Introduction

Rheumatoid arthritis (RA) is a systemic autoimmune disease that affects about 0.5–1% of the global population1. According to American College of Rheumatology/European League Against Rheumatism criteria, the disease is defined as inflammatory arthritis2. The pathogenesis of the disease is still unknown; however, it is considered a multistage process that linked genetic factors (e.g. HLA-DRB1 locus), environmental factors (e.g. bacterial and viral infections) and behavioral factors (e.g. smoking, physical activity)3,4,5,6.

The Peptidyl Arginine Deiminase 4 (PADI4) gene is located on chromosome 1 at location 17.308.195–17.364004, on the forward strand with reference to the assembly GRCh38. The gene encodes 5 splice variants and is associated with the development or function of the immune system, chromatin organization, and protein modification processes such as histone alternation (Ensembl database, release 103; 7). According to The Human Protein Atlas8 (available from: http://www.proteinatlas.org; release date: 2021.02.24) and Ensembl genome browser (version: 92.38), the gene is expressed primarily in the spleen, bone marrow, granulocytes, monocytes, and to a lesser extent is transcripted in T and B cells and dendritic cells. The PADI4 gene is associated with the formation of neutrophil extracellular traps (NETs) formation9,10. Previous reports have shown that NETs are related to a variety of autoimmune diseases, including RA. In this process, PADI4 is responsible for histone modification and promotes chromatin decondensation. Neutrophils secrete NETs that contain extracellular chromatin with histones and granular proteins. This mechanism leads to an antimicrobial effect, but may also be pathological due to the nonspecific nature of NETs that leads to uncontrolled pro-inflammation11. The PADI4 gene is a member of the family genes that encodes the PAD4 enzyme, which is responsible for the post-translational protein citrullination by converting arginine residues to citrulline. Antibodies against citrullinated proteins (ACPA) together with rheumatoid factor (RF) are important markers of RA. Furthermore, PAD4 is the target of anti-PAD4 antibodies (anti-PAD4), that are associated with a more unfavorable course of the disease12.

Single nucleotide polymorphisms (SNPs) have been related to susceptibility to RA 3. The first and most significant genetic locus associated with the development of the disease are class II human leukocyte antigen (HLA) genes and other non-HLA genes, including: PADI4, IL23R, PTPN22 and others 13. Previous studies have shown an association between SNPs in PADI4 gene and susceptibility to RA mainly in Asian population, but the results are inconsistent amongst Caucasians 14.

An SNP with rs2240340 G > A, PADI4_94, is located in the intronic region of the PADI4 gene with a minor allele frequency in the European population (HapMap CEU) of 0.43 15. The association between PADI4_94 and susceptibility to RA in the Japanese population was reported with an odds ratio (OR) of 1.22 16. On the contrary, this association has not been confirmed with respect to the European population 17,18,19. A SNP with rs1748033 T > C (A > G), PADI4_104, is located in the 5' untranslated region with a minor allele frequency in the European population (HapMap CEU) of 0.36. The nucleotide change T > C is responsible for the synonymous leucine change at codon 117 15. Some studies have shown the association between the aforementioned SNP and susceptibility to RA. In the Japanese population, the higher risk of RA development was related especially to male smokers 20. No relationship has been found between PADI4_104 and RA in Caucasians 19,21,22. It may appear that both SNPs are linked to the development of RA in Asian population, and there is no association amongst Caucasians, but the meta-analysis conducted by Lee et al. 14 has shown a new point of view on the above relation. The PADI4_94 and PADI4_104 variants were associated with susceptibility to RA in both populations after implementing a specific genetic model. The aim of this study was to investigate the association between the two polymorphisms PADI4_94 (rs2240340) and PADI4_104 (rs1748033) and the levels of anti-PAD4 antibodies in patients with RA, taking into account the different genetic models.

Results

The prevalence of the PADI4_94 and PADI4_104 minor allele (for a total of 122 RA patients) was 0.48: GG = 36 (29.5%), GA = 56 (45.9%), AA = 30 (24.6%) and 0.36: GG = 49 (40.2%), GA = 58 (47.5%), AA = 15 (12.3%), respectively. The distribution of genotypes in patients with RA was consistent with the Hardy–Weinberg equilibrium (p = 0.38 and 0.73, respectively). Regarding PADI4_94, in the anti-PAD4 positive RA group (n = 68), the prevalence of genotypes was as follows: GG = 17 (25%), GA = 33 (48.5%), AA = 18 (26.5%). In the anti-PAD4 negative group (n = 54) there were 19 homozygotes GG (35.2%), 23 heterozygotes GA (42.6%), and 12 homozygotes AA (22.2%). The distribution of PADI4_104 genotypes in anti-PAD4 positive patients with RA was as follows: GG = 28 (41.2%, GA = 33 (48.5%) and AA = 7 (10.3%). In the anti-PAD4 negative group there were 21 (38.9%) carriers of GG genotype, 25 (46.3%) carriers of GA and 8 carriers of AA (14.8%). Genotype distribution for the following groups: ACPA positive vs. ACPA negative, and RF positive vs. RF negative can be found in Supplementary Table S1.

In relation to PADI4_94 in the co-dominant 1 model there was no difference in the level of anti-PAD4 antibodies (p = 0.52) between carriers of the GG genotype—median: 558.17 U/ml [interquartile range: 368.94–1002.15] and carriers of the GA genotype 699.74 [368.68–1372.55].

The co-dominant 2 model (GG vs. AA genotypes) also showed no differences: 558.17 [368.94–1002.15] vs. 925.55 [278.16–1815.44], p = 0.54. The dominant model (GA + AA vs. GG) showed a lack of association (p = 0.47) between anti-PAD4 level and genotype distribution—727.81 [339.36–1398.8] vs. 558.17 [368.94–1002.15], respectively. There was no relationship to the over-dominant model (GA vs. GG + AA, p = 0.79): 699.74 [368.68–1372.55] vs. 679.16 [337.02–1286.2]. The recessive model (AA vs. GA + GG) also did not show an association (p = 0.66) between anti-PAD4 levels and genotype distribution—925.55 [278.16–1815.44] vs. 677.9 [368.68–1259.8], accordingly.

In relation to PADI4_104 in the co-dominant-1 model (GG vs. GA), the subgroups had similar levels of anti-PAD4 antibodies: 706.58 [412.17–1286.2] vs. 681.39 [336.22–1233.4]; p = 0.82.

In the co-dominant-2 model (genotypes GG vs. AA) there was no difference in the levels of anti-PAD4—706.58 [412.17–1286.2] vs. 564.96 [271.26–1910]; p = 0.85.

Dominant and over-dominant models (genotypes GA + AA vs. GG and GA vs. GG + AA, respectively) also showed an insignificant association between the groups: 680.86 [325.98–1288.1] vs. 706.58 [412.17–1286.2]; p = 0.8 and 681.39 [336.22–1233.4] vs. 691.49 [405.61–1372.55]; p = 0.86, accordingly. The recessive model (AA vs. GA + GG) did not show differences in anti-PAD4 levels–564.96 [271.26–1910] vs. 681.92 [346.2–1286.2]; p = 0.93. Furthermore, genetic models did not show differences in the levels of other antibodies (ACPA and RF) between seropositive and seronegative patients with RA. For details, please refer to Supplementary Tables S2-S8. Haplotype analysis was performed using the LDhap tool 23 (available from https://ldlink.nci.nih.gov/?tab=ldhap). Between PADI4_94 and PADI4_104, 3 functional haplotypes were estimated in the European population: G_G, A_A, and A_G. In our study, the A_G haplotype was represented by 3 cases, therefore it was omitted in further analysis. There were no associations between anti-PAD4, ACPA and RF levels and haplotypes (p = 0.99, p = 0.36 and p = 0.6, respectively).

Discussion

The present study shows the lack of association between anti-PAD4 antibody levels and the prevalence of genotypes with respect to different genetic models (codominant, dominant, overdominant, and recessive) for both single nucleotide polymorphisms: PADI4_94 (rs2240340) and PADI4_104 (rs1748033).

A total of five PAD enzyme isotypes were found, in which PAD2, PAD3, and PAD4 were responsible for autoimmune reactions in RA. PAD4 is the most characterized. It is found mainly in the nucleus of white blood cells, and its overexpression has been found specifically in neutrophils and monocytes in synovial tissue. SNPs in the PADI4 gene have also been linked to the development of RA in some populations, especially Asian, but amongst Europeans the results are inconclusive 24,25. The new light on this issue was presented in the study conducted by Lee et al. 14. They showed that the variants PADI4_94 and PADI4_104 may be associated with susceptibility to RA in Asian and Caucasian populations when homozygous contrast was used. Consequently, we assumed that the use of genetic models may be helpful in determining the relationship between SNPs and antibody levels. In addition to genetic factors associated with RA, epigenetic changes have been reported in the PADI4 gene. Increased methylation in the promoter region was associated with lower disease activity, lower levels of ACPA, and anti-PAD4 antibodies 26.

Anti-PAD4 antibodies are associated with structural damage of joints and a more severe clinical outcome; therefore, their evaluation may be of prognostic importance 25. The specificity of them was reported to be greater than 95% in patients with RA. Antibodies are not specific to RA, as they have been found in other rheumatic diseases, e.g. SLE, but at lower levels 27. Anti-PAD4 antibodies have an incidence ranging from 16.2 to 50% of RA patients and may be related to the study population and duration of the disease 12,24,28,29,30. In this study, anti-PAD4 positivity was approximately 56% and may be associated with the duration of the disease (the median duration of the disease was 10 years) and the relatively small size of the group derived from one medical center. Moreover, Reyes-Castillo et al. 24 showed that patients with RA, with the disease lasting more than 2 years (mean duration of the disease was 8 years), had higher levels of anti-PAD4 antibodies compared to patients with disease duration of less than 2 years.

We found no association between anti-PAD4 antibody levels and genotypes, which is consistent with recent reports 12,24. In the study conducted by Reyes-Castillo et al., three SNPs were tested: PADI4_89 G > A (rs11203366), PADI4_90 T > C (rs11203367) and PADI4_92 G > C (rs874881). They found no association between the susceptibility haplotype GTG and the levels of anti-PAD4 antibodies; however, carriers of the susceptibility haplotype demonstrated higher ACPA levels. On the other hand, a study conducted by Harris et al. 31 demonstrated an association between a susceptible haplotype and anti-PAD4 levels with an odds ratio (OR) of 2.59. They genotyped the same three SNPs as Reyes-Castillo et al. Furthermore, when diplotype analysis was applied, carriers of heterozygous genotypes, including both nonsusceptible and susceptible haplotypes, had increased antibody levels compared to patients homozygous in haplotype 1, with an OR of 4.02. It must be emphasized that the mean duration of the disease in RA patients (overall) was longer than in the study conducted by Reyes-Castillo et al. and was 12.5 years. In addition, a study by Guderud et al. 12 was focused on the relationship between autoantibody level and genetic factors. They found no association between PADI4 rs2240340 and rs1635579 and anti-PAD4 autoantibodies. On the other hand, Guderud et al. 12 demonstrated that ACPA- negative patients vs. healthy controls showed a weak association with RA morbidity in relation to the two PADI4 SNPs: rs2240340 (PADI4_94) and rs1635579. Both polymorphisms mentioned turned out to be associated with double negative patients with RA (ACPA negative and anti-PAD4 negative). The authors suggest that genetic risk factors should only be evaluated in relation to ACPA status, as also suggested in a previous study 24. In this study, we did not confirm the association between SNPs and ACPA, anti-PAD4 and RF antibody levels. Current diagnostic criteria include serological markers, such as RF and ACPA 2. ACPA shows positivity in the range of 60–80% of RA with a specificity in the range 90–95% 32. In our cohort, less than 15% of patients with RA were ACPA negative. As previously reported, ACPA has been associated with a higher risk of developing RA in healthy individuals and may occur before the clinical symptoms of RA. ACPA-positive RA patients have a less favorable prognosis, including more complicated structural damage and a worse response to therapy. In the present study, no relationship was found between ACPA concentration and genotypes, also in different genetic models.

We did not confirm the hypothesis that carriers of the genetic variants may have a different (higher) autoantibody status and therefore a less favorable clinical outcome, as previously suggested. The limitation of the study was the small size of the group, notably in the context of genotyping. Further investigations in a larger cohort of patients are necessary to confirm our data, especially the conclusions regarding ACPA-negative patients should be confirmed in the wider population. We believe that in relation to established RA, genetic testing, especially including common SNPs, may not be helpful. It seems to be helpful to measure and compare the anti-PAD4 levels with genotypes during the RA diagnosis or in pre-RA.

Methods

A total of 147 subjects, 122 patients with RA, 52 ± 12.3 aged, 84.4% women and 25 healthy controls, 53 ± 8.4 aged, 72% women were enrolled in the study. Written informed consent was obtained from every participant before entering the study and all research was performed in accordance with relevant guidelines. The study was conducted in accordance with the Declaration of Helsinki and the Ethics Committee of the Medical University of Lublin approved the study (protocol number KE-0254/7/2016). The characteristics of the individuals are presented in Table 1.

Table 1 Characteristics of the subjects.
Full size table

Whole blood was collected and stored at -80 °C until analysis. DNA was extracted from 200 µl of sample according to the manufacturer’s protocol using the GeneMATRIX Quick Blood DNA Purification Kit (Eurx, Poland).

Genotypes of PADI4_94 with rs2240340 (assay ID: C__16176717_10, Thermo Fisher Scientific, USA) and PADI4_104 with rs1748033 (assay ID: C___7541083_1, Thermo Fisher Scientific, USA) were evaluated using TaqMan Genotyping assays and the Endpoint Genotyping module of the COBAS z480 Real-Time PCR System (LightCycler 480 SW, version 1.5.1.62 SP2–UDF v.2.0.0, Roche, Germany). Antibodies levels were previously evaluated and the methodology was described 33. Briefly, ACPA and RF cut-offs were estimated according to manufacturer’s recommendations. Anti-PAD4 positivity was evaluated based on results in a control group as below 95th percentile. Additionally for PAD4 a receiver operating characteristic (ROC) curve with cut-off value (Supplementary Figure S1) was performed to confirm previously determined threshold. Both cut-offs are slightly different (615.24 vs. 628.5 U/mL), however it does not affect the results.

Quantitative values were presented as mean ± standard deviation (SD) or median [interquartile range]. Differences between two independent groups were compared using the Student's t test or the Mann–Whitney U test. Differences between more than two groups were evaluated using the Kruskal–Wallis ANOVA test. The qualitative parameters are given as numbers with percentages and were evaluated using the contingency tables with the χ2 test with Yates’s correction. A p-value of less than 0.05 was considered statistically significant. Logistic regression was used to estimate the odds ratio (OR) and confidence interval (CI). The analysis was performed with STATISTICA version 13 (Dell Inc. 2016). The ROC curve analysis was performed by MedCalc version 20.111 (MedCalc Software Ltd, Ostend, Belgium).