Detailed analysis of the variability of peptidylarginine deiminase type 4 in German patients with rheumatoid arthritis: a case–control study

Peptidylarginine deiminase type 4 (PADI4) genotypes were shown to influence susceptibility to rheumatoid arthritis (RA) in the Japanese population. Such an association could not previously be confirmed in different European populations. In the present study, we analysed exons 2–4 of PADI4 in 102 German RA patients and 102 healthy individuals to study the influence of PADI4 variability on RA susceptibility by means of haplotype-specific DNA sequencing. Analyses of the influence of PADI4 and HLA-DRB1 genotypes on disease activity and on levels of anti-cyclic citrullinated peptide antibodies were performed. Comparing the frequencies of PADI4 haplotype 4 (padi4_89*G, padi4_90*T, padi4_92*G, padi4_94*T, padi4_104*C, padi4_95*G, padi4_96*T) (patients, 14.7%; controls, 7.8%; odds ratio = 2.0, 95% confidence interval = 1.1–3.8) and carriers of this haplotype (patients, 27.5%; controls, 13.7%; odds ratio = 2.4, 95% confidence interval = 1.2–4.8), a significant positive association of PADI4 haplotype 4 with RA could be demonstrated. Other PADI4 haplotypes did not differ significantly between patients and controls. Regarding the individual PADI4 variants, padi4_89 (A→G), padi4_90 (C→T), and padi4_94 (C→T) were significantly associated with RA (patients, 49.5%; controls, 38.7%; odds ratio = 1.6, 95% confidence interval = 1.1–2.3). Considering novel PADI4 variants located in or near to exons 2, 3, and 4, no quantitative or qualitative differences between RA patients (8.8%) and healthy controls (10.8%) could be demonstrated. While the PADI4 genotype did not influence disease activity and the anti-cyclic citrullinated peptide antibody level, the presence of the HLA-DRB1 shared epitope was significantly associated with higher anti-cyclic citrullinated peptide antibody levels (P = 0.033). The results of this small case–control study support the hypothesis that variability of the PADI4 gene may influence susceptibility to RA in the German population. Quantitative or qualitative differences in previously undefined PADI4 variants between patients and controls could not be demonstrated.


Introduction
Peptidylarginine deiminases (EC 3.5.3.15) are enzymes involved in the post-translational deimination of protein-bound arginine to citrulline [1]. Five different types of peptidylarginine deiminases encoded by the genes PADI1-PADI4 and PADI6 are currently known [1]. The presence of citrulline-modified target epitopes for autoantibodies is a well-known phenomenon in rheumatoid arthritis (RA) [2,3]. Peptidylarginine deimi-nases were recently implicated in the generation of anti-cyclic citrullinated peptide antibodies (anti-CCP) detectable in early stages of the disease [2][3][4]. The process resulting in anti-CCP formation is thought to play a pivotal role in early stages of RA evolvement since it is detectable several years before the onset of symptoms [5]. Certain evidence suggests that deimination of arginine at those peptide side-chain positions that interact with the so-called shared epitope of some major hisanti-CCP = anti-cyclic citrullinated peptide antibodies; PADI4 = peptidylarginine deiminase type 4; PCR = polymerase chain reaction; RA = rheumatoid arthritis; SNP = single nucleotide polymorphism.
(page number not for citation purposes) tocompatibility complex class II molecules (for example, HLA-DRB1*0401) may result in the generation of high-affinity peptides, thus inducing a strong in-vitro T cell activation [4,6].
A Japanese research group recently identified a genomic region (1p36) containing the genes PADI1-PADI4, which were suspected to be associated with susceptibility to RA [7]. Peptidylarginine deiminase type 4 (PADI4) was identified as the gene actually responsible for the association with RA. PADI4 has at least five main haplotypes that differ at four exonic single nucleotide polymorphisms (SNPs) and three subsequent amino acid substitutions [7,8]. While the socalled susceptibility haplotypes 2, 3, and 4 were found to be significantly more frequent in Japanese individuals suffering from RA, the non-susceptibility haplotype 1 predominated in healthy individuals [7]. These results could be confirmed by a further Japanese study [9]. However, studies in different European countries did not reveal significantly different PADI4 haplotype distributions in RA patients and healthy individuals. Moreover, no influence of the PADI4 genotype on disease severity could be detected [10][11][12][13][14]. Thus, the relevance of PADI4 variability for susceptibility to RA is still unclear.
A recent analysis of our group characterising exons 2-4 of the PADI4 gene identified new variants and haplotypes by a novel haplotype-specific sequencing-based approach [8]. Importantly, three novel coding SNPs in exons 2, 3, and 4 and three SNPs in introns 2 and 3 located near the exon-intron boundaries were found in 11/102 individuals (10.8%). Moreover, a closely related novel haplotype (haplotype 1B) was found in 2.9% of healthy individuals, which differs from haplotype 1 by padi4_92*G/padi4_96*C [8]. Since this additional variability of the PADI4 gene has not been assessed by other studies, the aim of the present case-control study was to investigate the possible influence of PADI4 genotypes including previously unknown PADI4 variants on susceptibility to RA in a German population.

Subjects and clinical data
Blood samples were obtained from 102 consecutive healthy, unrelated blood donors presenting in our institution as described previously [8]. These samples were analysed in our previous study for genetic variability of exons 2, 3, and 4 of the PADI4 gene [8]. Samples from 102 RA patients were enrolled to this study from the Department of Rheumatology, Charité Berlin and from the Rheumatology Unit, Ludwig Maximilian University, Munich. RA patients fulfilled the American College of Rheumatology criteria for RA [15]. The study was approved by the local ethics committee. All individuals were included in this study after informed consent was obtained.
The median age at onset of RA was 47 years (range, 6-86 years). One of the patients (age at onset, six years; PADI4 haplotype constellation 1 + 2/3) presented with juvenile RA and later transformed to classical RA. Another patient (age at onset, 14 years; PADI4 haplotype constellation 2/3 + 2/3) presented with an early manifestation of classical RA. When excluding these two patients the median age at onset was 48 years (range, 17-86 years). Of the RA patients, 75% were women. The Disease Activity Score 28 was available in 77 cases (median, 5.2; range, 1.8-8.1). Anti-CCP antibodies were detectable in 47 of 75 cases (median, 100 U/ml; range, 0-1600 U/ml). The median age of the controls was 40 years (range, 19-64 years), and 58 (57%) were female.
After digestion of the remaining primers and dNTPs by ExoSAP-IT (Amersham Biosciences, Freiburg, Germany), the PCR products were sequenced. All primers were synthesised by TIB Molbiol (Berlin, Germany). The designations of the PADI4 haplotypes are in accordance with those of Suzuki and colleagues [7]. The positions of novel exonic or intronic PADI4 variants were designated relative to sequences NM_012387 and NT_034376.1, respectively.

Statistical analysis
Chi-square tests (odds ratio, 95% confidence interval) and Fisher's exact tests were performed using GraphPad Prism 4 (GraphPad Software, San Diego, CA, USA). Comparison of the serum anti-CCP levels and the Disease Activity Score 28 regarding dependence of the PADI4 and HLA-DRB1 genotypes was assessed by the Mann-Whitney U test (median and 25th-75th percentiles are presented).

Distribution of PADI4 haplotype combinations
The frequencies of the PADI4 haplotype combinations found in our study are presented in Table 1. A detailed description of the variability of exons 2-4 of the PADI4 gene in healthy individuals analysed by haplotype-specific DNA sequencing was given in our previous report [8]. PADI4 haplotype 1 was most frequently found in the homozygous form (34.3%) and in combination with haplotype 2/3 (34.3%) in normal controls. In contrast, PADI4 haplotype 1 occurred more frequently in combination with haplotype 2/3 (30.4%) than in the  homozygous form (24.5%) in patients with RA. Most strikingly, the frequency of combined PADI4 haplotype 1/haplotype 4 was significantly different between patients (19.6%) and controls (8.8%) (P < 0.05). Both in patients and controls the distributions of the PADI4 haplotype combinations were in accordance with Hardy-Weinberg equilibrium.

Frequencies of PADI4 SNPs and novel PADI4 variants
The haplotype-specific sequencing based approach used in this study covered the genomic regions of exons 2, 3, and 4 of PADI4 and included the SNPs padi4_89, padi4_90, padi4_92, padi4_94, padi4_104, padi4_95, and padi4_96. The approach used therefore allowed a very detailed analysis of this part of the PADI4 gene that was implicated in influencing RA susceptibility. Of these SNPs, the frequencies of padi4_89A→G, padi4_90C→T, and padi4_94C→T in the RA patients (49.5%) were significantly different from those in the controls (38.7%) ( Table 3). The resulting odds ratio was 1.6 (95% confidence interval = 1.1-2.3, P = 0.04).

Discussion
This study provides a hint that variability of the PADI4 gene is related to the susceptibility to RA in the German population, whereas certain differences of hitherto unknown PADI4 variants between patients and controls were not found. The impact of PADI4 genotypes on susceptibility to RA remains controversial [7,[9][10][11][12][13][14]. Until now, certain PADI4 genotypes (haplotypes 2, 3, and 4) have been implicated to be involved in the pathogenesis of RA only in Japanese populations [7,9]. No such association of PADI4 variability with RA prevalence and severity could be demonstrated in various European populations [10][11][12][13][14]. In our study, also, an influence of PADI4 genotype on disease activity or anti-CCP level could not be demonstrated. The mechanism by which PADI4 variability may influence the break of tolerance is still unknown. Initially, it was argued that detectable differences in mRNA stability could result in higher enzymatic activity in cases where the susceptibility haplotypes (2,3 and 4) of PADI4 are present, leading to the generation of larger amounts of citrullinated peptides [7]. Most recently, a close association of the production of anti-CCP antibodies and HLA-DRB1 has been described [6,11,13,19], indicating the importance of antigen presentation in the induction of autoimmunity. This finding clearly could be confirmed in our study.
With the exception of haplotype 4, the frequencies of all other PADI4 haplotypes in our control individuals were comparable with those reported by other groups [7,9,10,14]. While the frequency of PADI4 haplotype 4 in our study (7.8%) was similar to that reported by groups from the United Kingdom (9.4%, P = 0.51; here termed haplotype 3) [10], Spain (5.9%, P = 0.32; padi4_94*T, padi4_104*C) [14], and Japan (5.5%, P = 0.17) [9], it was statistically significant different from the frequency reported by the large initial Japanese study (4.0%, P = 0.013) [7]. All of our patients and healthy individuals were Caucasian. The fact that the PADI4 haplotype 4 frequency in our control population was significantly higher compared with one of the Japanese studies [7] may therefore be influenced by differences in the ethnic background.
In our study, a statistically significant positive association of PADI4 haplotype 4 with RA was observed (odds ratio = 2.0, 95% confidence interval = 1.1-3.8). The presence of this haplotype did not influence disease activity or the anti-CCP level. We did not found an association of RA and PADI4 haplotypes 2 and 3, which were described as the principal susceptibility haplotypes in the Japanese population [7]. However, we cannot exclude that this difference may be influenced by the size of our study population.
When analysing the distributions of those PADI4 SNPs covered by our genotyping approach, padi4_89A→G, padi4_90C→T, and padi4_94C→T were found to be significantly associated with RA. These SNPs are common with PADI4 haplotype 4 and haplotype 2/3, whereas padi4_104C→T, padi4_95G→C, and padi4_96T→C, which are common with PADI4 haplotype 4 and haplotype 1, exhibited no association with RA.
The present study identified uncommon PADI4 variants that are not typically included among the five main PADI4 haplotypes. Consistent with our previous findings in healthy individuals [8], this study also revealed additional variability in PADI4 exons 2-4 in RA patients. As a result of this study, the frequency of uncommon PADI4 variants as identified earlier [8] was apparently not different quantitatively or qualitatively between patients and controls.
Of note, a statistically significant association between certain PADI4 genotypes and RA was detected in our study, in contrast to reports from other European groups [10][11][12][13][14]. This puzzling discrepancy may be due to influencing factors, such as a homogeneous Caucasian population, although we cannot definitely exclude other selection biases.
The question of whether PADI4 variability alters the interactions between the enzyme and possible target proteins remains unclear [20]. Further studies are needed to characterise the influence of this variability on the repertoire of deiminated target proteins.

Conclusion
In summary, the PADI4 haplotype 4 and the SNPs padi4_89A→G, padi4_90C→T, and padi4_94C→T were found to be significantly associated with RA in a German pop- ulation. The genomic region of PADI4 exons 2-4 of RA patients exhibits additional variability, which is apparently not different quantitatively and qualitatively between RA patients and controls. While the PADI4 genotype did not influence disease activity or the anti-CCP level, the presence of the HLA-DRB1 shared epitope was associated with significantly higher anti-CCP levels.