Osteoarthritis (OA) is a common disease of the synovial joints characterised by articular cartilage thinning and loss, which is often accompanied by alteration in the normal function of other tissues of the joint.
As with most common diseases, genetic susceptibility is a major component in the development and progression of OA, although little is currently known about the exact nature of the genes involved. Nevertheless, some breakthroughs have been made. This review will focus on studies implicating genes whose encoded proteins are active in a well-defined biological pathway, TGF-b signalling.
The strategy used for the association mapping of susceptibility loci for common diseases has evolved in a clearly defined manner, dictated by technology, cost and outcomes (Fig. 1).
Fig. 1 The natural history of the association mapping of complex trait loci
In the candidate gene approach, relatively limited biological knowledge of disease aetiology is used to select a handful of genes that could plausibly be considered to harbour polymorphisms that modulate gene (or protein) activity and/or function. These could also influence the occurrence or outcome of the disease of interest. This hypothesis-driven approach can result in some spectacular successes. However, bearing in mind the publication bias working against negative studies, such that many are quietly buried, and the fact that most “positive” candidate association studies have p-values that are unconvincing and have never been subjected to independent replication, there are few success stories.
The hypothesis-free approach of the genome-wide association scan removes subjectivity (at least initially) and permits the discovery of the unknown, thus revealing new insights into disease aetiology that may change the understanding of how a particular disease is initiated and subsequently develops.
Initial genome scans used relatively porous polymorphism maps and small discovery samples; the number of cases and controls combined rarely exceeded 2000 individuals. These are the medium-powered studies already described and, although there were some successes, the porous maps and low sample numbers worked against comprehensive outcomes. There have been several examples of the medium-powered scans in OA1-3 but none of the loci reported have shown consistent replication.
Recently, OA moved into the realm of the high-powered genome scan, with high-density maps of over 500,000 polymorphisms and large cohorts that, after meta-analysis, exceeded 50,000 individuals. Examples include the Rotterdam scan published in early 20104 and the Arthritis Research UK funded arcOGEN study, which has recently completed. These are true discovery studies. For example, the main signal from the Rotterdam scan resides on chromosome 7q22 to a locus of high-linkage disequilibrium that stretches over 500kb and contains at least six known genes, none of which is an obvious OA candidate.
The agnostic genome-wide scan is clearly the future, with whole-genome sequencing being the likely next frontier. However, while medical science waits for robust, replicated loci to emerge from such scans, there is an understandable temptation to reflect on the outcomes from the candidate-gene approaches. It is also worth remembering that a flavour of the candidate philosophy will pervade the genome-scan arena, since in many instances investigators will be sorely tempted to prioritise genes from among those within the genome signals.
The candidate genes targeted can be grouped into three broad categories: those coding for structural components of the cartilage extracellular matrix, those coding for inflammatory molecules, and those coding for catabolic/anabolic regulators. Examples of each include COL2A1 (a1 chain of type II collagen), IL1RN (interleukin-1 receptor antagonist) and GDF5 (growth and differentiation factor 5), respectively. Of the three categories, the most compelling data have come from the catabolic/anabolic regulators and of these, members of the TGF-b super-family have proved to be particularly interesting.
It has been known for many years that TGF-b signalling plays a critical role in the homeostatic balance of articulating joint tissues. This pathway consists of the standard components for a cell signalling cascade: 1) small extracellular molecules that initiate the signal, often interacting with structural matrix molecules; 2) cell surface receptors to which the small molecules bind and which then transmit the signal into the cell; 2) intracellular molecules that, through a complex series of interactions, ultimately translocate the signal to the nucleus in order to initiate, enhance or inhibit gene transcription. For OA, many components of this pathway have been implicated in the disease and, from a genetics perspective, this article will focus on three of the more compelling.
Growth and differentiation factor-5 (GDF5) is also known as cartilage-derived morphogenetic protein-1 (CDMP1). It is an extracellular signalling molecule that participates in the development, maintenance and repair of bone, cartilage and other tissues of the synovial joint.5 Based on this functional knowledge the GDF5 gene was chosen by a Japanese group for examination as a potential OA susceptibility locus.6 A number of common GDF5 polymorphisms were genotyped in cases and in controls and association to the common single nucleotide polymorphism (SNP) rs143383, a T to C transition located in the 5' untranslated region of the gene, was observed. In the discovery part of this study the association easily surpassed genome-wide significance, with a p-value of 1.8x10-13, and the investigators also replicated the association in additional Asian cohorts. In subsequent studies, association with OA was observed in Europeans, demonstrating that the association had global relevance.7 Association of OA to rs143383 is currently the most compelling finding from the numerous OA genetic studies so far reported.
Functional studies have suggested that rs143383 may itself be the polymorphism that influences OA susceptibility, with the OA-associated T allele mediating reduced GDF5 transcription relative to the C allele.6,8 This effect appears to be modifiable by other polymorphisms in the gene and a candidate trans-acting factor, DEAF-1, has been proposed.9 Furthermore, association of rs143383 to other musculoskeletal phenotypes, including variation in normal height, Achilles tendon pathology, fracture risk and congenital dysplasia of the hip10-14 highlights the tendency of a common allele to influence multiple phenotypes. It also emphasises the potential role that developmental factors can have on conditions associated with the mature individual.
Asporin is an extracellular matrix (ECM) macromolecule belonging to the small leucine-rich proteoglycan (SLRP) protein family.15 SLRP family members are able to bind other structural components of the ECM, such as collagen, as well as growth factors that temporarily reside in the ECM, such as TGF-b. Asporin is present in a number of tissues, including adult articular cartilage.
The same Japanese group as performed the GDF5 investigation also performed the ASPN study, once again on the basis that this gene encodes a protein of functional relevance to joint tissues. In the discovery study, also performed in Asian cohorts, the investigators observed association of a triplet repeat within exon 2 of ASPN, coding for a polymorphic stretch of aspartic acid residues in the N-terminal region of the protein.16 This repeat polymorphism, termed the D-repeat after the one-letter code for aspartic acid, had ten alleles encoding 10–19 D residues. The D14 allele was more common in individuals with OA knee relative to controls, with a p-value of 0.00024, and in OA hip relative to controls, with a p-value of 0.0078. As well as the association to D14, the investigators also noticed that one allele, D13, was consistently under-represented in the OA cases. It appeared, therefore, that an OA-risk allele (D14) and an OA-protective allele (D13) had been detected. Furthermore, additional genetic studies in Asians have suggested a role for the D14 allele in lumbar disc disease, a common chronic condition that, like OA, is age-associated and degenerative.17
The investigators next conducted a number of functional studies.16,18,19 These implied that asporin interacts with TGF-b and attenuates its signalling effect by inhibiting its binding to its receptor. This leads to a reduction in the expression of genes coding for important structural components of cartilage, such as aggrecan and type-II collagen. This inhibitory effect was particularly strong for asporin encoded by the OA-risk allele, D14, and less so for the OA-protective allele, D13. These functional studies thus provided a plausible model of how the D-repeat polymorphism of ASPN influences susceptibility to OA: (1) asporin inhibits TGF-b signalling and, therefore, indirectly regulates the synthesis of critical components of articular cartilage ECM; (2) this inhibition is strongest for the D14 allele, leading to insufficient quantities of these proteins and to a cartilage that is structurally compromised; (3) D13-encoded asporin has the weakest TGF-b inhibitory effect, resulting in a structurally more resilient cartilage.
Subsequent genotyping of the ASPN D-repeat in European cohorts with OA has, however, failed to generate association signals to the levels seen in the Asian report, suggesting ethnic differences in the ability of the D-repeat to influence susceptibility to OA.20 This ethnic restriction has now been observed for a number of other statistically compelling loci that have emerged from Asian candidate-gene and GWAS studies, including for the genes CALM1, DVWA and EDG2, and for the HLA locus.21-27 These probably highlight differences between the ascertainment of OA cases in different geographical regions, which itself is a reflection of genuine differences in the presentation and occurrence of the disease. For example, in the context of the CALM1 association it has been suggested that OA of the hip in the Japanese may be principally explained by individuals who had congenital disease of that joint.28 Under such a scenario, replicating hits will prove hard across ethnicities and, where it does occur, for example with GDF5, the effect may be weaker in one group versus another: the phenotype is all important.
SMAD family member 3 (SMAD3) is an intracellular molecule that links the extracellular TGF-b signal with changes in gene transcription. There have been extensive studies investigating the role of SMAD3 in OA and it appears that reduction in activity of this protein leads to an OA phenotype in model systems.29 Based on these insights, a UK group genotyped ten SMAD3 SNPs in Europeans who had OA of the hip or knee, and in ethnically matched controls.30 They reported association of a SNP, rs12901499, from within the first intron of SMAD3, with a meta-analysis p-value of 4.5x10-6. Based on the functional data from the previously published models it is possible that the genetic data are highlighting a loss of function of SMAD3, possibly via quantitative cis-acting effects on the expression of its gene. Although the UK study did offer up replication of its original genetic find, independent replication is desirable. Furthermore, it will be intriguing to see if this association is observed in Asian cohorts or whether the ethnic restriction noted previously works both ways, at least for genetic signals within the TGF-b pathway.
It is perhaps fair to ask why the OA genetics community does not just remain with the candidate-gene approach, as this seems to have offered up so many rewards. It is true that the TGF-b pathway looks very promising; this pathway has received considerable focussed attention, so it is perhaps not unsurprising that some positives have emerged. What is disconcerting is that, on the whole, most OA association signals have low odds ratios (a measure of the effect size of the polymorphism on disease), typically <1.2, and many are relevant to only certain ethnic groups. This may be a reflection of OA as a particularly polygenic disease with little or no loci of moderate to high impact, or it may highlight heterogeneity, with genuine signals hidden by the background noise.
The focus on candidates has demonstrated how it is possible to be seen as either smart or lucky when investigating genetic susceptibility. However, the future is the genome-wide scan. If these are adequately powered (for OA this probably means tens of thousands of cases) then they may provide novel insights. More sophisticated functional tools are needed, however, if research is to move relatively easily from broad genome signals to the susceptibility alleles.
Supported by Arthritis Research UK and by the UK NIHR Biomedical Research Centre for Ageing and Age-related disease award to the Newcastle upon Tyne Hospitals NHS Foundation Trust.
1. Miyamoto Y, Shi D, Nakajima M, et al. Common variants in DVWA on chromosome 3p24.3 are associated with susceptibility to knee osteoarthritis. Nat Genet 2008;40:994-8.
2. Spector TD, Reneland RH, Mah S, et al. Association between a variation in LRCH1 and knee osteoarthritis: a genome-wide single-nucleotide polymorphism association study using DNA pooling. Arthritis Rheum 2006;54:524-32.
3. Valdes AM, Loughlin J, Timms KM, et al. Genome-wide association scan identifies a prostaglandin-endoperoxide synthase 2 variant involved in risk of knee osteoarthritis. Am J Hum Genet 2008;82:1231-40.
4. Kerkhof HJM, Lories RJ, Meulenbelt I, et al. A genome-wide association study identifies an osteoarthritis susceptibility locus on chromosome 7q22. Arthritis Rheum 2010;62:499-510.
5. Francis-West PH, Abdelfattah A, Chen P, et al. Mechanisms of GDF-5 action during skeletal development. Development 1999;126:1305-15.
6. Miyamoto Y, Mabuchi A, Shi D, et al. A functional polymorphism in the 5' UTR of GDF5 is associated with susceptibility to osteoarthritis. Nat Genet 2007;39:529-33.
7. Chapman K, Takahashi A, Meulenbelt I, et al. A meta-analysis of European and Asian cohorts reveals a global role of a functional SNP in the 5' UTR of GDF5 with osteoarthritis susceptibility. Hum Mol Genet 2008;17:1497-504.
8. Southam L, Rodriguez-Lopez J, Wilkins JM, et al. An SNP in the 5'-UTR of GDF5 is associated with osteoarthritis susceptibility in Europeans and with in vivo differences in allelic expression in articular cartilage. Hum Mol Genet 2007;16:2226-32.
9. Egli R, Southam L, Wilkins JM, et al. Functional analysis of the osteoarthritis susceptibility-associated GDF5 regulatory polymorphism. Arthritis Rheum 2009;60:2055-64.
10. Sanna S, Jackson AU, Nagaraja R, et al. Common variants in the GDF5-UGCC region are associated with variation in human height. Nat Genet 2008;40:198-203.
11. Posthumus M, Collins M, Cook J, et al. Components of the transforming growth factor-b family and the pathogenesis of human Achilles tendon pathology - a genetic association study. Rheumatology (Oxford) 2010 (in press).
12. Vaes RB, Rivadeneira F, Kerkhof JM, et al. Genetic variation in the GDF5 region is associated with osteoarthritis, height, hip axis length and fracture risk: the Rotterdam study. Ann Rheum Dis 2009;68:1754-60.
13. Dai J, Shi D, Zhu P, et al. Association of a single nucleotide polymorphism in growth differentiate factor 5 with congenital dysplasia of the hip: a case-control study. Arthritis Res Ther 2008;10:R126.
14. Rouault K, Scotet V, Autret S, et al. Evidence of association between GDF5 polymorphisms and congenital dislocation of the hip in a Caucasian population. Osteoarthritis Cartilage 2010 (in press).
15. Lorenzo P, Aspberg A, Onnerfjord P, et al. Identification and characterization of asporin. J Biol Chem 2001;276:12201-11.
16. Kizawa H, Kou I, Iida A, et al. An aspartic acid repeat polymorphism in asporin inhibits chondrogenesis and increases susceptibility to osteoarthritis. Nat Genet 2005;37:138-44.
17. Song YQ, Cheung KM, Ho DW, et al. Association of the asporin D14 allele with lumbar-disc degeneration in Asians. Am J Hum Genet 2008;82:744-7.
18. Nakajima M, Kizawa H, Saitoh M, et al. Mechanisms for asporin function and regulation in articular cartilage. J Biol Chem 2007;282:32185-92.
19. Kou I, Nakajima M, Ikegawa S. Binding characteristics of the osteoarthritis-associated protein asporin. J Bone Miner Metab 2010;28:395-402.
20. Nakamura T, Shi D, Tzetis M, et al. Meta-analysis of association between the ASPN D-repeat and osteoarthritis. Hum Mol Genet 2007;16:1676-81.
21. Mototani H, Mabuchi A, Saito S, et al. A functional single nucleotide polymorphism in the core promoter region of CALM1 is associated with hip osteoarthritis in Japanese. Hum Mole Genetic 2005;14:1009-17.
22. Loughlin J, Sinsheimer JS, Carr A, et al. The CALM1 core promoter polymorphism is not associated with hip osteoarthritis in a United Kingdom Caucasian population. Osteoarthritis Cartilage 2006;14:295-8.
23. Miyamoto Y, Shi D, Nakajima M, et al. Common variants in DVWA on chromosome 3p24.3 are associated with susceptibility to knee osteoarthritis. Nat Genet 2008;40:994-8.
24. Meulenbelt I, Chapman K, Dieguez-Gonzalez R, et al. Large replication study and meta-analyses of DVWA as an osteoarthritis susceptibility locus in European and Asian populations. Hum Mol Genet 2009;18:1518-23.
25. Mototani H, Iida A, Nakajima M, et al. A functional SNP in EDG2 increases susceptibility to knee osteoarthritis in Japanese. Hum Mol Genet 2008;17:1790-7.
26. Dieguez-Gonzalez R, Calaza M, Shi D, et al. Testing the druggable endothelial differentiation gene 2 knee osteoarthritis genetic factor for replication in a wide range of sample collections. Ann Rheum Dis 2009;68:1017-21.
27. Nakajima M, Takahashi A, Kou I, et al. New sequence variants in HLA class II/III region associated with susceptibility to knee osteoarthritis identified by genome-wide association study. PLoS One 2010;5:e9723.
28. Hoaglund FT. CALM1 promoter polymorphism gene and Japanese congenital hip disease. Osteoarthritis Cartilage 2007;15:593.
29. Yang X, Chen L, Xu X, et al. TGF-b/Smad3 signals repress chondrocyte hypertrophic differentiation and are required for maintaining articular cartilage. J Cell Biol 2001;153:35-46.
30. Valdes AM, Spector TD, Tamm A, et al. Genetic variation in the SMAD3 gene is associated with hip and knee osteoarthritis. Arthritis Rheum 2010;62:2347-52.