Cytokine-induced cysteine- serine-rich nuclear protein-1 (CSRNP1) selectively contributes to MMP1 expression in human chondrocytes

Irreversible cartilage collagen breakdown by the collagenolytic matrix metalloproteinases (MMPs)-1 and MMP-13 represents a key event in pathologies associated with tissue destruction such as arthritis. Inflammation is closely associated with such pathology and occurs in both rheumatoid and osteoarthritis making it highly relevant to the prevailing tissue damage that characterises these diseases. The inflammation-induced activating protein-1 (AP-1) transcription factor is an important regulator of both MMP1 and MMP13 genes with interplay between signalling pathways contributing to their expression. Here, we have examined the regulation of MMP1 expression, and using in vivo chromatin immunoprecipitation analyses we have demonstrated that cFos bound to the AP-1 cis element within the proximal MMP1 promoter only when the gene was transcriptionally silent as previously observed for MMP13. Subsequent small interfering RNA-mediated silencing confirmed however, that cFos significantly contributes to MMP1 expression. In contrast, silencing of ATF3 (a prime MMP13 modulator) did not affect MMP1 expression whilst silencing of the Wnt-associated regulator cysteine- serine-rich nuclear protein-1 (CSRNP1) resulted in substantial repression of MMP1 but not MMP13. Furthermore, following an early transient peak in expression of CSRNP1 at the mRNA and protein levels similar to that seen for cFOS, CSRNP1 expression subsequently persisted unlike cFOS. Finally, DNA binding assays indicated that the binding of CSRNP1 to the AP-1 consensus-like sequences within the proximal promoter regions of MMP1 and MMP13 was preferentially selective for MMP1 whilst activating transcription factor 3 (ATF3) binding was exclusive to MMP13. These data further extend our understanding of the previously reported differential regulation of these MMP genes, and strongly indicate that although cFos modulates the expression of MMP1/13, downstream factors such as CSRNP1 and ATF3 ultimately serve as transcriptional regulators in the context of an inflammatory stimulus for these potent collagenolytic MMPs.


Introduction
Degenerative joint diseases result in destruction of the articular cartilage, typically when a variety of pro-inflammatory mediators perpetuate the disease process (reviewed in [1]). There is now increasing evidence that such inflammatory mediators are involved the pathogenesis of both rheumatoid arthritis (RA) and osteoarthritis (OA) [1]. These mediators include interleukin (IL)-1, members of the IL-6 family (including IL-6 and oncostatin M (OSM)) as well as other inducers of signalling such as Wnt [2][3][4]. Cartilage collagen breakdown is well documented to be primarily due to the matrix metalloproteinases (MMPs), in particular the collagenases MMP-1 and MMP-13, which are responsible for the irreversible degradation of the collagenous matrix [5]. Although dogma suggests MMP-1 is the major collagenase in RA, whilst MMP-13 prevails in OA, the above mediators all drive the expression of both MMPs with MMP1 induction typically more marked [6,7]. Furthermore, we have previously reported differential regulatory mechanisms for cytokine-induced MMP1 and MMP13 expression in human chondrocytes [8,9], and recently confirmed that the most potent cytokine stimulus reported promotes many signalling pathways that are common to a wide variety of pro-inflammatory stimuli [10]. Indeed, we and others have demonstrated marked MMP expression in chondrocytes stimulated with numerous inflammatory mediators [11][12][13][14][15] indicating such potent expression occurs within most inflammatory situations associated with tissue destruction such as RA and OA. Consequently, the collagenases represent key therapeutic targets for disease-modifying agents [5,16].
Investigations of the molecular mechanisms by which MMPs are regulated [17,18] have highlighted a primary role for activator protein (AP)-1 transcription factor complexes in regulating MMP expression [9,[19][20][21]. AP-1 exists as a protein dimer found ubiquitously throughout different tissues and cell types and is comprised of FOS and JUN family heterodimers. Many studies [13,[22][23][24], including our own [9,17,25], have indicated that cFos/cJun AP-1 heterodimers act as critical regulators of MMP1 and MMP13 in chondrocytes following procatabolic stimuli such as IL-1+OSM. Interestingly, despite the importance of cFos (gene name FOS), differential regulation of MMP1 and MMP13 via specific cell signalling pathways [8, 9, 13-15, 24, 26] suggests more complex mechanism(s) underpin the transcriptional activation of these MMPs. In this context, we have recently shown that MMP13 expression is initially modulated by a transient expression of cFos, which subsequently results in the prolonged expression of activating transcription factor 3 (ATF3) to promote MMP13 expression via an AP-1 cis element within the MMP13 proximal promoter [10].
In order to expand our understanding of MMP1 regulatory mechanisms, MMP1 regulation was assessed following the same potent pro-catabolic stimulus as detailed above for MMP13. Data herein indeed confirm that cFos silencing was sufficient to significantly inhibit MMP1 expression; however, ATF3 silencing did not affect MMP1 expression. Importantly, IL-1 +OSM stimulation induced cysteine-serine-rich nuclear protein-1 (CSRNP1; also termed Axin upregulated-1 (AXUD1)), expression to selectively modulate MMP1 in human chondrocytes. CSRNP1 was first identified as being induced by the β-catenin regulator, AXIN1 [27], and has subsequently been characterised as a transcription factor having a role in the regulation of Wnt signalling [28]. In this respect the Wnt/β-catenin signalling cascade pathway has recently been shown to play an important regulator in the onset and progression of OA [29]. This is the first report indicating a role for CSRNP1 in modulating the expression of a chondrogenic collagenase due to an inflammatory insult, and may be indicative of diverse regulatory moieties affecting specific MMP gene transcription as a consequence of inflammation associated with OA progression.

Materials
All chemicals were of the highest purity available and obtained from Sigma Chemical Co (Poole, UK) unless otherwise stated. All cytokines were recombinant human. IL-1α was a generous gift from Dr Keith Ray (GlaxoSmithKline, Stevenage, UK). OSM was prepared in-house as described [30]. siRNA reagents were screened for toxicity using the Toxilight assay of adenylate kinase release (Lonza, Wokingham, UK).

Chondrocytes
Human chondrocytes were isolated by enzymatic digestion of macroscopically normal articular cartilage from OA patients undergoing joint replacement surgery as described [31]. All subjects gave informed consent and the study was approved by the Newcastle and North Tyneside Joint Ethics Committee. Chondrocytes were maintained in Dulbecco's modified Eagle's medium supplemented with 10% foetal bovine serum, 100 IU/mL penicillin, 100 μg/mL streptomycin, 40 U/mL nystatin.

Gene expression analyses
RNA was typically stabilised in cell lysates in a 96-well format and cDNA synthesised using the Cells-to-Signal kit (Life Technologies) as directed. SYBR Green real-time PCR or Taqman assays (using Universal Probe Library probes (Roche Applied Sciences)), referred to herein as qPCR, were conducted using primers and conditions described previously [8][9][10]

Chromatin immunoprecipitation (ChIP)
ChIP experiments were performed according to the standard protocol detailed in the EZ-ChIP kit (Merck-Millipore). Briefly, human chondrocytes were cultured until 70-80% confluent at which point they were treated with IL-1 (0.05 ng/mL) in combination with OSM (10 ng/mL) for various durations. Cells were then subject to cross-linking with formaldehyde for 5 min with agitation, and the reaction quenched with 0.125 M glycine for 5 min (with agitation). Cells were then washed twice and scraped into PBS supplemented with protease inhibitors. Cells were pelleted, resuspended in lysis buffer and lysates sonicated in a Bioruptor Plus (Diagenode, Liege, Belgium) sonicating water bath (15 cycles of 30 sec on, 30 sec off at full power) to shear the chromatin. Samples were then pre-cleared with the appropriate agarose beads for 1 h and subject to an overnight immunoprecipitation with relevant antibodies (10 μg) or corresponding isotype controls; anti-phospho-RNA polymerase II CTD repeats YSPTSPS (Ser5; pRNA Pol II) was from Abcam (Cambridge, UK); anti-acetyl(Lys 9/14 )-histone H3 (AcH3; cat no. 06-599) was from Millipore (Dundee, UK). Antibody/protein/DNA complexes were extracted with agarose beads which were washed sequentially with kit-supplied buffers. DNA was then eluted and crosslinks reversed (4 h incubation at 65˚C with 5 M NaCl, and then 1 h incubation at 37˚C with RNase (1 IU) and proteinase K (1 IU) to degrade RNA and protein, respectively). DNA was purified using spin columns and then used as input for SYBR green qPCR using primers outlined above. Data were normalised to background using isotype control antibodies and variations in amount of genomic DNA were normalised using Input.

Cell fractionation and immunoblotting
Chondrocyte lysates were prepared as described previously [8,31]. In some experiments chondrocytes were subjected to subcellular fractionation using NE-PER Nuclear and Cytoplasmic Protein Extraction Kit or Subcellular Protein Fractionation Kit (both from ThermoFisher Scientific, Loughborough, UK). Lysates or fractions were resolved by SDS-PAGE, transferred to PVDF membranes and subsequently probed with the following antibodies: ATF3 (sc188) and CSRNP1 (sc81191) (used on cellular fractionation and experiments utilising nuclear lysates) purchased from Santa Cruz Biotechnology (Santa Cruz, CA); and CSRNP1 (ab103708) (used on whole cell lysates) from Abcam (Cambridge, UK). The specificity of all antibodies was confirmed using chondrocyte lysates (see full-length blots in S1 Fig), whilst blots were cropped for clarity of comparison.

DNA affinity precipitation assays (DAPA)
Chondrocyte nuclear lysates were generated as described above. Nuclear protein (100-300 μg) was incubated with double stranded biotinylated oligonucleotides (45 pmol) containing either the proximal AP-1 binding site for MMP1 in 500 μl of binding buffer (12 mM HEPES pH 7.9, 4 mM Tris-HCl, 60 mM KCl, 5% (w/v) glycerol, 0.5 mM EDTA, 1 mM DTT and 1 mini protease inhibitor cocktail tablet/10 mL of buffer) for 1 h at 4 C with rotation. Streptavidin-coated magnetic dynabeads (ThermoFisher Scientific) were then added and incubated for 2 h with rotation at 4˚C. Dynabead-bound protein/DNA complexes were isolated using a Dynabead magnet (ThermoFisher Scientific). Beads were resuspended in 50 μl of SDS-PAGE sample buffer, incubated at 100˚C for 5 min, pelleted and supernatants removed for immunoblotting (as above) except that each individual blot was cut to provide two blots with proteins of molecular mass �35kDa or �35kDa for CSRNP1 and ATF3 probing, respectively. Specificity of binding was determined in the presence of 50x excess non-biotinylated AP-1 oligonucleotides.

Statistical analyses
Statistical differences between sample groups were assessed using one-way analysis of variance (ANOVA) with a post-hoc Bonferroni's multiple comparison test or Student's 2-tailed unpaired t-test, where ��� p<0.001, �� p<0.01, � p<0.05. For clarity, only selected comparisons are presented in some figures.

MMP1 gene expression is temporally delayed in cytokine-stimulated chondrocytes
We have previously shown that the potent cytokine combination of IL-1+OSM demonstrates a time-dependency for collagenolytic MMP induction, typically reaching maximal induction at 24 h [6,7,35]. This dependency is confirmed here where primary human articular chondrocytes expressed MMP1 mRNA within 6 h post-stimulation ( Fig 1A) in line with previous findings [7]. Importantly, nascent hnRNA mirrored the observed mRNA increase (Fig 1B) indicating MMP1 transcription occurred independently of mRNA synthesis/stability. Furthermore, previous data indicate that cFOS is rapidly, but transiently, expressed following IL-1+OSM stimulation [10,25]; employing ChIP assays herein cFos was observed to be enriched at the proximal AP-1 element of the MMP1 promoter at 1 h post-stimulation (Fig 2A and 2B). However, ChIP assays also confirmed enhanced enrichment of phosphorylated RNA Pol II to the proximal MMP1 promoter 24 h post-stimulation, but not at 1 h (Fig 2C and 2D), whilst the time-dependent recruitment of acetylated histone H3 to the MMP1 promoter (Fig 2E) reflected the observed time course of MMP1 mRNA expression. Our previous data indicated a requirement for new protein synthesis for MMP13 induction [10] by employing the protein synthesis inhibitor emetine. Similarly, in this study chondrocytes were stimulated for 24 h with IL-1+OSM and emetine added subsequently for varying durations throughout the stimulation which significantly reduced cytokine-induced MMP1 expression and persisted for at least 6-8 h post-stimulation (Fig 3).
Given the discord between the transient increase of cFos by IL-1+OSM [9,10,25] and the requirement for new protein synthesis at times after cFos expression in terms of MMP1 expression, these findings strongly indicated the requirement of additional transcriptional regulator(s) as shown for the expression of MMP13 [10]. Employing a GEO-genome-wide analysis of IL-1+OSM-induced genes in chondrocytes (accession file GSE86578), we further confirmed that the temporal expression profile of MMP1 was maximal at 24 h (Fig 4A) whilst qPCR confirmed significant expression at 24h in these populations (Fig 4B). Moreover, this gene array analysis identified a number of transcriptional regulatory genes with an expression profile similar to ATF3 which has been shown to be an important AP-1-binding factor for MMP13 expression [10]. A siRNA screen of these regulators, including ATF3, to assess MMP1 repression confirmed a requirement for both FOS (cFos) and JUN (cJun) but not ATF3 for MMP1 expression (Fig 4C). Notably, only silencing of CSRNP1 selectively and significantly repressed MMP1 with no significant inhibition of MMP13 (S2 Fig).

Sustained CSRNP1 expression contributes to MMP1 transcription
The siRNA data above implicated a selective role for CSRNP1 in contributing to the transcriptional activation of MMP1 (but not MMP13). Furthermore, expression profiling also revealed that CSRNP1 expression was evident by 1 h and peaked at 1.25 h (Fig 5A and 5B). In order to correlate CSRNP1 expression with FOS, a detailed time-course was performed which showed peak CSRNP1 expression to follow maximal FOS expression (Fig 5C). Importantly, CSRNP1 protein was present from 1 to 24 h (Fig 6A-6C) and appeared to be exclusively nuclear (both soluble and chromatin-bound fractions) following subcellular fractionation (Fig 6C), peaking at 3 h (Fig 6B and 6D).
In silico analyses using the ENCODE database [34] indicated that a consensus-like CSRNP1 motif AGAGTN [33] was present within the proximal ChIPSeq AP-1 (cJun) binding site within the MMP1 promoter, with a 1 bp mismatch. We therefore performed DAPA analyses which, importantly, confirmed specific CSRNP1 binding to this sequence, compared to the equivalent motif within MMP13 (Fig 7), thus further confirming the selective specificity of CSRNP1 for MMP1. Finally, we also assessed the ability of the MMP1/13 AP-1 sequences to bind ATF3 and only observed binding to the MMP13 AP-1 cis element (Fig 7) in line with previous findings [10].

Discussion
Recent evidence indicates that IL-1+OSM-induces MMP13 expression via transcriptional modulation by initial AP-1 (cFos) expression and then by subsequent ATF3 transcriptional activity [10]. In this context, the transcriptional regulation of MMP1, which is also potently induced by this stimulus [6][7][8][9][35][36][37], has remained undefined. Although MMP-13 is the most potent collagenolytic proteinase against type II collagen [38], the abundance of MMP-1 compared to MMP-13 following pro-inflammatory stimuli [35,37,39] also results in significant cleavage of type II collagen resulting in irreversible cartilage damage [40]. Since MMP1 and MMP13 appear to be differentially regulated following IL-1+OSM stimulation [8,9,26], in the study presented herein we sought to identify transcriptional modulators that are selective for MMP1 in comparison to our previous observations for MMP13 [10].
Employing ChIP assays with cFos, phosphorylated RNA Pol II and acetylated histone H3 antibodies, we demonstrated that cFos transcriptional control was similarly important for regulating MMP1 expression as observed for MMP13 [10]. cFOS was recruited to the MMP1 promoter 1 hour post-stimulation, however pRNA pol II was not detected at this timepoint, rather being recruited to the promoter 24 hours post-stimulation. These data indicate that despite cFOS recruitment to the MMP1 promoter, MMP1 transcription does not occur at this early time point. Furthermore, the inhibition of protein synthesis after IL-1+OSM stimulation demonstrates that when protein synthesis is inhibited within the first 4 hours of stimulation, no MMP1 expression is observed 24 hours post-stimulation. However if protein synthesis is not inhibited until after 6 hours of IL-1+OSM, MMP1 expression is observed following 24 hour stimulation. The requirement for the de novo synthesis of transcriptional regulators following cell activation has been previously reported in the literature [10,41,42]. Taken together, these data strongly suggest that MMP1 expression is dependent on de novo protein synthesis of additional factors post-FOS but prior to MMP1 expression.
In order to identify such factors, GEO-genome-wide microarray analyses with subsequent siRNA assays were performed. These studies identified several transcriptional regulators which had an elevated expression level 1-1.25 hours post IL-1+OSM stimulation. In comparing the effect of siRNA knock down of these transcriptional regulators on MMP expression [10], only CSRNP1 was found to be selective and specific for the regulation of MMP1 expression over MMP13. This study indicated that MMP1 expression was indeed dependent on cFos but that other FOS family members also affected MMP1 expression via promoting AP-1-dependent gene expression. In this respect, FOSL1 and FOSB were induced (see S3 Fig) and gene silencing of both resulted in significant MMP1 repression. These analyses are also in  [15,[43][44][45]. We show here that siRNA silencing of RELA leads to a significant decrease in MMP1 expression although to a much lesser extent than seen for FOS or FOSL1.  Little is known about the role of CSRNP1 as a transcriptional regulator. Here we provide the first evidence that CSRNP1 has an important modulatory role in affecting cytokineinduced MMP1 expression in human articular chondrocytes. The consensus sequence for CSRNP1 binding is an AP-1-like binding motif 5'-AGAGTN-3' [33] and both MMP1 and MMP13 have similar AP-1-like binding motifs within their proximal promoter regions. The MMP1 sequence is 5'-TGAGTCA-3' and the MMP13 sequence has a C instead of a G (underlined) resulting in 5'-TGACTCA-3'. This G to C switch indicates preferential binding of CSRNP1 to the MMP1 compared to the MMP13 AP-1-like cis element as observed via DAPA analyses. This may be indicative that the proximal MMP1 AP-1-like sequence is a specific cis element capable of binding CSRNP1 in chondrocytes. Furthermore, in silico analysis indicates the presence of other putative CSRNP1 binding sites located within the MMP1 proximal promoter, but not within MMP13, that may also contribute to the expression of MMP1. Since no evidence for reproducible ATF3 binding to the MMP1 AP-1 cis element was observed employing DAPAs in contrast to MMP13, and gene silencing of ATF3 resulted in no significant decrease in MMP1 expression, MMP1 appears unlikely to be an ATF3-responsive gene. Thus, CSRNP1 and ATF3 may respectively have mutually exclusive roles in the cytokine-mediated expression of MMP1 and MMP13 in chondrocytes.
In conclusion, our data presented here indicate that a potent catabolic stimulus such as IL-1 +OSM induces the expression of specific transcriptional regulators that contribute to MMP1 gene activation which may well reflect the markedly enhanced levels of expression that are observed compared to either cytokine alone [35,39]. Since multiple pro-inflammatory mediators are likely to be prevalent in most inflammatory situations, this study further highlights the complexity of MMP gene regulation that occurs in human chondrocytes when subjected to inflammatory cues.
Supporting information S1 Fig. Confirmation of antibody specificity using human chondrocyte lysates. The specificity of each antibody used in the study was confirmed using nuclear lysates prepared as described in the Methods from primary human articular chondrocytes either unstimulated or Cytokine-induced cysteine-serine-rich nuclear protein-1 (CSRNP1) and MMP1 expression stimulated with IL-1 (0.2 ng/mL) in combination with OSM (10 ng/mL). Following SDS-PAGE, proteins were transferred to PVDF membranes and probed with the indicated antibodies. Full-length blots are presented to highlight the specific immuno-reactivity of each antibody with an arrow indicating the expected molecular mass.