Association of the genetic variants in the endoplasmic reticulum aminopeptidase 2 gene with ankylosing spondylitis susceptibility

Genetic polymorphisms in the endoplasmic reticulum aminopeptidase gene ERAP2 has been attributed with the etiopathogenesis of ankylosing spondylitis (AS). Here we assessed the association of ERAP2 gene single nucleotide polymorphisms (SNPs) with AS predisposition in Iranian patients and determined their effect on the inflammatory state of the patients.


| INTRODUC TI ON
Ankylosing spondylitis (AS) is a chronic autoimmune disease and is characterized by involvement of spine and sacroiliac joints, which causes spinal deformities, increased disability, and mortality. 1 A bulk of research has suggested a significant role for genetic variations in the etiology and pathogenesis of AS, 2,3 in spite of the remarkable involvement of environmental factors as well as aberrant epigenetic regulations. [4][5][6][7][8] The pathogenesis of AS is complicated, and earlier research concentrated on the misfolding of the human leukocyte antigen (HLA)-B27 molecule in disease susceptibility; however, genetic studies have proposed that HLA-B27 accounts for a small part of the overall AS risk. 9 Epidemiological investigations has shown that the HLA-B27 gene is carried by almost 90% of AS patients, whereas only 1%-5% of individuals carrying the HLA-B27 gene will be affected with AS in the future. 10 These observations imply the involvement of non-HLA genes in AS risk.
Endoplasmic reticulum aminopeptidase (ERAP) 2 is an enzyme that belongs to the zinc-containing metallopeptidase family and the corresponding gene is located on the chromosome 5q15. This enzyme is found within the endoplasmic reticulum, which is involved in priming peptides during the antigen presentation pathway via the major histocompatibility complex class I. 11 In comparison to ERAP1, data are lacking about the attribution of ERAP2 polymorphisms with AS predisposition. 4,12,13 A number of genetic polymorphisms in the ERAP2 gene have been attributed to produce alterations in the protein structure and function. The AS-protective ERAP2 gene single nucleotide polymorphism (SNP) rs2248374 14 alters the splicing site in the exon 10, leading to synthesis of a lengthy exon 10 transcript. 15 As a loss of enzyme polymorphism, rs2248374 causes no protein expression of ERAP2 and has been attributed with downregulation of major histocompatibility complex class I molecule levels on the cell surface. 15 Another variant is rs2549782 SNP, which confers a modulation in the specificity as well as functional velocity of the enzymatic activity of ERAP2. 14,16 ERAP2 gene rs2549782 SNP shows a linkage disequilibrium (LD) with other ERAP2 SNPs, including rs2548538, rs2287988, rs1056893, and rs2248374, which are marker SNPs that constitute A and B haplotypes that are associated with protein expression of ERAP2. 15 In addition, ERAP2 gene rs17408150 leads to substitution of a T with an A at codon 669 (p.Leu-669Gln), which alters the leucine residue to glutamine, producing a significant effect on the ERAP2 enzyme function. 17 Association of ERAP1 SNPs with AS susceptibility in Iranian patients has already been reported in our previous studies. [18][19][20][21] Furthermore, we recently indicated the association of ERAP2 gene SNPs with AS susceptibility in HLA-B27-positive individuals. 22 With respect to the involvement of genetic variations in the alteration of ERAP2, it seems that evaluation of such SNPs is worthwhile.
Hence, this study aims to determine the associations of ERAP2 gene rs2548538, rs2287988, and rs17408150 SNPs, for the first time to the best of our knowledge, in an Iranian AS population. Furthermore, the possible role of these SNPs was investigated in the control of inflammatory and immunomodulatory mediators in AS.

| Study participants
In this investigation, 250 individuals with AS and 250 persons as healthy controls were included ( Nonetheless, rs2287988 and rs17408150 SNPs showed statistically significant association with susceptibility to the disease in those AS patients who were positive for human leukocyte antigen (HLA)-B27. Transcriptional level and serum concentration of IL-17A and IL-23 were higher, but those of IL-10 were lower in both AS patients and the HLA-B27-positive patient group relative to the control group. Nevertheless, ERAP2 gene SNPs in the HLA-B27-positive AS patients did not affect the transcription level and serum concentration of cytokines.
Conclusions: ERAP2 gene rs2287988 and rs17408150 SNPs are associated with susceptibility to AS, but they are probably not determining the levels of IL-17A, IL-23, and IL-10 in this disease.

| PBMC separation, RNA isolation, and cDNA synthesis
Peripheral blood mononuclear cells (PBMCs) were isolated from the peripheral blood using Ficoll/Hypaque (Lymphodex; Inno-Train, Kronbergim-Taunus, Germany) density-gradient centrifugation. Extraction of RNA from PBMCs was carried out using the Trizol total RNA extraction kit (GeneAll, Seoul, Korea) based on the producer's guidelines.
Synthesis of the complementary DNA (cDNA) from the extracted RNA samples was conducted using the BioFact™ RT Series cDNA Synthesis Kit (Daejeon, Korea), conforming to the company's protocols.

| Quantitative real-time PCR
In order to assess the mRNA expression levels of inflammatory and im-  Table S1. We randomly selected 80 AS patients and

| Serum concentration of cytokines
Like quantitative mRNA expression analyses, serum samples from the venous blood of the same individials were used to determine the concentration of the cytokines. Hence, the enzyme linked immunosorbent assay (ELISA) technique was applied to determine the serum concentrations of IL-23, IL-17A, IL-10, and TGFβ in 80 patients and 80 controls. The optical density was determined using commercial kit (Invitrogen, Thermo Fisher Scientific, Carlsbad, CA, USA) and determination of the optical density in each well by an ELISA reader (Tecan Spectra, Zürich, Switzerland).

| Statistical analysis
The baseline features of the patient and control groups were determined by descriptive statistical analysis. Determination of the

| Baseline features of the participants
The baseline and laboratory indexes of the study population are listed with details in the

| Association test of ERAP2 polymorphisms
The frequency distribution pattern of the genotypes for ERAP2 gene rs2548538 (P = 0.830), rs2287988 (P = 0.443), and rs17408150 (P = 0.165) SNPs in the healthy control group adhered to the Hardy-Weinberg equilibrium. In the total population analysis, results revealed that none of the three SNPs was significantly associated with risk of AS (Table 2). Notwithstanding this, when the HLA-B27-positive AS patients and the control group were compared, rs2287988 and rs17408150 SNPs showed statistically significant association with AS risk (Table 3). In the case of rs17408150, the C allele had a significant association with higher risk of AS in the HLA-B27-positive patients (OR

| Frequency of the haplotype
Regarding the haplotypic analysis (rs2548538 A/T, rs2287988 A/G, and rs17408150 A/T), four haplotypes had significant

F I G U R E 1
The linkage disequilibrium (LD) block structure under the ERAP2 gene rs2548538, rs2287988, and rs17408150 single nucleotide polymorphisms (SNPs). Two upper captures present the LD scores in the total population, whereas the AS patients positive for HLA-B27 and all healthy controls included in the analysis are illustrated in the lower captures. Inside each block, the scores from 0% to 100% imply the value of Dʹ or r 2 for SNP pairs. As the scores increase, the possibility of two SNP pairs being inherited simultaneously is higher  (Table 4).

| Linkage disequilibrium test
The structure of SNP pairs in the LD block according to the SNP sequence of rs2548538, rs2287988, and rs17408150 are shown in Figure 1. The Dʹ value between rs2548538 and rs2287988 SNPs was 55% and between rs2548538 and rs17408150 SNPs was 81%.
Nonetheless, a partially stronger linkage was detected when only the HLA-B27-positive patients were included in the analysis (Dʹ = 56% and 87%, respectively, for rs2548538-rs2287988 and rs2548538-rs17408150). However, the r 2 values indicated no remarkable LD between SNP pairs.

| mRNA expression of cytokines
The  Figure 2D). The mRNA expression of all four cytokines had no significant difference among AS patients with three genotypes for rs2287988 and rs17408150 SNPs ( Figure 2 and Table S2).
Transcription levels of IL-17A (FC 3.00, P = 0.0001, Figure 3A) and IL-23 (FC 1.70, P = 0.0084, Figure 3B Figure 3D). The mRNA expression of all cytokines had no significant difference among the AS patients positive for HLA-B27 carrying three different genotypes for rs2287988 and rs17408150 SNPs (Figure 3 and Table S2).

| Serum concentration of cytokines
The serum concentrations of IL-17A (46.50 ± 13.10 vs 19.56 ± 6.48, Figure 4A) and IL-23 (388.9 ± 38.78 vs 200.2 ± 24.88, Figure 4B) were increased in the AS group relative to the normal control group. However, the serum concentration of IL-10 was significantly lower in the AS group compared with the control group (1.89 ± 0.18 vs 3.68 ± 1.10, P = 0.0015, Figure 4C). The difference in the serum concentration of TGFβ was not statistically significant between AS group and the control group (20.4 ± 4.51 vs 23.4 ± 4.21, P > 0.05, Figure 4D). The serum concentrations of all cytokines had no significant difference among the AS cases positive for HLA-B27 and harboring three different genotypes for rs2287988 and rs17408150 SNPs ( Figure 4 and Table S3).
The serum concentrations of IL-17A (49.11 ± 14.21 vs 19.56 ± 6.48, P = 0.0001, Figure 5A) and IL-23 (395.4 ± 39.7 vs 200.2 ± 24.88, P = 0.0001, Figure 5B) were higher in the AS patients positive for HLA-B27 relative to normal controls. In contrast, the serum concentration of IL-10 was significantly lower in the HLA-B27-positive AS group compared with the control group (1.75 ± 0.14 vs 3.68 ± 1.10, P = 0.0017, Figure 5C). However, the difference in the serum concentration of TGFβ was not statistically significant between HLA-B27-positive AS cases and the control group (19.7 ± 4.47 vs 23.4 ± 4.21, P > 0.05, Figure 5D). The serum concentration of all cytokines had no significant difference among the AS cases positive for HLA-B27 carrying three different genotypes for rs2287988 and rs17408150 SNPs ( Figure 5 and Table S3).

| Association of the genotypes and clinical manifestations
The ERAP2 gene rs2548538, rs2287988, and rs17408150 polymor-

| D ISCUSS I ON
To date, large-scale analyses have found over 60 genetic loci for AS risk, even though most investigations have assigned a large proportion of genetic-related risk of AS to the HLA-B27 gene. 3 Association of the ERAP2 gene with AS was detected after identification of ERAP1 gene association with AS. 12 Both ERAP1 and ERAP2 genes are found on chromosome 5; they are structured in an inverse direction and share a common intergenic sequence. 28 Studies have found a haplotype in association with AS that is constituted with both ERAP1 and ERAP2 polymorphisms. 29 According to studies, the association of ERAP2 with AS seems to be independent of HLA-B27. 14,33 ERAP2 has been reported to impress directly the B*2705 peptidome, removing a number of ligands containing N-terminal basic residues, leading to promoted levels of nonamers by the increased activity of ERAP1. 34 Based on the level of ERAP1 trimming activity, the influences of ERAP2 on the B27 peptidome could be altered. 35 A study indicated that the ERAP2 gene rs2248374 SNP was particularly associated with risk of psoriatic arthritis in individuals who are negative for HLA-B27. 36 We observed that ERAP2 gene rs2287988 and rs17408150 polymorphisms were associated with increased risk of the disease in AS patients positive for HLA-B27. In addition, we previously reported association between ERAP2 gene rs2910686 SNP and AS risk in patients positive for HLA-B27. 22 Further evidence is required to conclusively determine the ERAP2 involvement in collaboration with HLA-B27 in AS pathogenesis.
The overall effect of ERAP2 on the conformation of HLA-B27 has not been fully determined. A study reported that ERAP2 expression did not significantly impress the expression of the folded as well as unfolded HLA-B27 molecules, markers of endoplasmic reticulum stress, and the expression of proinflammatory cytokines. 37 In contrast, a study demonstrated that lack of ERAP2 SNPs with AS risk, these ERAP2 polymorphisms might not be involved in the alteration of HLA-B27 and, hence, the inflammatory settings in the AS disease.

ACK N OWLED G M ENTS
The authors are grateful to the patients and the healthy individuals for their participation in the study.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest to report.  Abbreviations: AS, ankylosing spondylitis; ASQoL, ankylosing spondylitis quality of life; BASDAI, Bath ankylosing spondylitis disease activity index; BASFI, Bath ankylosing spondylitis functional index; BASG, Bath ankylosing spondylitis global score; CRP, C-reactive protein.

AUTH O R CO NTR I B UTI O N S
financial support, participated in manuscript preparation, and read the manuscript critically.

E TH I C S A PPROVA L A N D CO N S E NT TO PA RTI CI PATE
Approval of the study protocol was received from the local ethical review committee in Alborz University of Medical Sciences Written informed consent forms were obtained from patients and healthy controls before blood taking.