Functional ERAP1 Variants Distinctively Associate with Ankylosing Spondylitis Susceptibility under the Influence of HLA-B27 in Taiwanese

Epistasis of ERAP1 single nucleotide variations (SNVs) and HLA-B27 has been linked to ankylosing spondylitis susceptibility (AS). The current study examined how prevalent ERAP1 allelic variants (SNV haplotypes) in Taiwan affect ERAP1 functions and AS susceptibility in the presence or absence of HLA-B27. Sanger sequencing was used to discover all ERAP1 coding SNVs and common allelic variants in Taiwanese full-length cDNAs from 45 human patients. For the genetic association investigation, TaqMan genotyping assays were utilized to establish the genotypes of ERAP1 SNVs in 863 AS patients and 1438 healthy controls. Ex vivo biological analysis of peripheral blood mononuclear cells from homozygous donors of two common-risk ERAP1 allelic variants was performed. Two common-risk ERAP1 allelic variants were also cloned and functionally studied. In Taiwanese, eleven frequent ERAP1 SNVs and six major ERAP1 allelic variants were discovered. We discovered that in Taiwanese, the most prevalent ERAP1-001 variant with 56E, 127R, 276I, 349M, 528K, 575D, 725R, and 730Q interacting with HLA-B27 significantly contributed to the development of AS. In HLA-B27 negative group, however, the second most prevalent ERAP1-002 variant with 56E, 127P, 276M, 349M, 528R, 575D, 725R, and 730E was substantially related with an increased risk of AS. Ex vivo and in vitro research demonstrated that ERAP1 allelic variants have a significant impact on ERAP1 functions, suggesting that ERAP1 plays a role in the development of AS. In an HLA-B27-dependent manner, common ERAP1 allelic variants are related with AS susceptibility.


Introduction
Ankylosing spondylitis (AS) is a subtype of axial spondylopathy (AxSpA) that primarily affects the axial skeleton and is characterized by the formation of syndesmophytes and spinal ankylosis deformities [1]. The pathophysiology of AS, which has long been thought to be a highly familial and heritable disease [2,3], is complicated by a number of variables. Multiple genes may be involved in the development of AS, based on the heterogeneity of the disease's progression. Several genes and genomic areas have been linked to AS susceptibility and severity in genome-wide association studies (GWAS), demonstrating that both MHC and non-MHC genes have a role in the illness process [4][5][6][7][8]. Despite breakthroughs in recent years, the genetic and pathophysiology processes of AS remain poorly known.
Antigenic peptides on MHC molecules are scanned by the human immune system to discern infected or sick cells. Antigenic peptides degraded in proteasomes are then trimmed to appropriate lengths on the ER luminal side by aminopeptidases for correct presentation on MHC-I molecules [9][10][11][12][13]. Endoplasmic reticulum aminopeptidase 1 (ERAP1), a unique member of the M1 aminopeptidase family, trims antigen peptides to fit the antigen binding site on MHC-I during the final stage of antigen processing. Peptide-MHC-I complexes on the cell surface reflect the existing proteome of respective cells.
The function of ERAP1 has a major impact on the stability and immunological characteristics of MHC-I molecules because of its critical involvement in processing MHC-I ligands. Low or excessive ERAP1 activity could result in antigen peptides that are unsuitable for MHC-I presentation. As a result, effective ERAP1 function is required for optimal antigen presentation and health maintenance. ERAP1 is a non-MHC gene on chromosome 5q15, and genetic studies have shown that epistasis between ERAP1 SNVs and HLA-B27 is highly linked to AS vulnerability in different ethnic groups [6,14]. The biological functions of ERAP1 allelic variants on the other hand, are largely unexplored. The molecular processes underpinning the link between ERAP1 allelic variations and AS are still being investigated [15]. Furthermore, different combinations of ERAP1 SNVs (or allelic variants) may interact with HLA-B in a unique way to contribute to AS vulnerability in a single ethnic population. The breadth of ERAP1 SNVs and allelic variations in Taiwanese was investigated in this study. The ERAP1-001 variation was discovered to be a key risk factor for AS in Taiwanese people. Furthermore, in an HLA-B27-dependent manner, two prevalent ERAP1 allelic variants are strongly related with AS susceptibility.

Study Subjects
A total of 863 AS patients (717 males and 146 females; onset ages ranged from 8 to 72 with the mean onset age of 25.38-year-old) who fulfilled the 1984 revised New York diagnostic criteria for AS were recruited at the Chang Gung Memorial Hospital. HLA-B27 carriers accounted for 801 (92.8%) of the 863 AS patients. As normal healthy controls, a total of 1438 adult healthy blood donors (706 males and 732 females; ages varied from 18 to 65, with the mean age of 40.78-year-old) from the same geographical region were recruited. All human research protocols were approved by the ethics committee of Chang Gung Memorial Hospital, and all subjects gave their informed consent.

HLA-B27 Determination
Flow cytometry and/or PCR tests were used to determine HLA-B27 positive. Whole blood samples were analyzed by flow cytometry for HLA-B27 using the BDTM HLA-B27 kit (BD Biosciences, Franklin Lakes, NJ, USA), which combines fluorescein-conjugated anti-HLA-B27 and phycoerythrin-conjugated anti-CD3 antibodies.

HLA-B Allele Determination
On ABI 3730XL DNA Analyzers, HLA-B alleles were determined using a commercial sequencing-based typing (SBT) kit (TBG Biotechnology Corp., New Taipei City, Taiwan). The AccuType HLA SBT Analysis Software assigned HLA-B alleles (TBG Biotechnology Corp, New Taipei City, Taiwan).

Genomic DNA Isolation
As previously disclosed [16], genomic DNA was recovered from EDTA anticoagulated peripheral blood using the Puregene DNA isolation kit (Gentra Systems, Minneapolis, MN, USA).

RNA Isolation and cDNA Synthesis
TRIzol total RNA isolation reagent was used to isolate total RNA from human peripheral blood leukocytes (Invitrogen, Carlsbad, CA, USA). For deep sequencing investigation of ERAP1 coding SNVs and creation of ERAP1 expression constructs, the SuperScript Preamplification system (Invitrogen, Carlsbad, CA, USA) was utilized to synthesize peripheral blood leukocyte cDNA.

ERPA1 SNV Genotype Analyses
The common ERPA1 SNVs identified in Taiwanese population were genotyped using made-to-order allelic discrimination assays from Applied Biosystems (ABI, Foster City, CA, USA). SNV genotyping assays were carried out on a vi7 real-time PCR system (ABI). The SNV genotypes were determined using ABI TaqMan Genotyper software (Foster City, CA, USA) according to the vendor's instructions.

ERAP1 Expression Constructs
The Xho I/Xba I-flanked Lir-EGFP DNA fragment from pBacCMV-MCS-Lir-EGFP was amplified by PCR using the forward primer 5 -TTCTCGAGTTAAAACAGCCTGTGGGT-3 (the Xho I site is underlined) and the reverse primer 5 -GGCTCTAGATTACTTGTACAGCT CGTCCA-3 (the Xba I site is underlined) and the fragment was subsequently cloned into pcDNA3.1 vector (Invitrogen) at Xho I and Xba I restriction enzyme sites for the generation of pcDNA3.1-Lir-EGFP vector. The full-length ERAP1 cDNAs from ERAP1-001 homozygous donors and ERAP1-002 homozygous donors were amplified using the forward primer 5 -ACGGGAATTCCCACCATGGTGTTTCTGCCCCTCAA-3 (the EcoR I site is underlined) and the reverse primer 5 -AACGGGCTCGAGTCATCCTGTTGCGTCAGCTTCA-3 (the Xho I site is underlined). The ERAP1 cDNA inserts were cloned into pcDNA3.1-Lir-EGFP vector digested with EcoR I and Xho I restriction enzymes.

Peripheral Blood Mononuclear Cell (PBMC) Isolation and Culture
Heparin collection tubes were used to collect whole blood from donors. Centrifugation through FicollPaque Premium (GE Healthcare, Chicago, IL, USA) density gradients was used to isolate PBMCs. PBMCs were cultured for 48 h in RPMI-1640 with 10% FBS and 10 ng/mL phorbol 12-myristate 13-acetate (PMA) before being stimulated for 16 h with LPS (1 µg/mL). The supernatants of PBMC cultures were then collected for IL-23 analysis, and cells were extracted for Western blot analysis.

Transfection of CHO and U937 Cells
TransIT transfection reagent (Mirus, Madison, WI, USA) was used for transfection of ERAP1 expression constructs into cells. On 6-well plates, 2 × 10 6 cells were seeded in each well. Before transfection, the cells were washed twice with medium devoid of serum. ERAP1 expression construct plasmid DNA (2.5 g) was diluted in 400 µL serum-free DMEM medium and then combined with TransIT transfection reagent (5 µL) to generate DNA-TransIT complexes. The DNA-TransIT complexes were added to cells in 400 µL of serum-free media per well. After 5 h of incubation, the medium was discarded and replaced with 2 mL of new medium containing 10% fetal bovine serum. After 24 h of transfection, the efficacy of transfection was evaluated by fluorescence microscopy.

Western Blot Analysis
On a micro Protein III system, cell lysates were subjected to SDS-PAGE for protein separation (Bio-Rad, Hercules, CA, USA). Proteins on the gel were electrotransferred onto a polyvinylidene difluoride (PVDF) membrane following SDS-PAGE. Tris-buffered saline (TTBS; 100 mmol/L Tris (pH 7.4), 100 mmol/L NaCl, and 0.1 percent Tween 20) containing 5% (v/v) non-fat milk was used to block PVDF membrane at room temperature for 1 h with moderate shaking. The membrane was then shaken overnight at 4 • C with primary antibody diluted in TBS containing 0.5 percent (v/v) nonfat milk. The membrane was washed three times for ten minutes at room temperature with TBS solution to remove unbound antibodies. The membrane was then treated at room temperature for one hour with secondary antibodies coupled with horseradish peroxidase (HRP). Enhanced Chemiluminescence kit (Pierce, Rockford, IL, USA) and the imaging equipment were used to detect the HRP-labeled proteins on membrane (UVP ChemStudio PLUS Touch, Analytik Jena US LLC, Cambridge, UK). Band intensities were quantified from the digital image by densitometry using ImageJ software (NIH, Bethesda, MD, USA).

ERAP1 Protein Immunoprecipitation
Cell pellets (25 × 10 6 cells for lymphoblastoid B cells or 2 × 10 6 cells for PBMCs) were lysed on ice for 30 min in lysis buffer (M-PER™ Mammalian Protein Extraction Reagent, Thermo Scientific, Waltham, MA, USA). The lysate supernatants were extracted by centrifuging at 14,000 rpm for 10 min at 4 degrees Celsius. Using the BCA (bicinchoninic acid) technique, protein concentrations were determined prior to immunoprecipitation. The lysates were then precleared by incubation with glycine Sepharose beads prior to treatment with anti-ERAP1 4D2-mAb-conjugated Sepharose beads for at least 2 h. The beads were subsequently washed three times in buffer (50 mM Tris, 150 mM NaCl, pH 7.4) and resuspended in buffer (500 µL) prior to use in the activity experiment.

ERAP1 Enzyme Activity Determination
The effect of ERAP1 allelic variants on ERAP1 enzyme activity was analyzed using a continuous fluorigenic assay as described [17]. Enzymatic reactions were carried out at 37 • C. In all reactions, the substrate concentration exceeded the enzyme concentration by at least 100-fold. Typically, 10 to 1000 nM ERAP1 protein was combined with 1 to 100 µM peptide, and the fluorescence of the reaction was monitored. On a QuantaMaster 4 spectrofluorimeter (Photon Technology International, Birmingham, NJ, USA) and a Tecan SpectraFluor plate reader, fluorescence measurements were conducted. Excitation at 295 nm and emission at 350 nm were used to acquire fluorescence spectra.

Cytokine Determination
ELISA kits were obtained from BioLegend (San Diego, CA, USA). All assays were conducted according to the manufacturer's instructions. Absorbencies were measured using MAXlineMicroplate Readers at 450 nm (Molecular devices, San Jose, CA, USA). The data are presented as the mean ± standard deviation of technical replicates.

Statistical Analysis
Linear regression analysis and Pearson's coefficient of correlation were used to analyze the correlation between different parameters. The distributions of SNV alleles and genotypes in patients and controls were analyzed using single-locus analysis. These comparisons were made using three chi-square tests (the genotype case-control test, the allele case-control test, and the Cochran-Armitage trend test). Using the SAS/Genetics software package release 8.2 (SAS Institute, Cary, NC, USA), significant relationships of SNVs with characteristics (p < 0.05) were detected, and p-values, odds ratios (ORs), and 95 percent confidence intervals (CIs) were calculated. The reported probability of false-discovery rate (PFDR) was estimated using Q-VALUE software (http://genomics.princeton.edu/storeylab/qvalue accessed on 6 June 2022) to account for multiple simultaneous statistical comparisons. Haploview 4.2 (Broad Institute, Cambridge, MA, USA; http://www.broad.mit.edu/mpg/ haploview accessed on 6 June 2022) was used to calculate linkage disequilibrium (LD) between marker loci and to create haplotype blocks. We employed illness status (case vs. control) and HLA-B27 positivity to test for haplotype-trait association for markers within the same haplotype block (or allelic variations). To study genetic correlations, researchers utilized stepwise logistic regression analysis to account for sex. For all analyses, a 5% level of significance was employed for p-values.

Associations of ERAP1 SNVs with AS Susceptibility
By sequencing the full-length ERAP1 cDNAs from 45 Taiwanese subjects randomly selected for SNV discovery purpose, we identified nine common ERAP1 coding SNVs (cSNVs) including eight non-synonymous cSNVs that cause amino acid substitutions (Table S1). We subsequently carried out genotype analyses of nine ERAP1 cSNVs plus two ERAP1 intron SNVs that were previously reported to associate with AS susceptibility [18,19].

ERAP1 Variant Genotypes Affect Phenotypes
Taiwanese ERAP1 variations have a number of amino acid changes (Table S1). To examine the epistaxis effects if ERAP1 haplotypes influence phenotypes in HLA-B27 positive individuals, peripheral blood mononuclear cells (PBMCs) were extracted from 11 human participants (seven ERAP1-001 homozygous AS patients and four ERAP1-002 AS patients) having a single HLA-B2704 allele. As demonstrated in Figure 1, PBMCs from ERAP1-001 homozygous donors produced more ERAP1 protein, MHC class I free heavy chain (FHCs) and FHC dimers, unfolded protein response (UPR) markers (immunoglobulin heavy-chain binding protein (BiP), CCAAT-enhancer-binding protein homologous protein (CHOP) and X-box-binding protein 1 (XBP1), autophagy markers (Beclin-1, LC3 I and LC3 II), and inflammation markers (Caspase 1 and IL-1β) as compared to those from ERAP1-002 homozygous donors. Notably, PBMCs of seven ERAP1-001 homozygous donors produced significantly higher levels of IL-23 p19 and p40 than PBMCs from four ERAP1-002 homozygous donors. chain (FHCs) and FHC dimers, unfolded protein response (UPR) markers (immunoglobulin heavy-chain binding protein (BiP), CCAAT-enhancer-binding protein homologous protein (CHOP) and X-box-binding protein 1 (XBP1), autophagy markers (Beclin-1, LC3 I and LC3 II), and inflammation markers (Caspase 1 and IL-1β) as compared to those from ERAP1-002 homozygous donors. Notably, PBMCs of seven ERAP1-001 homozygous donors produced significantly higher levels of IL-23 p19 and p40 than PBMCs from four ERAP1-002 homozygous donors. Previous studies indicates that non-conservative ERAP1 SNVs, such as K528R and R725Q substitutions, may impact ERAP1 protein production and aminopeptidase activity [14]. We conducted in vitro experiments to evaluate the effects of ERAP1 variations on ERAP1 expression and aminopeptidase activity. The variant cDNAs of ERAP1-001 and Previous studies indicates that non-conservative ERAP1 SNVs, such as K528R and R725Q substitutions, may impact ERAP1 protein production and aminopeptidase activity [14]. We conducted in vitro experiments to evaluate the effects of ERAP1 variations on ERAP1 expression and aminopeptidase activity. The variant cDNAs of ERAP1-001 and ERAP1-002 were cloned into the bi-cistronic mammalian expression vector pcDNA-Lir-EGFP. Fluorescence microscopy was used to measure the transfection effectiveness of EGFP in CHO cells (Figure 2A). Western blot analysis demonstrated that ERAP1 expression in transfected CHO cells was efficient ( Figure 2B). Moreover, cells transfected with the ERAP1-001 variant produced a much greater amount of ERAP1 protein than those transfected with the ERAP1-002 variant ( Figure 2C), corresponding with the results obtained from the PBMCs of homozygous ERAP1-001 and ERAP1-002 human individuals ( Figure 1C). In addition, CHO cells expressing ERAP1-001 had considerably higher enzyme activity than ERAP1-002-expressing cells (p = 0.0105). ( Figure 2D). Our research demonstrated that ERAP1-001 exhibited considerably more ERAP1 protein expression and enzyme activity than ERAP1-002.

ERAP1 Allelic Variants Affect the Crosstalk of ER Stress (UPR), Autophagy, and Inflammation
Recent evidence suggests that the axis of ER stress, autophagy, and inflammasome plays a crucial role in the etiology of inflammatory illness. U937 cells were transfected with ERAP1 expression constructs in order to determine whether ERAP1 variations

ERAP1 Allelic Variants Affect the Crosstalk of ER Stress (UPR), Autophagy, and Inflammation
Recent evidence suggests that the axis of ER stress, autophagy, and inflammasome plays a crucial role in the etiology of inflammatory illness. U937 cells were transfected with ERAP1 expression constructs in order to determine whether ERAP1 variations impact the interaction between ER stress, autophagy, and inflammation. U937 cells transfected with ERAP1-001 displayed significantly higher levels of HLA-B27, ERAP1 protein, FHCs, FHC dimers, UPR markers (BiP, CHOP and XBP1), and inflammatory markers (Caspase 1 and IL-1) than U937 cells transfected with ERAP1-002. Moreover, U937 cells transfected with ERAP1-001 produced larger quantities of IL-23 in culture supernatants than U937 cells transfected with ERAP1-002. On the other hand, autophagy markers including Beclin 1 (I), LC3 I (J), LC3 II (K), caspase 1 (L), and IL-1β (M) were not significantly different between ERAP1-001 and ERAP1-002 cells.

Discussion
As a non-MHC susceptibility gene, ERAP1 was found to interact with HLA-B27 to increase the risk of AS [4,14]. The pathophysiological mechanism behind the impact of ERAP1 allelic variations on AS susceptibility, on the other hand, is unknown. The common ERAP1-001 allelic variant was discovered to be a significant risk factor for AS in Taiwanese in the current investigation. According to our findings, epistasis between ERAP1 allelic variants and HLA-B plays a crucial role in the development of AS in Taiwanese. ERAP1-001 induces a higher level of ERAP1 protein expression, enzyme activity, and IL-23 production than ERAP1-002. Increased ERAP1 enzyme activity appears to be a pathophysiological mechanism for the development of AS in HLA-B27 + individuals; therefore, inhibition of ERAP1 enzyme activity may be a viable treatment option for HLA-B27 + AS patients. In HLA-B27 − people, however, ERAP1-002 is a risk factor for AS. The importance of epistasis between ERAP1 variants and HLA-B27 + / HLA-B27 − in the development of AS in Taiwanese is demonstrated.
ERAP1 is a multifunctional zinc-metallopeptidase that trims antigen peptides to the proper length for MHC-I molecules to display them. Antigenic epitopes are both produced and destroyed by ERAP1. The length of the peptidome is determined by interactions between the N-terminus of the antigen peptide and the enzyme active site of ERAP1, as well as the C-terminus of the peptide and an ERAP1 regulatory site. By changing the sequence and length of antigenic peptides deposited onto the corresponding MHC Class I molecules, ERAP1 affects a major portion of the immunopeptidome [20]. ERAP1 trims peptide precursors in solution to adjust the immunopeptidome's optimal fit [21].
Peptidome studies of MHC-bound peptides in HLA-B27 positive cell lines expressing various ERAP1 variants reveal that ERAP1 and HLA-B27 have a functional relationship [22]. The structure, immunological effect, and peptide-trimming property of ERAP1 point to a mechanism or pathway for AS vulnerability that involves HLA-B27 [23,24]. ERAP1 directly affects peptide binding and presentation by HLA-B27, which could be a pathogenic mechanism for ERAP1's participation in AS [25]. According to in vitro peptide catalysis experiments, ERAP1 risk alleles with higher catalytic activity degrade HLA-B27 epitopes more efficiently, resulting in lower HLA-B27 presentation of the identical peptides [26]. mRNA stability, protein translation, and enzymatic activity may all be affected by combined effects of ERAP1 genetic variations [27,28]. Between ERAP-001 (127R, 276I, 528K, and 730Q) and ERAP-002 (R127P, I276M, K528R, and Q730E), there are four residue differences. The molecular basis of ERAP1 SNVs on N-terminal peptidase function was discovered using crystal structure research [29]. The ERAP1 SNV rs30187A > G (K528R) is found near the substrate pocket's entrance, which may affect the C-terminus stability and aminopeptidase activity [29]. The R127P and K528R alterations affect ERAP1 conformational transition, whereas the Q730E change in ERAP1's peptide-binding pocket may affect the peptides processed [30]. ERAP1 with 528R has lower enzymatic activity than ERAP1 with 528K, while ERAP1 with Q730E influences the lengths of antigenic peptides produced [22,25,28,[31][32][33]. Therefore, the enzyme activity of ERAP1-002 (127P, 276M, 528R, and 730E) will be reduced. Furthermore, lower surface FHC expression on monocytes and HLA-B27-expressing APCs in AS patients was linked to 528R and 730E. HLA-B27 FHC surface expression was reduced by ERAP1 silencing or inhibition in APCs, as was IL-2 production by KIR3DL2CD3-reporter cells, and Th17 growth and IL-17A secretion by CD4+ T cells from AS patients [34].
Reduced ERAP1 mRNA expression in the presence of a coding variant has been linked to Behçet's Disease (BD) [42]. With longer epitopes, the promiscuous sub-peptidome binding property of HLA-B51 indicates a genetic effect on dysregulated CD8+ T response and aberrant NK cell activation, leading to the development of BD [43]. Other HLA-B alleles have been involved in AS development in European populations [6], therefore ERAP1 epistasis may not be limited to HLA-B27. HLA-B40 increased susceptibility to AS in HLA-B27 − AS patients in Taiwanese [44], but not in Caucasian [45]. As shown in Table S5, HLA-B40:01 frequencies (phenotype frequency = 37.1%, allele frequency = 18.55%) among 62 HLA-B27 − AS patients were similar to those in general Taiwanese population, indicating that HLA-B40 may not play a unique role in AS. HLA-B27 − Taiwanese AS patients have lower ERAP1 enzyme activity, suggesting that decreased ERAP1 enzyme activity may play a role in AS. Understanding the pathophysiology of AS in people will require more research into the link between HLA-B alleles and ERAP1 variations.
Misfolding of HLA-B27 HC promotes ER stress and activation of the UPR. BiP and XBP1 are two UPR indicators that are increased upon the activation of the three major UPR pathways. Compared to cells expressing the ERAP1-002 variant, cells expressing the ERAP1-001 variant produced larger quantities of ERAP1 protein, FHCs, FHC dimers, BiP, XBP1, and IL23. The increased synthesis of BiP and XBP1 in cells expressing ERAP1-001 suggests that the overexpression of MHC class I FHCs and FHC dimers results in the activation of UPR, the elevation of IL-23 levels, and proinflammatory response. The enhanced expression of FHCs and FHC dimers in cells bearing HLA-B27 may explain why the link between ERAP1-001 and AS susceptibility is exclusive to HLA-B27 + people.
Autophagy is a lysosome-dependent degradation mechanism that destroys misfolded proteins and damaged ER. UPR not only enhances cytokine-mediated inflammatory responses leading to disease, but also promotes autophagy. The control of autophagy by reducing ER stress could therefore provide cytoprotection [46]. PBMCs expressing the ERAP1-001 variant displayed significantly higher levels of autophagy markers (Beclin 1, LC3 I, and LC3 II), inflammation markers (caspase 1 and IL-1), and IL-23 (p19 and p40) compared to PBMCs expressing the ERAP1-002 variant (Figure 1). Increased productions of BiP, CHOP, and XBP1 in cells expressing ERAP1-001 suggest that ERAP1-001 may cause the misfolding of HLA-B complexes or HLA-B27 associated with abnormal peptides, resulting in the activation of the UPR [47] and subsequent induction of autophagy (Beclin 1, LC3 I and LC3 II), inflammation (caspase 1 and IL-1β) [48]. Nonetheless, U937 cells transfected with ERAP1-001 displayed comparable levels of autophagy markers (Beclin 1, LC3 I and LC3 II) compared to those with ERAP1-002 ( Figure 3). In this regard, we speculated that endogenous ERAP1 may mask the effect of ERAP1-001 and ERAP1-002 variants on autophagy in U937 cells.

Conclusions
The present investigation definitively determined a functional role of ERAP1 allelic variants In AS. Ex vivo and in vitro research demonstrated that ERAP1 allelic variations have a substantial effect on ERAP1 expression and enzyme activity, FHCs, FHC dimers, and UPR markers (BiP, CHOP and XBP1). Two common ERAP1-001 and ERAP1-002 variants are associated with susceptibility to AS in Taiwanese in an HLA-B27-dependent manner.
Although the exact mechanism underlying IL-23 overexpression remains unknown, IL-23 was detectable in the peripheral blood and tissues of AS patients. In particular, it is unknown if the overproduction of IL-23 could be linked to the misfolding of HLA-B27 by inducing the UPR [49]. On the other hand, the proof that IL-23 is active at the surface of the intestinal mucosa [50] and the evidence that certain types of bacterial stimuli may alter its intestinal expression [51] suggest the role of microbes in the IL-23 overexpression observed in AS [52]. Multiple roles could be carried out by ERAP1 in the extracellular area [53]. ERAP1 produced into the circulation in response to viral or inflammatory stimuli that boost NO production may function as an acute-phase host defense protein [54]. Exosomes contain ERAP1, TNF-α, IFN-γ, and CCL3, which play a crucial role in the inflammatory process by activating macrophages [55]. The immunomodulation cytokine profiles of the IL-17/IL-23 pathway are affected by the ERAP1 SNPs rs30187, rs2287987 [56], and rs27038 [57]. Emerging immunological and genetic evidence validated the central function of the IL-17A/IL-23 axis, resulting in the continued development of new medications targeting distinct components, such as medicines blocking the p40 subunit of IL-12/IL-23 or the p19 subunit of IL-23 for therapeutic purposes. Our results indicate that the genetic relationship between HLA-B27 and ERAP1 is a critical predictor of IL-23 pathogenesis in AS.

Conclusions
The present investigation definitively determined a functional role of ERAP1 allelic variants In AS. Ex vivo and in vitro research demonstrated that ERAP1 allelic variations have a substantial effect on ERAP1 expression and enzyme activity, FHCs, FHC dimers, and UPR markers (BiP, CHOP and XBP1). Two common ERAP1-001 and ERAP1-002 variants are associated with susceptibility to AS in Taiwanese in an HLA-B27-dependent manner.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/cells11152427/s1, Figure S1: Pair-wise LD patterns with D' (left side) and r2 (right side) measures of eleven ERAP1 SNPs for healthy controls (A) and AS cases plus normal (B). Figure S2: ERAP1 allelic variants detected among different populations. Figure S3: The original Western blots of Figures 1A, 2B and 3A. Table S1: ERAP1 SNVs and allelic variants (SNV haplotypes) in Taiwanese healthy controls and AS patients. Table S2: Association of ERAP1 SNVs with AS susceptibility in Taiwanese. Table S3: Association of ERAP1 SNVs with HLA-B27 positivity among AS patients. Table S4: Distributions of ERAP1 allelic variants in HLA-B27 + AS patients (AS B27 + ) and HLA-B27 + healthy controls (Control B27 + ). Table S5  Institutional Review Board Statement: The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Review Board of Chang Gung Memorial Hospital (protocol code 201902162B0C101).

Informed Consent Statement:
Written informed consent has been obtained from the patients to publish this paper.

Data Availability Statement: Not applicable.
Acknowledgments: We greatly appreciate the Shin Chu Blood Donor Center for the collection of samples.

Conflicts of Interest:
The study was carried out in accordance with the relevant guidelines and regulations. The authors declare no conflict of interests and any potential financial conflicts of interest that any of the authors may have.

Abbreviations
AS ankylosing spondylitis; HLA-B27 human leukocyte antigen B27; FHCs MHC class I free heavy chain; UPR unfolded protein response; BiP immunoglobulin heavy-chain binding protein; CHOP CCAAT-enhancer-binding protein homologous protein and XBP1 X-box-binding protein 1.