Article Text
PDF
Abstract
This review focuses on the contribution of genetic markers to the severity of radiological damage in rheumatoid arthritis (RA). Currently available biomarkers of more severe disease include elevated erythrocyte sedimentation rates or C-reactive protein levels and rheumatoid factor (RF) or anticyclic citrullinated protein antibodies positivity; however, these biomarkers explain a relatively modest proportion of the variance in radiological damage. An important role of genetic factors on RA severity has recently emerged but studies to date have generally been of low statistical power and many have not been replicated. Genetic markers have a number of advantages over conventional biomarkers; genotypes are stable, measurable at disease onset, remain unchanged by treatment and are amenable to high-throughput assays. The recent advances in genome-wide genetic analysis should lead to a more comprehensive understanding of RA severity genes. This knowledge could be used, along with existing biomarkers, to therapeutically target subjects at risk of poor radiological outcome.
Statistics from Altmetric.com
Introduction
Rheumatoid arthritis (RA) is characterised by bone and cartilage loss in affected joints. Rheumatoid synovial cells produce a wide range of proinflamma-tory cytokines such as tumour necrosis factor (TNF), interleukin 1 (IL1) and IL6 that stimulate osteoclast-like cells to produce proteolytic enzymes such as matrix metalloproteinases (MMPs) and prostaglan-dins. IL1 stimulates synoviocytes and chondrocytes to release MMPs such as MMP-13 (collagenase) and MMP-3 (stromelysin) that degrade collagen, resulting in extracellular matrix degradation ultimately leading to bone and cartilage loss.1 2
The clinical course of RA shows greatly variable disease severity and there is much evidence that early, aggressive treatment of RA results in a more effective clinical, radiological and functional outcome. The use of anti-TNF and methotrexate3 4 or combinations of conventional disease-modifying agents5 has been shown to be more effective at slowing radiological progression than methotrexate alone. Furthermore, there is evidence of a prolonged benefit that is independent of subsequent antirheumatic treatment.5 These so-called ‘step-down’ approaches of intensive earlier treatment are consistent with the idea of the existence of a therapeutic ‘window of opportunity’ in early RA in which treatments administered early have greater efficacy than when used later6 and that this period may be as short as a few months.7
The usefulness of biomarker profiles at presentation in predicting radiological damage have been examined in several studies (reviewed by Harrison and Symmons8). One study reported on the value of measuring a range of laboratory markers (rheumatoid factor (RF) isotypes, anti-cyclic citullinated peptides (anti-CCP), DRB1 alleles, C-reactive protein (CRP), erythrocyte sedimentation rate and cartilage oligomeric matrix protein) at disease presentation. Elevated inflammatory indices at presentation are predictive of more severe disease, and total inflammatory load, as measured by the time-integrated CRP level, has been correlated with radiological damage9; however, several studies have shown this correlation to be modest.10 11 Levels of cartilage or bone turnover such as cartilage oligometric protein, collagen II epitopes and osteocalcin have been correlated with radiological damage in early RA.12 13 Levels of MMP-3 at presentation are correlated with x-ray progression14; however, this is not an independent marker but is strongly correlated with levels of inflammation as assessed by erythrocyte sedimentation rate and CRP.15 The variance of radiological hand and foot damage that could be explained by these measurements was 44% at 5 years and only 32% at 10 years.16 In established disease the presence of RFs and anti-citrullinated protein antigens (ACPA) are both independently associated with increased radiological damage,17 but the variance in radiological damage related to these markers is modest at 3% and 2%, respectively.17 These results highlight the need to identify additional biomarkers predictive of severe RA that could be used to target treatments more effectively.
Relatively few studies have investigated the role of genetic factors in RA severity and many have been underpowered and have not been validated. There is evidence of a significant impact on radiological damage that is independent of covariates, including disease duration and RF status.18 A recent study compared the variance in radiological hand damage in monozygotic and dizygotic twins and pairs of unrelated patients with RA. When a linear relationship between radiological progression and disease duration was assumed, the variation in joint destruction was highest in unrelated pairs, followed by dizygotic twins and was smallest between monzygotic twins, supporting a genetic input.19 The study was insufficiently powered to allow a calculation of heritability for this trait; however, the differences were independent of RF production. We performed a PubMed search to identify original articles, published before November 2008 that reported genetic studies of radiological severity of RA. The search terms were [genetics] and [rheumatoid arthritis] and [radiographic damage] or [radiological damage] or [severity]. The search was limited to human studies and original articles published in English.
The major histocompatibility complex (MHC) and RA severity
Carriage of the HLA-DRB1 shared epitope alleles is associated with higher levels of radiological damage,20 21 although this may be secondary to their association with ACPA.22 A prospective study conducted by Wagner et al reported that the HLA-DRB1 shared epitope *01 or *04 subtypes were significantly associated with higher x-ray damage at all points analysed (presentation, 2 and 4 years).20
The association of HLA-DRB1 with radiographic damage has also been examined using a recently proposed classification system.23 In this novel system the presence or absence of the RAA motif at positions 72-74 distinguishes susceptibility (S) from non-susceptibility alleles (X). The S alleles are then subdivided into four groups according to the amino acid at position 71: an alanine (A), a glutamic acid (E), a lysine (K) or an arginine (R). Based on this classification four different genotype groups are defined: S1 for ARAA or ERAA, S2 for KRAA, S3 for RRAA and X for those lacking the RAA-susceptibility motif. An aspartic acid (D) at position 70 has been proposed to confer protection against the development of RA, and therefore two additional allele groups have been defined based on the amino acids at position 70. S3D for those having a D at position 70 (DRRAA) and S3P for those having either a glutamine (Q) or an arginine (QRRAA or RRRAA). Both a cross-sectional and a prospective study in European populations have showed that the S2 alleles coding for HLA-DRB1 *0401 are associated with more severe structural damage in patients with RA as well as production of autoantibodies.24 25 The S2 alleles have been shown to affect disease severity in a gene-dose effect and have also been shown to affect production of both RF and anti-CCP.25 The French study also reported a protective effect conferred from S3D alleles.24 Overall, DRB1 alleles seem to be primarily associated with anti-CCP status rather than primarily with radiological damage.26 We have recently reported that carriage of S2 alleles is associated with more severe disease with a gene-dose effect, and also associated with the presence of both ACPA and RF. Carriage of S1 alleles was associated with less severe disease, however there was no association between S1 and either ACPA or RF, suggesting that the basis for this possibly protective effect was not related to autoantibody-producing B cells.
The MHC is a particularly gene-dense region and encodes the TNF loci approximately 900 kb telomeric of DRB1. In view of the importance of TNF in the pathogenesis of RA studies it represents a very plausible RA severity-determining locus and several studies have reported association of the −308A with radiological damage.18 27 However, as with all such studies of the MHC these studies were modest in size and did not robustly correct for linkage disequilibrium with the DRB1 locus.
Cytokines
Interleukin 1
The IL1 cluster is located on 2q14-21 and contains nine genes, including IL1A, IL1B and IL1RN. IL1A and IL1B encode proin-flammatory mediators—namely, IL1α and IL1β, while IL1RN encodes their natural competitive inhibitor, IL1 receptor antagonist (IL1Ra), which binds IL1 cell-surface receptor type I without inducing signal transduction.28
IL1 is a mediator of cartilage damage and bone destruction. In addition, it activates leucocytes, chondrocytes and synovio-cytes.29 Release of IL1 in synovium stimulates local proliferation of fibroblasts, contributing to the formation of pannus.30 It also stimulates chondrocytes to release collagenases and other degradative enzymes, leading to the loss of the extracellular matrix.30 It is responsible for the stimulation of osteo-clasts and suppression of osteoblasts through the activation of receptor activators for nuclear factor-κB (RANK) and osteopro-tegerin, resulting in the balance of bone formation shifting to bone loss.31
A number of studies have reported the association of IL1 polymorphisms with severity of radiographic damage in RA. In a study of 108 patients with early RA followed up for 2 years, Cantagrel et al demonstrated that the IL1B +3954 T (rs1 143634) allele increases the risk of developing erosions.32 Similar associations have also been reported in two other studies with the +3954 T allele being associated with more severe structural damage.33 34 The IL1B promoter polymorphism −511 (rs16944) has also been associated with radiographic progression.35 Although these results are encouraging, the different outcome variables across the studies make it difficult to quantify the value of IL1B genotypes in predicting radiological damage.
Interleukin 6
In contrast to IL1, the association of IL6 with x-ray damage has not been extensively studied. The −174G (rs1800795) allele has been associated with increased transcriptional activity, potentially leading to higher production of this proinflammatory cytokine.36 We recently reported an allele dose association with radiological damage in autoantibody-positive RA and showed that this genotype correlated with 1.2% of the variance of this trait as measured by the correlation coefficient.37 Haplotypes containing the protective −572G and −174C alleles have been associated with lower levels of markers of bone resorption such as C-terminal cross-linking of type I collagen.38 These data make the IL6 −572 marker an interesting variant that should be studied in combination with IL6 −174.
IL4/IL13
A central feature of RA is a relative imbalance between pro- and anti-inflammatory cytokines with higher levels of TNF, IL1 and IL6 but much lower levels of IL4 and IL13.39 The genes for IL4 and IL13 are located in a gene cluster at 5q31–33. Their effects are mediated by a heterodimeric receptor composed of the IL4Rα chain (16p12.1) and either the common γ chain or the IL13Rα subunit.40 A Swiss study reported association of an IL4 variable number of tandem repeat with lower radiographic damage, while the IL4 receptor (IL4R) +1902 (rs1 801275) variant was not correlated with radiological damage.35 Polish and German studies have reported associations of the IL4 −590 (rs2243250) polymorphism and of the IL4R I50V (rs1805010) variant with more severe joint damage41 42; however, these associations have not been replicated in independent cohorts.32
We recently performed a haplotype-tagging approach to investigate the role of the IL4/IL13 pathway as well as the role of IL4R in radiological damage of RA. This approach allowed us to examine the contribution of these genes to RA severity in a more robust manner than previously reported.43 In our cohort single nucleotide polymorphism (SNP) analysis as well as haplo-type analysis provided evidence of no association between RA severity and variation in the IL4/IL13 pathway; the IL4R I50V variant, which has been previously associated with increased radiological damage during the first 2 years of disease,42 was not associated with joint damage in our cohort.43 Such conflicting results might be due to differences in study design; our study was composed of a cross-sectional cohort with minimum disease duration of 3 years while Prots et al performed a prospective study with a follow-up of 2 years. It is therefore possible we have not detected a genetic effect that is important in the first 3 years of disease. Furthermore, the methods of assessing severity were different; we used the modified Larsen scores rather than comparing genotypes between patients with erosive and non-erosive RA.
Interleukin 10
IL10 is produced predominantly by monocytes and lymphocytes and has a range of anti-inflammatory and immunoregulatory properties, including inhibition of synthesis of proinflammatory molecules, including TNF, IL1 and IL6.44 Production of IL10 by rheumatoid synovial macrophages and T cells inhibits production of IL1 and TNF by synovial cells,45 and increased relative expression of IL10 has been reported in joints without erosion compared with those with erosions, suggesting a protective role in RA.46 Studies in animal models of RA have also demonstrated the anti-inflammatory role of IL10, inhibiting both the incidence and severity of the disease.47 48
Promoter variants and haplotypes are associated with different production levels of IL10.49 50 We have recently reported association of homozygosity for IL10 −592C (rs18000872) with higher Larsen scores in ACPA-negative and RF-negative disease and shown that this genotype correlated with 0.6% of the variance of this trait as measure by the correlation coefficient.37 Our data are consistent with those from a Dutch prospective study of 91 female patients with RA followed up for 12 years, in whom radiographic damage to the hands and feet progressed more rapidly in subjects homozygous for −1082G (rs1800896) than in those homozygous for −1082A.46 The same study also reported higher levels of IL10 mRNA in patients with non-destructive disease than in patients with destructive disease and also demonstrated lower production of IL10 in whole-blood cultures of subjects with the GG genotype compared with the AA genotype, suggesting that the association of IL10 −1082G with radiographic damage is related to lower IL10 production.46
Proteases
Genes encoding enzymes regulating degradation of cartilage and bone turnover are candidates for involvement in radiological severity. Collagenase (MMP-1) and stromelysin-1 (MMP-3) degrade key components of cartilage and bone matrix and are highly expressed in the serum and synovial fluid of patients with RA.51 52 Functional polymorphisms in the promoter region of these genes have been associated with more severe radiographic damage.53 54 Constantin et al and Matteyetalhave shown that the functional MMP-3 polymorphism at position −1171 (rs3025082) is associated with more severe radiographic damage.53 54 In both studies the 6A allele, which has six adenosines (A) compared with the 5A allele that has five, was associated with increased radiographic damage. An MMP-1 guanine insertion/deletion (1G/2G) functional polymorphism has been detected at position −1607 (rs17886084) with the G insertion creating an activation protein 1 transcription factor binding site.51 55 56 This study demonstrated that a novel haplotype containing the MMP-1 1G and MMP-3 5A alleles had a protective effect over a period of 15 years.51 This is of particular interest as both alleles are associated with a lower transcriptional activity.55 57
Protein tyrosine phosphatase, non-receptor type 22
The gene for protein tyrosine phosphatase, non-receptor type 22 (PTPN22) is located at 1p13 and encodes the intracellular protein LYP,58 which has been shown to be an important inhibitor of T-cell activation.59 Protein tyrosine phosphatases have an important role in immune responses as they control kinases involved in early stages of the T-cell receptor signalling. Mice deficient in PEP, the mouse orthologue of PTPN22, develop autoimmune abnormalities, including expansion of memory T cells and increased production of autoantibodies.60
The PTPN22 non-synonymous SNP (R620W) (rs2476601) lies in one of the four proline-rich Src homology 3 (SH3) binding sites of the protein tyrosine kinase Csk.58 61 The amino acid substitution of arginine (R) for tryptophan (W) has been shown to disrupt binding of LYP to Csk. As interaction between LYP and Csk inhibits T-cell activation; disruption of their interaction could result in hyper-reactive T-cell responses.62 Although the PTPN22 R620W variant has been associated with susceptibility to many autoimmune diseases, its effect in disease phenotype and the clinical outcome is less clear.
A non-significant trend towards a higher Larsen score in PTPN22 +1858T-positive patients over time has been reported63 and in those with erosive disease.64 Lie et al have recently demonstrated a marginal association between annual progression rate of Sharp score and carriage of the PTPN22 +1858T allele but this finding needs to be replicated in larger cohorts.65 We reported that patients with RA (n=964) carrying the PTPN22 risk allele tend to have higher x-ray damage than those lacking the +1858T allele.37 However, other studies did not report association of this variant with radiological damage,66 67 and further large studies will be required to clarify the role of this variant and radiological severity.
Other immune genes
For other immunorelevant genes, such as the natural resistant-associated macrophage protein-1, α2 macroglobulin (α2m) and cyclo-oxygenase-2 (COX-2), as shown in table 1, only a single study has been reported positive associations in small numbers of patients.68,–,70 These studies require replication in larger independent cohorts in order to confirm the validity of these results.
Results of studies of non-MHC SNPs with RA severity
The rs6920220 at 6q23 is an RA susceptibility gene.71 72 Scherer et al have recently reported the association of two polymorphisms in this block (rs675520 and rs9376293) with severity of joint damage in patients positive for ACPA.73 The linkage disequilibrium block encoding this SNP only contains part of the PTPN11 pseudogene but lies close to the TNFAIP3 gene that is a negative regulator of inflammatory cytokines. Sequencing and fine mapping studies of this region will be required to identify the most likely causal variant in this locus.
Discussion
Biological agents have proved very effective in the treatment of severe RA but their more widespread use is limited in many countries by cost. The identification of biomarkers of severity should facilitate therapeutic targeting of these agents with potential clinical and economic benefits.
Recent advances in high-throughput genetic screening and computational analysis have resulted in the identification of a large number of susceptibility genes for common diseases such as RA. Similar studies mapping genetics of disease severity traits such as radiological damage are feasible but require large well-characterised cohorts. Such a study will require careful selection of patients with RA in order to avoid disease heterogeneity. Disease heterogeneity can complicate these studies—for example, it is generally accepted that ACPA-positive and -negative RA are genetically distinct diseases. International collaborations and meta-analysis of published data will be key in the identification of genetic severity markers associated with disease phenotype. Such an approach would also provide increased power for the detection of susceptibility genes, with modest effects on the disease pathogenesis, and also allow the replication of positive findings across different populations.
Copy number variants are potentially additional biomarkers that can be interrogated using SNP whole-genome chips. Recent studies have shown that large segments of DNA, ranging from thousands to millions of DNA bases, may vary in copy number, leading to dosage imbalances.74 Low FCGRIIB copy number has been associated with glomerulonephritis in systemic lupus erythematosus (SLE),75 76 and low copy number of C4 protects against the development of SLE.77
In summary, although there is substantial evidence of a genetic contribution to radiological severity in RA, most of the studies to date have been of relatively low statistical power and many associations have not been replicated. The recent advances in genotyping technology and statistical analysis have resulted in the identification of a large number of RA susceptibility genes. Similar studies are required in large RA populations to map RA severity genes and determine both gene-gene and gene-environment interactions for this trait. The identification of genetic markers of poorer outcome should, in combination with established biomarkers, facilitate therapeutic targeting, leading to the development of prognostic algorithms and resultant earlier disease control and reduced joint damage, it may also provide insights into mediators of tissue damage that could potentially represent novel targets in RA.
References
Footnotes
-
Competing interests None.
-
Provenance and peer review Not commissioned; externally peer reviewed.