EuropeanSJ Gilbert Cells et al .and Materials Vol. 23 2012 (pages 41-57) DOI: PKR 10.22203/eCM.v023a04 mediates OSM and IL-1 signalling inISSN chondrocytes 1473-2262

PROTEIN KINASE R PLAYS A PIVOTAL ROLE IN ONCOSTATIN M AND INTERLEUKIN-1 SIGNALLING IN BOVINE ARTICULAR CHONDROCYTES S.J. Gilbert*, E.J. Blain, A. Al-Sabah, Y. Zhang, V.C. Duance and D.J. Mason

Arthritis Research UK Biomechanics and Bioengineering Centre, Pathophysiology and Repair Division, School of Biosciences, Cardiff University, Cardiff, CF10 3AX, UK

Abstract Abbreviations

This study investigated whether treatment of articular MMP = matrix ; ADAMTS = a cartilage chondrocytes with a combination of oncostatin M and metalloproteinase with (OSM) and interleukin-1 (IL-1) could induce a degradative motif; OSM = oncostatin M; IL-1 = interleukin-1; PKR phenotype that was mediated through the protein kinase R = protein kinase R; PKRi = PKR inhibitor; PKRi neg = (PKR) signalling pathway. High-density monolayer cultures PKR inactive inhibitor; TNF = tumour necrosis factor of full depth, bovine chondrocytes were treated with a ; eIF2 = eukaryotic initiation factor 2 alpha; IB = combination of OSM and IL-1 (OSM+IL-1) for 7 days. To the inhibitor of kappa B; NFB = nuclear factor kappa inhibit the activation of PKR, a pharmacological inhibitor of B; DMEM = Dulbecco’s Modifi ed Eagle’s Medium; ITS PKR was added to duplicate cultures. Pro- and active matrix = insulin-transferrin-sodium selenite; LDH = lactate metalloproteinase-9 (MMP9) and MMP9 mRNA were dehydrogenase; SDS = sodium dodecyl sulphate; TBST signifi cantly upregulated by OSM+IL-1 through a PKR = tris buffered saline with tween 20; GLM-ANOVA = dependent mechanism. ADAMTS4 and ADAMTS5 mRNA general linear model analysis of variance; CT = delta were also upregulated by OSM+IL-1. The upregulation of delta threshold cycle. ADAMTS4 and ADAMTS5 were, in part, mediated through PKR. OSM+IL-1 resulted in a loss of type II collagen, Introduction which could not be rescued by PKR inhibition. OSM+IL-1 reduced the expression of COL2A1 (type II collagen), Cartilage degradation occurs as a result of an imbalance COL9A1 (type IX collagen), COL11A1 (type XI collagen), between proteinases and their and ACAN () mRNAs. Expression of type II and inhibitors, in particular matrix XI collagen and aggrecan was reduced further when PKR (MMPs) and (; a disintegrin was inhibited. OSM+IL-1 resulted in an 11-fold increase in and metalloproteinase with thrombospondin motif) TNFα mRNA which was, in part, mediated through the PKR and the tissue inhibitors of metalloproteinases. These pathway. This study demonstrates, for the fi rst time, that a are induced in arthritic disease by a number number of catabolic and pro-infl ammatory effects known of proinfl ammatory cytokines such as tumour necrosis to be important in human arthritis and induced by OSM and factor  (TNF), interleukin-1 (IL-1), and oncostatin M IL-1, are mediated by the PKR signalling pathway. (OSM). OSM is a member of the interleukin-6 superfamily of cytokines and has been implicated in infl ammatory Keywords: Articular cartilage, chondrocytes, protein kinase joint disease (Cawston et al., 1998; Langdon et al., R, oncostatin M, interleukin-1, , 2000). Several studies have shown that OSM signals , tumour necrosis factor alpha, type II collagen. synergistically with IL-1 in cartilage, inducing collagen (Cawston et al., 1995) and aggrecan (Cawston et al., 1998; Durigova et al., 2011) degradation via up-regulation of MMPs (Blain et al., 2010; Cawston et al., 1998; Hui et al., 2003; Milner et al., 2006) and ADAMTSs (Hui et al., 2005; El Mabrouk et al., 2007; Durigova et al., 2008; Durigova et al., 2011), respectively. In particular, OSM potently induces MMP13 in human and bovine articular cartilage (Fearon et al., 2006; El Mabrouk et al., 2007; Blain et al., 2010) upregulation of which plays a crucial role in the pathology of degenerative joint disease ((Little et al., 2009) and reviewed in (Takaishi et al., 2008; van den Berg, 2011)). *Address for correspondence: Previously, we have shown that the TNF induced S.J. Gilbert, upregulation of MMP9 expression and activation in School of Biosciences, Cardiff University articular cartilage occurs through a novel mechanism Museum Avenue, Cardiff, CF10 3AX involving the interferon-induced, double-stranded RNA- Wales, UK activated protein kinase, PKR (Gilbert et al., 2004). PKR Telephone Number: +44 (0) 29 20874287 is expressed at constitutively low levels in a number FAX: +44 (0) 29 20874594 of cell types, including chondrocytes, and mediates a E-mail: [email protected] stress-induced signalling pathway that is activated in

www.ecmjournal.org 41 SJ Gilbert et al. PKR mediates OSM and IL-1 signalling in chondrocytes early in vivo (reviewed in (Gilbert et al., phenotype and prevent serum withdrawal activation of 2006b)). PKR expression and activation is upregulated signalling pathways (Patel et al., 2000). by proinflammatory signals including TNF (Gilbert et al., 2002) and IL-1 (Tam et al., 2007) resulting in Biochemical treatments degradation (Gilbert et al., 2004; Tam et al., After 48 h, the media was replaced and chondrocytes 2007), phosphorylation of the eukaryotic initiation factor cultured for 24 h prior to the direct addition of treatments 2 alpha (eIF2) (Gilbert et al., 2002) and down-regulation to the existing media to prevent activation of the MAPK of protein synthesis (reviewed in Gilbert et al., 2006b). signalling pathway by changing the media (Gilbert et This inhibition of protein synthesis has been shown to be al., 2008). Initial experiments (confocal microscopy and important in controlling cell growth (reviewed in Clemens protein concentration) examined whether PKR could be and Elia, 1997) and in increasing the cells susceptibility activated in vitro by IL-1 (5 ng/mL) and OSM (10 ng/mL) to apoptosis (Lee and Esteban, 1994). Interestingly, work either alone or in combination. From these experiments, the by Osman et al. has uncovered a role for PKR in the combination of IL-1 (5 ng/mL)and OSM (10 ng/mL) was regulation of TNF (Osman et al., 1999). A cis-acting chosen as the treatment regime for subsequent experiments element within the 3’UTR of the TNFα mRNA binds to since it had the most potent effect. Chondrocytes were and activates PKR in response to external stimuli, which in treated with cytokines for 7 days. Unstimulated, control turn increases TNFα pre-mRNA splicing effi ciency up to chondrocytes were treated with media alone. To inhibit 20-fold when the amount of PKR increases (Osman et al., the activation of PKR, a potent selective small molecule 1999). Therefore, since the cis-acting element is capable inhibitor (PKRi; 1 μM; Merck Chemicals, Nottingham, of acting as a sensor, responding to the levels of PKR in UK) which binds to the ATP-binding site was added to the cell, a positive feedback loop is established for TNF replicate cultures 30 min prior to the addition of treatments production within the cell (reviewed in (Kaempfer, 2003; (Jammi et al., 2003; Tam et al., 2007). In addition, an Gilbert et al., 2006b)). Thus, splicing of TNFα mRNA inactive version of the PKR inhibitor was added to replicate precursors proceeds effi ciently in response to external cultures as a negative control (PKRi neg; 1 M; Merck stimuli in early infl ammation when PKR is available and Chemicals). All treatments and media were warmed to is subject to regulation by PKR (Chiu and Yang, 2007). 37 C before being added gently to cells. The media Since IL-1 activates PKR in chondrocytes and IL-1 and remained in place for the 7-day culture and no additional OSM act synergistically to induce cartilage catabolism, treatments were added during this period. At the end of the aim of this study was to delineate the role of PKR the culture period, cells were lysed with 100 μL ice-cold in mediating the cellular responses induced by the Triton X-100 (0.9 % v/v) containing inhibitors proinfl ammatory cytokine combination of OSM and IL-1 (Protease Inhibitor Cocktail Set I; Merck Chemicals) and in bovine articular cartilage chondrocytes. phosphatase inhibitors (Phosphatase inhibitor cocktail Set III; Merck Chemicals). Cell extracts were stored at -80 C Materials and Methods and media at -20 C.

Materials Total protein assay All chemicals were obtained from Sigma (Poole, UK) The total protein concentration of the cell extracts at unless otherwise stated and were of analytical grade. the end of the 7-day culture period was analysed using Recombinant human IL-1α and OSM were purchased from the Pierce bicinchoninic acid (BCA) assay according to Peprotech EC (London, UK). Culture medium consisted manufacturer’s instructions (Thermo Fisher Scientifi c, of Dulbecco’s Modifi ed Eagle’s Medium (1:1 mixture of Cramlington, UK). DMEM-Glutamax-I™ and Ham’s F12 media) containing 10 mM HEPES pH 7.4, 100 U/mL penicillin, 100 g/ Cell number and toxicity mL streptomycin, 50 μg/mL ascorbate-2-phosphate and Cell number and cell death over the culture period were supplemented with 1x Insulin-transferrin-sodium selenite determined using the CytoTox 96® assay according to (ITS) (Invitrogen, Paisley, UK). manufacturer’s instructions (Promega, Southampton, UK), which quantitatively measures lactate dehydrogenase Primary chondrocyte culture (LDH) present in the culture media that has been released Full depth articular cartilage was taken from the upon death of cells during the culture period. This assay metacarpophalangeal joints (from ≥ 4 legs) of 7 day-old can also be used to measure indirectly the LDH activity bovine calves within 6 h of slaughter using a scalpel. present in the cytoplasm of cells that are intact at the end Chondrocytes were isolated by enzymatic digestion (0.1 % of the culture period. Cell quantifi cation, therefore, occurs (w/v) (Roche Applied Science, Burgess Hill, UK) following lysis of the cells by the addition of extract buffer. in media containing 5 % foetal calf serum (FCS) for 30 min The number of cells present is directly proportional to followed by overnight incubation in 0.04 % (w/v) type II the absorbance value, which represents LDH activity. (Invitrogen) in media containing 5 % FCS at Absorbance at 490 nm was recorded for cell number and

37 C in a humidifi ed atmosphere of 5 % CO2, 95 % air. cell death after 10 and 30 min, respectively. Following isolation, unless stated otherwise, chondrocytes were seeded at 1.0x106 cells per well of a 24-well plate and Gelatin substrate zymography cultured for 48 h at 37 C in 1 mL serum-free media. Media MMP9 activity was detected in media samples by was supplemented with 1x ITS to maintain chondrocyte gelatin substrate zymography as described previously

www.ecmjournal.org 42 SJ Gilbert et al. PKR mediates OSM and IL-1 signalling in chondrocytes

(Gilbert et al., 2004). Briefl y, gelatin (porcine skin; VWR at 4 ºC with a rabbit polyclonal phosphospecifi c primary International, Lutterworth, UK) was incorporated into 7.5 antibody to active PKR [pT446] (Abcam, Cambridge, UK) % (w/v) sodium dodecyl sulphate (SDS)-polyacrylamide diluted 1:100, cells were washed before incubating for gels at a fi nal substrate concentration of 1.0 mg/mL. Media 1 h with FITC-conjugated anti-rabbit antibody diluted samples (diluted 1:1 with 2x sample buffer), a recombinant 1:160. Cells were washed and blocked as before prior to MMP9 standard (Merck Chemicals), and protein molecular incubation with a monoclonal antibody to Golgi matrix weight markers (Mr 14-200 kDa; BioRad Laboratories, protein (anti-GM130; BD Transduction Laboratories, Hemel Hempstead, UK) were loaded onto the gels and Oxford, UK) diluted 1:500. Cells were washed and a goat resolved by electrophoresis. Following electrophoresis, anti-mouse Alexa 594 conjugated secondary antibody the gels were washed, with agitation, in 2.5 % (v/v) Triton (1:400 dilution; Molecular probes, Invitrogen) applied to X-100 for 1 h with at least 3 changes of solution and the cells for 1 h. Finally, after washing, cells were mounted subsequently incubated for 16-20 h at 37 ˚C in proteolysis in VECTASHIELD® Mounting Medium containing DAPI buffer. Gels were stained with Coomassie® Brilliant Blue (1.5 μg/mL) to counterstain DNA (Vector Laboratories, R 250 (2 g/L) in distilled water, methanol and acetic acid Peterborough, UK). Representative cells from multiple (4.5:4.5:1) for at least 1 h, and destained in methanol, acetic fi elds of view were scanned with a Leica TCS SP2 confocal acid, water (1 : 0.75 : 8.25) until the zones of proteolysis microscope (Leica, Heidelberg, Germany) using a 63x had cleared. Molecular weight markers and standards oil immersion objective lens (x7.5 zoom) as described facilitated identification of the enzymes and allowed previously (Blain et al., 2006). To eliminate the possibility comparisons between gels. Relative quantities of MMPs of spectral bleed-through between fl uorescent probes, were measured by scanning densitometry (UMAX magic representative regions were scanned using appropriate scan; Umax Colour Scanner, Willich, Germany), using excitation and emission settings for sequential recordings image analysis software (NIH Image; National Institutes of DAPI (ex max 358 nm; em max: 461 nm), FITC of Health, Bethesda, MD, USA; online at: http://rsb.info. (excitation max: 494 nm; emission max: 518 nm) and nih.gov/nih-image/) and expressed as absorbance units per Alexa 594 (excitation max: 594 nm; emission max: 617 cell (Vaughan-Thomas et al., 2000). nm). Maximum intensity, 3D projections were obtained by processing stacks of optical sections through the full Western blot analysis depth of chondrocytes at a spacing of 0.5 μm. Since PKR is Protein extracts, from an equivalent cell number, were constitutively activated in sub-populations of chondrocytes reduced (2.5 % v/v -mercaptoethanol) and resolved on in vitro (Gilbert et al., 2002; Blalock et al., 2009), untreated 10 % (w/v) SDS-polyacrylamide gels. For each analysis, cultures served as positive controls. Negative controls control and treated samples were run on the same gel where the primary antibody was omitted were devoid of for direct comparison. Protein was transferred to PVDF fl uorescent signal (data not shown). membranes (Immobilon; Millipore, Watford, UK) and blocked for 1 h at room temperature in blocking buffer Quantitative PCR analysis of expression (Tris Buffered Saline with Tween 20 (TBST); 0.15 M Immediately after isolation, chondrocytes were seeded NaCl, 0.05 M Tris/HCl pH 8.0, 3 % (w/v) skimmed milk at 0.5 x 106 cells per well of a 48 well plate and left for powder, 0.05 % (v/v) Tween 20). After washing (TBST), 48 h prior to stimulation as described above. At the end membranes were incubated for 2 h at room temperature of the experiment, media were replaced with 0.5 mL of with a monoclonal antibody to type II collagen (AVT6E3 TRIZOL® (Invitrogen) and total RNA extracted according (Vaughan-Thomas et al., 2008)) diluted in TBST. to the manufacturer’s instructions with the following Membranes were washed (TBST) and then incubated exceptions. After the addition of chloroform, the entire for 1 h at room temperature in anti-mouse horseradish RNA extraction mix was transferred to a tube containing peroxidase conjugated IgG. Bands were detected using Heavy Phase-lock Gel™ (Eppendorf®; VWR International) enhanced chemiluminescence reagents and Hyperfi lm- and centrifuged at 13,000 g for 2 min at 4 C. The upper ECL (GE Healthcare Life Sciences, Little Chalfont, UK) aqueous layer was removed to a new Eppendorf tube, an and blots scanned (Umax Magic scan) and analysed by equal volume of isopropanol added and the RNA left to densitometry (NIH Image) (Vaughan-Thomas et al., 2000). precipitate overnight at -20 C. At the end of the extraction protocol, the RNA was DNase treated to remove genomic Immunocytochemistry DNA (Ambion; Applied Biosystems, Warrington, UK) Immediately after isolation, chondrocytes were seeded onto and re-suspended in 50 L RNase-free water. cDNA was 8 well glass chamber slides (0.5 x 106 cells/well) and left generated in a 20 L reaction from 11 L RNA using for 48 h prior to stimulation as described above. Following 250 ng random hexamers (0.5 mg/mL; Promega) and treatment, cells were fi xed in 2 % (w/v) paraformaldehyde Superscript III reverse transcriptase (200 units; Invitrogen). and permeabilised in 0.2 % (v/v) Triton-X100. Each was measured by SYBR green subsequent step was performed at room temperature unless quantitative PCR (qPCR) using the MX3000P™ stated otherwise and between each incubation step, sections qPCR system according to manufacturer’s instructions were washed 3x 5min in 0.01 M phosphate buffered saline (Stratagene®; Agilent Technologies UK, Stockport, UK) (PBS, pH 7.4) containing 0.001 % Tween 20 (wash buffer). with 200 nM forward and reverse primers (Table 1) All antibodies were diluted in wash buffer. Cells were annealing at 60 C (except MMP13 and MMP9 which washed before blocking in 2 % (v/v) normal goat serum annealed at 58 C and 62 C, respectively). Fold changes (Dako UK, Ely, UK) for 1 h. After overnight incubation relative to untreated cells were calculated using the

www.ecmjournal.org 43 SJ Gilbert et al. PKR mediates OSM and IL-1 signalling in chondrocytes

Table 1. Quantitative PCR primers

Gene Strand Sequence Reference/Accession No. RN18S1 Forward 5’-GCAATTATTCCCCATGAACG-3’ (Frye et al. 2005) (18s ribosomal 1) Reverse 5’- GGCCTCACTAAACCATCCAA-3’ COL2A1 Forward 5’-AACGGTGGCTTCCACTTC-3’ (Darling and Athanasiou 2005) (Type II collagen) Reverse 5’-GCAGGAAGGTCATCTGGA-3’ COL9A1 Forward 5’-CGCCCAAGGTTCCCTGTT-3’ AF419343 (Type IX collagen) Reverse 5’-ACTTGTAAGCCACCTGTAAGGCA-3’ COL11A1 Forward 5’-AAAACCTGCCCCAGAAGACTATC-3’ AY275644 (Type XI collagen) Reverse 5’-AAAATCCTTGTCCCAAAAACCATA-3’ ACAN Forward 5’-GCTACCCTGACCCTTCAT-3’ (Darling and Athanasiou 2005) (Aggrecan) Reverse 5’-AAGCTTTCTGGGATGTCCAC-3’ Forward 5’-AGTACCCGCACCTGCACAAC-3’ SOX9 AF278703 Reverse 5’-TCGTTCAGCAGTCTCCAGAGC-3’ Forward 5’-CATCCTGTCTGCCATCAAGA-3’ TNFα NM173966 Reverse 5’-GGCGATGATCCCAAAGTAGA-3’ Forward 5’-TAGCACGCACGACATCTTTC-3’ MMP9 NM174744 Reverse 5’-GAAGGTCACGTAGCCCACAT-3’ Forward 5’-CCCTTGATGCCATAACCAGT-3’ MMP13 (Blain et al. 2010) Reverse 5’-GCCCAAAATTTTCTGCCTCT-3’ Forward 5’-CTCCATGACAACTCGAAGCA-3’ ADAMTS4 (Blain et al. 2010) Reverse 5’-CTAGGAGACAGTGCCCGAAG-3’ Forward 5’-CTCCCATGACGATTCCAAGT-3’ ADAMTS5 (Blain et al. 2010) Reverse 5’-TACCGTGACCATCATCCAGA-3’

∆∆CT method with RN18S1 (18S ribosomal) RNA as chondrocytes diffuse staining for phosphorylated PKR was the reference gene (Frye et al., 2005). All primers were observed throughout the cytoplasm (Fig. 1a). In addition, purchased from MWG and validated using a standard curve a perinuclear pool of active PKR was also observed in of fi ve serial dilutions so that all primer effi ciencies were some cells (Fig. 1b), which co-localised with staining for between 90-110 % (Taylor et al., 2010). the Golgi marker, GM130 (Fig. 1c). The localisation of active PKR in the cytoplasm was similar in cells that had Data analysis been cultured in the presence of IL-1 (Fig. 1d-f), OSM (Fig. Data are presented as mean SEM and tested for normality 1g-i) or a combination of OSM+IL-1 (Fig. 1j-l). and equal variances prior to analysis (Minitab). Data were analysed by general linear model analysis of variance Il-1 and OSM, in combination, potently reduces (GLM ANOVA) followed by Tukey’s post hoc test. cellular protein concentration through a PKR Comparisons were made to untreated, control cells unless dependent mechanism otherwise stated and differences considered signifi cant To determine whether each cytokine treatment activated at p ≤ 0.05. Each experiment consisted of 4-6 replicates PKR and resulted in reduced translation, the protein per treatment where one replicate was a single well of a concentration of cell extracts was determined in the tissue culture plate. To ensure a heterogeneous population presence or absence of the small molecule PKR inhibitor of cells for each experiment, cells were extracted from a after 7 days (Fig. 2). Only the combined treatment with pool of cartilage taken from a minimum of 4 bovine legs. IL-1+OSM signifi cantly reduced protein concentration Experiments were repeated once to ensure reproducibility (p ≤ 0.01 vs. untreated control cells). Inhibition of PKR of results and representative data from 1 experiment shown. reversed the effect of OSM+IL-1, returning the protein Since the qPCR data for ADAMTS5 mRNA was more concentration to that of untreated, control cells (p ≤ 0.01). variable than other data sets, data from the 2 experiments Subsequent experiments used only the combination of were combined for analysis. OSM+IL-1 as the treatment regime since was the only signifi cant PKR mediated effect.

Results Inhibition of PKR enhances the OSM and IL-1 induced increase in chondrocyte proliferation Active PKR is localised within the cytoplasm of After 7-days in culture, OSM+IL-1 treatment resulted in articular chondrocytes a trend towards increased chondrocyte proliferation (Fig. After 7-days of culture, chondrocytes were fi xed and labelled 3a; p = 0.07). To investigate whether this increase occurred for phosphorylated (active) PKR (Fig. 1). In unstimulated through activation of PKR, duplicate cultures were treated

www.ecmjournal.org 44 SJ Gilbert et al. PKR mediates OSM and IL-1 signalling in chondrocytes

Fig. 1. Active PKR is localised within the cytoplasm of articular chondrocytes. The subcellular localisation of phosphorylated PKR (FITC green) and Golgi matrix protein (AlexaFluor red) was determined by immunocytochemistry. Representative cells from multiple fi elds of view were imaged by confocal microscopy with appropriate settings for FITC (green), AlexaFluor 594 (red) and DAPI nuclear counterstain (blue). Maximum intensity 3D reconstructions are shown. Staining for phosphorylated PKR was observed throughout the cytoplasm of untreated, control cells (a) and within peri-nuclear pools (b). Areas of green PKR and red Golgi co-localisation can be observed as yellow staining (c). Treatment with IL-1 (d-f), OSM (g-i) or a combination of OSM+IL-1 (j-l) did not alter the localisation of active PKR.

with the small molecule PKR inhibitor. Inhibition of PKR Increased chondrocyte death induced by combined resulted in an increase in basal proliferation levels (p < treatment of OSM and IL-1 is not due to activation 0.05). Pre-treatment with this inhibitor prior to the addition of PKR of OSM+IL-1 greatly enhanced the cytokine-induced Media was assessed for the level of LDH release as an increase in proliferation (p < 0.001 vs. untreated controls; indicator of cell death over the 7-day culture period (Fig. p < 0.01 vs. OSM/IL-1 alone). 3b). Treatment with OSM+IL-1 for 7-days resulted in a signifi cant increase in chondrocyte death (p < 0.001 vs.

www.ecmjournal.org 45 SJ Gilbert et al. PKR mediates OSM and IL-1 signalling in chondrocytes

Fig. 2. The combination of OSM+IL-1 potently reduces cellular protein concentration through a PKR dependent mechanism. At the end of the 7-day culture period, cells were lysed and the protein concentration of the cell extract determined by BCA assay. Data is presented normalised to cell number equivalent assessed by LDH assay (mean ± SEM; n = 6 per treatment) and analysed by GLM ANOVA and Tukey’s post hoc test: **p < 0.01 vs. untreated controls unless shown otherwise.

untreated controls). Inhibition of PKR had no signifi cant expression 8.9-fold over untreated controls (p < 0.01 effect on basal or cytokine-induced cell death. following ranking of data). The PKR inhibitor reduced basal MMP9 mRNA expression by 24-fold (p < 0.001 OSM and IL-1 increase the expression and activation vs. untreated controls following ranking of data) and of MMP9 in primary articular chondrocytes through signifi cantly reduced the OSM+IL-1 treated group to activation of PKR 8.9-fold below the basal expression level (p < 0.001 vs. The cytokine combination of OSM and IL-1 has untreated controls; p < 0.001 vs. OSM+IL-1 following previously been shown to up-regulate MMP9 expression ranking of data). The PKR inhibitor negative control (PKRi and activation in nasal cartilage (Milner et al., 2006) neg) had no effect on cytokine-induced MMP9 mRNA and articular cartilage (Blain et al., 2010). We therefore levels. assessed whether these cytokines could increase MMP9 expression and/or activation in articular chondrocytes OSM and IL-1 increase the expression of MMP13 via a mechanism involving PKR. Media was analysed mRNA in articular chondrocytes through a PKR for MMP9 by gelatin substrate zymography (Fig. 4a,b). independent mechanism Low levels of proMMP9 were detected in the media from Several studies have shown that treatment of chondrocytes untreated, control chondrocytes after 7 days in culture. In with OSM and IL-1 results in the up-regulation of MMP13 contrast, OSM+IL-1 resulted in a 7.8-fold up-regulation mRNA expression (Koshy et al., 2002; Barksby et al., of proMMP9 (p < 0.001 vs. untreated controls following 2006; Milner et al., 2006; Blain et al., 2010). To determine log transformation of data). Pre-treatment with the PKR whether this increase was mediated through the PKR inhibitor completely blocked the basal expression of pathway, the expression of MMP13 mRNA was assessed by MMP9 (p < 0.001 vs. untreated controls following log qPCR following OSM+IL-1 treatment and PKR inhibition transformation of data). Addition of the inhibitor prior to (Fig. 5). OSM+IL-1 treatment resulted in a 73-fold increase treatment with OSM+IL-1 resulted in a 2.5-fold reduction in MMP13 mRNA over untreated controls (p < 0.001 in proMMP9 expression (p = 0.05 vs. OSM+IL-1 alone following log transformation of data). Inhibition of PKR following log transformation of data) although levels had no signifi cant effect on basal or cytokine-induced remained 3-fold higher than those detected in control MMP13 mRNA expression. The PKR inhibitor negative cultures (p = 0.01 following log transformation of data). No control (PKRi neg) had no effect. active MMP9 was detected in untreated, control cultures. In contrast, high levels of MMP9 activation were observed in The OSM and IL-1 induced increase in the cultures following OSM+IL-1 treatment. This OSM+IL-1- expression of ADAMTS4 and ADAMTS5 mRNAs in induced activation of MMP9 was completely abolished by primary articular chondrocytes is mediated, through addition of the PKR inhibitor. The PKR inhibitor negative activation of PKR control (PKRi neg) had no effect on cytokine-induced Since OSM and IL-1 treatment of chondrocytes increases MMP9 levels (data not shown). the expression of ADAMTS4 and 5 (Koshy et al., 2002; MMP9 mRNA expression was also analysed by qPCR Young et al., 2005; Milner et al., 2006; Durigova et al., (Fig. 4c). OSM+IL-1 treatment increased MMP9 mRNA 2008; Durigova et al., 2011), we determined whether this

www.ecmjournal.org 46 SJ Gilbert et al. PKR mediates OSM and IL-1 signalling in chondrocytes

Fig. 3. Inhibition of PKR enhances the effect of OSM and IL-1 on chondrocyte proliferation but not cell death. (a) At the end of the culture period, cells were lysed and the relative cell number determined by measuring LDH activity. The number of cells present is directly proportional to the absorbance value, which represents LDH activity. (b) Media were assessed for the level of LDH released over the culture period as a measure of cell death. Data are presented as mean ±SEM (n = 4 per treatment) and analysed by GLM ANOVA and Tukey’s post hoc test: *p ≤ 0.05; **p < 0.01; ***p < 0.001 vs. untreated controls unless shown otherwise.

involved PKR, by assessing the expression of ADAMTS4 0.001). The PKR inhibitor negative control (PKRi neg) and 5 following OSM+IL-1 treatment and PKR inhibition had no effect on cytokine-induced ADAMTS4 (Fig. 6a) or (Fig. 6; data ranked prior to analysis). OSM+IL-1 treatment ADAMTS5 (data not shown) levels. resulted in a 754-fold increase in ADAMTS4 (p < 0.001, Fig. 6a) and a 40.8-fold increase in ADAMTS5 mRNAs The OSM and IL-1 down-regulation of type II (p < 0.001, Fig. 6b) compared to untreated control cells. collagen in articular chondrocytes is enhanced by Inhibition of PKR signifi cantly reduced basal ADAMTS4 PKR inhibition and ADAMTS5 (p < 0.05 Fig. 6). Inhibition of PKR in Cell extracts were examined for type II collagen expression cytokine treated cultures resulted in an 8.5-fold reduction in by qPCR (Fig. 7a) and Western blot (Fig. 7b) analysis. ADAMTS4 mRNA compared to OSM+IL-1 alone (Fig. 6a; After 7-days in culture, inhibiting PKR had no effect on p < 0.01) although ADAMTS4 still remained 88-fold higher the basal level of COL2A1 mRNA. Cytokine treatment than control expression levels (p < 0.001). PKR inhibition resulted in a signifi cant, 11-fold reduction in the expression also signifi cantly reduced ADAMTS5 mRNA expression in of COL2A1 mRNA compared to untreated control cells OSM+IL-1 treated cultures (Fig. 6b; 3.8-fold reduction p (Fig. 7a; p < 0.001 following log transformation of data). < 0.001 compared with OSM+IL-1 alone) although levels Inhibition of PKR in cytokine treated cultures resulted in a still remained 10.7-fold higher than controls (Fig. 6b; p < 41-fold reduction in the level of COL2A1 mRNA compared

www.ecmjournal.org 47 SJ Gilbert et al. PKR mediates OSM and IL-1 signalling in chondrocytes

Fig. 4. OSM and IL-1 treatment increases the expression and activation of MMP9 via activation of PKR. (a) Media were collected at the end of the 7-day treatment period and analysed by gelatin substrate zymography. A proMMP9 standard (s) is indicated. (b) The area (absorbance units) of substrate gel cleared by pro- and active MMP9 was measured by scanning densitometry. Data shown are absorbance units per cell, expressed as means SEM (n = 4 per treatment) and analysed by GLM ANOVA and Tukey’s post hoc test: *** p < 0.001; **p < 0.01; *p ≤ 0.05 vs. untreated controls unless shown otherwise. (c) RNA from cells (n = 6 per treatment) was analysed by quantitative PCR to determine the relative expression of MMP9 mRNA. Data are presented as fold change relative to untreated

control cells calculated using the∆∆CT method with RN18S1 (18S ribosomal) RNA as the reference gene and plotted as mean ±SEM. Data were ranked prior to analysis and signifi cant differences detected by GLM ANOVA and Tukey’s post hoc test: *** p < 0.001; ** p < 0.01 vs. untreated controls unless shown otherwise.

www.ecmjournal.org 48 SJ Gilbert et al. PKR mediates OSM and IL-1 signalling in chondrocytes

Fig. 5. OSM and IL-1 treatment increases the expression of MMP-13 through a PKR independent mechanism. RNA from cells (n = 3-6 per treatment) was analysed by quantitative PCR to determine the relative expression of

MMP13 mRNA. Data are presented as fold change relative to untreated control cells calculated using the∆∆CT method with RN18S1 RNA as the reference gene and plotted as mean ±SEM. Signifi cant differences were detected by GLM ANOVA and Tukey’s post hoc test on log transformed data: *** p < 0.001 vs. untreated controls.

to untreated control cultures (p < 0.001 following log was detected in cells treated with cytokines which was transformation of data) and 3.5-fold reduction compared reduced a further 10-fold following inhibition of PKR to cytokine treatment alone (p < 0.01 following log (Fig. 8c; p < 0.001 after log transformation of data). The transformation of data); a fi nding that was reproduced fi nding that PKR inhibition further reduced COL11A1 and to the same level in the experimental repeat. The PKR ACAN gene expression was reproduced to the same extent inhibitor negative control (PKRi neg) had no effect on in both experimental repeats. No signifi cant difference was cytokine treated cells. observed in the level of SOX9 mRNA following any of the A complete loss of type II collagen protein was treatments (data not shown). The PKR inhibitor negative consistently observed in all samples analysed following control (PKRi neg) did not alter the response of the cells OSM+IL-1 treatment (n = 4 samples; 2 independent to OSM+IL-1. experiments, Fig. 7b). Inhibiting PKR had no effect on the basal level of type II collagen protein or that observed Cytokine-induced increases in TNFα expression in OSM+IL-1 treated chondrocytes. The PKR inhibitor occurs through a mechanism involving PKR negative control (PKRi neg) had no effect (data not shown). Since PKR regulates TNFα transcription (Kaempfer, 2003; Gilbert et al., 2006b) we determined whether TNFα mRNA Treatment of chondrocytes with OSM and IL-1 expression was affected by the combined treatment of decrease mRNA expression of that maintain OSM and IL-1 and if so, whether this occurred through the chondrocytic phenotype a mechanism involving PKR (Fig. 9). OSM and IL-1 The expression level of COL9A1 and COL11A1, stimulation of articular chondrocytes resulted in a 10.6-fold and ACAN mRNAs were assessed by qPCR (Fig. 8). increase in TNFα mRNA compared to untreated control OSM+IL-1 treatment resulted in a 3500-fold reduction cells (p < 0.001 following log transformation of data). The in the expression of COL9A1 mRNA (Fig. 8a; p < 0.001 basal expression level of TNFα mRNA was signifi cantly following log transformation of data). Inhibition of reduced 1.9-fold following inhibition of PKR (p < 0.05 PKR had no signifi cant effect on the level of COL9A1 following log transformation of data). In addition, TNFα mRNA. Cytokine treatment also signifi cantly reduced mRNA expression in cytokine treated cultures was reduced the expression of COL11A 15-fold compared to control 3.5-fold by inhibiting PKR (p < 0.001 following log levels (Fig. 8b; p < 0.001 following log transformation transformation of data). Despite this signifi cant reduction, of data). When cytokine stimulated cells were incubated levels still remained 3-fold higher than that detected in with the PKR inhibitor, a further 4-fold reduction in the control cells (p < 0.001 following log transformation of amount of COL11A1 mRNA was detected compared with data). The PKR inhibitor negative control (PKRi neg) did OSM+IL-1 alone (p = 0.001 following log transformation not alter the response of the cells to OSM+IL-1 (data not of data). A 3.8-fold reduction in the level of ACAN mRNA shown).

www.ecmjournal.org 49 SJ Gilbert et al. PKR mediates OSM and IL-1 signalling in chondrocytes

Fig. 6. OSM and IL-1 treatment increases the expression of ADAMTS-4 and ADAMTS-5 via activation of PKR. RNA from cells (n = 6 per treatment) was analysed by quantitative PCR to determine the relative expression of ADAMTS4 (a) and ADAMTS5 (b) mRNA. Data are presented as fold change relative to untreated control cells

calculated using the ∆∆CT method with RN18S1 RNA as the reference gene and plotted as mean ±SEM. Signifi cant differences were detected by GLM ANOVA and Tukey’s post hoc test: *** p < 0.001; ** p < 0.01; * p ≤ 0.05 vs. untreated controls unless shown otherwise. Data were ranked prior to analysis.

Discussion the OSM and IL-1 induced catabolic phenotype in primary articular chondrocytes. This study investigated the mechanism behind the well Immunohistochemistry using an antibody to the documented synergistic effect of OSM and IL-1 on phosphorylated (active) form of PKR revealed that in cartilage destruction. The stress-induced PKR signalling unstimulated chondrocytes, active PKR was localised pathway is activated in early osteoarthritis and is involved diffusely throughout the cytoplasm and within distinct in the propagation of proinfl ammatory signals in articular peri-nuclear pools that co-localised with the Golgi. cartilage (Gilbert et al., 2002; Gilbert et al., 2004; Tam et The localisation of active PKR did not alter following al., 2007). Therefore, we investigated the role of PKR in cytokine treatment. PKR is largely associated with

www.ecmjournal.org 50 SJ Gilbert et al. PKR mediates OSM and IL-1 signalling in chondrocytes

Fig. 7. The effect of cytokine treatment on type II collagen expression is enhanced following inhibition of PKR. Chondrocytes were treated for 7 days with OSM and IL-1 in the presence or absence of the PKR inhibitor. (a) Relative expression of COL2A1 (type II collagen) mRNA, determined by quantitative PCR, is presented as fold

change relative to untreated control cells calculated using the ∆∆CT method with RN18S1 RNA as the reference gene. Data are plotted as mean ±SEM (n = 6 replicates per treatment). Signifi cant differences were detected by GLM ANOVA and Tukey’s post hoc test on log transformed data: *** p < 0.001; ** p < 0.01. (b) Protein from an equivalent cell number was analysed by Western blotting for type II collagen (n = 4 replicates per treatment, representative data shown from 2 independent experiments). 1(II) = alpha 1 chain type II collagen. Blots were stripped and re-probed for total ERK1/2 to confi rm equivalent loading.

ribosomes in the cytoplasm and the rough endoplasmic the time point analysed, we did not detect phosphorylated reticulum (MacQuillan et al., 2009) but also localises to PKR in the nucleus after OSM or IL-1 individual or the juxta-nuclear cytoplasmic layer, the nuclear envelope combined treatment. Interferon treatment of peripheral and within the nucleus itself (Jeffrey et al., 1995; Besse blood mononuclear cells for 24 h increases the amount of et al., 1998). Localisation of active PKR within the active PKR in the nucleus (MacQuillan et al., 2009). The cytoplasm is necessary for access to its substrates such as role of PKR in the nucleus is less well defi ned but may eIF2 associated with the ribosomes (Zhu et al., 1997) involve ribosome biogenesis (Jimenez-Garcia et al., 1993), and the IKK complex that releases IB, the inhibitor regulation of the amount of RNA molecules released into of kappa Bto allow the translocation of nuclear factor the cytoplasm (Besse et al., 1998), and enhancement of kappa B (NFB) to the nucleus (Kumar et al., 1994). At TNF precursor transcript splicing (Osman et al., 1999).

www.ecmjournal.org 51 SJ Gilbert et al. PKR mediates OSM and IL-1 signalling in chondrocytes

Fig. 8. Combined treatment of chondrocytes with OSM and IL-1 results in a decrease in the expression of matrix genes that are important in maintaining a chondrocytic phenotype. Quantitative PCR was used to determine the relative expression of (a) COL9A1 (type IX collagen), (b) COL11A1 (type XI collagen), and (c) ACAN (aggrecan)

mRNA. Data are presented as fold change relative to control cells calculated using the∆∆CT method with RN18S1 RNA as the reference gene and plotted as mean ±SEM (n = 6 replicates per treatment). Signifi cant differences were detected by GLM ANOVA and Tukey’s post hoc test on log transformed data: *** p < 0.001 vs. untreated controls.

Future studies are clearly warranted, assessing earlier 7 days (Fig. 2). Since active PKR blocks translation in time points and co-localising active PKR with specifi c response to stress signals (Clemens et al., 2000), the subcellular compartments in order to shed light on the amount of total protein within the cell can be measured roles of PKR following cytokine treatments. following cytokine treatment in the presence/absence of a To indicate the extent of PKR activation after individual PKR inhibitor to provide an indicator of the extent of PKR or combination cytokine treatments, we measured total activation (Gilbert et al., 2006a; Scheuner et al., 2006). protein concentration under each treatment regime after These data clearly show that only the combination of

www.ecmjournal.org 52 SJ Gilbert et al. PKR mediates OSM and IL-1 signalling in chondrocytes

Fig. 9. Cytokine-induced increases in the expression of TNFα mRNA occurs through a mechanism involving PKR. TNFα mRNA expression analysed by quantitative PCR presented as fold change relative to control cells

calculated using the ∆∆CT method with RN18S1 RNA as the reference gene and plotted as mean ±SEM (n = 6 per treatment). Signifi cant differences were detected by GLM ANOVA and Tukey’s post hoc test on log transformed data: *** p < 0.001; * p ≤ 0.05 vs. untreated controls unless shown otherwise.

OSM+IL-1 signifi cantly reduced the protein concentration induced MMP9 mRNA and protein expression, as well in a PKR-dependent manner. Thus, subsequent experiments as its activation, implicates PKR, for the fi rst time, in the focused on the mechanisms underlying this combined regulation of this cytokine-induced catabolic pathway. effect of OSM and IL-1. Our previous data using 2-aminopurine as an inhibitor OSM+IL-1 treatment increased cell death, however the of PKR found that activation of the PKR signalling mean cell number at the end of 7-days of culture was also pathway was critical in TNF induced MMP9 expression slightly increased suggesting an increase in chondrocyte and activation (Gilbert et al., 2004). This was supported proliferation. OSM treatment of foetal liver cells stimulates by the studies of Takada et al. who demonstrated that cellular proliferation (Ehashi et al., 2005) but there have TNF-induced MMP9 expression was abolished in PKR been no studies, to our knowledge, investigating the role of -/- cells (Radtke et al., 2002). Taken together, these data PKR in this process. Inhibition of PKR increased basal cell demonstrate that the PKR pathway is pivotal in mediating proliferation as well as further increasing proliferation of the pro-infl ammatory cytokine-induced MMP9 synthesis OSM+IL-1 treated cells. This is in keeping with the known and activation pathways known to be critical in arthritic role of PKR as a suppressor of cell growth (Koromilas et cartilage degeneration. al., 1992). Mammalian cells normally have a low level Our fi nding that OSM+IL-1 increased both ADAMTS4 of active PKR, which phosphorylates a small fraction and 5 mRNA expression is consistent with previous studies of eIF2, enough to hold the synthesis of cell growth using bovine cartilage explants (Milner et al., 2006; Blain promoting proteins in check but not to inhibit overall et al., 2010; Durigova et al., 2011) and human primary protein synthesis (Kim et al., 2000). chondrocytes (Young et al., 2005). In addition, Koshy et Our data, showing that OSM+IL-1 potently induced the al. (Koshy et al., 2002) reported that OSM+IL-1 induced expression of MMP9 mRNA supports the work of others ADAMTS4 but not ADAMTS5 in T/C28a4 cells (Koshy (Milner et al., 2006; Blain et al., 2010). Extending these et al., 2002). Our data show that although ADAMTS5 was fi ndings, our current study has shown that OSM+IL-1 more highly expressed than ADAMTS4, the OSM+IL-1- treatment of articular chondrocytes additionally results induced increases were far greater for ADAMTS4 gene in an increase in the activation of proMMP9, a key factor expression compared to ADAMTS5 (830-fold vs. 40- that contributes to joint destruction and infl ammation in fold). These data, in conjunction with previous fi ndings the arthritic joint. Our fi nding that OSM+IL-1 induced (Koshy et al., 2002; Durigova et al., 2011), suggest that MMP9 mRNA is opposite to the fi ndings of Koshy et al. ADAMTS4 is the primary aggrecanase responsible for the (Koshy et al., 2002). The reason for this discrepancy may synergistic action of OSM+IL-1 on aggrecan degradation. be due to the use of a cell line by Koshy et al. rather than We have shown for the fi rst time that OSM+IL-1 induced primary chondrocytes here. Importantly, our observation up-regulation of ADAMTS4 and ADAMTS5 is, in part, that the PKR inhibitor signifi cantly reduces OSM+IL-1 mediated through PKR. Since the inhibition of PKR only

www.ecmjournal.org 53 SJ Gilbert et al. PKR mediates OSM and IL-1 signalling in chondrocytes partly decreased the level of ADAMTS4 and ADAMTS5 in the expression of TNFα mRNA, which importantly, mRNAs, an additional signalling mechanism must be is partly mediated by PKR. This demonstrates that activated following OSM+IL-1 treatment and is likely OSM+IL-1 propagate the inflammatory response in to occur via a complex interplay of signal transduction chondrocytes via activation of the PKR signalling pathway pathways including the MAPK/ERK, PI3K/Akt, Jak/STAT, and induction of the pro-infl ammatory cytokine, TNF. JNK1/2 and p38 kinase pathways (Blanchard et al., 2001; The OSM+IL-1 induced, PKR mediated upregulation Catterall et al., 2001; Radtke et al., 2002; Barksby et al., of TNF may occur via NFB transcriptional activation 2006; El Mabrouk et al., 2007). (Kumar et al., 1994) or by direct binding of activated PKR This is the first report, to our knowledge, that to a cis-acting element in the TNFα pre-mRNA resulting OSM+IL-1 treatment of bovine articular chondrocytes in an increase in TNFα splicing and production in the cell also decreases the expression of cartilage matrix genes (Osman et al., 1999; Kaempfer, 2003). PKR may also (COL2A1, COL9A1, COL11A1 and ACAN). OSM+IL-1 infl uence the OSM+IL-1 induced increases in expression thus profoundly alter chondrocyte homeostasis by of enzymes such as ADAM10 (Koshy et al., 2002) or decreasing matrix synthesis in conjunction with increasing MMP12 (own unpublished data) (Barksby et al., 2006; matrix catabolism. OSM+IL-1 treatment resulted in Milner et al., 2006) which process proTNF to its active a complete loss of type II collagen protein, consistent form (Rosendahl et al., 1997; Matsumoto et al., 1998). An with the transcriptional down regulation we report as important question that remains to be elucidated is which of well as other studies showing that OSM+IL-1 induces the PKR mediated effects of OSM+IL-1 treatment we have collagenolysis (Milner et al., 2006; Litherland et al., observed in the current study are via PKR’s transcriptional 2008; Lakey and Cawston, 2009). A PKR independent effects on TNF. mechanism for type II collagen breakdown is implicated since it was not prevented by PKR inhibition. Somewhat Conclusions surprisingly, the effect of OSM+IL-1 on down-regulating the transcription of type II (COL2A1) and XI collagen In summary, we have clearly demonstrated for the fi rst time (COL11A1) and aggrecan (ACAN) was enhanced when that combined OSM and IL-1 treatment increased MMP9 PKR was inhibited. One explanation is that OSM+IL-1 expression and activation and ADAMTS4, ADAMTS5 treatment increases the amount of active PKR within the and TNFα mRNA expression through a PKR dependant cell to a level that is high enough to block the translation mechanism. Since these catabolic and pro-infl ammatory of proteins involved in the transcriptional control of these effects are increased in human arthritis (Sandy et al., specifi c genes. Thus removing active PKR allows the 1992; Yoshihara et al., 2000a), where TNF helps drives translation of these OSM+IL-1 induced proteins to occur cartilage degradation (Feldmann and Maini, 1999), these which then increases the negative effect on transcription new roles for PKR are important. Targeting this PKR of COL2A1, COL11A1 and ACAN although this remains pathway may therefore have the potential to suppress to be confi rmed. The precise mechanism of action at work MMP and ADAMTS expression during the infl ammatory here clearly requires further study. Inhibition of PKR did phase of cartilage destruction and thus presents a potential not signifi cantly infl uence the effect of OSM+IL-1 on target for pharmaceutical intervention in arthritic disease. the expression of COL11A1 mRNA indicating that the OSM+IL-1 mediated down-regulation of the transcription of this gene maybe independent of the PKR signalling Acknowledgements pathway. Several studies have shown that treatment of This work was funded by Arthritis Research UK (grant chondrocytes with OSM+IL-1 increases the expression numbers: 18221, 16436, and 18461). The authors thank Dr of MMP13, which plays a pivotal role, along with Peter Watson for the antibody to GM130, Dr Ilyas Khan MMP1, in the breakdown of type II collagen and is thus for the provision of the SOX9 and ADAMTS5 primers, Dr likely to contribute to the loss of type II collagen shown Siyuan Li for the ADAMTS4 primers, and the late Dr Ann here. MMP13 mRNA expression was also increased by Vaughan-Thomas for the COL9A1 and COL11A1 primers. OSM+IL-1 treatment in our study and this, indeed occurred through a mechanism that was independent of PKR. It is clear that a complex array of signalling interactions References infl uence type II collagen breakdown following OSM treatment. These include ERK/MAPK mediated induction Barksby HE, Hui W, Wappler I, Peters HH, Milner of ADAMTS4 and MMP13 (El Mabrouk et al., 2007), JM, Richards CD, Cawston TE, Rowan AD (2006) PI3K/Akt mediated induction of MMP1 and 13 (Yoshihara Interleukin-1 in combination with oncostatin M up- et al., 2000b) and the JAK/STAT, ERK1/2, JNK1/2, and regulates multiple genes in chondrocytes: implications p38 kinase pathways (Blanchard et al., 2001; Catterall for cartilage destruction and repair. Arthritis Rheum 54: et al., 2001; Radtke et al., 2002; Barksby et al., 2006). 540-550. The nature of the cross-talk between PKR and these other Besse S, Rebouillat D, Marie I, Puvion-Dutilleul F, signalling pathways remains to be elucidated. Hovanessian AG (1998) Ultrastructural localization of Finally, our data reveal that OSM+IL-1 treatment of interferon-inducible double-stranded RNA-activated bovine articular chondrocytes results in a 10-fold increase enzymes in human cells. Exp Cell Res 239: 379-392.

www.ecmjournal.org 54 SJ Gilbert et al. PKR mediates OSM and IL-1 signalling in chondrocytes

Blain EJ, Gilbert SJ, Hayes AJ, Duance VC (2006) El Mabrouk M, Sylvester J, Zafarullah M (2007) Disassembly of the vimentin cytoskeleton disrupts articular Signaling pathways implicated in oncostatin M-induced cartilage chondrocyte homeostasis. Matrix Biol 25: 398- aggrecanase-1 and matrix metalloproteinase-13 expression 408. in human articular chondrocytes. Biochim Biophys Acta Blain EJ, Ali AY, Duance VC (2010) Boswellia 1773: 309-320. frereana (frankincense) suppresses cytokine-induced Fearon U, Mullan R, Markham T, Connolly M, matrix metalloproteinase expression and production of pro- Sullivan S, Poole AR, FitzGerald O, Bresnihan B, Veale infl ammatory molecules in articular cartilage. Phytother DJ (2006) Oncostatin M induces angiogenesis and cartilage Res 24: 905-912. degradation in rheumatoid arthritis synovial tissue and Blalock WL, Grimaldi C, Fala F, Follo M, Horn S, human cartilage cocultures. Arthritis Rheum 54: 3152- Basecke J, Martinelli G, Cocco L, Martelli AM (2009) PKR 3162. activity is required for acute leukemic cell maintenance Feldmann M, Maini RN (1999) The role of cytokines and growth: a role for PKR-mediated phosphatase activity in the pathogenesis of rheumatoid arthritis. Rheumatology to regulate GSK-3 phosphorylation. J Cell Physiol 221: (Oxford) 38 Suppl 2: 3-7. 232-241. Frye SR, Yee A, Eskin SG, Guerra R, Cong X, McIntire Blanchard F, Wang Y, Kinzie E, Duplomb L, Godard LV (2005) cDNA microarray analysis of endothelial cells A, Baumann H (2001) Oncostatin M regulates the subjected to cyclic mechanical strain: importance of motion synthesis and turnover of gp130, leukemia inhibitory factor control. Physiol. Genomics 21: 124-130. receptor alpha, and oncostatin M receptor beta by distinct Gilbert SJ, Duance VC, Mason DJ (2002) Tumour mechanisms. J Biol Chem 276: 47038-47045. necrosis factor alpha up-regulates protein kinase R (PKR)- Catterall JB, Carrere S, Koshy PJ, Degnan BA, activating protein (PACT) and increases phosphorylation Shingleton WD, Brinckerhoff CE, Rutter J, Cawston of PKR and eukaryotic initiation factor 2-alpha in articular TE, Rowan AD (2001) Synergistic induction of matrix chondrocytes. Biochem Soc Trans 30: 886-889. metalloproteinase 1 by interleukin-1alpha and oncostatin Gilbert SJ, Duance VC, Mason DJ (2004) Does M in human chondrocytes involves signal transducer protein kinase R mediate TNF-alpha- and ceramide- and activator of transcription and activator protein 1 induced increases in expression and activation of matrix transcription factors via a novel mechanism. Arthritis metalloproteinases in articular cartilage by a novel Rheum 44: 2296-2310. mechanism? Arthritis Res Ther 6: R46-R55. Cawston TE, Ellis AJ, Humm G, Lean E, Ward D, Curry Gilbert SJ, Blain EJ, Jones P, Duance VC, Mason DJ V (1995) Interleukin-1 and oncostatin M in combination (2006a) Exogenous sphingomyelinase increases collagen promote the release of collagen fragments from bovine and sulphated glycosaminoglycan production by primary nasal cartilage in culture. Biochem Biophys Res Commun articular chondrocytes: an in vitro study. Arthritis Res Ther 215: 377-385. 8: R89. Cawston TE, Curry VA, Summers CA, Clark IM, Riley Gilbert SJ, Duance VC, Mason D (2006b) Protein GP, Life PF, Spaull JR, Goldring MB, Koshy PJ, Rowan kinase R: a novel mediator of articular cartilage degradation AD, Shingleton WD (1998) The role of oncostatin M in in arthritis. Curr Rheumatol Rev 2: 9-21. animal and human connective tissue collagen turnover Gilbert SJ, Blain EJ, Duance VC, Mason DJ (2008) and its localization within the rheumatoid joint. Arthritis Sphingomyelinase decreases type II collagen expression Rheum 41: 1760-1771. in bovine articular cartilage chondrocytes via the ERK Chiu S-C, Yang N-S (2007) Inhibition of tumor necrosis signaling pathway. Arthritis Rheum 58: 209-220. factor-{alpha} through selective blockade of pre-mRNA Hui W, Barksby HE, Young DA, Cawston TE, McKie splicing by shikonin. Mol Pharmacol 71: 1640-1645. N, Rowan AD (2005) Oncostatin M in combination with Clemens MJ, Bushell M, Jeffrey IW, Pain VM, Morley tumour necrosis factor {alpha} induces a chondrocyte SJ (2000) Translation initiation factor modifi cations and membrane associated aggrecanase that is distinct from the regulation of protein synthesis in apoptotic cells. Cell ADAMTS aggrecanase-1 or -2. Ann Rheum Dis 64: 1624- Death Differ 7: 603-615. 1632. Clemens MJ, Elia A (1997) The double-stranded RNA- Hui W, Rowan AD, Richards CD, Cawston TE dependent protein kinase PKR: structure and function. J (2003) Oncostatin M in combination with tumor necrosis Interferon Cytokine Res 17: 503-524. factor alpha induces cartilage damage and matrix Durigova M, Soucy P, Fushimi K, Nagase H, Mort JS, metalloproteinase expression in vitro and in vivo. Arthritis Roughley PJ (2008) Characterization of an ADAMTS-5- Rheum 48: 3404-3418. mediated cleavage site in aggrecan in OSM-stimulated Jammi NV, Whitby LR, Beal PA (2003) Small molecule bovine cartilage. Osteoarthritis Cartilage 16: 1245-1252. inhibitors of the RNA-dependent protein kinase. Biochem Durigova M, Troeberg L, Nagase H, Roughley PJ, Mort Biophys Res Commun 308: 50-57. JS (2011) Involvement of ADAMTS5 and hyaluronidase Jeffrey IW, Kadereit S, Meurs EF, Metzger T, in aggrecan degradation and release from OSM-stimulated Bachmann M, Schwemmle M, Hovanessian AG, Clemens cartilage. Eur Cell Mater 21: 31-45. MJ (1995) Nuclear localization of the interferon-inducible Ehashi T, Miyoshi H, Ohshima N (2005) Oncostatin M protein kinase PKR in human cells and transfected mouse stimulates proliferation and functions of mouse fetal liver cells. Exp Cell Res 218: 17-27. cells in three-dimensional cultures. J Cell Physiol 202: Jimenez-Garcia LF, Green SR, Mathews MB, Spector 698-706. DL (1993) Organization of the double-stranded RNA-

www.ecmjournal.org 55 SJ Gilbert et al. PKR mediates OSM and IL-1 signalling in chondrocytes activated protein kinase DAI and virus-associated VA resorbing cartilage reveals pro-collagenase activation as a RNAI in adenovirus-2-infected HeLa cells. J Cell Sci 106: critical step for collagenolysis. Arthritis Res Ther 8: R142. 11-22. Osman F, Jarrous N, Ben-Asouli Y, Kaempfer R Kaempfer R (2003) RNA sensors: novel regulators of (1999) A cis-acting element in the 3’-untranslated region gene expression. EMBO Rep 4: 1043-1047. of human TNF-alpha mRNA renders splicing dependent Kim SH, Forman AP, Mathews MB, Gunnery S (2000) on the activation of protein kinase PKR. Genes Dev 13: Human breast cancer cells contain elevated levels and 3280-3293. activity of the protein kinase, PKR. Oncogene 19: 3086- Patel CV, Handy I, Goldsmith T, Patel RC (2000) PACT, 3094. a stress-modulated cellular activator of interferon-induced Koromilas AE, Roy S, Barber GN, Katze MG, double-stranded RNA-activated protein kinase, PKR. J Sonenberg N (1992) Malignant transformation by a mutant Biol Chem 275: 37993-37998. of the IFN-inducible dsRNA-dependent protein kinase. Radtke S, Hermanns HM, Haan C, Schmitz-Van De Science 257: 1685-1689. Leur H, Gascan H, Heinrich PC, Behrmann I (2002) Novel Koshy PJ, Lundy CJ, Rowan AD, Porter S, Edwards DR, role of Janus kinase 1 in the regulation of oncostatin M Hogan A, Clark IM, Cawston TE (2002) The modulation receptor surface expression. J Biol Chem 277: 11297- of matrix metalloproteinase and ADAM gene expression 11305. in human chondrocytes by interleukin-1 and oncostatin M: Rosendahl MS, Ko SC, Long DL, Brewer MT, a time-course study using real-time quantitative reverse Rosenzweig B, Hedl E, Anderson L, Pyle SM, Moreland transcription-polymerase chain reaction. Arthritis Rheum J, Meyers MA, Kohno T, Lyons D, Lichenstein HS (1997) 46: 961-967. Identifi cation and characterization of a pro-tumor necrosis Kumar A, Haque J, Lacoste J, Hiscott J, Williams factor-alpha-processing from the ADAM family BR (1994) Double-stranded RNA-dependent protein of zinc metalloproteases. J Biol Chem 272: 24588-24593. kinase activates transcription factor NF-kappa B by Sandy JD, Flannery CR, Neame PJ, Lohmander phosphorylating I kappa B. Proc Natl Acad Sci USA 91: LS (1992) The structure of aggrecan fragments in 6288-6292. human synovial fl uid. Evidence for the involvement in Lakey RL, Cawston TE (2009) Sulfasalazine blocks osteoarthritis of a novel proteinase which cleaves the Glu the release of proteoglycan and collagen from cytokine 373-Ala 374 bond of the interglobular domain. J Clin Invest stimulated cartilage and down-regulates metalloproteinases. 89: 1512-1516. Rheumatology (Oxford) 48: 1208-1212. Scheuner D, Patel R, Wang F, Lee K, Kumar K, Wu J, Langdon C, Kerr C, Hassen M, Hara T, Arsenault AL, Nilsson A, Karin M, Kaufman RJ (2006) Double-stranded Richards CD (2000) Murine oncostatin M stimulates mouse RNA-dependent protein kinase phosphorylation of the synovial fi broblasts in vitro and induces infl ammation and alpha-subunit of eukaryotic translation initiation factor 2 destruction in mouse joints in vivo. Am J Pathol 157: 1187- mediates apoptosis. J Biol Chem 281: 21458-21468. 1196. Takaishi H, Kimura T, Dalal S, Okada Y, D’Armiento Lee SB, Esteban M (1994) The interferon-induced J (2008) Joint diseases and matrix metalloproteinases: a double-stranded RNA-activated protein kinase induces role for MMP-13. Curr Pharm Biotechnol 9: 47-54. apoptosis. Virology 199: 491-496. Tam CL, Hofbauer M, Towle CA (2007) Requirement Litherland GJ, Dixon C, Lakey RL, Robson T, Jones for protein kinase R in interleukin-1alpha-stimulated D, Young DA, Cawston TE, Rowan AD (2008) Synergistic effects in cartilage. Biochem Pharmacol 74: 1636-1641. collagenase expression and cartilage collagenolysis are Taylor S, Wakem M, Dijkman G, Alsarraj M, Nguyen phosphatidylinositol 3-kinase/Akt signaling-dependent. J M (2010) A practical approach to RT-qPCR-Publishing data Biol Chem 283: 14221-14229. that conform to the MIQE guidelines. Methods 50: S1-5. Little CB, Barai A, Burkhardt D, Smith SM, Fosang van den Berg WB (2011) Osteoarthritis year 2010 in AJ, Werb Z, Shah M, Thompson EW (2009) Matrix review: pathomechanisms. Osteoarthritis Cartilage 19: metalloproteinase 13-deficient mice are resistant to 338-341. osteoarthritic cartilage erosion but not chondrocyte Vaughan-Thomas A, Dudhia J, Bayliss MT, Kadler hypertrophy or osteophyte development. Arthritis Rheum KE, Duance VC (2008) Modifi cation of the composition 60: 3723-3733. of articular cartilage collagen fi brils with increasing age. MacQuillan G, Caterina P, de Boer B, Allan J, Platten Connect Tissue Res 49: 374-382. M, Reed W, Jeffrey G (2009) Ultra-structural localisation Vaughan-Thomas A, Gilbert SJ, Duance VC (2000) of hepatocellular PKR protein using immuno-gold labelling Elevated levels of proteolytic enzymes in the aging human in chronic hepatitis C virus disease. J Mol Histol 40: 171- vitreous. Invest Ophthalmol Vis Sci 41: 3299-3304. 176. Yoshihara Y, Nakamura H, Obata K, Yamada Matsumoto S, Kobayashi T, Katoh M, Saito S, Ikeda Y, H, Hayakawa T, Fujikawa K, Okada Y (2000a) Kobori M, Masuho Y, Watanabe T (1998) Expression and Matrix metalloproteinases and tissue inhibitors of localization of matrix metalloproteinase-12 in the aorta of metalloproteinases in synovial fl uids from patients with cholesterol-fed rabbits: relationship to lesion development. rheumatoid arthritis or osteoarthritis. Ann Rheum Dis 59: Am J Pathol 153: 109-119. 455-461. Milner JM, Rowan AD, Cawston TE, Young DA (2006) Yoshihara Y, Nakamura H, Obata Ki, Yamada Metalloproteinase and inhibitor expression profi ling of H, Hayakawa T, Fujikawa K, Okada Y (2000b)

www.ecmjournal.org 56 SJ Gilbert et al. PKR mediates OSM and IL-1 signalling in chondrocytes

Matrix metalloproteinases and tissue inhibitors of gene expression in chondrocytes and block cartilage metalloproteinases in synovial fl uids from patients with resorption. Arthritis Res Ther 7: R503-512. rheumatoid arthritis or osteoarthritis. Ann Rheum Dis 59: Zhu S, Romano PR, Wek RC (1997) Ribosome targeting 455-461. of PKR is mediated by two double-stranded RNA-binding Young DA, Lakey RL, Pennington CJ, Jones D, domains and facilitates in vivo phosphorylation of Kevorkian L, Edwards DR, Cawston TE, Clark IM (2005) eukaryotic initiation factor-2. J Biol Chem 272: 14434- Histone deacetylase inhibitors modulate metalloproteinase 14441.

www.ecmjournal.org 57