PTPN22: a Confirmed Rheumatoid Arthritis Susceptibility Gene?
Total Page:16
File Type:pdf, Size:1020Kb
For reprint orders, please contact: EDITORIAL [email protected] PTPN22: a confirmed rheumatoid arthritis susceptibility gene? for association with RA, or were located in link- ‘It will be important...to verify age regions identified in whole genome screens. whether the function of the A missense SNP within the PTPN22 gene was (PTPN22) gene in autoimmune found to be associated with RA in their initial disease susceptibility is the same discovery data set (p = 0.0007, OR: 1.65, 95% CI: 1.23–2.20) and also in a second repli- for all diseases.’ cation data set (p = 2.1 x 10-8, OR: 1.97, 2005 was a landmark year for research into the 95% CI: 1.55–2.5) [7]. The polymorphism is a genetics of rheumatoid arthritis (RA). It was C→T substitution (rs2476601) at nucleotide the year that a new confirmed susceptibility gene position 1858 that results in a tryptophan (W) Anne Hinks was identified, along with a new pathological for arginine (R) transition at codon 620. Inter- University of Manchester, pathway important in susceptibility to RA. The estingly, an earlier independent association study Arthritis Research Campaign protein tyrosine phosphatase, nonreceptor on a different autoimmune disease, Type 1 dia- Epidemiology Unit, Manchester, M13 9PT, UK type 22 (PTPN22) gene is a member of the pro- betes (T1D), demonstrated an association with Tel.: +44 161 275 1669; tein tyrosine phosphatase (PTP) family of pro- the same SNP [8]. Fax: +44 161 275 5043; teins, which are key regulatory proteins of the The original association in RA has subse- Anne.Hinks@manchester. immune system. The PTPs work in conjunction quently been replicated by many groups in all ac.uk with protein tyrosine kinases (PTKs), regulate populations studied, including British [9,10], Span- the reversible phosphorylation of tyrosine resi- ish [11], Norwegian [12], Dutch [13], Canadian [14], dues and are fundamental in controlling many New Zealand [15] and Finnish [16] populations. In physiological processes [1,2]. all studies, the direction of the association is the RA, similar to other autoimmune diseases, is same; there is an increase in T allele frequency in a complex genetic disease and dissecting the cases compared with controls, with the ORs for contributing genetic variants is challenging. RA conferred by carriage of the T allele ranging The association of the shared epitope alleles of from 1.38 to 2.04. Together, these data provide major histocompatability complex, class II, DRβ1 compelling evidence that PTPN22 is a genuine (HLA-DRB1) with RA is well established and, susceptibility gene for RA. until this year, has been the only consistently rep- All of the PTPN22 and RA association studies licated RA susceptibility gene and the only region have examined whether there are any potential to be identified in all four whole genome screens interactions between PTPN22 and HLA-DRB1. for RA [3–6]. The HLA-DRB1 gene is thought to In the first study, Begovich and colleagues per- account for approximately 40% of the genetic formed conditional logistic regression to adjust contribution to RA and, therefore, genes account- for the HLA-DRB1 genotype and found that it ing for a large proportion of the genetic risk had little impact on risk estimates [7]. Many remain to be identified. The search for non-HLA other studies have also stratified the data accord- susceptibility genes has been in progress for a ing to shared epitope (SE) status and in all cases, number of years; however, the task is complicated no significant difference in allele or genotype fre- by the fact that the remaining proportion of risk is quencies between SE-positive and -negative cases predicted to be due to multiple genes, each with a has been observed [9,10]. All of these analyses sug- modest effect on overall RA susceptibility (odds gest that PTPN22 is acting independently ratio [OR] < 2.0). There have been several associ- of HLA-DRB1. ations of RA with various genes, although replica- Many studies have also examined the effect of tion of some of these associations in independent gender, age at disease-onset, rheumatoid factor data sets has generally been elusive. (RF) status, family history of disease and severity of However, in 2005, Begovich and colleagues disease course upon the RA–PTPN22 association. performed a case–control association study of All of the studies found no consistent gender potential functional single nucleotide poly- difference in the association of PTPN22 with RA. morphisms (SNPs) located in candidate genes Some studies have shown that the association is 10.2217/17460816.1.2.153 © 2006 Future Medicine Ltd ISSN 1746-0816 Future Rheumatol. (2006) 1(2), 153–158 153 EDITORIAL – Hinks stronger in RA cases with a young age at disease- that have observed an increased statistical risk onset [9,10], whilst others found this had no effect that an individual with an autoimmune disease on the association [11,14]. will have a blood relative with either the same or The issue that appears to be most contra- another autoimmune disease [30–33]. The hypo- dictory across the different studies is from the thesis has been strengthened further in results of the stratification according to RF sta- recent years, following whole genome screens of tus. The initial association study by Begovich different autoimmune diseases. Analysis of the and colleagues suggested that the association results of autoimmune disease whole genome with PTPN22 was only found in the subgroup screens identified 18 distinct clusters, with evi- of patients that were RF positive [7]. This was dence of linkage to two or more autoimmune or also borne out when the researchers further immune-related diseases [34]. This overlap of loci increased their sample size [17] and in another has also been observed in mouse models of RA cohort [10]. Conversely, a number of other autoimmunity [35]. studies have demonstrated an association in both RF-positive and -negative subgroups [9,14,15]. ‘The association of the PTPN22 gene with The issue here is that, as RF is present in RA is the first widely replicated genetic 75–80% of RA cases, the RF-negative subgroup association with this disease since HLA.’ sample sizes in the different studies are modest. However, the RF-negative subgroup sample size However, it is also now becoming clear that was also used by Lee and coworkers and should the PTPN22 missense SNP is not universally have had sufficient power to detect an associa- associated with autoimmune diseases and a bet- tion with PTPN22. It is possible that there has ter picture of which diseases are and are not been a misclassification of RF-positive RA cases associated with PTPN22 will provide a better into the RF-negative RA case subgroup in the understanding of different autoimmune disease studies where there is an association with pathways. Diseases that, to date, do not appear PTPN22, possibly due to different RF detection to be associated with PTPN22 include multiple methodologies, or to the time points, in terms of sclerosis (MS) [9,28,36,37], psoriasis [9,38], psoriatic disease duration, where these tests were per- arthritis [9], Sjögren’s syndrome [39], celiac dis- formed. Alternatively, these differences may ease [40,41], Crohn’s disease [14] and systemic reflect clinical heterogeneity across different sclerosis [S Parameshwar & J Worthington, Unpublished Data]. populations. Additional studies of larger cohorts The function of the protein encoded by of RA cases and controls, as well as combined PTPN22 may provide clues to the role of data or meta-analyses, will be required to give a PTPN22 in autoimmune diseases and the path- better understanding of the association of ways that might be important. PTPN22 encodes PTPN22 with RA and its clinical presentation. the protein lymphoid-specific phosphatase Since the original associations in both T1D (Lyp), which is composed of an N-terminal and RA, there has been a flurry of reports of the phosphatase domain and a noncatalytic C-ter- association of PTPN22 with other autoimmune minal end containing several proline-rich diseases. Diseases that show evidence for associ- motifs. It is expressed in hemopoietic tissues, ation with PTPN22 include RA [7,9–12,14–17], thymus, spleen and bone marrow, as well as in all juvenile idiopathic arthritis (JIA) [9,12,16], subtypes of peripheral blood mononuclear cells, T1D [8,18–23], Graves’ disease [21,24,25], systemic including T and B cells, monocytes, neutrophils lupus erythematosus (SLE) [11,26,27], Hashi- and natural killer cells [7]. Initial studies of the moto’s thyroiditis [28], autoimmune Addison’s function of the protein have largely been deter- disease [24] and generalized vitiligo [29]. mined through experiments on the mouse The fact that PTPN22 is associated with homolog, PEST domain-enriched tyrosine numerous autoimmune diseases lends support to phosphatase (PEP), which is encoded by the the hypothesis that there are common patho- gene PTPN8. Cloutier and Veilette showed that physiological pathways underlying certain there is a synergistic relationship between a PTP autoimmune diseases. The hypothesis of over- (PEP) and a PTK (C-terminal Src tyrosine lapping susceptibility to autoimmune diseases is kinase [Csk]), which together inhibit signaling not a new one, as it was originally proposed fol- of Src family kinases, such as Lyk, Fyn and lowing observations of clustering of different ZAP-20, and thus mediate T-cell inactiva- autoimmune diseases within families. This has tion [42]. They demonstrated that PEP inhibited been confirmed through epidemiological studies T-cell receptor signaling by dephosphorylation 154 Future Rheumatol. (2006) 1(2) PTPN22: a confirmed rheumatoid arthritis susceptibility gene? – EDITORIAL of the Src family kinases, but this dephosphory- signaling may lead to weaker signaling and a fail- lation was dependent on the interaction of PEP ure to delete autoreactive T cells during thymic with the SH3 domain of Csk.