Purpose: Previously, we identified a high-impact, missense mutation (C518F) in the gelatin-binding domain of FN1 that was causal to the early-onset osteoarthritis (OA) phenotype in an extended early-onset OA family. By introducing this high-impact mutation in human induced pluripotent stem cells (hiPSCs) using CRISPR/Cas9 gene editing and using them in an established in vitro organoid cartilage model, we showed that the chondrogenic potential and deposition of neo-cartilage of the FN1 mutant chondrocytes was decreased. Moreover, we demonstrated that the underlying pathogenic mechanism of the mutation is caused by a decreased binding of fibronectin to the surrounding extracellular matrix (ECM) protein collagen type II. Since fibronectin functions as a transducer of biomechanical signals from the ECM to chondrocytes via integrins, the decreased binding to collagen type II likely results in changed interactions between ECM and chondrocytes. Therefore, we aimed to investigate how the FN1 mutation affects the response of chondrocytes to hyper-physiological mechanical loading in our in vitro organoid cartilage model and further elucidate the function of fibronectin as mechanotransducer.

Objective To identify FN1 transcripts associated with OA pathophysiology and investigate the downstream effects of modulating FN1 expression and relative transcript ratio. Methods FN1 transcriptomic data was obtained from our previously assessed RNA-seq dataset of lesioned and preserved OA cartilage samples from the Research osteoArthritis Articular Cartilage (RAAK) study. Differential transcript expression analysis was performed on all 27 FN1 transcripts annotated in the Ensembl database. Human primary chondrocytes were transduced with lentiviral particles containing short hairpin RNA (shRNA) targeting full-length FN1 transcripts or non-targeting shRNA. Subsequently, matrix deposition was induced in our 3D in vitro neo-cartilage model. Effects of changes in the FN1 transcript ratio on sulphated glycosaminoglycan (sGAG) deposition were investigated by Alcian blue staining and dimethylmethylene blue assay. Moreover, gene expression levels of 17 cartilage-relevant markers were determined by reverse transcription quantitative polymerase chain reaction. Results We identified 16 FN1 transcripts differentially expressed between lesioned and preserved cartilage. FN1-208, encoding migration-stimulating factor, was the most significantly differentially expressed protein coding transcript. Downregulation of full-length FN1 and a concomitant increased FN1-208 ratio resulted in decreased sGAG deposition as well as decreased ACAN and COL2A1 and increased ADAMTS-5, ITGB1 and ITGB5 gene expression levels. Conclusion We show that full-length FN1 downregulation and concomitant relative FN1-208 upregulation was unbeneficial for deposition of cartilage matrix, likely due to decreased availability of the classical RGD (Arg-Gly-Asp) integrin-binding site of fibronectin.

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 (| r| > 0.7) were identified. Among these genes, we identified GJA1, GDF10, STC2, WDR1, and WNK4. Sub-sequent 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, upregu-lation 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 detri-mental 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.(c) 2022 The Author(s). Published by Elsevier Ltd on behalf of Osteoarthritis Research Society International. This is an open access article under the CC BY license (http://creativecommons.org/ licenses/by/4.0/).

Purpose: To explore the co-expression network of the osteoarthritis (OA) risk gene WWP2 in articular cartilage and study gene and protein expression when mimicking the upregulated effect of risk allele rs1052429-A on WWP2 expression in 3D in vitro neo-cartilage organoids.

Purpose: Founded by the hypothesis that the OA disease pathway of the previously identified high impact gain-of-function mutation in TNFRSF11B could be extrapolated to common, age-related, OA disease phenotypes, we set out to functionally study its effect. To this end, we established in vitro hiPS derived spherical chondrogenic and osteogenic pellets and compared TNFRSF11B mutated cells generated from affected family members to CRISPR/Cas9 repaired isogenic control cells. Given that the mutant OPG decoy receptor more efficiently antagonizes osteoclastogenesis, we expect that the mutation affects matrix mineralization hence inflicting the observed chondrocalcinosis phenotype. Methods: Using skin fibroblasts, hiPSCs of an affected family member were created (FOA) and differentiated to induced mesenchymal stromal stem cells (iMSCs). These were differentiated for 6 weeks to deposit neo-cartilaginous and neo-osseous pellets or neo-bone in parallel with a CRISPR-Cas9 repaired hiPSC line (FOA-R) serving as an isogenic control. Presence of glycosaminoglycans was determined by Alcian blue while mineralization was determined by Alizarin red staining. Gene expression analyses were performed to study anabolic markers and hypertrophy (ie, COL2A1, COL1A1, COL10A1) and matrix mineralization (ie, ALPL, MGP). Results: As demonstrated by the intensity of Alcian blue and Alizarin red staining of spherical neo-cartilaginous and neo-osseous pellets, respectively (Figure 1) the bi-directional differentiation via iPSC derived iMSCs appeared successful. Additionally, while only subtle differences in intensity were observed for the Alcian blue staining, the Alizarin red staining intensity in FOA mutated neo-osseous pellets was clearly higher as that of the FOA-repaired pellets, indicating increased matrix mineralization (Figure 1). Gene expression analyses of neo-cartilaginous pellets, showed, however, that in the presence of the FOA TNFRSF11B mutation, MGP levels were considerably decreased (FD=0.14 P=1.6x10-13), indicating cell signaling towards matrix mineralization also in the mutated neo-cartilaginous tissues. In line with this, neo-cartilaginous tissues from FOA mutated cells compared to FOA repaired cells expressed lower levels of COL2A1 (FD=-6.8, P-value=3.9x10-3) concomitant with higher levels of COL1A1 (FD=1.7, P-value=7.4x10-3). Additionally, in the FOA mutated neo-osseous pellets gene expression analysis showed a higher expression of RUNX2 (FD=2.6, P-value= 8.29x10-3) and a trend towards a higher expression of COL1A1 (FD=1.54, P-value= 0.079) and ALPL (FD=1.57, P-value=0.055) relative to FOA repaired neo-osseous pellets.supplements. Conclusions: Employing hiPS-cells from early-onset OA family members (FOA) and CRISPR/Cas9 repaired isogenic controls (FOA-R) in an established spherical 3D in vitro chondrogenic and osteogenic pellet model confirmed that the previously identified gain of function mutation in TNFRSF11B encoding OPG acts via increased matrix mineralization of cartilage and bone. In cartilage, likely, this is modulated via MGP expression as suggested by the significant lower levels in the mutant cells. This may raise the possibility to slow down OA progression in early-onset family members with the use of vitamin K.

Purpose: Osteoarthritis (OA) has a considerable genetic component and genetic research has resulted in many insights into underlying disease pathways. Most recent large-scale genomic association studies have identified TGFA , MGP, CHADL as compelling genes involved in the aetiology of OA. The functions of these genes confirm that deviations in both cartilage and bone maintenance processes are major pathways underlying OA pathology in humans. Moreover, follow-up studies have shown that risk SNPs frequently modulate pathology due to altering transcription of the genes in cis both in bone and cartilage. Nonetheless, follow up study of these genes towards underlying biological mechanisms, preclinical studies on target discovery and eventually drug testing have thus far focused on gene expression profiling in articular cartilage. The aim of the current study was to characterize interactive pathophysiological processes of subchondral bone and articular cartilage in human osteoarthritis (OA) in large human cohort RNA sequencing. Methods: We performed RNA sequencing on macroscopically preserved and lesioned OA subchondral bone of patients that underwent a joint replacement surgery due to OA (N=24 pairs; 6 hips, 18 knees, RAAK-study). Unsupervised hierarchical clustering and differential expression analysis were performed on the data. Gene enrichment was performed using the online tool DAVID. Moreover, we compared the differentially expressed genes identified here to our previously reported differentially expressed genes in OA articular cartilage. Results: Upon performing cluster analysis we identified two clusters; one representing hips and the other representing knees. Due to the low hip sample size, differential expression analysis was performed on knee samples only. As shown in Figure 1, we identified 1757 significantly differentially expressed genes between lesioned and preserved subchondral bone. The most significantly downregulated gene was FRZB (FC=0.48, FDR=3.07x10-12), encoding the frizzled receptor protein and is a well-known cartilage OA gene. The most significantly upregulated gene was CNTNAP2 (FC=2.90, FDR=3.38x10-6), encoding the contactin-associated protein-like 2 protein (CASPR2). Among the 1757 differentially expressed genes, 102 genes showed an absolute foldchange of 2 or higher. The highest upregulated gene was STMN2 (FC=23.0, FDR=1.8x10-4), encoding stathmin 2, while the most downregulated gene was CHRDL2 (FC=0.09, FDR=4.13x10-6), encoding chordin-like protein 2 and is also known to be involved in OA in cartilage. Enrichment of the differentially expressed genes resulted in significant enrichment of GO-terms regarding the extracellular matrix. To investigate whether the subchondral bone and the articular cartilage show similarities in the OA pathophysiological processes, we compared the results presented here with our previously reported results on differentially expressed genes in OA articular cartilage. As a result, we identified 398 genes that were differentially expressed in both tissues, of which 361 showed similar directions of effect, including the genes IL11, encoding Interleukin 11, and CHADL, encoding Chondroadherin-Like Protein. Conclusions: To our knowledge, we are the first to report on the differential expression pattern of OA subchondral bone using RNA sequencing on a large sample size while integrating these with differential expression patterns of cartilage of the same individuals. By performing cluster analysis on our RNA sequencing data, we identified two clusters representing hips and knees, respectively. These clusters indicate distinct OA pathophysiological processes in the subchondral bone between the two joint sites, which was not previously seen with similar analyses of the cartilage. Among the most consistently differentially expressed genes with OA pathophysiology in bone and cartilage are IL11, and CHADL. Notably, these genes were recently found in large GWA studies on OA indicating also their causal involvement in the OA pathophysiology. Moreover, IL11 was recognized as being targeted by FDA approved drugs. As such, we advocate that IL11 could be an attractive potential drug for OA patients with a therapeutic effect both in cartilage and subchondral.

Purpose: Induced pluripotent stem cells (iPSCs) technology has been proposed as an excellent method for engineering of cartilage and bone. Currently, iPSCs are being extensively studied in regenerative medicine strategies, disease modelling and drug screening in OA. These studies, however, still show a large variability and, more importantly, expansion of the cells following differentiation is shown to rapidly decrease quality of deposited cartilage extracellular matrix (ECM). We set out to compare molecular and histological differences between two approaches, aiming at establishing a controllable protocol ensuring deposition of high-quality cartilage. Methods: A step-wise method was applied to differentiate iPSCs towards chondroprogenitor cells (CPCs) directly followed by chondrocytes depositing cartilage ECM. Alternatively iPSC were differentiated first towards mesenchymal stromal cells (iMSCs) followed by chondrogenesis. Cells were characterized by flow cytometry for mesenchymal stromal cell (MSC) associated markers (e.g. CD90, CD73, CD31, CD45, CD105). Chondrogenic differentiations were performed during 3 and 5 weeks in a 3D pellet culture system and compared to cartilage generated from human MSCs and from human primary chondrocytes. Subsequently, pellets were analysed by histology (Alcian Blue, collagen type 2), and by RT-qPCR. Results: Flow cytometry analyses confirmed similarity of the iMSCs with bone marrow derived MSCs while chondrocytes generated by the two protocols significantly differed in their characteristics. Specifically some iMSCs could only differentiate into chondrocytes while others did not deposit any ECM. However, when applying the step-wise direct protocol, the chondrogenic differentiation appeared more consistent while simultaneously showing deposition of high-quality cartilage ECM (Figure 1). Gene expression analyses show a significant difference between the matrix genes COL2A1, COL1A1 and COL10A1 in cartilage deposited by chondrocytes generated with the two methods. COL2A1 is higher expressed in chondrocytes generated by CPCs, while the iMSCs-derived chondrocytes have a higher expression of COL1A1 and COL10A1, suggesting a more hypertrophic cartilage. The generated iMSCs and CPCs show similar expression levels in these genes compared to MSCs and autologous neo-cartilage respectively. Conclusions: Overall, it was observed that cartilage deposited by chondrocytes from the direct protocol was highly comparable to that from autologous neo-cartilage, while chondrocytes from iMSC were more comparable to those generated from MSCs. This suggests their high resemblance and makes them a good candidate to research different disease models. Additionally, the step-wise chondroprogenitor protocol generates the optimal Results with high levels of deposition of ECM when compared to their iMSC counterpart, indicating this method is preferable to further explore its applications for regenerative medicine purposes.

Purpose: One of the hallmarks of osteoarthritis (OA) is cartilage degeneration, which has been demonstrated to be marked by aberrant gene expression regulation, i.e. via changes in epigenetic control mechanisms. Frequently studied epigenetic mechanisms in OA include DNA methylation at CpG sites, histone modifications, and expression of microRNAs (miRNAs < 20 nt). More recently, long non-coding RNAs (lncRNAs > 200 nt) gained attention, where studies have confirmed the involvement of lncRNAs in the pathophysiology of OA. Nonetheless, these lncRNA studies have focused primarily on candidate and microarray studies and are therefore limited to pre-existing probe designs, hence frequently based on readily detected expression levels. RNA-sequencing can overcome these limitations, but studies employing RNA-seq experiments so far were performed with relative low sample size, or have investigated primarily long intergenic non-coding RNAs. The goal of this study was to identify all lncRNAs associated with OA pathophysiology in a large human cohort of OA cartilage using RNA sequencing and explore their function in the pathophysiology process of OA. Methods: RNA sequencing was performed on macroscopically preserved and lesioned OA cartilage of patients who underwent a joint replacement surgery due to end stage OA (N = 32 pairs, RAAK study). Differential expression (DE) analysis was performed on lncRNAs that are annotated in GENCODE, by performing a paired Wald test on the normalized counts, FDR < 0.05 was considered significant. RT-qPCR was performed to confirm direction of effect of selected DE lncRNAs in 20 paired samples, of which 10 overlapped with the RNA-seq samples. Pearson’s correlations were calculated between the identified DE lncRNAs and our previously reported differentially expressed genes in OA cartilage to identify potential functional targets of the lncRNAs. Modulation of lncRNA chondrocyte expression was achieved by using LNA GapmeRs, subsequently expression of correlated genes were analyzed to investigate their functional effect. Results: Differential expression (DE) analysis between lesioned and preserved OA cartilage resulted in 191 FDR significant lncRNAs, which included those previously associated to OA, such as MEG3, PART1, and LINC01614, as well as novel OA-related lncRNAs, such as P3H2-AS1 and AC090877.2. Nonetheless, previously identified lncRNAs in OA cartilage such as HOTAIR and GAS5 were not found FDR significantly expressed. Five of the most significant DE lncRNAs with the highest absolute FC were replicated by RT-qPCR. We observed equal direction of effect for MEG3 (FC = 0.63, FDR = 9x10-3), AC025370.1 (FC = 2.02, FDR = 4x10-3), AC090877.2 (FC = 0.33, FDR = 6x10-5), P3H2-AS1 (FC = 2.7, FDR = 4x10-4), and TBILA (FC = 3.48, FDR = 6x10-5). Subsequently, P3H2-AS1 was selected as a proof of concept for functional validation in an in vitro primary chondrocyte culture model. Prior to the in vitro experiment we identified potential interactions between P3H2-AS1 and DE OA-related protein-coding genes, based on Pearson’s correlations. Notably, the highest correlation of P3H2-AS1 was with P3H2 (r = 0.66), which is the positional gene of P3H2-AS1 and encodes the P3H2 protein, which plays a critical role in collagen chain assembly, stability and cross-linking. To confirm the identified interaction, P3H2-AS1 expression was downregulated by transfecting primary chondrocytes with antisense LNA GapmeRs targeting P3H2-AS1 (FC = 0.22, P = 0.02). Moreover, P3H2 expression levels were measured, which showed that P3H2 expression was significantly downregulated compared to cells transfected with non-targeting LNA GapmeRs (FC = 0.37, P = 0.04, 2 donors). The direction of effect of the expression levels is in accordance with their positive correlation, thereby confirming our hypothesis that P3H2 is a downstream target of P3H2-AS1. Conclusions: To our knowledge, we are the first to report on the transcriptome wide differential expression pattern of OA lncRNAs using RNA-seq on a large sample size, while integrating differential mRNA expression levels within the same samples. P3H2-AS1 was selected for functional follow-up, where we showed that P3H2 is a downstream target of P3H2-AS1. Since P3H2 was found to be involved in the OA process, it is relevant to investigate whether P3H2-AS1 could be used as a potential preclinical target, resulting in modulation of P3H2 expression levels. To conclude, studies based on RNA-seq for detecting lncRNAs show that this is a powerful tool, which should be complemented by functional validation. This could be done for example by modulating lncRNA expression levels using antisense LNA GapmeRs.

Purpose: Upon challenging environmental changes, such as diseases or mechanical stress, articular chondrocytes need to dynamically alter their gene expression for which they highly depend on epigenetic mechanisms. During osteoarthritis (OA), it has been consistently shown that control of epigenetically regulated transcription is lost, also for canonical Wnt signaling, which plays an important role in cartilage matrix homeostasis. This was further exemplified in our previous work by a highly significant upregulation in the epigenetically regulated transcription of Wnt-1-induced signaling protein 1 (WISP1). We showed that expression levels of WISP1 correlated with DNA methylation of 2 positional CpG sites (cg10191240 and cg26617637) residing within the WISP1 gene. Here, we set out to establish a model in which we can modulate WISP1 expression through mechanical stress, causal to development of OA. Subsequently, this model was applied to investigate epigenetic regulation of WISP1. Methods: Human primary articular chondrocytes were isolated from joints of patients that underwent total joint replacement at end stage OA (n=9, RAAK study). Cells were cultured and pellets were created by centrifugation. On day 11, pellets were subjected twice to mechanical stress (20% compression, 5 Hz, 10 minutes, with a 10 min recovery interval). After 2 (short-term effect) and 12 (long-term effect) hours, pellets were collected and WISP1 mRNA and protein expression were determined. DNA methylation of CpG sites cg10191240 and cg26617637 were assessed by bisulfite pyrosequencing. In addition, to determine the effects of mechanical stress on the cartilage, mRNA expression of catabolic genes MMP3, MMP13 and ADAMTS5, anabolic genes COL2A1 and ACAN, cartilage marker SOX9 and hypertrophic marker COL10A1 were investigated. Furthermore, damage to the cartilage was assessed by Alcian Blue staining and by measuring the release of sulfated glycosaminoglycans (sGAG) into the medium using the Dimethylmethylene Blue (DMMB) assay. Results: WISP1 expression significantly increased (FC=1.43, p=0.009, Table 1) in human chondrocyte pellets 2 hours after applying mechanical stress, similarly to the changes in expression levels observed in lesioned vs preserved OA-cartilage. Nonetheless, WISP1 protein expression seemed to decrease. Methylation levels of cg10191240 and cg26617637 were not significantly changed within 2 hours. Expression of MMP3 and ADAMTS5 significantly decreased in loaded pellets, while MMP13, COL2A1, ACAN, SOX9 and COL10A1) did not change (Table 1). We observed a significant increase in sGAG release into the medium in the loaded pellets compared to control pellets (1.44±0.50 μg/ml and 0.68±0.20 μg/ml respectively, p<0.001). 12 hours after loading, we observed a significant decrease in WISP1 expression (FC=0.51, p<0.001, Table 1), while no differences were observed in WISP1 protein expression. In line with the previously observed inverse correlation between WISP1 expression and methylation of CpG site cg10191240, this CpG site showed an increase in methylation. However, no changes were observed in DNA methylation of cg26617637. ADAMTS5 expression significantly increased and COL10A1 expression significantly decreased, but there were no changes in expression of the other genes (Table 1). sGAG release into the medium did not increase in loaded pellets compared to control (2.50±0.60 μg/ml and 2.02±0.90 μg/ml respectively, p=0.131). Overall, no changes in Alcian Blue intensity were observed. Conclusions: Mechanical stress initially triggered upregulation of WISP1 in human chondrocyte pellets, but WISP1 expression decreased significantly within 12 hours. We hypothesize that the mechanically induced changes in WISP1 expression could be an important event resulting in the loss of control of the canonical Wnt signaling pathway, and that identifying signaling pathways downstream of WISP1 may shed light on the contribution of WISP1 to the pathophysiological process of OA. In agreement with changes in lesioned vs preserved OA cartilage, we now show that mechanical loading of pellets changes the methylation of cg10191240. Currently, we are investigating whether methylation changes for cg10191240 affects transcription factor binding. Although the increased sGAG release and the upregulation of ADAMTS5 suggest that the mechanical stress we applied may result in OA-like damage, further optimization to robustly study OA-like changes is required.Tabled 1Table 1. Effects of mechanical stress on mRNA expression in human articular chondrocyte pelletsGeneTime after applying mechanical load2 hours (n=9)12 hours (n=5)Fold Changep-valueFold Changep-valueWISP11.430.0090.51<0.001CatabolicMMP130.750.3120.810.576MMP30.630.0211.380.722ADAMTS50.720.0385.01<0.001AnabolicCOL2A10.790.2081.140.087ACAN1.070.9680.790.152Cartilage markerSOX90.980.6841.710.187HypertrophyCOL10A11.690.4690.35<0.001 Open table in a new tab

Purpose: By applying a genome wide screen on allelic imbalanced (AI) expression of genes in cartilage, we identified rs1052429, located in the 3’UTR of WWP2, that showed highly significant AI with allele rs1052429-A (frequency∼0.7) marking higher expression of WWP2 relative to rs1052429-G. Moreover, rs1052429-A conferred risk to reduced minimal joint space width (P=0.0028) whereas WWP2 expression was downregulated in lesioned as compared to macroscopically preserved OA cartilage (FDR=0.005). Notably, in a recent large meta-analyse of Icelandic and UK datasets WWP2 was also identified as OA susceptibility gene. Here the risk allele rs34195470_G with high linkage disequilibrium to rs1052429_A (r2=0.60) showed genome wide significant association to knee OA. Together, these data imply that increased WWP2 expression confers robust risk to OA. In the current study, we set out to functionally investigate the effect of WWP2 upregulation in a 3D in vitro model of primary chondrocytes. Methods: We generated a lentiviral particle mediated WWP2 overexpression in primary chondrocytes (N=5 donors, RAAK-study). Transduced cells were kept in 2D to expand for 2 passages and cells were counted at consecutive days. Subsequently, pellet cultures were made and harvested at 7, 14, and 21 days. When harvesting the cell pellets the surface areas were measured. The effects of the WWP2 overexpression on cartilage matrix deposition was assessed by performing RT-qPCR. With the qPCR we measured the expression levels of chondrocyte hypertrophy genes, such as MMP13, COL1A1, RUNX2, COL10A1, EPAS1, and ADAMTS5 and the expression levels of anabolic genes, such as COL2A1, SOX9, and AGC. To robustly analyse the effect of WWP2 overexpression generalized estimating equation (GEE) analyses was performed and corrected for donors. Results: As shown in Figure-1 (left panel), WWP2 overexpressed cells had less capacity to expand and showed smaller surface areas after 21 days of pellet culture (right panel). Upon comparing the qPCR data of controls and WWP2 overexpressed pellets at all timepoints, a consistent decreased expression was observed for COL1A1, AGC, SOX9, EPAS1 and COL10A1, and increased expression for RUNX2. Nonetheless, only the effect of COL1A1 showed a significant effect (P=0.012). Conclusions: Although WWP2 overexpressing cells had decreased capacity to expand and deposit matrix, we did not observe a tendency of these cells to undergo hypertrophy or exhibit OA signalling pattern. We advocate that WWP2 is maintaining maturational arrest of chondrocytes whilst the OA risk allele rs1052429-A may act during skeletal development, which we are currently addressing in a more appropriate model system. Moreover, we are also analysing the effects of WWP2 knock down on cartilage matrix deposition.

Purpose: Osteoarthritis (OA) is a joint disease characterized by cartilage degeneration and bone spur formation. Due to the fact that there are no disease modifying drugs, OA is placing a high burden on society occurring during diseases. By applying genetic analyses and functional follow up research, our group demonstrated that the upregulation of the type II iodothyronine deiodinase (D2) gene (DIO2), likely enhancing thyroid signaling, affects propensity of joint tissues to engage on osteoarthritis (OA) state, particularly upon mechanical loading. Conversely, it was shown that inhibition of D2 using a dio2 knock out mouse or iopanoic acid (IOP) during in vitro chondrogenesis contributed to prolonged tissue homeostasis. In the current study we aimed to set up a human pre-clinical OA explant model in which mechanical loading induces damage and to test the long term effect of the pharmacological agent IOP. Methods: Full thickness osteochondral explants were isolated from 17 human knees from the RAAK study. From day 0 to 5 explants were kept under standardized conditions in chondrogenic medium to normalize expression levels. Mechanical loading was subsequently applied during four days, with a regime of 1 hz at 65% strain, during 10 minutes using a Mach-1 mechanical testing system (Biomomentum). To study the long term effect explants were harvested four days after the last loading was applied. To investigate the application of our model for drug research, we inhibited the D2 activity by adding 100 μM IOP to the explant culture medium from day 3 onwards. RT-PCR of COL2A1, AGC (anabolism), MMP13 (catabolism), COL10A1, ALPL (hypertrophy) and COL1A1 (mineralization) was performed as primary output measure. To determine the effect of loading on extracellular matrix (ECM) composition, sulphated glycosaminoglycans (sGAG) content of cartilage at end point and sGAG released was measured using dimethylmethylene blue assay (DMMB). To robustly analyse the effect of mechanical loading and IOP treatment we performed generalized estimating equation (GEE) analyses and corrected for donors. Results: In our human explant model, a strain of 65% resulted in a significant and consistent upregulation of the catabolic gene MMP13 (P<0.000, n=63), downregulation of COL2A1 (P=0.017, n=68) and a trend towards decreased AGC (P=0.067, n=68) expression. Additionally, mechanical loading increased the amount of sGAG (P=0.004, n=32) released to the explant culture medium. Strikingly, we observed that after treatment with IOP in combination with mechanical loading at a strain of 65% resulted in a significantly decreased expression of MMP13 (P=0.004, n=36) and COL10A1 (P=0.022, n=27) and an increased expression of COL2A1 (P=0.009, n=40) compared to explants treated with only a 65% strain. Conclusions: By applying mechanical loading at a strain of 65% to autologous aged cartilage we have set up a human ex vivo pre-clinical model in which we can investigate the effect of pharmaceutical agents on preventing damage induced by overloading. In our model we have measured an increase of the catabolic marker of collagen breakdown, MMP13, and decrease of the anabolic markers of aggrecan and collagen type 2 production. Additionally, an increased release of sGAG from the explants suggests that mechanical loading at a strain of 65% induces cartilage degradation. To investigate the application of our model for pre-clinical testing, we inhibited the D2 activity by adding IOP to the model and observed abolishment of MMP13 induction and COL2A1 reduction by mechanical loading. These results show that IOP is able to prevent prolonged induction of cartilage breakdown upon mechanical loading and even leads to a less hypertrophic environment as seen by a reduced COL10A1 expression.

Purpose: TO IDENTIFY AND VALIDATE CIRCULATING MICRO RNAs (miRNAs) that mark ongoing osteoarthritis (OA) pathophysiology in articular cartilage. Osteoarthritis affects over 10% of the elderly population, yet, no disease modifying drugs are available. In part, development of such drugs is hindered due to lack of sensitive biomarkers that enable timely diagnosis, even before damage becomes visible with X-ray, and successful monitoring of disease activity over time. To this purpose, we focused on the use of miRNAs as promising small molecules that are secreted from (diseased) tissues into the circulation thereby reflecting ongoing pathophysiological processes (Figure 1). Methods: Cartilage (N=103) was collected within the ongoing RAAK study (Research in Articulair Artrose Kraakbeen), and plasma was collected within the RAAK study (N=31), and healthy controls (N=31; studies and its consent procedure are approved by the institutional ethics review committee 'Commissie Medische Ethiek' of the Leiden University Medical Center). RNA (miRNA and mRNA) were isolated, and high quality next generation RNA sequencing datasets were established of human preserved and lesioned OA cartilage as well as for plasma of healthy controls and OA patients. Notably, within the RAAK study the miRNAs in OA cartilage and plasma were determined for the same individual patients. Differential miRNA profiles of cartilage and plasma were assessed and correlated to gene expression of preserved cartilage. For this, 998 genes were selected, previously shown to be significant differentially expressed with at least 1.5-fold difference in cartilage of hip OA patients as compared to non-OA cartilage. Results: We identified two genes (Echinoderm Microtubule Associated Protein Like 1 or EML1 and Poliovirus Receptor or PVR) with differential expression in OA cartilage for which the expression significantly correlated with 16 miRNAs in plasma and that were also significantly different expressed in plasma of OA patients as compared to healthy controls. We were able to replicate nine out of the sixteen miRNAs in plasma of the GARP study (N=72) with similar size and direction of effects when compared to healthy controls. Of note, among the nine miRNAs we identified miR-451a. This miRNA is higher expressed in plasma of OA patients as compared to healthy controls (for RAAK and GARP, respectively 1.8-fold with P=2.3x10-2 and 1.4-fold with P=5.0x10-3). We have recently demonstrated that miR-451a is the most significant upregulated miRNA in lesioned as compared to preserved OA cartilage (2.3-fold; P=1.2x10-6). Moreover, this appeared to be mainly driven by hip OA as compared to knee OA cartilage. Conclusions: Here, we determined circulating micro RNAs correlating with specific OA-associated changes in gene expression as promising molecular biomarkers reflecting ongoing OA pathophysiology in articular cartilage. EML1 is increased in OA cartilage (1.3-fold; P=1.8x-2) while PVR is downregulated in cartilage of hip OA patients when compared to non-OA cartilage and subsequently upregulated in preserved compared to lesioned cartilage (1.3-fold; P=5.8x10-4). Therefore, miRNAs correlating with PVR may be highly suitable for early diagnosis. Currently we are analyzing expression of the 16 miRNAs in plasma of participants of the CHECK study at baseline for which we have available the 5-year follow-up progression data that is also available for the GARP study. With this, we will be able to establish applicability of our candidate miRNAs as markers for early diagnosis.

Purpose: We sought to identify a pathogenic mutation in an extended early onset osteoarthritis (OA) family. The value of such causal mutations lies in the functional characterization of the underlying biological mechanisms driving OA which, via extrapolation, can also give insight into common OA pathogenesis. Methods: Exome sequencing was applied to a relevant affected member of an extended family with early-onset primary OA without syndromic features. A prioritization scheme to identify the pathogenic mutation was applied, followed by de novo genotyping and linkage analysis across the family. Conformational changes in the protein were predicated by the online tool RaptorX (www.uchicago.raptorx.edu). To investigate the gene in common OA pathophysiology, expression analysis was performed in preserved and osteoarthritic cartilage of 33 independent patients undergoing joint replacement surgery (RAAK study). Results: Exome sequencing a relevant affected individual of the early-onset OA family resulted in the identification of a high impact, likely damaging, missense mutation in FN1, encoding fibronectin. De novo genotyping and linkage analysis confirmed that the FN1 mutation was carried in affected but not in un-affected family members (LOD score > 3), thereby being likely causal to the early onset of OA in this family. The identified mutation is located in the gelatin-binding domain and structure prediction indicated a significant change in the conformation of the protein. Gene expression analysis in preserved and affected cartilage in OA patients showed that the FN1 gene was among the highest upregulated in OA lesioned tissue compared to preserved tissue (2.3-fold, P-value = 1.13x10-5). Conclusions: We have demonstrated that a high impact mutation in FN1 is likely causal to OA in an extended family, thereby confirming that proper cellular function of fibronectin is crucial to maintain articular cartilage tissue homeostasis, whereas aberrant function of fibronectin potentially modulates underlying pathophysiology of OA. Due to their strong effect, high impact mutations can provide direct clues to underlying disease mechanisms as they are actionable to express a particular disease state in experimental in vitro models. Since the underlying pathway in which fibronectin is functioning is likely extrapolated to confer risk to common OA, the identified mutation provides a unique opportunity to dissect underlying mechanisms of the FN1 pathway in OA in the general population.

Purpose: By applying next generation sequencing to members of an extended early onset osteoarthritis (OA) family a gain of function mutation was identified within the TNFRSF11B gene which encoding osteoprotegerin (OPG). OPG is a decoy receptor which inhibits osteoclastogenesis by competing with RANK (Receptor activator of the nuclear factor KB) for the binding of RANKL (nuclear factor KB ligand). Of note is that expression of TNFRSF11B is highly increased during common OA pathophysiology. To better understand its mechanism of action we generated human in vitro 3D pellets and administered exogenously increasing doses of OPG. Methods: Human bone marrow derived stromal cells (hBMSC) were collected within the Research Arthritis and Articular Cartilage (RAAK) study, designed at collecting joint materials. Cells were differentiated into chondrocytes by applying a 3D pellet culture. Recombinant OPG was administered at different doses (10, 40, 100 and 500 ng/μL). Pellets were collected at 3, 5 and 7 weeks for analysis by Alcian Blue staining and qRT-PCR. Surface area of the pellets was determined as a measure of matrix deposition. Results: Upon adding OPG to the medium, an increase in pellet size was observed in a dose dependent manner indicating increased extracellular matrix deposition. Subsequent, qRT-PCR analyses showed a dose dependent increase in gene expression of COL2A1 and COL10A. The latter specifically at week 7 indicating increased cartilage hypertrophy. Histology with Alcian Blue did not show an apparent dose response increased staining with OPG. Conclusions: Our study demonstrates that upregulation of OPG levels during in vitro chondrogenesis enhances hypertrophic matrix deposition. Altogether our results indicate TNFRSF11B upregulation drives OA pathology by affecting the chondrocytes to engage an unbeneficial hypertrophic state.

Purpose: By applying 3D in vitro chondrogenesis in a model using human bone marrow mesenchymal stem cells (hBMSC), it was demonstrated that excess thyroid hormone (T3) and the upregulation of deiodinase iodothyronine type-2 (D2) gene (DIO2), had similar detrimental effects on cartilage matrix deposition. Moreover, an in vivo rat model further indicated that this unfavourable effect of DIO2 upregulation on articular cartilage is likely modulated by mechanical loading. In the current study our aim was to investigate the effect of excess T3 in a human ex vivo explant model in interaction with mechanical loading. Methods: Full thickness osteochondral explants were isolated from the macroscopically preserved condyles of three human osteoarthritic (OA) donors included in the RAAK study. From day 0 to 5 the explants were kept under standardized conditions in chondrogenic medium to normalize expression levels. All explants were loaded from day 6 to 9, with a regime of 1 hz (Mach-1 mechanical testing system, Biomomentum) at 30% strain, during 10 min. To determine the effect of thyroid hormone, 10 nM T3 was added to the culture medium. As primary output, RT-PCR was performed to measure changes in COL2A1 (anabolism), MMP13 (catabolism) and EPAS1 (hypertrophy) gene expression. To determine structural damage, paraffin sections of the cartilage were stained with hematoxylin and eosin (HE), Safranin O and Alcian blue staining. Results: In our human explants, excess T3 resulted in a significant upregulation of EPAS1 (P = 0.037) whereas for MMP13 and COL2A1 suggestive evidence for upregulation was found (P = 0.065 and P = 0.065, respectively). Additionally, staining with Safranin O, confirmed that excess T3 was detrimental to cartilage homeostasis as reflected by a loss of proteoglycans. Conclusions: Similar to our 3D in vitro chondrogenesis model with hBMSC, we here show that excessive T3 while applying 30% strain is detrimental to cartilage matrix homeostasis in a human explant culture model. The DIO2 gene, resulting in excessive T3 levels, has previously been identified as a susceptibility gene for OA and our results here further recognize the thyroid homeostasis as a potential therapeutic target of OA.