If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Address correspondence and reprint requests to: D. Chan, School of Biomedical Sciences, Faculty of Medicine, The University of Hong Kong, 21 Sassoon Road, Hong Kong. Tel: 852-3917-9482.
This “year in review” provides a summary of the research findings on the topic of genetics, genomics and epigenetics for osteoarthritis (OA) between Mar 2021 and Apr 2022. A search routine of the literature in PubMed for the keyword, osteoarthritis, together with topics on genetics, genomics, epigenetics, polymorphism, DNA methylation, noncoding RNA, lncRNA, proteomics, and single cell RNA sequencing, returned key research articles and relevant reviews. Following filtering of duplicates across search routines, 695 unique research articles and 112 reviews were identified. We manually curated these articles and selected 90 as references for this review. However, we were unable to refer to all these articles, and only used selected articles to highlight key outcomes and trends. The trend in genetics is on the meta-analysis of existing cohorts with comparable genetic and phenotype characterization of OA; in particular, clear definition of endophenotypes to enhance the genetic power. Further, many researchers are realizing the power of big data and multi-omics approaches to gain molecular insights for OA, and this has opened innovative approaches to include transcriptomics and epigenetics data as quantitative trait loci (QTLs). Given that most of the genetic loci for OA are not located within coding regions of genes, implying the impact is likely to be on gene regulation, epigenetics is a hot topic, and there is a surge in studies relating to the role of miRNA and long non-coding RNA on cartilage biology and pathology. The findings are exciting and new insights are provided in this review to summarize a year of research and the road map to capture all new innovations to achieve the desired goal in OA prevention and treatment.
Osteoarthritis (OA) as a complex degenerative disease with many interacting factors. OA not only leads to cartilage degeneration causing pain, but changes within the joint and surrounding tissues. Understanding the genetic contribution to the development and progression of OA will provide molecular insights to the disease and thus therapeutic and preventative measures. However, as there are many sub-phenotypes or traits to consider, including weight loading, posing major challenges in genetic studies. Over the years, many potential genetic risk factors have been identified, but the effect size is not large, with odds ratios mostly around 1.3 or under
. Further, the individual cohort studies are generally not large; thus, the genetic power can be limiting, and the available resources in genetic studies of OA not fully utilized.
A current trend is to increase the genetic power by bringing together different cohorts to perform meta-analyses with accurate characterization of the OA phenotype with subclassifications (endophenotypes). With a larger combined cohort size, this presents opportunities for more in-depth assessment, reliable and reproducible findings. Many of the associated genetic changes are not within coding regions of genes, suggesting changes in potential regulatory elements alter the expression pattern of a risk factor. Further, there is a need to correlate genetics with epigenetic changes and the global expression of genes in normal and diseased states. Thus, it is necessary to synergize genetic findings with Omics innovations to derive meaningful integration of data for translational outcomes.
This “2022 year in review” looks at the research outputs in OA genetics, genomics, epigenetics, and related topics. A search of these key words with OA returned numerous articles and reviews in the last 12 months, showing a strong interest and research activities in this area. A search routine (Fig. 1) between Apr 2021 and Mar 2022 in PubMed resulted in 695 unique publications for the topics indicated. We then manually curated 90 of these and selected relevant articles within the focus of this review.
Fig. 1Flow chart of the PubMed search routine and result. Workflow of literature search for Genetics, epigenetics, omics, and expression in OA related publications during the reviewed period. The outcome is presented in a graphical format depicting the number of original research articles and reviews.
. The authors provided an account of 124 single nucleotide polymorphisms (SNPs) or variants (SNVs) that have been identified, with the variances accounting for only about 20% of the estimated hereditability. Further, the authors systematically documented these SNPs and correlated their relative positions in the genome with respect to the immediate or nearest protein coding gene(s). Only 9.8% of the associated SNPs are nonsynonymous changes (missense or nonsense mutations) within the coding regions of genes, 1.7% and 4.9% in the 5′UTR and 3′UTR regions of a gene, respectively; and majority are intronic (54.1%) or intergenic (29.5%), supporting that many of the changes are likely to affect gene expression level or pattern of expression.
highlighted the importance of performing fine mapping analyses to identify accompanying variances in linkage disequilibrium (LD) and apply posterior probabilities (PP) to identify credible causal SNPs, with reference to studies that have applied this approach to several genome wide association studies (GWAS) of OA
. Also provided is a road map for further assessments of risk variants and in silico analyses that are currently applied in other genetic studies with complex traits, linking to quantitative trait loci (QTL); and other Omic data and in vitro functional analyses for translation.
A good example of integration and mining of post-GWAS data study is an association of phenotypes, based on clustering patterns of hand OA that prioritized WNT9A as an OA gene, with the study centered around the most consistent SNP, rs109161199, associated with thumb OA, leading to the identification of a causal variant, rs1158850, in a regulatory element of the WNT9A locus
. Key to the success is the stratification of the cohort into three hand OA endophenotypes (thumb, finger, and hand), highlighting the importance of accurate and consistent phenotype characteristics to enhance the power of genetic studies. Tools to assess the functional relationship between risk factors with diverse functions is through in silico interrogation of potential network of protein interactions such as STRING
(ALDH1A2, COLGALT2, GDF5, MGP, PLEC, RUNX2, RWDD2B and TGFB1), six (bold) can be located within a STRING protein–protein interaction network. Such interaction provides knowledge on the biochemical interactions, and thus information on the “liability threshold” of polygenic diseases involving multiple risk alleles, and an opportunity to bring in unconfirmed casual genes as OA risk loci.
Validation of OA risk factors
The 2021 year in review on genetics of OA in this topic
provided a summary of the recent discoveries and examination of genetic variants up to the first quarter of 2021. A message was that there are less GWAS studies but a focus on consolidation of previous identified OA candidate genes. This is indeed still the case. This year, a further replication/meta-analysis on the rs143383 polymorphism of GDF5 on Caucasian cohorts firmly established this functional variant with knee OA
. This variant was first shown to be strongly associated with both knee and hip OA in Japanese and Han Chinese populations, with the major (T) allele causing a reduced expression level of GDF5 for an increasing risk (OR 1.13, 95% CI 1.06–1.20)
Correlation between growth differentiation factor 5 (rs143383) gene polymorphism and knee osteoarthritis: an updated systematic review and meta-analysis.
Association between growth differentiation factor 5 rs143383 genetic polymorphism and the risk of knee osteoarthritis among Caucasian but not Asian: a meta-analysis.
Identification of functional outcomes of risk alleles is of major importance, and expression assessment can assist to narrow down the effector genes at the OA locus. Several recent studies
Multi-tissue epigenetic and gene expression analysis combined with epigenome modulation identifies RWDD2B as a target of osteoarthritis susceptibility.
Prioritization of PLEC and GRINA as osteoarthritis risk genes through the identification and characterization of novel methylation quantitative trait loci.
have demonstrated the value of allelic expression imbalance (AEI) analysis in quantifying the relative ratio of mRNA transcripts from each allele of the SNP, and in the case of the matrix Gla protein gene (MGP), confirming this gene as the mediator of OA with the locus at rs4764133
. The study design included a control group conformed to Hardy–Weinberg equilibrium (HWE), generalizing the finding to a normal population. rs6426749 was shown to affect LINC00339 and CDC42 expression, possibly via transcription factor TFAP2A, for a motif with the C risk allele, allowing a stronger binding
. A hypothesis is that rs6426749 increases CDC42 expression, which is linked to increased WNT4 expression that would promote osteoblast differentiation. This known inverse relationship between OA and osteoporosis is complex and remains unresolved
RNA sequencing reveals interacting key determinants of osteoarthritis acting in subchondral bone and articular cartilage: identification of IL11 and CHADL as attractive treatment targets.
. Indeed, this association was confirmed in the Chinese population through a combination of discovery and replication cohorts, with the susceptibility and clinical severity of OA
. It is also considered as a target for treatment and prognostic prediction in OA. It is important to note that while rs1126464 is mapped to DPEP1, it does not seem to affect its expression, and that 15 surrounding genes could serve as candidate loci for OA. Further, a limitation of the study is the difference of phenotypes in the discovery (knee) and replication (hip) cohorts, although the study from the UK Biobank is based on mixed OA patients. Lastly, a study of 32 OA risk factors in another Chinese cohort showed TP63 to be associated with an endemic OA, Kashin-beck disease, with clinical manifestations and radiographic features in adult patients similar to OA, indicating that rare and common OA conditions are linked
Discovering novel genetic loci with OA endophenotypes
The importance of large cohort sizes and uniform traits for the genetic study of complex diseases is clearly illustrated in a comprehensive assessment of OA risk factors, with one publication dominating this year's genetic studies
, demonstrating the need for multi-center international collaborations. Given the small effective size of most OA risk variants, huge cohort sizes are needed for statistical power. This meta-analysis of GWAS data consisted of 177,517 OA patients, and 649,173 controls, from 9 populations, doubling the largest cohort studied so far. It featured 11 OA sub phenotypes in knee, hip, hand, and spine. From a total of 13 combined cohorts, 11 were of European descent, and 2 were from East Asia. With a GWAS significance set at P < 1.3 × 10−8 to account for the effective number of independent tests, 11,897 significant SNVs were found. Joint replacements for knee and hip OA were included as endophenotypes, and hand OA was subclassified into thumb and finger OA.
Applying conditional analyses for associations that do not overlap across disease phenotype definitions, 100 unique and independent variants were identified, with 60 associated with more than one sub phenotype, and 52 novel variants that have not been reported before
. When defining the lead SNVs with the strongest statistical evidence as the risk variant, 6 are missense changes in coding regions of a gene, 59 are within non-coding regions, and 35 are intergenic, consistent with the overall distribution assessment prior to this study
. Importantly, 62% of the signals had supporting evidence in an East-Asia ancestry-only analysis, albeit patient numbers were small, suggesting relevance of the findings across ethnicities.
Gender specific variants were also identified; 3 were female-specific or with effect of opposite directions
, of which 2 showed significant differences in effect size between gender. One variant, rs11612221, showed significance in the female-only hip replacement phenotype. This variant is located in a region containing long intergenic non-coding RNA, with FANCL as the nearest protein coding gene
. The other 2 variants, rs10282983 and rs10453201 are located in an intron of C8orf34 and 5′ region of UBAP2, respectively. Of note, both these variants have been shown to be associated with heel bone mineral density
. A new variant, rs148693048, associated with early onset OA was also identified, with two proximity genes, NEFM and DOCK5, showing different expression levels between intact and degrading cartilage
. In support of DOCK5 as a causal gene, intronic variations in this gene is also associated with other bone phenotypes such as heel bone mineral density and adolescent idiopathic scoliosis
Identification of 613 new loci associated with heel bone mineral density and a polygenic risk score for bone mineral density, osteoporosis and fracture.
The coexistence of copy number variations (CNVs) and single nucleotide polymorphisms (SNPs) at a locus can result in distorted calculations of the significance in associating SNPs to disease.
Specific and common variants between weight bearing (knee, hip and spine) and non-weight bearing (hand) joints can inform underlying mechanisms in OA. In a multi-population cohort study
, 40 and 4 SNVs were specific to weight-bearing and non-weight bearing OA joint, with 42 that were common. Thus, in addition to variants in GDF5 from previous studies, new SNVs, rs3771501 (TGFA), rs3993110 (TEAD1/DKK3), rs72979233 (CHRDL2), and rs7967762 (PFKM/WNT10B) were associated across these joint sites
Critical to OA is the management of symptomatic pain. This area is understudied due to the complexity of pain pathways and the perception of pain. Of interest, insights were obtained in reference to SNVs with larger effect size in patients receiving knee or hip replacements, likely linked to pain and inflammation. Thus, some of these variants related to more than just OA pathology. Here, the variants and inferred genes included rs76340814 (PTCH1) and rs28929474 (SERPINA) in neurological and inflammatory pathways for painful OA, therefore the need for joint replacements
. There is also a strong correlation between OA and sciatica, fibromyalgia, headache, and other back pain phenotypes, with the strongest correlation with spine OA
. Thus, cross disease studies will provide useful information and we need to take note of related findings, such as a rare variant in SLC13A1 showing the role of sulphate in intervertebral disc function and pain
, with implication for synovial joint biology and pathology. Further, this relationship is supported with shared genetics between major depression and OA pain
A road map for molecular insights toward OA management and treatment is an integration of genetics with other Omic datasets such as transcriptomics, proteomics and epigenomics. A recent review provided a summary of the insights into the molecular landscape of human OA tissues
. An early example of a multi-omic approach is a study by Steinberg and colleagues, examining mRNA and protein profiles of low-grade and high-grade OA cartilage, and synovium tissue of 115 patients with total joint replacement which combined the data with patient genotypes
. The authors examined differentially expressed genes with eQTLs as well as hallmarks of cartilage degradation in high-grade OA cartilage and identified 36 markers as high value targets for drug development and repurposing. Indeed, 19 potential therapeutic compounds were proposed to alter expression of these genes in high-grade cartilage. IL-11, associated with inflammation and increased risk of hip OA and knee OA, was highlighted in this study as a point of intervention
RNA sequencing reveals interacting key determinants of osteoarthritis acting in subchondral bone and articular cartilage: identification of IL11 and CHADL as attractive treatment targets.
. In resolving GWAS signals, the authors looked for high posterior probability for co-localization with molecular QTLs. Strong evidence was identified for 5 OA GWAS signals; rs10502437 (NPC1), rs11732213 (FAM53A), rs12901372 (SMAD3), rs1560707 (SLC44A2) and rs4775006 (ALDH1A2). As the index variants are all in non-coding regions, the likely effector gene (as indicated) is one that is closest to the variant or predicted.
Molecular signatures and effector genes for OA
An important translational outcome would be the identification of molecular signatures for diagnostic and prognostic predictions. Clustering and multi-omic factor analysis of OA cartilage and synovial transcriptome profiles from 113 OA patients showed that low-grade OA cartilage and synovium share biological processes but also identified independent clusters
. Thus, changes in OA cartilage and synovium are not the same. However, a seven-gene classifier of inflammatory markers was proposed, where the expression of genes MMP1, MMP2, MMP13, IL6, CYTL1, APOD, and C15orf48, improved generalizability to predict probabilities of knee low-grade OA cartilage
. Of interest, a study showed higher inflammation in the OA cartilage cluster with the female sex, which is in line with the disproportionate increase in the occurrence of OA in women post-menopause, possibly due to the lower concentration of estrogen and androgens
(COL6A3, COL1A2, COL1A1, MMP2, COL3A1, POST, and FN1) diagnostic signature of OA was proposed. These gene sets are quite variable, indicating heterogeneity, and further integration of data is needed for better genotype to phenotype correlation.
Following the comprehensive and informative multi-center study and cohorts to discover genetic loci
, here, the authors used the lead variants to discover candidate genes, and their expression enrichment in joint tissues to discover eQTLs, and together with available phenotype from human and animal models, arrived at a set of 77 potential effector genes from 637 candidates. Of note, there is a large contribution of genes and pathways involved in skeletal and neuronal development and function, leading to joint degeneration and symptomatic outcomes. These included the ECM niche of a chondrocyte, and OA implicated pathways such as FGF, TGFβ, Hedgehog and Wnt signaling, as well as mechanosensing.
Epigenetics and OA
Given that most of the genetic variants for OA are in non-coding regions of the genome, it is reasonable to assume that majority of the variants affect gene regulation altering protein expression, rather than protein sequence altering function. In addition to transcriptional regulation through cis regulatory elements, epigenetics is a mechanism by which a cell can change its expression characteristics with no changes in the primary sequence. This can be achieved via chemical modification of the nucleotides, as in the methylation of cytosine-guanine dinucleotides (CpGs), post-translational modifications of histones, or expression of non-coding regulatory RNAs. Changes in DNA methylation in human OA cartilage have been reported
, was used to functionally test 1,605 single nucleotide variants from 35 known OA risk loci that identified 6 SNPs with differential regulatory activity between the major and minor alleles, with rs4730222 as most significant, likely by impacting expression of an alternative isoform of HBP1
. Using a deactivated Cas9 (dCas9)-TET1 system to specifically edit targeted CpG methylation sites in a cell model to detect methylation QTLs (mQTLs), methylation at CpG cg20220242, correlating to the rs6516886 genotype was identified, acting on the gene, RWDD2B
Multi-tissue epigenetic and gene expression analysis combined with epigenome modulation identifies RWDD2B as a target of osteoarthritis susceptibility.
This year, several studies addressed DNA methylation in OA. Using available normal and OA cartilage datasets for methylation (GSE63695), and mRNA expression (GSE114007 and GSE113825) profiling datasets from the Gene Expression Omnibus (GEO) database, differentially methylated genes (DMG) and differentially methylated sites (DMS) for OA were assessed
. Close to 2,000 differentially expressed genes (DEGs), and 574 DMS that contained 394 DMGs were identified between the normal and OA cartilage, and mapped to potential hypermethylated and hypomethylated genes in OA
. A study also addressed the genetic (rs11583641) and epigenetic interplay within an enhancer of COLGALT2 associated with OA, correlated with DNA methylation at three CpGs, while the risk C allele of rs11583641 corresponded to reduced levels of methylation
. COLGALT2 encodes for a galactosyltransferase for the glycosylation of collagens and is therefore a potential OA target. Leveraging on a mouse model of posttraumatic OA that showed pronounced DNA methylation (5mC) changes in chondrocytes in OA progression
, Singh and colleagues showed that Lrrc15 is differentially methylated and expressed in OA cartilage, linked to the cellular stress response, and dysregulated ECM remodeling. For histone modification in OA, a recent review provided an excellent account of the histone-modifying proteins involved in the pathogenesis and progression of OA, with examples given that could promote or suppress the expression of SOX9 and COL2A1
Identification and correlation of non-coding RNA to OA is gaining ground on mechanism and regulatory controls, as biomarkers in OA pathogenesis, or targets as therapeutic options. These include short nucleotide transcripts as micro-RNA ((miRNAs) of less than 200bp, long linear noncoding (lncRNA) or circular (circRNA) RNA of more than 200bp transcripts. miRNAs are more prominent and in general, repressive, by destabilizing targeted mRNAs, or inhibition of protein translation [Fig. 2(A)]. CircRNA is of particular interest for clinical application as they act as a sponge of miRNA in fine tuning gene expression that can inhibit activity of one or multiple miRNAs via direct interaction
. Here, we will add recent articles on micRNA, lcnRNA and circRNA in OA related studies, building on the proposed relationship and interactions, regulating inflammation, cartilage degradation, ECM synthesis, cellular proliferation, and apoptosis [Fig. 2(B)].
Fig. 2Role of non-coding RNA in OA biology and pathology. (A) Diagrammatic representation on the interaction of non-coding RNAs and their role in regulation the level of protein coding mRNA and thus protein function for the various cellular events known to be affected in cartilage biology and pathology. (B) Recent findings of non-coding RNAs in cartilage homeostasis and diseases. The proposed long linear and circular noncoding RNAs, the miRNA that it regulates, and the target mRNA are depicted for each of the axis described in this review, and the arrows point to the various cellular processes that the axis regulates. The axes presented in the 2021 year in review (Young et al., 2021)
. Activation of the NOD-like receptor containing a pyrin domain 3 (NLRP3) inflammasome pathway leads to production of IL-1β, inducing pyroptosis. Zhang and colleagues showed that Cathepsin B (CTSB) as a regulator of the activation of NLRP3 inflammasome can be suppressed by miR-140-5p, reducing chondrocyte pyroptosis in OA
MicroRNA-140-5p represses chondrocyte pyroptosis and relieves cartilage injury in osteoarthritis by inhibiting cathepsin B/Nod-like receptor protein 3.
. Thus, increasing the level of miR-140-5p can be considered as therapeutic option. The NLRP3 inflammasome pathway interacts with the oxidative stress pathway sensing the release of reactive oxygen species (ROS) from mitochondrial damage
On the topic of cellular stress pathways, Ito and colleagues showed that both microRNA-455 (miR-455), -5p and -3p, interact with the hypoxia pathway, by suppressing HIF-2α
. Functionally, the authors showed that mice with inactivation of miR-455 developed cartilage degeneration mimicking OA, and HIF-2α is a direct target of both miR-455-5p and -3p, a catabolic factor necessary cartilage homeostasis
. Further, miR-455-5p and -3p are upregulated by SOX9, and thus would be interesting to assess whether the unfolded protein response (UPR) or integrated stress response (ISR)
Previously, miR-128-3p was shown to be associated with inflammation and ECM degradation, interacting with SIRT1 to mediate TNF-α in an inflammatory response
MiR-128-3p post-transcriptionally inhibits WISP1 to suppress apoptosis and inflammation in human articular chondrocytes via the PI3K/AKT/NF-kappaB signaling pathway.
. Here, the authors showed that miR-128-3p expression was negatively correlated with ZEB1, and that ZEB1 is a direct target of miR-128-3p. This explains the inverse relationship in OA biology where miR-128-3p is low and ZEB1 high, leading to increased ECM degradation and chondrocyte apoptosis
. Serum/glucocorticoid regulated kinase 1, SGK1, is also highly expressed in OA cartilage, mediating an IL-1β-induced anabolic and catabolic imbalance in human chondrocytes
Knockdown of SGK1 alleviates the IL-1beta-induced chondrocyte anabolic and catabolic imbalance by activating FoxO1-mediated autophagy in human chondrocytes.
. Interestingly, the level of miR-503-5p can be downregulated by the binding of histone deacetylases (HDACs) to the promoter of miR-503-5p and inhibiting H3K27ac expression
Histone deacetylases and NF-kB signaling coordinate expression of CX3CL1 in epithelial cells in response to microbial challenge by suppressing miR-424 and miR-503.
. Based on these premises, a study by Wang and colleagues showed that SGK1 is a target of miR-503-5p, and that HDAC2 could promote OA through miR-503-5p/SGK1 axis
, suggesting an inter-connection. Most recently, Zhang and colleagues showed dual functions of miR-17 in maintaining cartilage homeostasis and protection against OA
. Here, miR-17 through induction of GDF5 maintains cartilage homeostasis and protects against OA progression which is achieved through the targeting of pathological catabolic factors, matrix metallopeptidase-3/13 (MMP3/13), aggrecanase-2 (ADAMTS5), and nitric oxide synthase-2 (NOS2). This is consistent with decreased expression of miR-17 in OA contributing to disease progression
. Adding to these and function is the lncRNA HOX antisense intergenic RNA myeloid 1 (HOTAIRM1), first identified as a myeloid-specific regulator in the HOXA gene family
Downregulation of long noncoding RNA HOTAIRM1 variant 1 contributes to osteoarthritis via regulating miR-125b/BMPR2 axis and activating JNK/MAPK/ERK pathway.
. This year, a study by Liu and colleagues confirmed that HOTAIRM1-1 and miR-125b are inversely expressed, that are up and down regulated in OA cartilage, respectively
. Current data suggest that miR-125b influences proliferation and apoptosis of chondrocytes in OA development, and its level is regulated by a direct interaction with HOTAIRM1-1, confirming a hypothesis that miR-125b could be a target of HOTAIRM1
. While the precise mechanism is still unclear, down regulation of HOTAIRM1-1 leads to increased expression of MMP-13 and IL-10 expression, while collagen II and aggrecan protein expression were decreased
. It was shown that HOTAIR expression was enhanced while miR-107 was reduced in OA cartilage. Mechanistically, HOTAIR acts as a sponge for miR-107, and CXCL12 is a target of miR-107, resulting in cartilage ECM degradation and chondrocyte proliferation and apoptosis.
In an assessment of differentially expressed lncRNAs between mild and severe OA, lncRNA LINC01385 expression was shown to be increased
. To rationalize the significance of LINC01385, Wang and colleagues showed a relationship in OA progression via interaction with miRNA-140-3p, of which TLR4 is a target. Thus, reducing LINC01385 or the use of miR-140-3p mimics can reduce the level of inflammatory factors in IL-1β-induced human articular chondrocytes, and promoting cell survival
circRNAs constitute a large subset of ncRNAs involved in cartilage homeostasis, regulating chondrocyte proliferation, apoptosis, and degradation of articular cartilage. Their abnormal expression if closely associated with OA development and progress
. Such diversity is further illustrated in recent findings for protective or risk outcomes for OA elicited through various circRNA/miRNA/gene axes.
Levels of circSEC24A (hsa_circ_0005105) are induced in chondrocytes that have been stimulated with IL-1β. circSEC24A antagonize the function of miR-26a, and its target (NAMPT) is increased, causing an imbalance in ECM synthesis and degradation
. Recently, Shi and colleagues identified another function of circSEC24A regulating chondrocyte proliferation, apoptosis, and inflammation through a circSEC24A/miR-142-5p/SOX5 axis
. SOX5 is a target of miR-142-5p, and circSEC24A functions as a sponge to reduce the level of miR-142-5p. However, the role of SOX5 in OA progression is not clear. It is known that SOX5 partners with SOX9 and SOX6 in regulating chondrocyte differentiation
. Similarly, circ_0001598 is upregulated in chondrocytes treated with IL-1β, and in cartilage tissue from destabilization OA mouse model induced from anterior cruciate ligament transection (ACLT)
. Consequently, the level of miR-127-3p is lowered, with reduced chondrocyte proliferation and increased apoptosis. However, the precise mechanism of this circ_0001598/miR-127-3p axis in OA pathology is unclear.
Two other circRNA/miRNA/gene axes promoting OA progression included a circ_0128846/miR-127-5p/NAMPT axis
. In this study, circ_0128846 was confirmed to be upregulated in OA cartilage, and reduced levels in OA chondrocytes had a positive effect on cell viability, while reducing apoptotic, inflammation and ECM degradation processes. NAMPT is nicotinamide phosphoribosyltransferase, with known catabolic
Crucial role of visfatin/pre-B cell colony-enhancing factor in matrix degradation and prostaglandin E2 synthesis in chondrocytes: possible influence on osteoarthritis.
. In OA cartilage, the expression of circRHOT1 and CCND1 were elevated, while miR-142-5p was reduced. Similar findings were observed in a rat OA model, and OA characteristics such as ECM degradation were suppressed with shcircRHOT1 treatment. An association with autophagy was proposed, but the relationship leading to OA is still not clear.
On protective roles of cicrRNAs that regulate ECM homeostasis in cartilage and chondrocyte survival, circSLC7A2 was found to be downregulated in OA cartilage
. CircSLC7A2 appears to act as a sponge for miR-4498, with TIMP3 as a potential target of miR-4498. Increased levels of miR-4498 as a result of circSLC7A2 down-regulation could enhance MMP activities that include MMP3, MMP13 and ADMATS5, and thus degradation of collagen II and aggrecan in OA pathology
. HSA_circ_0005567 inhibits synovial M1 but promotes M2 macrophage polarization in suppressing OA development, acting through a miR-492/SOCS2 axis. Given hsa_circ_0005567 levels are downregulated in synovial tissues of OA patients, SCOCS2 levels would be increased. Thus, the effect on chondrocyte viability and apoptosis is by promoting M2 type macrophage polarization
Many of the findings arising from the OA genetics and -omic integrations will need validation but support the concept that minor changes in the cartilage and bone shape, ECM and cell content in early development, could impact initiation and progression of degenerative changes later in life because of genetic variants. By combining multiple lines of evidence, it is possible to further pinpoint likely causal genes as drug targets
. Many have existing FDA approval, providing a steppingstone to the repurposing of drugs for fast-track clinical OA applications. The resource paper by Boar, Hatzikotoulas, Southam and colleagues
, will be invaluable for future integration of new genetic, QTLs and functional data. The road map in realizing the potential of Omic analyses to patient outcomes is still a long way off (Fig. 3), but the progress is rapid. There is a need to bring in more genetic cohorts from Asian populations for confirmation or meta-analyses. Clear accurate OA endophenotypes need to be maintained. This together with more available single cell data of OA cartilage tissues in development and progression, and with better understanding of the complex regulatory mechanisms through ATAC-Seq and non-coding RNA data, we can begin to build integrative information assisted by artificial intelligence and high throughput screening protocols for gene and drug functions. The outlook is optimistic for OA prevention and treatment.
Fig. 3Road map from genetics to clinical translation. Genetics and omics analyses have enriched our understanding of cartilage homeostasis and OA pathogenesis. Based on the availability of novel information and technologies, we are moving a comprehensive understanding of the genetics of OA endophenotypes, their interactions with the epigenome in regulation gene function, and thus transcriptomic, proteomic, and metabolomic changes at a single cell level, as we move forward in development of diagnostic, predictive and the goal of prevention and therapeutic measures for OA.
All authors were involved in the design of the review, analysis of the literature and interpretation of studies included. MK, AR and DC assessed the literature to be included. WCWC provided the artistic content of the figures. All authors are involved in the drafting and revisions of the manuscript and approved the final version for submission.
Conflict of interest
All authors declare no conflicts of interest.
Acknowledgments
This work was supported by Collaborative Research Grant (C7030-18GF) from the Research Council of Hong Kong, and the S.Y. and H.Y. Cheng Professor endowment in Stem Cell Biology and Regenerative Medicine to DC. We thank Dr Vivian Tam for proof-reading of the manuscript. Figures are created with BioRender.com.
Correlation between growth differentiation factor 5 (rs143383) gene polymorphism and knee osteoarthritis: an updated systematic review and meta-analysis.
Association between growth differentiation factor 5 rs143383 genetic polymorphism and the risk of knee osteoarthritis among Caucasian but not Asian: a meta-analysis.
Multi-tissue epigenetic and gene expression analysis combined with epigenome modulation identifies RWDD2B as a target of osteoarthritis susceptibility.
Prioritization of PLEC and GRINA as osteoarthritis risk genes through the identification and characterization of novel methylation quantitative trait loci.
RNA sequencing reveals interacting key determinants of osteoarthritis acting in subchondral bone and articular cartilage: identification of IL11 and CHADL as attractive treatment targets.
Identification of 613 new loci associated with heel bone mineral density and a polygenic risk score for bone mineral density, osteoporosis and fracture.
The coexistence of copy number variations (CNVs) and single nucleotide polymorphisms (SNPs) at a locus can result in distorted calculations of the significance in associating SNPs to disease.
MicroRNA-140-5p represses chondrocyte pyroptosis and relieves cartilage injury in osteoarthritis by inhibiting cathepsin B/Nod-like receptor protein 3.
MiR-128-3p post-transcriptionally inhibits WISP1 to suppress apoptosis and inflammation in human articular chondrocytes via the PI3K/AKT/NF-kappaB signaling pathway.
Knockdown of SGK1 alleviates the IL-1beta-induced chondrocyte anabolic and catabolic imbalance by activating FoxO1-mediated autophagy in human chondrocytes.
Histone deacetylases and NF-kB signaling coordinate expression of CX3CL1 in epithelial cells in response to microbial challenge by suppressing miR-424 and miR-503.
Downregulation of long noncoding RNA HOTAIRM1 variant 1 contributes to osteoarthritis via regulating miR-125b/BMPR2 axis and activating JNK/MAPK/ERK pathway.
Crucial role of visfatin/pre-B cell colony-enhancing factor in matrix degradation and prostaglandin E2 synthesis in chondrocytes: possible influence on osteoarthritis.
00725-2&id=gr1.jpg){kind=link}
00725-2&id=gr2.jpg){kind=link}
00725-2&id=gr3.jpg){kind=link}