The R620W Polymorphism of the Protein Tyrosine Phosphatase PTPN22 Is Not Associated with Multiple Sclerosis

To the Editor:

We recently reported the association of the minor allele of a missense SNP (1858C→T [dbSNP accession number rs2476601]) in the gene for hematopoietic-specific intracellular protein tyrosine phosphatase (PTPN22) with susceptibility to both rheumatoid arthritis (RA [MIM 180300]) (Begovich et al. 2004) and systemic lupus erythematosus (SLE [MIM 152700]) (Kyogoku et al. 2004). Independently, Bottini et al. (2004) reported that the same risk allele is also strongly associated with type 1 diabetes mellitus (T1D [MIM 222100]). This SNP results in the substitution of a highly conserved arginine with tryptophan (R620W) in the proximal proline-rich SH3-binding domain of PTPN22, which is important for interaction with the c-Src tyrosine kinase, Csk (Cloutier and Veillette 1996; Gregorieff et al. 1998), and for down modulation of T cell receptor signaling via phosphorylation of regulatory tyrosines of another Src family kinase, Lck (Cloutier and Veillette 1999; Gjorloff-Wingren et al. 1999). In vitro experiments show that the W620 variant of PTPN22 binds less efficiently to Csk (Begovich et al. 2004; Bottini et al. 2004), suggesting that T cells expressing this allele may be hyperresponsive. Correspondingly, knockout mice deficient in the murine orthologue of PTPN22, PEP, show selective disregulation of the effector/memory T cell compartment, with enhanced activation of Lck, hyperproliferation, and exaggerated early signaling responses in restimulated T cells (Hasegawa et al. 2004). These mice also spontaneously develop germinal centers and increased serum levels of certain immunoglobulin isotypes; however, they do not display overt signs of autoimmunity.

Together, these genetic and functional data suggest that the W620 variant of this phosphatase is a significant genetic risk factor for both organ-specific (T1D) and systemic (RA and SLE) autoimmune syndromes, perhaps because of enhanced activation of memory/effector T cells. These findings support the hypothesis that there are common genetic variants contributing to general immune dysregulation and susceptibility to autoimmunity (Marrack et al. 2001; Wandstrat and Wakeland 2001) and suggest that this SNP should be studied in other autoimmune disorders.

Multiple sclerosis (MS [MIM 126200]) is a severe autoimmune disorder of the CNS characterized by chronic inflammation, myelin loss, gliosis, varying degrees of axonal pathology, and progressive neurological dysfunction (Hauser and Goodin 2001). A large body of research supports a multifactorial etiology for MS, with an underlying complex genetic component likely acting in concert with undefined environmental exposures. HLA-DR, within the MHC on chromosome 6p21, is the only locus that has consistently demonstrated linkage in studies of families with MS (Haines et al. 1996; Sawcer et al. 1996; Barcellos et al. 2002), and both family-based and case-control MS studies have convincingly shown a specific association between the HLA-DR2 (DRB1*1501, DQB1*0602) haplotype and disease susceptibility (Fogdell-Hahn et al. 2000; Rubio et al. 2002; Barcellos et al. 2003). HLA has been estimated to account for 17%–62% of the genetic etiology as calculated from sibling relative risk (Haines et al. 1998); hence, much of the genetic effect in MS remains to be explained.

To determine whether the R620W SNP in PTPN22 also plays a role in susceptibility to MS, we genotyped two large, well-characterized, family-based data sets of 748 MS-prone families (563 single-case and 185 multicase families) comprising 3,251 individuals, including 1,086 affected individuals and 2,165 unaffected family members. All individuals and their known ancestors were non-Hispanic whites of European descent. Diagnostic criteria, ascertainment protocols, and clinical and demographic characteristics are summarized elsewhere (Barcellos et al. 2002, 2003). A third African American data set was ascertained using identical clinical criteria and consisted of 398 cases and 209 controls (primarily spouses of patients) (Oksenberg et al. 2004). The appropriate institutional review boards approved all studies, and written informed consent was obtained from all participants. Genotyping was performed using a Taqman Assay By Design (Applied Biosystems). Primer and probe sequences were as follows: forward primer, 5′-CAACTGCTCCAAGGATAGATGATGA-3′; reverse primer, 5′-CCAGCTTCCTCAACCACAATAAATG-3′; probe for C allele, 5′-FAM-TCAGGTGTCCGTACAGG-3′; and probe for T allele, 5′-VIC-TCAGGTGTCCATACAGG-3′. This assay was validated using samples of known genotype. All family genotypes were examined for Mendelian inconsistencies by use of PEDCHECK (O’Connell and Weeks 1998), and any discrepancies were addressed. Family-based association analysis was performed using the pedigree disequilibrium test (PDT), version 4.0 (Martin et al. 2000). P values for χ² or Fisher’s exact tests of allele or genotype case-control comparisons were derived using SAS, version 9.0 (SAS Institute). No deviations from Hardy-Weinberg equilibrium were observed for PTPN22 R620W genotypes in any of the patient groups, controls, or unrelated family founders.

We performed an extensive evaluation of this polymorphism, using large family-based and case-control comparisons. First, we compared the W620 allele frequencies of the white probands with MS from each of the two family-based studies with the results generated from 1,961 white controls (Begovich et al. 2004). The frequency of the T allele in controls was 8.7%, compared with 9.8% in the simplex probands (P=.22), 8% in the multiplex probands (P=.65), and 9.4% in the combined patient group (P=.40). The frequency of this SNP allele in 398 African American patients with MS was 2.9%, compared with 1.4% in 209 spousal controls (P=.11) and 2.4% in 409 randomly chosen African American controls (P=.58) (Begovich et al. 2004).

Results from PDT analyses for both the single-case and multicase families are shown in table 1. There was no evidence of transmission distortion in either data set considered individually (P=.48 and .28, respectively) or when combined for analysis (P=.69). Similar results were also observed for PDT analyses of PTPN22 genotypes in families (P=.56; data not shown) (Martin et al. 2003). Families were stratified by the presence or absence of HLA-DR2 in affected individuals, to identify potential genetic interactions with the PTPN22 R620W SNP. No difference in transmission of either allele was observed in either DR2-positive or DR2-negative families (P=.91 and .22, respectively). To identify a genetic relationship between PTPN22 R620W and clinical presentation of MS, families were also categorized by whether cases had an initial relapsing-remitting (RR) course or had a primary progressive or progressive relapsing (PP or PR) course. Significant results for PTPN22 R620W were not observed in either clinical subset (P=.59 in one and .69 in the other). Furthermore, no evidence of association was present in the African American data set when cases were stratified by disease course (RR at onset vs. PP or PR at onset) and then compared (P=.18; data not shown). Overall, we observed no evidence of genetic association between the PTPN22 R620W polymorphism and MS susceptibility or disease course in both the white and African American data sets.

Table 1.

P Values from PDT Analysis of White Families with MS

Families	n	P
All	748	.69
Single-case	563	.48
Multicase	185	.28
With disease phenotypea:
RR course at onset	652	.59
Progressive course at onset	36	.69
With HLA-DR2 statusb:
DR2 positive	324	.91
DR2 negative	231	.22

^aFor multicase families, only those families in which all members were concordant for either phenotype (RR course at onset or progressive course at onset) were used for analysis; phenotypically discordant families were not used.

^bHLA data were available for 555 families. Multicase families in which all affected members were not concordant for DR2-positive or DR2negative classifications were not used in the subgroup analyses.

We calculated power for PDT analyses in the families with MS under a multiplicative model (Risch and Merikangas 1996), assuming an allele frequency of 9.0% (derived from white controls used in this study) and a type 1 error rate of 5%. On the basis of these assumptions, we estimate that we have close to 90% power to detect a very modest genotype relative risk of 1.5 for disease susceptibility in the family data set. This estimate is very conservative, in that a combined total of 475 independent trios and 375 independent discordant sib pairs (DSPs) were used for calculation of power. A total of 582 trios and 1,467 DSPs were actually available for the analyses, and, although this larger number of trios and DSPs were not derived from independent families, their inclusion is expected to increase statistical power substantially (Martin et al. 2000). It is important to note that, owing to the significantly lower PTPN22 SNP allele frequency in the African American controls, compared with that in white controls (1.4%–2.4% vs. 8.7%), there is less power to detect modest genetic effects in the African American case-control analysis. Under the assumption of a type 1 error rate of 5%, there is 80% power to detect a relative risk of ∼3.0–4.0. To rule out a smaller genetic contribution to disease susceptibility for PTPN22 R620W in this population, a larger data set is needed.

The observation that MS is not associated with an increased frequency of the PTPN22 W620 variant, whereas RA, T1D, and SLE are, suggests that the latter three diseases share a common underlying mechanism that may not be integral to MS pathogenesis. All three diseases associated with the PTPN22 W620 variant are characterized by the presence of autoantibodies (anti-citrulline antibodies and rheumatoid factor in RA, anti-GAD (anti–glutamate decarboxylase) antibodies in T1D, and an array of autoantibodies in SLE) that, more often than not, predate clinical disease. This has led to the proposal that the PTPN22 W620 variant may predispose individuals to autoimmunity by facilitating the generation of certain disease-associated autoantibodies, thereby contributing to disease onset and progression (Kyogoku et al. 2004). This hypothesis is supported by the observation that the W620 variant is strongly associated with rheumatoid factor–positive RA, whereas there is no significant association with rheumatoid factor–negative RA (Begovich et al. 2004; Lee et al., in press).

Interestingly, MS is thought to be primarily a T cell–mediated disease, and, although there are reports of anti-myelin antibodies in the sera and cerebrospinal fluid (CSF) of patients with MS, these antibodies appear to be of low affinity, and their role in disease pathogenesis and progression is not well understood. Genain et al. (1999) have observed myelin-specific infiltrating B cells in the brains of patients with MS, and the CSF and brain of an affected individual can show an elevated frequency of clonally expanded B cells with properties of postgerminal center memory or antibody-forming lymphocytes (Baranzini et al. 1999). This apparent dichotomy can be explained by the observation that MS has distinctive patterns of presentation; this heterogeneity is also reflected in the neuropathology of MS. Recent work has demonstrated the existence of multiple histologic phenotypes in which the core components of the lesion (inflammation, demyelination, and oligodendrocyte pathology) are present in different degrees (Lucchinetti et al. 2000). For example, some hyperacute forms are characterized by substantial CNS inflammation, which suggests cellular mechanisms of disease, whereas immunoglobulin deposition is prominent in other demyelinating forms. The well-documented effect of locus heterogeneity in MS raises the question of whether distinct phenotypes are associated with distinct underlying genotypes. Hence, although the experimental tools to accurately assess humoral endophenotypes in MS are not yet available, the possibility of PTPN22 variation being a susceptibility factor in a subset of patients with a dominant antibody component warrants further investigation.

In summary, the W620 variant of PTPN22 does not appear to play a major role in MS pathogenesis in our data sets. Analyses of the role of this SNP in additional autoimmune disorders should provide hints about the underlying mechanisms of these complex diseases as well as a framework to begin to understand how they are related and why certain autoimmune diseases can cosegregate in the same families.