of Ponceau Red staining.
Chondrocytes were grown as primary cultures on 4×1 cm2 glass slides (Nunc, Strasbourg, France). Cells at 70% to 80% confluence were treated or not with IL1β (2.5 ng/ml) for various periods. Cells were then rinsed twice with PBS, fixed with 4% paraformaldehyde for 10 min at room temperature, permeated with 0.2% Triton X-100 in PBS, saturated with 1% bovine serum albumin in PBS and incubated with specific primary antibodies (1 µg/ml) as described in the text, or preimmune rabbit immunoglobulins (1 µg/ml), overnight at 4°C. After two washes in PBS, the cells were incubated with Alexa 488-conjugated secondary antibody (1:200) in PBS for 2 h at room temperature. After three washes in PBS, cover slips were mounted with Slow-Fade Gold mounting medium (Invitrogen) and observed under a Nikon Diaphot 300 microscope (Nikon, Tokyo, Japan). Label localisation was assessed on 3 sets of 100 cells from triplicate cover slips.
Real-time quantitative (q)PCR
Total RNA was extracted from IMACs with use of the RNAeasy kit (Qiagen, Courtaboeuf, France) according to the manufacturer's instructions. Concentrations were determined by spectrophotometry. Aliquots of 1–2 µg total RNA were analysed by use of the High Capacity cDNA Reverse Transcription (RT) kit (Applied Biosystems, Foster City, California, USA). Two RTs were performed for each RNA to prevent intraexperimental variations. Amplification was performed with 1 or 2 µg cDNA and involved a thermal cycler (ABI prism 7900 HT; Applied Biosystems), the SYBR green fluorescence method and specific oligonucleotides. Results were analysed with the SDS 2.1 real-time detection system software. Quantification of RNA involved comparison of the number of cycles through the (δ−δCt) method. Sequences used for amplification standardisation were 36B4 for mouse mRNA and RPL13A for human mRNA.
Specific cDNA primers (Eurogentec, France) were designed with use of Oligo Explorer 1.1 (http://www.genelink.com/tools/gl-downloads.asp) and verified with use of the BLAST online engine. Mouse primers were as follows: 36b4: upper 5′-TACACCTTCCCACTTGCTG-3′, lower 5′-TCTGATTCCTCCGACTCTTC-3′; Hes1: upper 5′-ATTTGCCTTTCTCATCCCC-3′, lower 5′-GTTCGTTTTTAGTGTCCGTCAG-3′; Hey1: upper 5′-TCTCAGCCTTCCCCTTTTC-3′, lower 5′-CTTTCCCCTCCCTTGTTCTAC-3′; aggrecan: upper 5′-GACCAGGAAGGGAGGAGTAG-3′, lower 5′-CAGCCGAGAAATGACACC-3′; type II collagen: upper 5′-TGGCTTAGGGCAGAGAGAG-3′, lower 5′-GGTGGCAGAGTTTCAGGTC-3′; iNOS: upper 5′-ACGCTTCACTTCCAATGC-3′, lower 5′-GACAATCCACAACTCGCTC-3′; MMP1: upper 5′-AGGTTTGGGGGGTGATG-3′, lower 5′-TGGCTGGATGGGATTTG-3′; MMP3: upper 5′-TGACGATGATGAACGATGG-3′, lower 5′-AGAGATGGAAACGGGACAAG-3′; MMP13: upper 5′-CACAGCAAGCCAGAATAAAG-3′, lower 5′-CACACATCAGTAAGCACCAAG-3′; ADAMTS5: upper 5′-AGACCTACCACGGAAGCAG-3′, lower 5′-CACCACAGCACACCACAG-3′; c-Jun: upper 5′-AGCGAACTGGGGAGGAG-3′, lower 5′-CGGACTGGAGGAACGAG-3′; cyclin D1: upper 5′-TACCCTGACACCAATCTCCTC-3′, lower 5′-CCTCCTCTTCGCACTTCTG-3′; p21: upper 5′-ACAGGGATGGCAGTTAGGAC-3′, lower 5′-CAGAAGGGGAAGTATGGGG-3′; Runx2: upper 5′-AACTTCCTGTGCTCCGTG-3′, lower 5′-AACTCTTGCCTCGTCCG-3′; Sox9: upper 5′-AGGAAGTCGGTGAAGAACG-3′, lower 5′-TGAGATTGCCCAGAGTGC-3′; and Indian hedgehog (Ihh): upper 5′-ACAATCCCGACATCATCTTC-3′, lower 5′-CATCTTCATCCCAGCCTTC-3′. Human primers were as follows: Hes1: upper 5′-AGGCGGCTAAGGTGTTTG-3′, lower 5′-GAAGAGAGGTGGGTTGGG-3′; RPL13A: upper 5′-CCTGGAGGAGAAGAGGAAAGAGA-3′, lower 5′-GAGGACCTCTGTGTATTTGTCAA-3′.
siRNA transient transfection
Four Hes1-specific siRNA oligoribonucleotides pairs were designed by Xeragon (Qiagen). Lyophilised oligoribonucleotides were resuspended in RNase-free annealing buffer (100 mM potassium acetate, 30 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) (pH 7.4), 2 mM magnesium acetate) to obtain a 300 mg/litre solution, heated for 1 min at 90°C, incubated for 1 h at 37°C to allow duplex formation and disrupt larger aggregates, and stored at −20°C. Cells were transfected at 70% to 80% confluence by use of HiPerFect Transfection reagent (Qiagen) according to the manufacturer's instructions. Cells were then treated or not for 20 h with IL1β, and RNA was extracted and processed for real-time qPCR. The control non-silencing siRNA used in this study was 5′-AATTCTCCGAACGTGTCACGT-3′. A siRNA corresponding to positions (+963) to (+983) relative to the transcription start site of the mouse Hes1 cDNA (5′-CGGCCCGGTCATCCCGGTCTA-3′) was identified as the most efficient and used in all ulterior experiments.
Results are reported as means±SD. Experimental analysis was performed in triplicate with a minimum of three independent experiments. Data analysis involved analysis of variance (ANOVA) followed by the Tukey test. A p<0.05 was considered statistically significant.
IL1β targets Hes1 in chondrocytes in vitro
We analysed Notch1, Hes1 and Hey1 expression in IMACs in basal conditions and after IL1β (2.5 ng/ml) or fetal calf serum (5%) treatment. Western blot analysis revealed a Notch1 species at 63 kDa in accordance with previous estimations of the cleaved NICD molecular weight (figure 1A, arrow). IL1β treatment did not modify the protein pattern. Immunocytochemical analysis showed Notch1 segregation in the nucleus. IL1β treatment had no effect on Notch1 subcellular localisation (figure 1A). Conversely, the 38-kDa Hes1 protein, identified by western blot (figure 1B), was significantly increased in cells treated with IL1β over control cells. Hes1 was mainly detected in the cytoplasm of untreated chondrocytes (75±8%) but exclusively in the nucleus of IL1β-treated cells. The anti-Hey H-120 antibody could not detect a clear labelling in these experiments (data not shown). A control immunostaining using preimmune rabbit IgG yielded no detectable labelling (data not shown).
IMACs cultured in the presence of 5% FCS displayed an oscillatory pattern of expression of Hes1 mRNA as shown by qPCR analysis of a 20 min interval timecourse experiment. Hey1 mRNA was barely detectable in basal conditions (fivefold to sixfold lower than Hes1 mRNA) and did not significantly vary during the 6 h experiment (figure 2A). In sharp contrast, IL1β stimulated Hes1 mRNA level as a single peak of expression (2.5-fold over control) at 1 h. Hes1 mRNA expression then rapidly decreased after 90 min down to 50% and 75% lower than control after 6 h and 20 h incubation, respectively (figure 2B,C). IL1β had no significant effect on Hey1 mRNA expression. IL1β-induced Hes1 mRNA expression was not modified in chondrocytes pretreated with a γ-secretase inhibitor (DAPT), suggesting independency from the canonical Notch signalling pathway (figure 2C).
A similar study was performed on human chondrocytes in primary culture with or without IL1β (figure 3). Hes1 mRNA was increased twofold after 1 h in the presence of IL1β as compared to basal conditions (upper panel). Immunocytochemistry (lower panel) showed that the Hes1 protein was cytoplasmic or cytonuclear in untreated chondrocytes (left micrograph). Upon IL1β addition, Hes1 became essentially nuclear between 60 and 90 min. At 120 min, Hes1 labelling was again mainly cytoplasmic (data not shown).
To ascertain the nature of the IL1β effect on Hes1 mRNA expression, cells were treated with the transcriptional inhibitor 5,6-dichloro-1-β-D-ribofuranosylbenzimidazole (DRB; 5 µg/ml) and exposed to IL1β for 1 h (figure 4A). DRB had no effect on Hes1 mRNA level in untreated cells, but abolished IL1β-induced Hes1 expression. An 18 h treatment of chondrocytes with the protein synthesis inhibitor cycloheximide with or without IL1β strongly enhanced basal Hes1 mRNA expression but not its induction by IL1β (figure 4B). Finally, the proteasome inhibitor MG132 decreased Hes1 basal expression and abolished the IL1β-induced Hes1 peak expression (figure 4C).
NFκB mediates IL1β effect on Hes1 expression
We used the NFκB inhibitor BAY 1170-82 and inhibitors for ERK1/2 (PD98059), p38MEK (SB203580) and JNK II to investigate the pathways for IL1β-induced Hes1 expression. BAY 1170-82 decreased Hes1 basal expression by 50% and abolished IL1β-induced Hes1 mRNA expression (figure 5A). In contrast, none of the three kinase inhibitors modulated Hes1 expression, whether basal or induced (figure 5B–D).
A specific siRNA inhibits Hes1 induction by IL1β in chondrocytes
We inhibited Hes1 expression using an anti-Hes1 siRNA versus a non-specific siRNA to identify Hes1-dependent IL1β effects in chondrocytes, stimulated or not with IL1β for 20 h. A preliminary screening of several anti-Hes1 siRNAs in chondrocytes revealed two efficient ones, yielding, respectively, 70% and 50% reductions in Hes1 mRNA levels. Subsequent experiments were performed with the most efficient.
Silencing Hes1 expression did not affect chondrogenic markers (type II collagen, aggrecan) in naive or IL1β treated cells (figure 6A,B). The Sox9 regulator was equally unmodified (data not shown). In sharp contrast, Hes1 inhibition decreased IL1β induced MMP13 and ADAMTS5 expression by 65% and 51%, respectively (figure 6C,D). Finally, Hes1 silencing did not influence inflammation or degradative markers levels (iNOS, c-Jun, MMP1, MMP3) or the expression of regulatory proteins controlling proliferation (p21, cyclin D1) or hypertrophic transition (Ihh, Runx2) in naive or IL1β treated cells (data not shown). The inhibition of Hes1 protein expression by the anti-Hes1 siRNA is shown as a control in figure 6E.
Recent studies on the Notch network in cartilage have opened an avenue of research in the mechanisms of chondrocyte proliferation and differentiation. In a previous collaborative work, an in vitro model of immature mouse chondrocytes, involving repetitive passages in the presence of serum, was used to investigate the influence of Notch signalling during chondrocyte dedifferentiation.3 Notch expression did not change during successive passages, while Notch ligands and Hes1 expression progressively increased in parallel with a decrease in phenotypic markers and increase in MMP13 expression. MMP13 induction was Notch dependent.3
The perturbations of chondrocyte phenotype and differentiation processes observed in OA suggest that cartilage cells do attempt but fail to repair OA lesions. In the present study, we investigated the potentiality that OA-associated inflammation may contribute to this failure through anomalous modulation of Notch network members by IL1β. Notch1 appeared to be constitutively activated in IMACs: Notch1 labelling was restricted to the nucleus, and a 63-kDa protein resembling the truncated NICD was the major Notch1 molecular species. Such a constitutive activation of the Notch pathway has been previously described.22 23
Our observation that Hes1 expression in serum-treated chondrocytes is oscillatory concurs with other models of serum-dependent Hes1 oscillatory expression controlled by negative feedback.8 24 Thus, the Hes1 protein may also repress the expression of its own gene in chondrocytes. In sharp contrast, IL1β elicited a transient induction followed by a rapid decrease of Hes1 and suppressed the oscillatory pattern. The negative feedback hypothesis was further supported in IMACs by our observation that the proteasome inhibitor MG132 decreased basal and IL1β-induced Hes1 mRNA expression (figure 4C).
We also show that Hey1 mRNA expression is at the limit of detection in IMACs (five times lower than Hes1 transcripts) and is not modified by 5% FCS or IL1β. The discrepancy between expression and modulation of Hes1 and Hey1 suggests their heterodimerisation in differentiated chondrocytes is unlikely. Accordingly, they should not be able to compete against Sox9 on the Col2a1 enhancer as shown in human mesenchymal stem cells, where they repress chondrogenesis in a Notch-dependent pathway.9 In further support of this theory, the transient IL1β-induced upregulation of Hes1 expression in IMACs was insensitive to the γ-secretase inhibitor DAPT, suggesting the absence of de novo Notch activation. Notch-independent Hes1 induction was previously described for mitogen-activated protein kinases (MAPKs),25 JNK,26 ERK1/2 and growth factors27 or the Wnt/β-catenin pathway.28 IL1β effects required the NFκB pathway in our experiments, concurring with previous reports to suggest a complex array of relations between the Notch network and NFκB proteins.29 30
What is the part played by Hes1 in regulating IL1β target genes? Hes1 silencing by a specific siRNA did not modify basal or IL1β-repressed type II collagen, aggrecan or Sox9 expression suggesting the phenotypic alterations in IL1β-treated chondrocytes do not involve Hes1. By contrast, Hes1 silencing decreased the IL1β-induced expression of MMP13 and ADAMTS5. Hes1 exerts its function through binding to N boxes (CACNAG) or more specific ESE boxes in target genes.7 As shown in table 1, mouse, rat MMP13 and human ADAMTS5 genes contain such sequences.
Where and when might this regulation occur in vivo? MMP13 and ADAMTS5 are involved in cartilage hypertrophic differentiation as well as ECM remodelling during normal growth. MMP13 and ADAMTS5 have been described as under the control of the Wnt/β-catenin pathway in mature chondrocytes during endochondral ossification32 and in articular chondrocytes.33 MMP13 and ADAMTS5 are increased during OA.34 Knowing the stimulatory effects of Wnt3A on oscillatory genes in mesenchymal stem cells, including Hes1,35 the model developed by Rodríguez-González et al, becomes very attractive.36 According to these authors, Notch and Wnt networks are organised in a loop through Axin2 and Hes1 genes to control tissular development. It is tempting to speculate that chondrocytes cannot repair the OA matrix because inflammatory cytokines impair their molecular clock.
In conclusion, Hes1 is a novel IL1β target gene in chondrocytes, influencing genes linked to ECM remodelling and hypertrophic maturation of chondrocytes, independently of the Notch signalling pathway. This may pertain to the abortive repair effects observed in OA. Future work will investigate a possible crosstalk between Hes proteins and Wnt cascade members. Hes1 expression should be considered a new parameter of matrix degradation in cartilage and therefore a potential therapeutic target.
We thank Professor C Glorion (Necker Hospital, Paris, France) for providing human cartilage surgical samples. We also thank L Heraty for copyediting work.
AF and ZNH contributed equally to this work.
Funding This work was supported by INSERM (Programme National de Recherche sur les Maladies Osteo-Articulaires (PRO-A)) and grants from ‘Association Rhumatisme et travail’, ‘Fonds d'Etude et de Recherche du Corps Medical’ and ‘Societe Française de Rhumatologie’.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
Patient consent Obtained.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.