Introduction
TNFAIP3 (tumor necrosis factor α-induced protein 3) encodes a ubiquitin-editing protein, A20, known as an inhibitor of nuclear factor-κB (NF-κB). Several adaptor molecules are thought to associate with A20 and be involved in inhibition of NF-κB [
1]. TNIP1 (TNFAIP3 interacting protein 1), also known as ABIN (A20-binding inhibitor of NF-κB)-1, is one such adaptor molecule binding to A20.
TNIP1 mRNA is strongly expressed in peripheral blood lymphocytes, spleen and skeletal muscle, and the expression is also detected in kidney [
2].
TNIP1 expression is induced by NF-κB, and in turn, overexpression of
TNIP1 inhibits NF-κB activation by TNF [
1], although deficiency of
TNIP1 has few effects on NF-κB inhibition [
3]. Thus, TNIP1 appears to play a role in NF-κB inhibition, at least partly by interacting with A20. In addition, TNIP1 was shown to inhibit TNF-induced apoptosis independently of A20 [
3].
TNFAIP3, located at 6q23, has been identified as a susceptibility gene for both systemic lupus erythematosus (SLE) and rheumatoid arthritis (RA) in Caucasian and Asian populations [
4‐
8]. Recently, Shimane et al. [
9] replicated association of
TNFAIP3 single nucleotide polymorphisms (SNPs) with SLE and RA in a Japanese population. We also detected association of
TNFAIP3 rs2230926 with Japanese SLE patients in an independent study [
10].
Recently a genome-wide association study (GWAS) reported association of
TNIP1 (5q32-q33.1) as well as
TNFAIP3 SNPs with psoriasis in the Caucasian populations [
11]. Subsequently, two recent GWAS revealed association of
TNIP1 intronic SNPs rs7708392 and rs10036748, which are in strong linkage disequilibrium (LD) with SLE in the Caucasian (European-American and Swedish) and Chinese Han populations, respectively [
8,
12]. These observations underscored the role of the pathway involving
TNFAIP3-TNIP1 in the genetic predisposition to SLE. The association of
TNIP1 with SLE needs to be further confirmed.
Recently, it has become increasingly clear that SLE and RA share a number of susceptibility genes.
TNFAIP3 [
4‐
10],
STAT4 [
13,
14] and
BLK [
15,
16] represent such shared susceptibility genes.
TNIP1 has been shown to be upregulated in synovial tissues from RA [
17], raising a possibility that
TNIP1 may also play a role in the pathogenesis of RA. To date, association of RA with
TNIP1 has not been reported.
This study was conducted to examine whether TNIP1 was associated with SLE and RA in a Japanese population.
Discussion
In the present study, we replicated the association of TNIP1 rs7708392 with SLE in a Japanese population, which indicated that TNIP1, as well as TNFAIP3, is a susceptibility gene to SLE shared by the Caucasian and Asian populations. Because both TNIP1 and A20 are thought to be involved in the inhibition of NF-κB activation, genetic association of these genes implicates a causal role of NF-κB regulation pathway in the pathogenesis of SLE.
Kalergis et al. [
24] demonstrated that expression of IκB-α, an inhibitor of NF-κB, was decreased in Fcγ receptor IIb-deficient mice which present lupus-like symptoms, and the symptoms were reduced by treatment with NF-κB inhibitors. Previous studies demonstrated that
TNFAIP3 risk allele rs2230926G (127Cys) leads to reduced inhibitory activity of NF-κB activation [
6] or reduced mRNA level of
TNFAIP3 [
10]. In view of these observations, it is speculated that the risk allele of
TNIP1 may also be associated with reduced inhibitory activity of
TNFAIP3-TNIP1 pathway.
TNIP1 rs7708392 is located in intron 1. Expression analysis using the GENEVAR [
25] and the International HapMap databases [
26] as previously described [
27] did not show significant effect of rs7708392 genotypes on the mRNA level of
TNIP1 (data not shown). Although the direct molecular mechanism of the risk allele to cause SLE remains unclear, it is possible that the risk allele may be associated with the selection of splicing variant. To date, at least 11 splice variants of
TNIP1 have been identified [
1]. Presence of alternative exon 1A and 1B, as well as splice variants lacking exon 2, has been described. Because rs7708392 is located between exon 1B and exon 2, it is possible that this SNP may influence the usage of the splicing isoform. It is also possible that other causative SNPs in tight LD with rs7708392 may exist. Such a possibility would be addressed by resequencing the entire
TNIP1 gene.
Interestingly, in sharp contrast to the Caucasian populations, the risk rs7708392C constituted the major allele in the Japanese population. This resulted in substantially higher PAR% in the latter. We previously reported similar findings in
STAT4 and
BLK SNPs [
14,
15]. In Chinese, a SNP rs10036748, which is in tight LD with rs7708392 in Japanese (
r
2
= 0.81, HapMap database [
26]), has been shown to be associated with SLE. The frequencies of rs10036748 risk allele in Chinese (cases 79.7%, controls 66.1%) [
8] are similar to those of rs7708392 in Japanese (Table
2). It should be noted that, because the information used to estimate the PAR% was based on the data from a variant that has not been shown to be the causal variant in
TNIP1, and the estimates of the allele frequency and OR (as an approximation for RR) were taken from a rather small case-control study, the PAR% values shown here represent rough estimates. Nevertheless, the data suggest that the significance of
TNIP1 in the genetic background of SLE may be substantially greater in the Asian than in the Caucasian populations.
In the association analysis with the clinical subsets, none of the case-only comparisons (cases with each clinical phenotype versus those without) reached statistical significance, partly because of the insufficient statistical power caused by the small sample size due to stratification. However, preferential association of
TNIP1 with renal disorder and anti-dsDNA antibody was suggested by comparison with healthy controls. In our subjects, preferential association with renal disorder was also observed for
TNFAIP3 [
10].
On the other hand, association was not observed with the SLE subsets having neurological disease, serositis, anti-Sm antibody and age of onset <20. It is interesting to note that renal disorder and presence of anti-dsDNA are significantly correlated in SLE, while neurologic disorders are not, suggesting that these clinical features might represent different clinical subsets of SLE [
28]. In view of this, our findings could be interpreted such that polymorphisms in
TNIP1-
TNFAIP3 pathway might play a significant role in the subset of SLE characterized by renal disorder and anti-dsDNA antibody, but not in the subset with neurologic disease. Such a hypothesis should be validated in future large-scale studies.
No strong evidence for association of rs7708392 with RA was obtained in this study. The sample size in this study (553 RA patients and 513 controls) provides 80% power to detect associations with genotype relative risk of 1.32 or greater, but we cannot rule out a possibility of weak association. Recently published meta-analysis of GWAS in Caucasians also failed to demonstrate statistically significant association of
TNIP1 SNP with RA, although similarly to our observation, a tendency for association was detected [
29]. Thus, while a role of
TNFAIP3 is observed both in SLE and RA genetics,
TNIP1 appears to play a major role in SLE, but not in RA. Such a difference might possibly imply that the molecular mechanism of
TNIP1 association might not be fully explained by A20 modification. In support of this possibility, TNIP1 has been shown to block TNF-induced programmed cell death in
TNFAIP3 deficient cells, indicating that TNIP1 does not always require A20 to perform its anti-apoptotic function [
3]. Thus, further analysis on the molecular mechanisms involving these molecules is required.
Competing interests
RRG and TWB are employees of Genentech, Inc. (South San Francisco, CA, USA). The other authors declare that they have no competing interests.
Authors' contributions
AK participated in the study design, carried out all genotyping and statistical analyses, and wrote the manuscript. JO carried out statistical analysis with AK and helped in the manuscript preparation. SI, HF, TH, DG, IM, MK, KM, ST, YT, HH and TS recruited Japanese patients with SLE and collected clinical information. RRG and TWB provided Caucasian data. NT designed and coordinated the study and helped in the manuscript preparation. All authors read and approved the final manuscript.