Advertisement

WWP2 confers risk to osteoarthritis by affecting cartilage matrix deposition via hypoxia associated genes

Open AccessPublished:October 05, 2022DOI:https://doi.org/10.1016/j.joca.2022.09.009

      Summary

      Objective

      To explore the co-expression network of the osteoarthritis (OA) risk gene WWP2 in articular cartilage and study cartilage characteristics when mimicking the effect of OA risk allele rs1052429-A on WWP2 expression in a human 3D in vitro model of cartilage.

      Method

      Co-expression behavior of WWP2 with genes expressed in lesioned OA articular cartilage (N = 35 samples) was explored. By applying lentiviral particle mediated WWP2 upregulation in 3D in vitro pellet cultures of human primary chondrocytes (N = 8 donors) the effects of upregulation on cartilage matrix deposition was evaluated. Finally, we transfected primary chondrocytes with miR-140 mimics to evaluate whether miR-140 and WWP2 are involved in similar pathways.

      Results

      Upon performing Spearman correlations in lesioned OA cartilage, 98 highly correlating genes (|ρ| > 0.7) were identified. Among these genes, we identified GJA1, GDF10, STC2, WDR1, and WNK4. Subsequent upregulation of WWP2 on 3D chondrocyte pellet cultures resulted in a decreased expression of COL2A1 and ACAN and an increase in EPAS1 expression. Additionally, we observed a decreased expression of GDF10, STC2, and GJA1. Proteomics analysis identified 42 proteins being differentially expressed with WWP2 upregulation, which were enriched for ubiquitin conjugating enzyme activity. Finally, upregulation of miR-140 in 2D chondrocytes resulted in significant upregulation of WWP2 and WDR1.

      Conclusions

      Mimicking the effect of OA risk allele rs1052429-A on WWP2 expression initiates detrimental processes in the cartilage shown by a response in hypoxia associated genes EPAS1, GDF10, and GJA1 and a decrease in anabolic markers, COL2A1 and ACAN.

      Keywords

      Introduction

      Globally, osteoarthritis (OA) is a highly prevalent and disabling joint disease which confers high social and economic burden to society. Risk factors for OA include sex, abnormal joint loading, obesity, metabolic diseases, and genetic factors
      • Vina E.R.
      • Kwoh C.K.
      Epidemiology of osteoarthritis: literature update.
      . To discover genes and underlying disease pathways, large genome wide association meta-analyses have been performed and multiple robust single nucleotide polymorphisms (SNPs) were identified significantly conferring risk to initiation and progression of OA
      • Styrkarsdottir U.
      • Lund S.H.
      • Thorleifsson G.
      • Zink F.
      • Stefansson O.A.
      • Sigurdsson J.K.
      • et al.
      Meta-analysis of Icelandic and UK data sets identifies missense variants in SMO, IL11, COL11A1 and 13 more new loci associated with osteoarthritis.
      • Zengini E.
      • Hatzikotoulas K.
      • Tachmazidou I.
      • Steinberg J.
      • Hartwig F.P.
      • Southam L.
      • et al.
      Genome-wide analyses using UK Biobank data provide insights into the genetic architecture of osteoarthritis.
      • Boer C.G.
      • Hatzikotoulas K.
      • Southam L.
      • Stefánsdóttir L.
      • Zhang Y.
      • Coutinho de Almeida R.
      • et al.
      Deciphering osteoarthritis genetics across 826,690 individuals from 9 populations.
      . Similar to other complex traits, these risk alleles have subsequently been found to affect expression of positional genes in cis in disease relevant tissues, also known as allelic imbalance (AI)
      • Gee F.
      • Clubbs C.F.
      • Raine E.V.
      • Reynard L.N.
      • Loughlin J.
      Allelic expression analysis of the osteoarthritis susceptibility locus that maps to chromosome 3p21 reveals cis-acting eQTLs at GNL3 and SPCS1.
      ,
      • Raine E.V.
      • Dodd A.W.
      • Reynard L.N.
      • Loughlin J.
      Allelic expression analysis of the osteoarthritis susceptibility gene COL11A1 in human joint tissues.
      . Founded by this mechanism, we previously used RNA sequencing data of OA articular cartilage to report on genome-wide AI expression of SNPs in cartilage specific genes, as such, providing an AI expression database to in silico check functional aspects of identified and/or future OA risk SNPs
      • den Hollander W.
      • Pulyakhina I.
      • Boer C.
      • Bomer N.
      • van der Breggen R.
      • Arindrarto W.
      • et al.
      Annotating transcriptional effects of genetic variants in disease-relevant tissue: transcriptome-wide allelic imbalance in osteoarthritic cartilage.
      . One of the top findings was SNP rs1052429 located in the 3′UTR of the WWP2 gene showing highly significant AI, with risk allele rs1052429-A marking higher expression of WWP2 relative to rs1052429-G. Among the OA risk SNPs identified in a large genome-wide meta-analysis of Icelandic and UK knee OA patients was rs34195470, located in WWP2 gene and a proxy of our AI SNP rs1052429 (r2 = 0.6)
      • Styrkarsdottir U.
      • Lund S.H.
      • Thorleifsson G.
      • Zink F.
      • Stefansson O.A.
      • Sigurdsson J.K.
      • et al.
      Meta-analysis of Icelandic and UK data sets identifies missense variants in SMO, IL11, COL11A1 and 13 more new loci associated with osteoarthritis.
      . Recently, rs34195470 was confirmed being OA risk SNP in the largest genome-wide meta-analysis so far, including individuals from 9 populations
      • Boer C.G.
      • Hatzikotoulas K.
      • Southam L.
      • Stefánsdóttir L.
      • Zhang Y.
      • Coutinho de Almeida R.
      • et al.
      Deciphering osteoarthritis genetics across 826,690 individuals from 9 populations.
      . Based on these data, we could make a firm hypothesis that WWP2, with risk alleles rs34195470-G and rs1052429-A, confers robust risk to human OA which is marked by increased expression of WWP2. We also previously identified transcription of WWP2 in cartilage being epigenetically regulated
      • den Hollander W.
      • Ramos Y.F.
      • Bomer N.
      • Elzinga S.
      • van der Breggen R.
      • Lakenberg N.
      • et al.
      Transcriptional associations of osteoarthritis-mediated loss of epigenetic control in articular cartilage.
      , as well as being responsive in the OA pathophysiological process
      • Coutinho de Almeida R.
      • Ramos Y.F.M.
      • Mahfouz A.
      • den Hollander W.
      • Lakenberg N.
      • Houtman E.
      • et al.
      RNA sequencing data integration reveals an miRNA interactome of osteoarthritis cartilage.
      . Moreover, WWP2 was previously shown to be a marker for hypertrophic chondrocytes in OA knee joints
      • Ji Q.
      • Zheng Y.
      • Zhang G.
      • Hu Y.
      • Fan X.
      • Hou Y.
      • et al.
      Single-cell RNA-seq analysis reveals the progression of human osteoarthritis.
      .
      WWP2 is a member of the Nedd4 superfamily, a small group within the E3 ubiquitin ligase enzymes and is involved in post-translational modifications. The WWP2 protein contains four double tryptophan (WW) domains, which allow specific protein–protein interactions and it is expressed in multiple organs throughout the body
      • Chantry A.
      WWP2 ubiquitin ligase and its isoforms: new biological insight and promising disease targets.
      . More specifically to cartilage, Nakamura et al.
      • Nakamura Y.
      • Yamamoto K.
      • He X.
      • Otsuki B.
      • Kim Y.
      • Murao H.
      • et al.
      Wwp2 is essential for palatogenesis mediated by the interaction between Sox9 and mediator subunit 25.
      showed that WWP2 interacts with SOX9 to form a complex that facilitates nuclear translocation of SOX9, as such enabling SOX9 transcriptional activity. Despite the association between the risk allele and higher expression levels of WWP2 in human cartilage, the effect of WWP2 knockout (KO) in mice with age-related and surgically induced models of OA showed that lack of WWP2 expression resulted in increased expression of catabolic cartilage markers RUNX2 and ADAMTS5
      • Mokuda S.
      • Nakamichi R.
      • Matsuzaki T.
      • Ito Y.
      • Sato T.
      • Miyata K.
      • et al.
      Wwp2 maintains cartilage homeostasis through regulation of Adamts5.
      . In a different context, WWP2 was found to be a host gene for microRNA-140 (miR-140), a miRNA highly expressed in cartilage and shown to be differentially expressed between preserved and lesioned OA cartilage
      • Coutinho de Almeida R.
      • Ramos Y.F.M.
      • Mahfouz A.
      • den Hollander W.
      • Lakenberg N.
      • Houtman E.
      • et al.
      RNA sequencing data integration reveals an miRNA interactome of osteoarthritis cartilage.
      . As such, it was suggested that expression of miR-140 and the C-terminal transcript of WWP2 (WWP2-C, also called WWP2 isoform 2) are co-regulated
      • Yang J.
      • Qin S.
      • Yi C.
      • Ma G.
      • Zhu H.
      • Zhou W.
      • et al.
      MiR-140 is co-expressed with Wwp2-C transcript and activated by Sox9 to target Sp1 in maintaining the chondrocyte proliferation.
      ,
      • Rice S.J.
      • Beier F.
      • Young D.A.
      • Loughlin J.
      Interplay between genetics and epigenetics in osteoarthritis.
      .
      In the current study, we set out to explore how increased levels of WWP2, conferring risk to OA, affect cartilage matrix. To get gain insight in the WWP2 pathway, we started with exploring a WWP2 co-expression network in our previous whole-transcriptome OA cartilage dataset
      • Coutinho de Almeida R.
      • Ramos Y.F.M.
      • Mahfouz A.
      • den Hollander W.
      • Lakenberg N.
      • Houtman E.
      • et al.
      RNA sequencing data integration reveals an miRNA interactome of osteoarthritis cartilage.
      . Moreover, to study the effect of the genetic risk allele (increased levels of WWP2), we functionally assessed the effect of lentiviral-mediated upregulation of WWP2 in a 3D in vitro model using primary human chondrocytes. Apart from conventional anabolic and catabolic cartilage markers, genes identified in the WWP2 co-expression network were used as a read-out to evaluate the effect of WWP2 upregulation. Since WWP2 is involved in post-translational modifications, we explored the effect of WWP2 upregulation on protein level by performing proteomic analysis. Finally, we explored the effects of upregulation of miR-140 in primary chondrocytes by transfection with miR-140 mimics.

      Methods

      Sample description

      All material included in this study is obtained as part of the Research Arthritis and Articular Cartilage (RAAK) study. The RAAK-study is aimed at biobanking of joint materials of patients who underwent a total joint replacement surgery due to OA. Classification of macroscopically preserved and lesioned OA cartilage was done as described previously
      • Ramos Y.F.
      • den Hollander W.
      • Bovee J.V.
      • Bomer N.
      • van der Breggen R.
      • Lakenberg N.
      • et al.
      Genes involved in the osteoarthritis process identified through genome wide expression analysis in articular cartilage; the RAAK study.
      . For all sample characteristics see Table S1. The RAAK-study is approved by the medical ethics committee of the Leiden University Medical Center (P08.239/P19.013).

      RNA-sequencing

      Lesioned OA cartilage was collected from hip and knee joints (N = 35 samples), snap frozen in liquid nitrogen, pulverized and homogenized in TRIzol (Invitrogen), and RNA was isolated using Qiagen RNeasy Mini Kit (Qiagen). Paired-end 2 × 100 bp RNA-sequencing (Illumina HiSeq2000 and Illumina HiSeq4000) was performed. Data from both Illumina platforms were integrated and analyzed with the same in-house pipeline. Additional details are described in Supplementary methods.

      Creating a co-expression network

      We explored co-expression behavior of WWP2 with progression of OA by correlating (Spearman correlation) WWP2 expression levels in our RNA sequencing dataset with expression levels of all genes expressed in OA articular cartilage (N = 20,048 genes)
      • Coutinho de Almeida R.
      • Ramos Y.F.M.
      • Mahfouz A.
      • den Hollander W.
      • Lakenberg N.
      • Houtman E.
      • et al.
      RNA sequencing data integration reveals an miRNA interactome of osteoarthritis cartilage.
      . To correct for multiple testing, the Benjamini-Hochberg method was used, as indicated by the false discovery rate (FDR), with a significance cutoff value of 0.05. To include the most informative genes a threshold of |ρ| > 0.7 and FDR < 0.05 were selected, corresponding to approximately the top 1% of the total significant correlations.

      Lentiviral transduction

      The full length WWP2 plasmid was digested and inserted into the XhoI/XbaI sites of the pLV-CMV-IRES-eGFP lentiviral backbone (kindly provided by Prof. Dr. Hoeben). The pLV-CMV-IRES-eGFP lentiviral backbone without the WWP2 insert was used as a control. Additional details are available in the Supplementary methods.

      In vitro 3D pellet cultures

      3D pellet cultures were formed by adding 2.5 × 105 cells in their expansion medium to a 15 ml Falcon tube and subsequently expose them to centrifugal forces (1,200 rpm, 4 min). Chondrogenesis was initiated in serum-free chondrogenic differentiation medium. Additional details are available in the Supplementary methods.

      RT-qPCR

      RNA was isolated from the samples using the RNeasy Mini Kit (Qiagen). cDNA synthesis was performed using the First Strand cDNA Synthesis Kit (Roche Applied Science). Subsequently, RT-qPCR was performed adjusting for housekeeping genes GAPDH and SDHA. Additional details are available in the Supplementary methods.

      Quantitative proteomics using TMT labeling

      Lysis, digestion, TMT labeling and mass spectrometry analysis was essentially performed as described previously
      • den Hollander W.
      • Ramos Y.F.
      • Bomer N.
      • Elzinga S.
      • van der Breggen R.
      • Lakenberg N.
      • et al.
      Transcriptional associations of osteoarthritis-mediated loss of epigenetic control in articular cartilage.
      . All searches and subsequent data analysis, including Percolator and abundance ratio calculation, were performed using Proteome Discoverer 2.4 (Thermo Scientific). Additional details are available in the Supplementary methods.

      Histochemistry

      Sections of the 3D chondrocyte pellet culutes were stained for glycosaminoglycan (GAG) deposition using the Alcian Blue staining. The staining was quantified using Fiji. Additional details are available in Supplementary methods.

      Transfection with miR-140 mimics

      Primary chondrocytes were transfected with hsa-miR-3p mimic (Invitrogen) or a control mimic at 5 nM final concentration using Opti-MEM (Gibco) and Lipofectamine RNAiMax Transfection reagent (Invitrogen) according to manufacturer's protocol. Additional details are available in Supplementary methods.
      An overview of the applied strategy can be seen in Fig. 1. The RNA sequencing data of the articular cartilage is deposited at ArrayExpress (E-MTAB-7313). Further data generated and used in this study is not openly available due to reasons of sensitivity and are available from the corresponding author upon reasonable request.
      Fig. 1
      Fig. 1Schematic overview of applied strategy.
      Fig. 2
      Fig. 2mRNA expression levels of cartilage matrix markers (A) and cartilage degeneration markers (B) for pellets with WWP2 upregulation and their controls at day 7 (N = 11–12 pellet cultures, N = 7 donors). (C) Alcian Blue staining visualizing GAGs deposition in WWP2 overexpressed pellets and their controls after 7 days of culturing (N = 26 pellet cultures, N = 8 donors). The scale bar indicates 50 um. Ns: not significant, ∗P < 0.05, ∗∗P < 0.005 upon performing a Paired sample t-test.

      Results

      Co-expression network of WWP2

      To identify genes that are regulated by, or co-expressed with, WWP2 in OA cartilage, we used RNA sequencing data of lesioned OA cartilage [N = 35 samples, Table S1(A)] to perform Spearman correlation between expression levels of WWP2 and genes expressed in cartilage [N = 20,048 genes, Table S2, Fig. 1(A)]. We identified 98 genes highly correlating (|ρ| > 0.7) to WWP2. These 98 genes were significantly enriched for, amongst others, GO-terms Extracellular exosome (GO:0070062, 36 genes), characterized by expression of GJA1 (encoding gap junction alpha 1), SMO (encoding smoothened frizzled class receptor), and WDR1 (encoding WD repeat domain 1), and Myelin sheath (GO:0043209, 10 genes), characterized by expression of WDR1, RALA (encoding RAS like proto-oncogene A), and CCT5 (encoding chaperonin containing TCP1 subunit 5) (Table S3). As shown in Fig. S1, genes highly correlating to WWP2 (N = 98) formed a highly interconnected network, i.e., genes that are all highly correlating with each other. In this network we identified direct and indirect relations with WWP2, including GJA1 (ρ = −0.81, 70 connections, i.e., highly correlating to 70 genes in the network), WNK4 (encoding WNK lysine deficient protein kinase 4, ρ = 0.81, 37 connections), ACAN (encoding aggrecan, ρ = 0.78, 16 connections), and STC2 (encoding stanniocalcin 2, ρ = 0.77, 17 connections).

      Lentiviral particle-mediated upregulation of WWP2

      The effect of upregulation of WWP2 was studied on cartilaginous matrix deposition in in vitro 3D chondrocyte pellet cultures, by creating a lentiviral particle mediated upregulation of WWP2. 3D pellet cultures were harvested after three or seven days of culturing and gene (N = 16 pellet cultures of N = 8 donors) and protein (N = 16 pellet cultures of N = 4 donors) expression levels were measured [Table S1(B)]. First, we confirmed whether WWP2 upregulation was successful by measuring both gene and protein expression levels at day zero of the 3D chondrocyte pellet culture, and we observed a significant increase in WWP2 gene expression levels [P = 1.0 × 10−5, Figs. S2(A)–(B)], which was confirmed on protein level [P = 3.2 × 10−6, Fig. S2(C)].

      Effect of WWP2 upregulation on cartilage matrix deposition

      Next, we evaluated effect of WWP2 upregulation on expression levels of conventional cartilage genes during 3D pellet culture of seven days [Fig. 1(B)]. As shown in Fig. 2(A), we found significant reduced gene expression of ACAN (FC = 0.80, P = 0.04) and COL2A1 (encoding collagen type 2 alpha chain 1, FC = 0.77, P = 0.01), in WWP2 upregulated pellets compared to their controls at day seven (Table S4). Moreover, we showed significant increased gene expression of degeneration markers EPAS1 (encoding endothelial PAS domain protein 1, FC = 1.56, P = 0.004) [Fig. 2(B)]. Notably, SOX9, ADAMTS5¸ and RUNX2, which were previously linked to WWP2 function, were not consistently changed upon WWP2 upregulation. Moreover, we stained 3D pellet cultures for presence of glycosaminoglycans (GAGs) using Alcian Blue staining, and observed a trend towards decreased Alcian Blue intensity when comparing WWP2 upregulated pellets with controls [Fig. 2(C)].

      Effect of WWP2 upregulation on genes correlated to WWP2

      To investigate functional relationships between WWP2 and identified correlating and highly interconnected genes, we selected GDF10 (encoding growth differentiation factor 10, ρ = 0.72, 18 connections), STC2 (ρ = 0.77, 17 connections), GJA1 (ρ = −0.81, 70 connections), WDR1 (encoding WD repeat-containing protein 1, ρ = −0.70, 5 connections), and WNK4 (ρ = 0.81, 37 connections) from the network (Fig. 1) to use as read-out of WWP2 upregulation in 3D chondrocyte pellet cultures [Fig. 1(B)]. As shown in Fig. 3, we observed significant decreased gene expression of GDF10 (FC = 0.62, P = 0.002) and STC2 (FC = 0.73, P = 0.04) with upregulation of WWP2. Albeit not significant, gene expression of GJA1 (FC = 0.76, P = 0.08) was also consistently lower in WWP2 upregulated pellets. Together, these data suggest that GDF10, STC2, and GJA1 are downstream of WWP2 either by direct or indirect activity. In contrast, WNK4 and WDR1 did not show consistent changes in expression with upregulation of WWP2, suggesting WNK4 and WDR1 are rather upstream in the pathway of WWP2.
      Fig. 3
      Fig. 3mRNA expression levels of genes correlating with WWP2 in WWP2 upregulated 3D chondrocyte pellet cultures compared to their controls after 7 days of culturing (N = 6–12 pellet cultures, N = 7 donors). Ns: not significant, ∗P < 0.05, ∗∗P < 0.05 upon performing a Paired sample t-test.

      Proteomics

      To study the extent to which gene expression levels translate to protein levels, we performed proteomics analysis. Prior to differential expression analysis of pellet cultures with and without WWP2 upregulation, we explored protein expression levels of cartilage markers in our control pellet cultures at day three and day seven of 3D pellet culture. Upon comparing day three and day seven with day zero of control pellets, we observed increased protein expression of cartilage markers COL2A1 (FC = 2.68 and FC = 32.94, respectively), ACAN (FC = 5.28 and FC = 13.75, respectively), COMP (cartilage oligomeric matrix protein, FC = 5.31 and FC = 20.25, respectively), and FN1 (fibronectin, FC = 2.12 and FC = 3.51, respectively) (Table S5, Fig. S3). Moreover, mesenchymal markers CD44 (FC = 0.83 and FC = 0.46, respectively) and CD166 (FC = 0.79 and FC = 0.62, respectively) and IGFBP3 (insulin growth factor binding protein 3, FC = 0.10 and FC = 0.10, respectively) were downregulated on both days. Together, this indicates that cartilage-like matrix is produced by chondrocytes already at day three, but is increasing towards day seven. Notably, SOX9 was not detected in the proteomics analysis.
      Next, we evaluated the effect of WWP2 upregulation on cartilage matrix deposition on protein level [Fig. 1(B)]. Since we observed increased protein expression of cartilage markers in control pellet cultures already on day three [Fig. S3(A)], we pooled day three and day seven for further analysis to increase power. Upon comparing pellet cultures with and without WWP2 upregulation, we found WWP2 still being significantly upregulated after three and seven days of culturing. Furthermore, we found 42 proteins significantly differentially expressed [Fig. 4(A), Table S6], of which GJA1 (FC = 0.54) was most significantly downregulated in WWP2 upregulated pellet cultures, confirming the downregulation observed on gene expression level (FC = 0.76). Oppositely, the observed changes in ACAN and COL2A1 gene expression levels were not confirmed on protein level. Proteins encoded by SOX9, EPAS1, RUNX2, and ADAMTS5 were either not identified or did not show unique peptides. Upon performing enrichment analysis on the 42 differentially expressed proteins, we found significant enrichment for ubiquitin conjugating enzyme activity (5 proteins, FDR = 0.002) and ubiquitin-protein transferase activity (6 proteins, FDR = 0.03), both terms characterized by expression of, amongst others, UBE2D4, UBE2L3, and UBE2D1. Furthermore, these 42 proteins showed significant protein–protein interactions [P = 0.02, Fig. 4(B)], also representing ubiquitin conjugating enzyme activity.
      Fig. 4
      Fig. 4(A) Volcano plot of proteins differentially expressed between WWP2 transduced 3D pellet cultures and their controls at day three and day seven together (N = 16 pellets, N = 4 donors). The red dots indicate the significantly differentially expressed proteins. (B) Protein–protein interaction network in STRING.

      miR-140-3p and WWP2

      Since it has been suggested that WWP2 and miR-140 are co-expressed
      • Yang J.
      • Qin S.
      • Yi C.
      • Ma G.
      • Zhu H.
      • Zhou W.
      • et al.
      MiR-140 is co-expressed with Wwp2-C transcript and activated by Sox9 to target Sp1 in maintaining the chondrocyte proliferation.
      ,
      • Rice S.J.
      • Beier F.
      • Young D.A.
      • Loughlin J.
      Interplay between genetics and epigenetics in osteoarthritis.
      , we transfected primary chondrocytes with miR-140-3p mimics, to assess whether this miRNA regulates WWP2 expression or similar genes as involved in the WWP2 co-expression network [N = 7, Table S1(B), Fig. 1(C)]. To investigate whether miR-140-3p regulates WWP2, we first evaluated the effects of miR-140-3p mimic on expression levels of WWP2 full length and WWP2 splice variants isoform 2, isoform 4, and isoform 6 (Fig. S4). MiR-140 is suggested to be co-expressed with splice variant WWP2 isoform 2, also called WWP2-C, as they share the promotor
      • Rice S.J.
      • Beier F.
      • Young D.A.
      • Loughlin J.
      Interplay between genetics and epigenetics in osteoarthritis.
      . Isoform 4 is also known as WWP2-N and is a transcript that does not contain miR-140 (Fig. S4). As shown in Fig. 5(A), we observed significant increased expression levels of WWP2 (FC = 1.22, P = 0.02) with upregulation of miR-140-3p. Moreover, we observed consistent increased expression of WWP2 isoform 6 (FC = 1.29, P = 0.06), while significant decreased expression of isoform 4 (FC = 0.63, P = 0.02) (Table S7). Notably, we did not see effect on expression levels of WWP2 isoform 2, also called WWP2-C. With respect to highly correlated genes, we observed increased expression of WDR1 (FC = 1.79, P = 1.00 × 10−3), one of the genes that was not consistently changed with WWP2 upregulation [Fig. 5(B)]. Albeit not significant, we observed an increased expression of STC2 (FC = 1.55, P = 0.08), which is also contradictory to the effects of WWP2 upregulation. Moreover, we did not observe consistent effects on GJA1 expression levels.
      Fig. 5
      Fig. 5mRNA expression levels upon transfection with miR-140-3p. (A) expression levels of WWP2 and its isoforms. (B) Expression levels of genes correlated to WWP2 (N = 8 wells, N = 4 donors). Ns: not significant, ∗P < 0.05, ∗∗P < 0.005 upon performing a Paired sample t-test.

      Discussion

      By combining a genome-wide screen for cartilage specific AI
      • den Hollander W.
      • Pulyakhina I.
      • Boer C.
      • Bomer N.
      • van der Breggen R.
      • Arindrarto W.
      • et al.
      Annotating transcriptional effects of genetic variants in disease-relevant tissue: transcriptome-wide allelic imbalance in osteoarthritic cartilage.
      and large scale GWAS
      • Styrkarsdottir U.
      • Lund S.H.
      • Thorleifsson G.
      • Zink F.
      • Stefansson O.A.
      • Sigurdsson J.K.
      • et al.
      Meta-analysis of Icelandic and UK data sets identifies missense variants in SMO, IL11, COL11A1 and 13 more new loci associated with osteoarthritis.
      ,
      • Boer C.G.
      • Hatzikotoulas K.
      • Southam L.
      • Stefánsdóttir L.
      • Zhang Y.
      • Coutinho de Almeida R.
      • et al.
      Deciphering osteoarthritis genetics across 826,690 individuals from 9 populations.
      , we hypothesized that upregulated expression of WWP2 confers robust risk to OA. Here, we set out to functionally investigate the role of WWP2 in cartilage by exploring the WWP2 co-expression network in a previously assessed RNA sequencing dataset
      • Coutinho de Almeida R.
      • Ramos Y.F.M.
      • Mahfouz A.
      • den Hollander W.
      • Lakenberg N.
      • Houtman E.
      • et al.
      RNA sequencing data integration reveals an miRNA interactome of osteoarthritis cartilage.
      . Moreover, lentiviral-mediated upregulation of WWP2 was shown to have detrimental effects on cartilage matrix deposition, as shown by downregulation of COL2A1 and ACAN and upregulation of EPAS1. Apart from conventional anabolic and catabolic cartilage markers, genes identified in the WWP2 co-expression network were used as read-out, showing GDF10, STC2, and GJA1 being responsive to WWP2 upregulation. Furthermore, to explore effects of miR-140-3p, that was suggested to be co-regulated with WWP2, we transfected primary chondrocytes with miR-140-3p mimics.
      Based on AI expression of the OA risk SNP rs1052429, we hypothesized that WWP2 confers risk to OA onset by upregulated expression, whereas WWP2 exhibited FDR significantly lower expression in lesioned compared to preserved cartilage (FC = 0.78, FDR = 5.3 × 10−3), together suggesting that lower expression levels of WWP2 in lesioned OA cartilage are rather an attempt of chondrocytes to reverse the OA state than a cause to the OA process
      • Reynard L.N.
      Analysis of genetics and DNA methylation in osteoarthritis: what have we learnt about the disease?.
      ,
      • Rice S.J.
      • Cheung K.
      • Reynard L.N.
      • Loughlin J.
      Discovery and analysis of methylation quantitative trait loci (mQTLs) mapping to novel osteoarthritis genetic risk signals.
      . Concomitantly, co-expression network analyses showed 98 highly and significantly (|ρ| > 0.7, FDR < 0.05) correlating genes to WWP2, including GJA1 (ρ = −0.81), WNK4 (ρ = 0.81), ACAN (ρ = 0.78), and STC2 (ρ = 0.77) (Table S2). Previously, it was shown that WWP2 interacts with SOX9 and that it regulates SOX9 transcriptional activity
      • Nakamura Y.
      • Yamamoto K.
      • He X.
      • Otsuki B.
      • Kim Y.
      • Murao H.
      • et al.
      Wwp2 is essential for palatogenesis mediated by the interaction between Sox9 and mediator subunit 25.
      . Although SOX9 is highly expressed in cartilage, SOX9 was not among the high and significant correlations (ρ = 0.5). On the other hand, SOX9 was previously shown to regulate the expression of, amongst others ACAN
      • Sekiya I.
      • Tsuji K.
      • Koopman P.
      • Watanabe H.
      • Yamada Y.
      • Shinomiya K.
      • et al.
      SOX9 enhances aggrecan gene promoter/enhancer activity and is up-regulated by retinoic acid in a cartilage-derived cell line, TC6.
      ,
      • Nishimura R.
      • Hata K.
      • Takahata Y.
      • Murakami T.
      • Nakamura E.
      • Yagi H.
      Regulation of cartilage development and diseases by transcription factors.
      , which was here shown to be highly correlated to WWP2 (ρ = 0.78).
      Upon studying the effect of upregulation of WWP2, we found EPAS1 and GDF10 being genes that had most consistent and significant changed levels of gene expression. EPAS1, encoding hypoxia-inducible factor 2 alpha, is known for its role in endochondral ossification and is a known cartilage degradation marker in OA
      • Yang S.
      • Kim J.
      • Ryu J.H.
      • Oh H.
      • Chun C.H.
      • Kim B.J.
      • et al.
      Hypoxia-inducible factor-2alpha is a catabolic regulator of osteoarthritic cartilage destruction.
      ,
      • Saito T.
      • Fukai A.
      • Mabuchi A.
      • Ikeda T.
      • Yano F.
      • Ohba S.
      • et al.
      Transcriptional regulation of endochondral ossification by HIF-2alpha during skeletal growth and osteoarthritis development.
      . GDF10, also known as bone morphogenic protein 3, is involved in osteogenesis as it inhibits osteoblast differentiation via SMAD2 and SMAD3
      • Rice S.J.
      • Cheung K.
      • Reynard L.N.
      • Loughlin J.
      Discovery and analysis of methylation quantitative trait loci (mQTLs) mapping to novel osteoarthritis genetic risk signals.
      ,
      • Matsumoto Y.
      • Otsuka F.
      • Hino J.
      • Miyoshi T.
      • Takano M.
      • Miyazato M.
      • et al.
      Bone morphogenetic protein-3b (BMP-3b) inhibits osteoblast differentiation via Smad2/3 pathway by counteracting Smad1/5/8 signaling.
      . The latter also being previously identified as OA susceptibility gene
      • Styrkarsdottir U.
      • Lund S.H.
      • Thorleifsson G.
      • Zink F.
      • Stefansson O.A.
      • Sigurdsson J.K.
      • et al.
      Meta-analysis of Icelandic and UK data sets identifies missense variants in SMO, IL11, COL11A1 and 13 more new loci associated with osteoarthritis.
      . Furthermore, lower expression level of GDF10 was associated with OA severity in both bone and cartilage
      • Chou C.H.
      • Lee C.H.
      • Lu L.S.
      • Song I.W.
      • Chuang H.P.
      • Kuo S.Y.
      • et al.
      Direct assessment of articular cartilage and underlying subchondral bone reveals a progressive gene expression change in human osteoarthritic knees.
      . Interestingly, GDF10 was shown to be a hypoxia inducible gene, like EPAS1, which is regulated by SOX9 and was identified as marker for differentiated chondrocytes as it inhibits adipogenesis and osteogenesis
      • Lafont J.E.
      • Talma S.
      • Hopfgarten C.
      • Murphy C.L.
      Hypoxia promotes the differentiated human articular chondrocyte phenotype through SOX9-dependent and -independent pathways.
      . Both EPAS1 and GDF10 were not in the proteomics analysis, either because they were not measured (EPAS1) or they did not show unique peptides (GDF10). Additionally, we observed a decrease in STC2 and GJA1 expression. Downregulation of GJA1 did not reach statistical significance on gene expression level, while on protein level GJA1 was the most significantly downregulated protein. STC2 is a glycoprotein and upregulation of STC2 in mice has been shown to delay endochondral ossification
      • Gagliardi A.D.
      • Kuo E.Y.
      • Raulic S.
      • Wagner G.F.
      • DiMattia G.E.
      Human stanniocalcin-2 exhibits potent growth-suppressive properties in transgenic mice independently of growth hormone and IGFs.
      ,
      • Chang A.C.
      • Hook J.
      • Lemckert F.A.
      • McDonald M.M.
      • Nguyen M.A.
      • Hardeman E.C.
      • et al.
      The murine stanniocalcin 2 gene is a negative regulator of postnatal growth.
      . Moreover, it was shown that STC2 was higher expressed in healthy cartilage compared to osteophytic cartilage
      • Gelse K.
      • Ekici A.B.
      • Cipa F.
      • Swoboda B.
      • Carl H.D.
      • Olk A.
      • et al.
      Molecular differentiation between osteophytic and articular cartilage--clues for a transient and permanent chondrocyte phenotype.
      , suggesting its potential role in initiation and progression of OA in presence of higher WWP2 expression. GJA1, also known as connexin 43, is a major protein of functional gap junctions which allows for cell–cell communication. More specific to cartilage, connexin 43 is essential in mechanotransduction
      • Gago-Fuentes R.
      • Bechberger J.F.
      • Varela-Eirin M.
      • Varela-Vazquez A.
      • Acea B.
      • Fonseca E.
      • et al.
      The C-terminal domain of connexin43 modulates cartilage structure via chondrocyte phenotypic changes.
      . Alterations in connexin 43 expression and localization affects this cell–cell communication, by which homeostasis to maintain cartilage tissue gets disturbed
      • Mayan M.D.
      • Gago-Fuentes R.
      • Carpintero-Fernandez P.
      • Fernandez-Puente P.
      • Filgueira-Fernandez P.
      • Goyanes N.
      • et al.
      Articular chondrocyte network mediated by gap junctions: role in metabolic cartilage homeostasis.
      . Notably, like EPAS1 and GDF10, the function of connexin 43 is regulated by oxygen levels
      • Knight M.M.
      • McGlashan S.R.
      • Garcia M.
      • Jensen C.G.
      • Poole C.A.
      Articular chondrocytes express connexin 43 hemichannels and P2 receptors - a putative mechanoreceptor complex involving the primary cilium?.
      . Upregulation of EPAS1 and downregulation of GDF10, STC2 and GJA1 suggests that increased level of WWP2 has detrimental effects on cartilage matrix deposition, which acts via hypoxia associated chondrocyte dedifferentiation. This is in line with decreased gene expression levels of COL2A1 and ACAN, two major cartilage markers.
      To evaluate the effects of WWP2 upregulation on cartilage matrix deposition on protein level, we performed proteomics analysis. We did confirm upregulation of WWP2 on day zero, which was still present at day three and seven (Fig. 4, Fig. S2). Moreover, we found 42 significantly differentially expressed proteins upon comparing pellet cultures with and without WWP2 upregulation, of which GJA1 was most significantly differentially expressed and showing the highest fold change (FC = 0.49). The observed differences in gene expression levels of COL2A1 and ACAN with WWP2 upregulation were not observed on protein level (Table S6), which might be due to the relatively low sample size (N = 4 donors) or due to suboptimal timepoint chosen to evaluate the effect of WWP2 on either gene or protein expression level. Since WWP2 is a E3 ubiquitin ligase and the differentially expressed proteins were significantly enriched for ubiquitin conjugating enzyme activity and ubiquitin-protein transferase activity, upregulation of WWP2 could also affect proteins cellular location, activity, and protein–protein interactions without changing expression levels itself, which is not captured by our read-outs. Moreover, it should be noted that the here observed fold differences were relatively low and additional validation and replication are necessary.
      Since it has been suggested that miR-140 is co-expressed with WWP2-C
      • Yang J.
      • Qin S.
      • Yi C.
      • Ma G.
      • Zhu H.
      • Zhou W.
      • et al.
      MiR-140 is co-expressed with Wwp2-C transcript and activated by Sox9 to target Sp1 in maintaining the chondrocyte proliferation.
      , WWP2 isoform 2, we generated upregulation of miR-140 in 2D primary chondrocytes to explore whether WWP2 and miR-140 are involved in similar pathways. In the miR-140 upregulated cells, we observed significant increased expression levels of WWP2, indicating miR-140 indeed targets WWP2. Nonetheless, given the predicted WWP2 target site of miR-140 (3′UTR of WWP2 isoform 6
      • Agarwal V.
      • Bell G.W.
      • Nam J.-W.
      • Bartel D.P.
      Predicting effective microRNA target sites in mammalian mRNAs.
      ) and the absence of SNPs in this region in linkage disequilibrium with the OA risk SNP rs1052429, the genetic WWP2 risk nor the AI is brought about via an aberrant miR-140 binding to WWP2. The fact that we did not observe consistent changes in expression levels of WWP2 isoform 2¸ suggests that WWP2 isoform 2 and miR-140 indeed share the intron 10 (WWP2 full length) promotor as hypothesized previously by Rice et al.
      • Rice S.J.
      • Beier F.
      • Young D.A.
      • Loughlin J.
      Interplay between genetics and epigenetics in osteoarthritis.
      . Additional research is required to fully understand the role of miR-140 in the WWP2 pathway and in OA pathophysiology in general.
      Although lower expression of WWP2 was observed in lesioned compared to preserved cartilage in our previous study
      • Coutinho de Almeida R.
      • Ramos Y.F.M.
      • Mahfouz A.
      • den Hollander W.
      • Lakenberg N.
      • Houtman E.
      • et al.
      RNA sequencing data integration reveals an miRNA interactome of osteoarthritis cartilage.
      , genetic evidence suggests that higher expression of WWP2 predisposes to development of OA, indicating that downregulation in OA pathophysiology is merely a response to the pathophysiological process and a beneficial attempt of chondrocytes to reverse the OA state
      • den Hollander W.
      • Pulyakhina I.
      • Boer C.
      • Bomer N.
      • van der Breggen R.
      • Arindrarto W.
      • et al.
      Annotating transcriptional effects of genetic variants in disease-relevant tissue: transcriptome-wide allelic imbalance in osteoarthritic cartilage.
      . The latter shows that genes identified being differentially expressed between preserved and lesioned OA articular cartilage are a response to the OA pathophysiological process and not necessarily causal to the OA pathophysiological process. To identify genes causal to OA, genetic studies have to be performed. To our surprise, Styrkarsdottir et al.
      • Styrkarsdottir U.
      • Lund S.H.
      • Thorleifsson G.
      • Zink F.
      • Stefansson O.A.
      • Sigurdsson J.K.
      • et al.
      Meta-analysis of Icelandic and UK data sets identifies missense variants in SMO, IL11, COL11A1 and 13 more new loci associated with osteoarthritis.
      reported on WWP2 expression in adipose tissue as function of SNP rs4985453-G, a proxy of their identified OA risk allele rs34195470-G (R2 = 0.79) and our AI SNP rs1052429-A (R2 = 0.77), highlighting the OA risk allele being associated with lower expression levels of WWP2. Upon investigating the GTEx eQTL data of WWP2 with the highlighted SNPs
      The Genotype-Tissue Expression (GTEx) project.
      , we found only data showing consistently higher expression of WWP2 as function of OA risk alleles of the respective SNPs across multiple tissues (Fig. S5), underscoring the aberrant effects observed here with WWP2 upregulation. Although Mokuda et al.
      • Mokuda S.
      • Nakamichi R.
      • Matsuzaki T.
      • Ito Y.
      • Sato T.
      • Miyata K.
      • et al.
      Wwp2 maintains cartilage homeostasis through regulation of Adamts5.
      showed that lack of WWP2 expression in mice resulted in increased expression of RUNX2 and ADAMTS5, we here did not observe consistent changes in expression of RUNX2 or ADAMTS5 upon upregulation of WWP2 in our human chondrocyte pellet cultures and culturing for seven days. This difference could be due to translational limitations from mice to humans. Alternatively, we here create neocartilage, and the effect on RUNX2 and ADAMTS5 may be a temporal or time-dependent effect, which we do not observe at day seven of culturing.
      In conclusion, our data provide support to our hypothesis that high levels of WWP2 have detrimental effects on cartilage homeostasis. We identified EPAS1, GJA1, GDF10, and STC2, all genes involved in chondrocyte dedifferentiation, to be involved in the WWP2 pathway. Moreover, we showed that miR-140 is likely involved in similar pathways as WWP2 and miR-140 might play a role in regulating WWP2 expression. Together these data contribute to a better understanding of how WWP2 confers risk to OA and is a step towards translation from bench to bedside.

      Data availability

      The RNA sequencing data of the articular cartilage is deposited at ArrayExpress (E-MTAB-7313). Further data generated and used in this study is not openly available due to reasons of sensitivity and are available from the corresponding author upon reasonable request.

      Contributions

      All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by M. Tuerlings, G.M.C. Janssen, P.A. van Veelen, Y.F.M. Ramos, and I. Meulenbelt. The first draft of the manuscript was written by M. Tuerlings, Y.F.M. Ramos, and I. Meulenbelt and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

      Competing interests

      The authors have no relevant financial or non-financial interests to disclose.

      Funding

      The study was funded by the Dutch Scientific Research council NWO/ZonMW VICI scheme (nr 91816631/528 ), Dutch Arthritis Society ( DAA_10_1-402 ), BBMRI-NL complementation project ( CP2013-83 ), European Commission Seventh Framework programme (TreatOA, 200800 ), and Ana fonds ( O2015-27 ).

      Acknowledgments

      We thank all the participants of the RAAK study. The LUMC has and is supporting the RAAK study. We thank all the members of our group for valuable discussion and feedback. We also thank Enrike van der Linden, Demiën Broekhuis, Peter van Schie, Shaho Hasan, Maartje Meijer, Daisy Latijnhouwers, Anika Rabelink-Hoogenstraaten, and Geert Spierenburg for collecting the RAAK material. We thank the Sequence Analysis Support Core (SASC) of the Leiden University Medical Center for their support. The study was funded by the Dutch Scientific Research council NWO/ZonMW VICI scheme (nr 91816631/528), Dutch Arthritis Society (DAA_10_1-402), BBMRI-NL complementation project (CP2013-83), European Commission Seventh Framework programme (TreatOA, 200800), and Ana fonds (O2015-27). Data is generated within the scope of the Medical Delta programs Regenerative Medicine 4D: Generating complex tissues with stem cells and printing technology and Improving Mobility with Technology.

      Appendix A. Supplementary data

      The following are the Supplementary data to this article:

      References

        • Vina E.R.
        • Kwoh C.K.
        Epidemiology of osteoarthritis: literature update.
        Curr Opin Rheumatol. 2018; 30 ([published Online First: 2017/12/12]): 160-167https://doi.org/10.1097/bor.0000000000000479
        • Styrkarsdottir U.
        • Lund S.H.
        • Thorleifsson G.
        • Zink F.
        • Stefansson O.A.
        • Sigurdsson J.K.
        • et al.
        Meta-analysis of Icelandic and UK data sets identifies missense variants in SMO, IL11, COL11A1 and 13 more new loci associated with osteoarthritis.
        Nat Genet. 2018; 50 ([published Online First: 2018/10/31]): 1681-1687https://doi.org/10.1038/s41588-018-0247-0
        • Zengini E.
        • Hatzikotoulas K.
        • Tachmazidou I.
        • Steinberg J.
        • Hartwig F.P.
        • Southam L.
        • et al.
        Genome-wide analyses using UK Biobank data provide insights into the genetic architecture of osteoarthritis.
        Nat Genet. 2018; 50 ([published Online First: 2018/03/22]): 549-558https://doi.org/10.1038/s41588-018-0079-y
        • Boer C.G.
        • Hatzikotoulas K.
        • Southam L.
        • Stefánsdóttir L.
        • Zhang Y.
        • Coutinho de Almeida R.
        • et al.
        Deciphering osteoarthritis genetics across 826,690 individuals from 9 populations.
        Cell. 2021; ([published Online First: 2021/08/28])https://doi.org/10.1016/j.cell.2021.07.038
        • Gee F.
        • Clubbs C.F.
        • Raine E.V.
        • Reynard L.N.
        • Loughlin J.
        Allelic expression analysis of the osteoarthritis susceptibility locus that maps to chromosome 3p21 reveals cis-acting eQTLs at GNL3 and SPCS1.
        BMC Med Genet. 2014; 15 ([published Online First: 2014/06/03]): 53https://doi.org/10.1186/1471-2350-15-53
        • Raine E.V.
        • Dodd A.W.
        • Reynard L.N.
        • Loughlin J.
        Allelic expression analysis of the osteoarthritis susceptibility gene COL11A1 in human joint tissues.
        BMC Musculoskelet Disord. 2013; 14 ([published Online First: 2013/03/19]): 85https://doi.org/10.1186/1471-2474-14-85
        • den Hollander W.
        • Pulyakhina I.
        • Boer C.
        • Bomer N.
        • van der Breggen R.
        • Arindrarto W.
        • et al.
        Annotating transcriptional effects of genetic variants in disease-relevant tissue: transcriptome-wide allelic imbalance in osteoarthritic cartilage.
        Arthritis Rheumatol. 2019; 71 ([published Online First: 2018/10/10]): 561-570https://doi.org/10.1002/art.40748
        • den Hollander W.
        • Ramos Y.F.
        • Bomer N.
        • Elzinga S.
        • van der Breggen R.
        • Lakenberg N.
        • et al.
        Transcriptional associations of osteoarthritis-mediated loss of epigenetic control in articular cartilage.
        Arthritis Rheumatol. 2015; 67 ([published Online First: 2015/04/22]): 2108-2116https://doi.org/10.1002/art.39162
        • Coutinho de Almeida R.
        • Ramos Y.F.M.
        • Mahfouz A.
        • den Hollander W.
        • Lakenberg N.
        • Houtman E.
        • et al.
        RNA sequencing data integration reveals an miRNA interactome of osteoarthritis cartilage.
        Ann Rheum Dis. 2019; 78 ([published Online First: 2018/12/07]): 270-277https://doi.org/10.1136/annrheumdis-2018-213882
        • Ji Q.
        • Zheng Y.
        • Zhang G.
        • Hu Y.
        • Fan X.
        • Hou Y.
        • et al.
        Single-cell RNA-seq analysis reveals the progression of human osteoarthritis.
        Ann Rheum Dis. 2019; 78: 100-110https://doi.org/10.1136/annrheumdis-2017-212863
        • Chantry A.
        WWP2 ubiquitin ligase and its isoforms: new biological insight and promising disease targets.
        Cell Cycle. 2011; 10 ([published Online First: 2011/07/14]): 2437-2439https://doi.org/10.4161/cc.10.15.16874
        • Nakamura Y.
        • Yamamoto K.
        • He X.
        • Otsuki B.
        • Kim Y.
        • Murao H.
        • et al.
        Wwp2 is essential for palatogenesis mediated by the interaction between Sox9 and mediator subunit 25.
        Nat Commun. 2011; 2 ([published Online First: 2011/03/24]): 251https://doi.org/10.1038/ncomms1242
        • Mokuda S.
        • Nakamichi R.
        • Matsuzaki T.
        • Ito Y.
        • Sato T.
        • Miyata K.
        • et al.
        Wwp2 maintains cartilage homeostasis through regulation of Adamts5.
        Nat Commun. 2019; 10: 2429https://doi.org/10.1038/s41467-019-10177-1
        • Yang J.
        • Qin S.
        • Yi C.
        • Ma G.
        • Zhu H.
        • Zhou W.
        • et al.
        MiR-140 is co-expressed with Wwp2-C transcript and activated by Sox9 to target Sp1 in maintaining the chondrocyte proliferation.
        FEBS Lett. 2011; 585 ([published Online First: 2011/08/30]): 2992-2997https://doi.org/10.1016/j.febslet.2011.08.013
        • Rice S.J.
        • Beier F.
        • Young D.A.
        • Loughlin J.
        Interplay between genetics and epigenetics in osteoarthritis.
        Nat Rev Rheumatol. 2020; 16: 268-281https://doi.org/10.1038/s41584-020-0407-3
        • Ramos Y.F.
        • den Hollander W.
        • Bovee J.V.
        • Bomer N.
        • van der Breggen R.
        • Lakenberg N.
        • et al.
        Genes involved in the osteoarthritis process identified through genome wide expression analysis in articular cartilage; the RAAK study.
        PLoS One. 2014; 9 ([published Online First: 2014/07/24])e103056https://doi.org/10.1371/journal.pone.0103056
        • Reynard L.N.
        Analysis of genetics and DNA methylation in osteoarthritis: what have we learnt about the disease?.
        Semin Cell Dev Biol. 2017; 62 ([published Online First: 2016/05/01]): 57-66https://doi.org/10.1016/j.semcdb.2016.04.017
        • Rice S.J.
        • Cheung K.
        • Reynard L.N.
        • Loughlin J.
        Discovery and analysis of methylation quantitative trait loci (mQTLs) mapping to novel osteoarthritis genetic risk signals.
        Osteoarthritis Cartilage. 2019; 27 ([published Online First: 2019/06/08]): 1545-1556https://doi.org/10.1016/j.joca.2019.05.017
        • Sekiya I.
        • Tsuji K.
        • Koopman P.
        • Watanabe H.
        • Yamada Y.
        • Shinomiya K.
        • et al.
        SOX9 enhances aggrecan gene promoter/enhancer activity and is up-regulated by retinoic acid in a cartilage-derived cell line, TC6.
        J Biol Chem. 2000; 275 ([published Online First: 2001/02/07]): 10738-10744https://doi.org/10.1074/jbc.275.15.10738
        • Nishimura R.
        • Hata K.
        • Takahata Y.
        • Murakami T.
        • Nakamura E.
        • Yagi H.
        Regulation of cartilage development and diseases by transcription factors.
        J Bone Metab. 2017; 24 ([published Online First: 2017/09/29]): 147-153https://doi.org/10.11005/jbm.2017.24.3.147
        • Yang S.
        • Kim J.
        • Ryu J.H.
        • Oh H.
        • Chun C.H.
        • Kim B.J.
        • et al.
        Hypoxia-inducible factor-2alpha is a catabolic regulator of osteoarthritic cartilage destruction.
        Nat Med. 2010; 16 ([published Online First: 2010/05/25]): 687-693https://doi.org/10.1038/nm.2153
        • Saito T.
        • Fukai A.
        • Mabuchi A.
        • Ikeda T.
        • Yano F.
        • Ohba S.
        • et al.
        Transcriptional regulation of endochondral ossification by HIF-2alpha during skeletal growth and osteoarthritis development.
        Nat Med. 2010; 16 ([published Online First: 2010/05/25]): 678-686https://doi.org/10.1038/nm.2146
        • Matsumoto Y.
        • Otsuka F.
        • Hino J.
        • Miyoshi T.
        • Takano M.
        • Miyazato M.
        • et al.
        Bone morphogenetic protein-3b (BMP-3b) inhibits osteoblast differentiation via Smad2/3 pathway by counteracting Smad1/5/8 signaling.
        Mol Cell Endocrinol. 2012; 350 ([published Online First: 2011/12/14]): 78-86https://doi.org/10.1016/j.mce.2011.11.023
        • Chou C.H.
        • Lee C.H.
        • Lu L.S.
        • Song I.W.
        • Chuang H.P.
        • Kuo S.Y.
        • et al.
        Direct assessment of articular cartilage and underlying subchondral bone reveals a progressive gene expression change in human osteoarthritic knees.
        Osteoarthritis Cartilage. 2013; 21 ([published Online First: 2012/12/12]): 450-461https://doi.org/10.1016/j.joca.2012.11.016
        • Lafont J.E.
        • Talma S.
        • Hopfgarten C.
        • Murphy C.L.
        Hypoxia promotes the differentiated human articular chondrocyte phenotype through SOX9-dependent and -independent pathways.
        J Biol Chem. 2008; 283 ([published Online First: 2007/12/14]): 4778-4786https://doi.org/10.1074/jbc.M707729200
        • Gagliardi A.D.
        • Kuo E.Y.
        • Raulic S.
        • Wagner G.F.
        • DiMattia G.E.
        Human stanniocalcin-2 exhibits potent growth-suppressive properties in transgenic mice independently of growth hormone and IGFs.
        Am J Physiol Endocrinol Metab. 2005; 288 ([published Online First: 2004/09/16]): E92-E105https://doi.org/10.1152/ajpendo.00268.2004
        • Chang A.C.
        • Hook J.
        • Lemckert F.A.
        • McDonald M.M.
        • Nguyen M.A.
        • Hardeman E.C.
        • et al.
        The murine stanniocalcin 2 gene is a negative regulator of postnatal growth.
        Endocrinology. 2008; 149 ([published Online First: 2008/02/09]): 2403-2410https://doi.org/10.1210/en.2007-1219
        • Gelse K.
        • Ekici A.B.
        • Cipa F.
        • Swoboda B.
        • Carl H.D.
        • Olk A.
        • et al.
        Molecular differentiation between osteophytic and articular cartilage--clues for a transient and permanent chondrocyte phenotype.
        Osteoarthritis Cartilage. 2012; 20 ([published Online First: 2012/01/03]): 162-171https://doi.org/10.1016/j.joca.2011.12.004
        • Gago-Fuentes R.
        • Bechberger J.F.
        • Varela-Eirin M.
        • Varela-Vazquez A.
        • Acea B.
        • Fonseca E.
        • et al.
        The C-terminal domain of connexin43 modulates cartilage structure via chondrocyte phenotypic changes.
        Oncotarget. 2016; 7 ([published Online First: 2016/09/30]): 73055-73067https://doi.org/10.18632/oncotarget.12197
        • Mayan M.D.
        • Gago-Fuentes R.
        • Carpintero-Fernandez P.
        • Fernandez-Puente P.
        • Filgueira-Fernandez P.
        • Goyanes N.
        • et al.
        Articular chondrocyte network mediated by gap junctions: role in metabolic cartilage homeostasis.
        Ann Rheum Dis. 2015; 74 ([published Online First: 2013/11/15]): 275-284https://doi.org/10.1136/annrheumdis-2013-204244
        • Knight M.M.
        • McGlashan S.R.
        • Garcia M.
        • Jensen C.G.
        • Poole C.A.
        Articular chondrocytes express connexin 43 hemichannels and P2 receptors - a putative mechanoreceptor complex involving the primary cilium?.
        J Anat. 2009; 214 ([published Online First: 2009/02/12]): 275-283https://doi.org/10.1111/j.1469-7580.2008.01021.x
        • Agarwal V.
        • Bell G.W.
        • Nam J.-W.
        • Bartel D.P.
        Predicting effective microRNA target sites in mammalian mRNAs.
        Elife. 2015; 4e05005https://doi.org/10.7554/eLife.05005
      1. The Genotype-Tissue Expression (GTEx) project.
        Nat Genet. 2013; 45 ([published Online First: 2013/05/30]): 580-585https://doi.org/10.1038/ng.2653