Spondyloarthritis refers to a group of inflammatory rheumatic diseases that share common clinical and genetic characteristics. Due to the rapid advances in technology and computational genetics, there is now an increasing list of well-validated genes in spondyloarthritis. The newest genetic associations are of modest magnitude and have been identified as a result of analysing thousands of samples, using genome-wide association scans or targeted candidate-gene association studies. In this article, we will highlight the genes associated with spondyloarthritis, with an emphasis on the recent candidate genes that have been identified in ankylosing spondylitis and psoriatic arthritis. If applicable, we will also discuss their potential relevance to the clinical rheumatologist.
Spondyloarthritis
Spondyloarthritis (SpA) is an inflammatory disease that is best characterised by inflammation of sacroiliac joints and the apophyseal joints of the spine. The prevalence of SpA approaches that of rheumatoid arthritis (1%) and the disability and economic impact of these two chronic diseases is comparable. The four major diseases that comprise SpA are ankylosing spondylitis (AS), psoriatic arthritis (PsA), reactive arthritis (ReA) and enteropathic arthritis (EA). The aetiology of SpA is still largely unknown. The pathogenesis is likely to be a result of genetic, immunological and environmental factors. There is convincing evidence for the genetic basis of SpA from genetic epidemiological studies that document high familial aggregation and some robust candidate genes that have reached genome-wide significance.
Genetics of AS
Genetic epidemiology
For over four decades, AS has been known to run in families. The recurrence risk ratio for AS in siblings of probands with AS ranges from 50 to 80 and the heritability from twin studies is estimated to be over 90% (as reviewed in ref. 1). This ranks among the highest heritability of any complex rheumatic disease. AS does not demonstrate a clear Mendelian pattern of inheritance (dominant or recessive), but rather an oligogenic model, where genes interact in a multiplicative fashion . The recurrence risk in families has been shown to differ between genders of the probands with increased risk of AS in relatives of young female probands . Approximately, 20% of HLA-B27 positive relatives of AS patients develop SpA . Genetic factors also appear to play a role in disease expression in AS, as cross-sectional studies have demonstrated that disease severity and radiological damage appear to be heritable . Due to the strong heritability of AS, the treating physician must inquire about a family history of SpA when assessing AS patients.
Major histocompatibility complex (MHC) region and AS
The major histocompatibility complex (MHC) region on chromosome 6 is about four megabases and is associated with disease susceptibility in most complex auto-immune diseases. The MHC is a gene-dense region that comprises of three clusters: class I, II and III. The MHC region is characterised by extreme linkage disequilibrium (LD), which has resulted in a major impediment for the identification of the actual or causative allele in this region.
The association of HLA-B27 and AS was noted in the early 1970s and continues to be one of the strongest disease associations for an auto-immune disease . HLA-B27 is present in over 90% of AS patients and in up to 50% in patients with other types of spondyloarthritis (PsA, EA and ReA). However, less than 5% of HLA-B27 -positive individuals in the general population develop AS. The overall contribution of HLA-B27 to AS is considerable and is estimated to be between 16% and 30%, based on linkage and association studies within the MHC region . There are now at least 57 coding variants reported of HLA-B27 (as reviewed in ref. 6). These subtypes differ from the parent and predominant subtype (B*2705) by 1–12 residues . Other common subtypes such as B*2702, B*2704 and B*2707 are also implicated in disease susceptibility. Exceptions to this rule include the B*2706 subtype (found in native Indonesians) and B*2709 (found in Sardinia and parts of mainland Italy) .
There is also substantial evidence that HLA-B27 is not the sole susceptibility factor within the MHC region in AS. The entire genetic contribution of the MHC region in AS is estimated to be 33%, of which the HLA-B27 accounts for about 16% . This has led to a plethora of association studies, searching for candidate genes within the MHC region. However, only a few have been consistently replicated, the most notable of which is HLA-B60 . Associations with MHC Class II alleles and AS have been reported for susceptibility and disease expression. However, these associations have not been widely replicated .
Multiple non-HLA-B27 genes within the MHC region have also been proposed to be associated with AS. These include TAP alleles, low molecular weight proteosome (LMP), tumour necrosis factor (TNF)-α and Hsp-70 (as summarised in 2). LMP-2 has been implicated in juvenile AS and in AS patients with uveitis . Whether these genes represent additional genetic effects above and beyond HLA-B27 is not clear due to the extensive LD in this region.
From a clinical perspective, HLA-B27 has recently become an important factor in the diagnosis of axial SpA . Presently, axial SpA is classified on the basis of sacroiliitis (via radiography or magnetic resonance imaging (MRI)) along with one clinical SpA feature or presence of HLA-B27 and two clinical SpA features. HLA-B27 also appears to be important in prognosticating radiographic damage in AS, as a cross-sectional study has shown that HLA-B27 patients appear to have a higher spinal damage score .
Linkage studies and AS
Three genome-wide linkage studies (from the United Kingdom, North America and France) have been completed in AS . Laval et al., assessed 255 affected AS sibling pairs form 185 British families. The strongest logarithm of odds (LOD) score was in the MHC region (15.6). The most significant linkage outside the MHC region was noted in chromosome 16q (extending from 99 to 106 cM from the p-telomere). Suggestive LOD scores were noted in chromosomes 2q, 9q, 10q and 19q. The North American Spondylitis Consortium (NASC) cohort assessed 244 sibling pairs from 180 families . Linkage was noted at the MHC region and suggestive linkage at 6q and chromosome 11. The French Task Force for genetic research on spondyloarthropathies performed a genome-wide linkage on 151 affected sibling pairs from 65 multiplex families . Significant linkage was found in MHC region and suggestive linkage in chromosomes 5q, 9q, 13q and 17q. The patient population from the French linkage study was different from the other two genome-wide linkage scans. The French Task Force included patients with SpA as defined by the Amor and/or the European Spondyloarthropathy Study Group (ESSG) criteria, while the British and North American studies assessed AS patients that satisfied the modified New York criteria.
In general, replication of linkage sites for AS has been difficult due to small effect sizes, inadequate sample sizes, poor marker density, computational challenges and differences in study design. The International Genetics of Ankylosing Spondylitis Consortium (IGAS) combined data from the three whole genome linkage scans for AS . Other than the convincing linkage at the MHC region, suggestive linkage was noted on chromosomes 10q and 16q and nominal linkage was noted on chromosomes 1q, 3q, 5q, 6q, 9q, 17q and 19q.
A smaller linkage study has been completed in 76 affected sibling pairs with acute anterior uveitis. Strong linkage was seen at a region on chromosome 9p21–9p24, with an LOD score of 3.7. When the authors compared the acute anterior uveitis sibling pairs with a companion cohort of AS families, the linkage at this region was found in association with acute anterior uveitis, but not with AS.
Association-based studies and AS involving non-MHC genes
Technological and computational advances have led to the successful execution of Genome-wide Association Scans (GWAS). A genome-wide association study is defined as any study of genetic variation across the entire human genome that is designed to identify genetic associations with observable traits or the presence or absence of a disease or condition. These studies have led to identification of four non-MHC related loci that have reached genome-wide significance ( IL-23R, endoplasmic reticulum aminopeptidase ( ERAP1) and two gene deserts 2p15 and 21q22). These loci will be discussed along other genes that have strong evidence for association with AS.
Genome-wide significant association
Endoplasmic reticulum aminopeptidase (ERAP1)
A genome-wide association study of 14 500 non-synonymous single-nucleotide polymorphism (nsSNP) in 1500 Caucasian controls and 1000 Caucasian AS patients identified ERAP1 as a candidate gene in AS. ERAP1 , also known as aminopeptidase regulating tumour necrosis factor receptor (TNFR) shedding 1 (ARTS-1) , is the strongest non-MHC gene associated with AS, with an estimated population attributable risk of 26%. The association with AS has been found in other Caucasian populations as well as in Koreans . In humans, there are two known functions of ERAP1. In the endoplasmic reticulum, ERAP1 has the potential to trim peptide antigens to optimal length for binding to MHC class I molecules . ERAP1 also modulates the shedding of interleukin (IL)-1RII, IL-6Rα, and tumour necrosis factor receptor (TNFR1), thus down-regulating pro-inflammatory cytokines signalling . However, the mechanism of ERAP1 association with AS is yet to be fully elucidated.
Interleukin 23 receptor (IL23R)
The association of IL-23R variants was first noted in the genome-wide pooling study for psoriasis. In that study, the common IL-23R haplotype (involving SNPs rs753051 and rs11209026) was associated with psoriasis in two replication sample sets and in the combined analysis across all three sample sets . At about the same time, IL-23R variants were also noted to be associated with inflammatory bowel disease, as a result of a GWAS .
IL-23R was first reported in AS as a result of a combined analysis of the IL-23R gene involving eight SNPs from AS patients in the Wellcome Trust Case Control Consortium (WTCCC) and NASC cohort . A peak association in the WTCCC cohort was noted for rs11209032 ( p = 6 × 10 −6 ) and in the North American cohort the peak was with rs1343151 ( p = 2 × 10 −5 ) . The peak association in the combined data set was with SNP rs11209032 (odds ratio (OR) 1.3; 95% confidence interval (CI) 1.2–1.4; p = 3 × 10 −8 ). This IL-2 3R association has been confirmed in case-control studies from Canada and Spain . Interestingly, there was no association of IL-23R in Korean AS patients when they examined similar SNPs to the Canadian study within the IL-23R gene .
Gene deserts at chromosome 2p15 and 21q22
The recently published TASC GWAS studied over 2000 AS patients of European descent and 5000 ethnically matched controls . The samples were genotyped with approximately 288 000 SNPs from the Illumina HumHap370 mapping set . A strong association was noted with the MHC and two well-established AS genes ERAP1 and IL-23R . Genome-wide significance was also noted in two gene deserts 2p15 and 21q22. Gene deserts are large tracts of the human genome that are devoid of protein-coding genes. Regarding 2p15, the closest gene is 100 kb away (B3GNT2) and there is no known disease association with this gene. Regarding 21q22, the nearest gene to the most strongly associated SNP is 82 kb away ( PSMG1 ). The authors report that PSMG1 was not differentially expressed in peripheral blood mononuclear cells (PBMCs) from cases with active AS compared to healthy controls .
The significance of the gene deserts is not clear. It has been estimated that approximately a quarter of the human genome consists of gene deserts. Some of the gene deserts appear to be non-essential to genome function as they can be deleted without significant phenotypic effect. However, other gene deserts have been shown to contain regulatory sequences that can act at large distances to control expression of neighbouring genes .
Candidate genes with modest evidence for association with AS
IL- 1 gene cluster
British, North American and Taiwanese cohorts have noted significant associations with the IL-1 gene cluster . The IL-1 family gene cluster spans, approximately, 360 kb and contains the following genes: IL-1A, IL-1B, IL-1F7, IL-1F9, IL-1F6, IL-1F8, IL-1F5, IL-1F10 and IL-1RN. A strong association and transmission of SNPs and haplotypes broadly within the IL-1 family gene cluster was first noted in white British families . The most significant associations were noted for markers in the IL1B and IL1F10 genes . Further, fine mapping from the British group suggested that the primary association in the IL-1B gene may be better explained by its LD with IL-1A . These studies were followed by a meta-analysis from the IGAS consortium where nine variants within the IL-1 gene cluster in 2675 cases and 2592 controls were examined. IL1 A was the only consistent genetic association that was noted and it had a population attributable risk of 4–6% among Caucasians with AS .
The recent TASC GWAS has identified additional genes suggestive of association. This includes IL1R2 , ANTXR2 and TNFR1 . IL1R2 is a decoy receptor interfering impeding the binding of IL-1 to IL-1R1 . Regarding ANTXR2, this maps to chromosome 4q21 and mutations in this gene have been associated with juvenile hyaline fibromatosis and infantile systemic hyalinosis . The significance of these two genes with respect to AS is unclear at present. Meanwhile, TNFR1 encodes TNF 1 and it has been proposed that TNFR1–ERAP1 molecular complex represents a novel mechanism by which TNFR1 shedding is regulated .
Other candidate genes that have been previously reported but still require larger replication studies include ANKH and CYP2D6 . ANKH has been implicated in AS and familial autosomal dominant chondrocalcinosis. With respect to AS, two novel genetic variants, one in the 5′ non-coding region and the other in the promoter region of ANKH, were found to be associated with AS . However, Timms et al. were unable to replicate the association between ANKH and AS with respect to susceptibility or clinical manifestations . Two independent case-control studies from the UK have reported an association between homozygosity for CYP2D6*4 and AS . However, the biological and clinical significance of this association is not clear and this association was not noted in the French population .
Genetics of AS
Genetic epidemiology
For over four decades, AS has been known to run in families. The recurrence risk ratio for AS in siblings of probands with AS ranges from 50 to 80 and the heritability from twin studies is estimated to be over 90% (as reviewed in ref. 1). This ranks among the highest heritability of any complex rheumatic disease. AS does not demonstrate a clear Mendelian pattern of inheritance (dominant or recessive), but rather an oligogenic model, where genes interact in a multiplicative fashion . The recurrence risk in families has been shown to differ between genders of the probands with increased risk of AS in relatives of young female probands . Approximately, 20% of HLA-B27 positive relatives of AS patients develop SpA . Genetic factors also appear to play a role in disease expression in AS, as cross-sectional studies have demonstrated that disease severity and radiological damage appear to be heritable . Due to the strong heritability of AS, the treating physician must inquire about a family history of SpA when assessing AS patients.
Major histocompatibility complex (MHC) region and AS
The major histocompatibility complex (MHC) region on chromosome 6 is about four megabases and is associated with disease susceptibility in most complex auto-immune diseases. The MHC is a gene-dense region that comprises of three clusters: class I, II and III. The MHC region is characterised by extreme linkage disequilibrium (LD), which has resulted in a major impediment for the identification of the actual or causative allele in this region.
The association of HLA-B27 and AS was noted in the early 1970s and continues to be one of the strongest disease associations for an auto-immune disease . HLA-B27 is present in over 90% of AS patients and in up to 50% in patients with other types of spondyloarthritis (PsA, EA and ReA). However, less than 5% of HLA-B27 -positive individuals in the general population develop AS. The overall contribution of HLA-B27 to AS is considerable and is estimated to be between 16% and 30%, based on linkage and association studies within the MHC region . There are now at least 57 coding variants reported of HLA-B27 (as reviewed in ref. 6). These subtypes differ from the parent and predominant subtype (B*2705) by 1–12 residues . Other common subtypes such as B*2702, B*2704 and B*2707 are also implicated in disease susceptibility. Exceptions to this rule include the B*2706 subtype (found in native Indonesians) and B*2709 (found in Sardinia and parts of mainland Italy) .
There is also substantial evidence that HLA-B27 is not the sole susceptibility factor within the MHC region in AS. The entire genetic contribution of the MHC region in AS is estimated to be 33%, of which the HLA-B27 accounts for about 16% . This has led to a plethora of association studies, searching for candidate genes within the MHC region. However, only a few have been consistently replicated, the most notable of which is HLA-B60 . Associations with MHC Class II alleles and AS have been reported for susceptibility and disease expression. However, these associations have not been widely replicated .
Multiple non-HLA-B27 genes within the MHC region have also been proposed to be associated with AS. These include TAP alleles, low molecular weight proteosome (LMP), tumour necrosis factor (TNF)-α and Hsp-70 (as summarised in 2). LMP-2 has been implicated in juvenile AS and in AS patients with uveitis . Whether these genes represent additional genetic effects above and beyond HLA-B27 is not clear due to the extensive LD in this region.
From a clinical perspective, HLA-B27 has recently become an important factor in the diagnosis of axial SpA . Presently, axial SpA is classified on the basis of sacroiliitis (via radiography or magnetic resonance imaging (MRI)) along with one clinical SpA feature or presence of HLA-B27 and two clinical SpA features. HLA-B27 also appears to be important in prognosticating radiographic damage in AS, as a cross-sectional study has shown that HLA-B27 patients appear to have a higher spinal damage score .
Linkage studies and AS
Three genome-wide linkage studies (from the United Kingdom, North America and France) have been completed in AS . Laval et al., assessed 255 affected AS sibling pairs form 185 British families. The strongest logarithm of odds (LOD) score was in the MHC region (15.6). The most significant linkage outside the MHC region was noted in chromosome 16q (extending from 99 to 106 cM from the p-telomere). Suggestive LOD scores were noted in chromosomes 2q, 9q, 10q and 19q. The North American Spondylitis Consortium (NASC) cohort assessed 244 sibling pairs from 180 families . Linkage was noted at the MHC region and suggestive linkage at 6q and chromosome 11. The French Task Force for genetic research on spondyloarthropathies performed a genome-wide linkage on 151 affected sibling pairs from 65 multiplex families . Significant linkage was found in MHC region and suggestive linkage in chromosomes 5q, 9q, 13q and 17q. The patient population from the French linkage study was different from the other two genome-wide linkage scans. The French Task Force included patients with SpA as defined by the Amor and/or the European Spondyloarthropathy Study Group (ESSG) criteria, while the British and North American studies assessed AS patients that satisfied the modified New York criteria.
In general, replication of linkage sites for AS has been difficult due to small effect sizes, inadequate sample sizes, poor marker density, computational challenges and differences in study design. The International Genetics of Ankylosing Spondylitis Consortium (IGAS) combined data from the three whole genome linkage scans for AS . Other than the convincing linkage at the MHC region, suggestive linkage was noted on chromosomes 10q and 16q and nominal linkage was noted on chromosomes 1q, 3q, 5q, 6q, 9q, 17q and 19q.
A smaller linkage study has been completed in 76 affected sibling pairs with acute anterior uveitis. Strong linkage was seen at a region on chromosome 9p21–9p24, with an LOD score of 3.7. When the authors compared the acute anterior uveitis sibling pairs with a companion cohort of AS families, the linkage at this region was found in association with acute anterior uveitis, but not with AS.
Association-based studies and AS involving non-MHC genes
Technological and computational advances have led to the successful execution of Genome-wide Association Scans (GWAS). A genome-wide association study is defined as any study of genetic variation across the entire human genome that is designed to identify genetic associations with observable traits or the presence or absence of a disease or condition. These studies have led to identification of four non-MHC related loci that have reached genome-wide significance ( IL-23R, endoplasmic reticulum aminopeptidase ( ERAP1) and two gene deserts 2p15 and 21q22). These loci will be discussed along other genes that have strong evidence for association with AS.
Genome-wide significant association
Endoplasmic reticulum aminopeptidase (ERAP1)
A genome-wide association study of 14 500 non-synonymous single-nucleotide polymorphism (nsSNP) in 1500 Caucasian controls and 1000 Caucasian AS patients identified ERAP1 as a candidate gene in AS. ERAP1 , also known as aminopeptidase regulating tumour necrosis factor receptor (TNFR) shedding 1 (ARTS-1) , is the strongest non-MHC gene associated with AS, with an estimated population attributable risk of 26%. The association with AS has been found in other Caucasian populations as well as in Koreans . In humans, there are two known functions of ERAP1. In the endoplasmic reticulum, ERAP1 has the potential to trim peptide antigens to optimal length for binding to MHC class I molecules . ERAP1 also modulates the shedding of interleukin (IL)-1RII, IL-6Rα, and tumour necrosis factor receptor (TNFR1), thus down-regulating pro-inflammatory cytokines signalling . However, the mechanism of ERAP1 association with AS is yet to be fully elucidated.
Interleukin 23 receptor (IL23R)
The association of IL-23R variants was first noted in the genome-wide pooling study for psoriasis. In that study, the common IL-23R haplotype (involving SNPs rs753051 and rs11209026) was associated with psoriasis in two replication sample sets and in the combined analysis across all three sample sets . At about the same time, IL-23R variants were also noted to be associated with inflammatory bowel disease, as a result of a GWAS .
IL-23R was first reported in AS as a result of a combined analysis of the IL-23R gene involving eight SNPs from AS patients in the Wellcome Trust Case Control Consortium (WTCCC) and NASC cohort . A peak association in the WTCCC cohort was noted for rs11209032 ( p = 6 × 10 −6 ) and in the North American cohort the peak was with rs1343151 ( p = 2 × 10 −5 ) . The peak association in the combined data set was with SNP rs11209032 (odds ratio (OR) 1.3; 95% confidence interval (CI) 1.2–1.4; p = 3 × 10 −8 ). This IL-2 3R association has been confirmed in case-control studies from Canada and Spain . Interestingly, there was no association of IL-23R in Korean AS patients when they examined similar SNPs to the Canadian study within the IL-23R gene .
Gene deserts at chromosome 2p15 and 21q22
The recently published TASC GWAS studied over 2000 AS patients of European descent and 5000 ethnically matched controls . The samples were genotyped with approximately 288 000 SNPs from the Illumina HumHap370 mapping set . A strong association was noted with the MHC and two well-established AS genes ERAP1 and IL-23R . Genome-wide significance was also noted in two gene deserts 2p15 and 21q22. Gene deserts are large tracts of the human genome that are devoid of protein-coding genes. Regarding 2p15, the closest gene is 100 kb away (B3GNT2) and there is no known disease association with this gene. Regarding 21q22, the nearest gene to the most strongly associated SNP is 82 kb away ( PSMG1 ). The authors report that PSMG1 was not differentially expressed in peripheral blood mononuclear cells (PBMCs) from cases with active AS compared to healthy controls .
The significance of the gene deserts is not clear. It has been estimated that approximately a quarter of the human genome consists of gene deserts. Some of the gene deserts appear to be non-essential to genome function as they can be deleted without significant phenotypic effect. However, other gene deserts have been shown to contain regulatory sequences that can act at large distances to control expression of neighbouring genes .
Candidate genes with modest evidence for association with AS
IL- 1 gene cluster
British, North American and Taiwanese cohorts have noted significant associations with the IL-1 gene cluster . The IL-1 family gene cluster spans, approximately, 360 kb and contains the following genes: IL-1A, IL-1B, IL-1F7, IL-1F9, IL-1F6, IL-1F8, IL-1F5, IL-1F10 and IL-1RN. A strong association and transmission of SNPs and haplotypes broadly within the IL-1 family gene cluster was first noted in white British families . The most significant associations were noted for markers in the IL1B and IL1F10 genes . Further, fine mapping from the British group suggested that the primary association in the IL-1B gene may be better explained by its LD with IL-1A . These studies were followed by a meta-analysis from the IGAS consortium where nine variants within the IL-1 gene cluster in 2675 cases and 2592 controls were examined. IL1 A was the only consistent genetic association that was noted and it had a population attributable risk of 4–6% among Caucasians with AS .
The recent TASC GWAS has identified additional genes suggestive of association. This includes IL1R2 , ANTXR2 and TNFR1 . IL1R2 is a decoy receptor interfering impeding the binding of IL-1 to IL-1R1 . Regarding ANTXR2, this maps to chromosome 4q21 and mutations in this gene have been associated with juvenile hyaline fibromatosis and infantile systemic hyalinosis . The significance of these two genes with respect to AS is unclear at present. Meanwhile, TNFR1 encodes TNF 1 and it has been proposed that TNFR1–ERAP1 molecular complex represents a novel mechanism by which TNFR1 shedding is regulated .
Other candidate genes that have been previously reported but still require larger replication studies include ANKH and CYP2D6 . ANKH has been implicated in AS and familial autosomal dominant chondrocalcinosis. With respect to AS, two novel genetic variants, one in the 5′ non-coding region and the other in the promoter region of ANKH, were found to be associated with AS . However, Timms et al. were unable to replicate the association between ANKH and AS with respect to susceptibility or clinical manifestations . Two independent case-control studies from the UK have reported an association between homozygosity for CYP2D6*4 and AS . However, the biological and clinical significance of this association is not clear and this association was not noted in the French population .
Related posts:
What are the spondyloarthropathies?
Classification criteria for psoriatic arthritis and ankylosing spondylitis/axial spondyloarthritis
Measurement: function and mobility (focussing on the ICF framework)
Soluble and Tissue Biomarkers in Ankylosing Spondylitis
Therapeutic considerations in spondyloarthritis patients who fail tumour necrosis factor antagonists
Bacterial agents in spondyloarthritis: a destiny from diversity?
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
