ORIGINAL RESEARCH published: 29 November 2019 doi: 10.3389/fimmu.2019.02755 Association of Functional Polymorphisms of KIR3DL1/DS1 With Behçet’s Disease Ángel Castaño-Núñez 1 , Marco-Antonio Montes-Cano 1 , José-Raúl García-Lozano 1 , Norberto Ortego-Centeno 2 , Francisco-José García-Hernández 3 , Gerard Espinosa 4 , Genaro Graña-Gil 5 , Juan Sánchez-Bursón 6 , María-Rosa Juliá 7 , Roser Solans 8 , Ricardo Blanco 9 , Ana-Celia Barnosi-Marín 10 , Ricardo Gómez de la Torre 11 , Patricia Fanlo 12 , Mónica Rodríguez-Carballeira 13 , Luis Rodríguez-Rodríguez 14 , Teresa Camps 15 , Santos Castañeda 16 , Juan-Jose Alegre-Sancho 17 , Javier Martín 18 and María-Francisca González-Escribano 1* Edited by: Kutty Selva Nandakumar, 1 Department of Immunology, Hospital Universitario Virgen del Rocío, Instituto de Biomedicina de Sevilla (IBiS), CSIC, Southern Medical University, China Universidad de Sevilla, Seville, Spain, 2 Department of Internal Medicine, Hospital Clínico San Cecilio, Granada, Spain, 3 Department of Internal Medicine, Hospital Universitario Virgen del Rocío, Seville, Spain, 4 Department Autoimmune Reviewed by: Diseases, Hospital Universitari Clínic, Barcelona, Spain, 5 Department of Rheumatology, Complejo Hospitalario Universitario Piotr Kusnierczyk, A Coruña, A Coruña, Spain, 6 Department of Rheumatology, Hospital Universitario de Valme, Seville, Spain, 7 Department of Ludwik Hirszfeld Institute of Immunology, Hospital Universitari Son Espases, Palma de Mallorca, Spain, 8 Department of Internal Medicine, Autoimmune Immunology and Experimental Systemic Diseases Unit, Hospital Vall d’Hebron, Universidad Autonoma de Barcelona, Barcelona, Spain, 9 Department of Therapy (PAN), Poland Rheumatology, Hospital Universitario Marqués de Valdecilla, Santander, Spain, 10 Department of Internal Medicine, Complejo Mirko Spiroski, Hospitalario Torrecárdenas, Almería, Spain, 11 Department of Internal Medicine, Hospital Universitario Central de Asturias, Scientific Foundation SPIROSKI, Asturias, Spain, 12 Department of Internal Medicine, Hospital Virgen del Camino, Pamplona, Spain, 13 Deparment of Internal North Macedonia Medicine, Hospital Universitari Mútua Terrassa, Terrassa, Spain, 14 Department of Rheumatology, Hospital Clínico San Carlos, *Correspondence: Madrid, Spain, 15 Department of Internal Medicine, Hospital Regional Universitario de Málaga, Málaga, Spain, 16 Department María-Francisca González-Escribano of Rheumatology, Hospital de la Princesa, IIS-Princesa, Madrid, Spain, 17 Department of Rheumatology, Hospital Universitario mariaf.gonzalez.sspa@ Doctor Peset, Valencia, Spain, 18 Instituto de Parasitología y Biomedicina “López-Neyra”, CSIC, PTS Granada, Granada, juntadeandalucia.es Spain Specialty section: Behçet’s disease (BD) is an immune-mediated vasculitis related to imbalances between This article was submitted to Autoimmune and Autoinflammatory the innate and adaptive immune response. Infectious agents or environmental factors Disorders, may trigger the disease in genetically predisposed individuals. HLA-B51 is the genetic a section of the journal Frontiers in Immunology factor stronger associated with the disease, although the bases of this association remain elusive. NK cells have also been implicated in the etiopathogenesis of BD. A family Received: 19 August 2019 Accepted: 11 November 2019 of NK receptors, Killer-cell Immunoglobulin-like Receptor (KIR), with a very complex Published: 29 November 2019 organization, is very important in the education and control of the NK cells by the union Citation: to their ligands, most of them, HLA class I molecules. This study aimed to investigate the Castaño-Núñez Á, Montes-Cano M-A, contribution of certain KIR functional polymorphisms to the susceptibility to BD. A total García-Lozano J-R, Ortego-Centeno N, of 466 BD patients and 444 healthy individuals were genotyped in HLA class I (A, B, and García-Hernández F-J, Espinosa G, C). The set of KIR genes and the functional variants of KIR3DL1/DS1 and KIR2DS4 were Graña-Gil G, Sánchez-Bursón J, also determined. Frequency of KIR3DL1∗ 004 was lower in patients than in controls (0.15 Juliá M-R, Solans R, Blanco R, Barnosi-Marín A-C, Gómez de la vs. 0.20, P = 0.005, Pc = 0.015; OR = 0.70; 95% CI 0.54–0.90) in both B51 positive Torre R, Fanlo P, and negative individuals. KIR3DL1∗ 004, which encodes a misfolded protein, is included Rodríguez-Carballeira M, Rodríguez-Rodríguez L, Camps T, in a common telomeric haplotype with only one functional KIR gene, KIR3DL2. Both, Castañeda S, Alegre-Sancho J-J, KIR3DL1 and KIR3DL2 sense pathogen-associated molecular patterns but they have Martín J and González-Escribano M-F different capacities to eliminate them. The education of the NK cells depending on the (2019) Association of Functional Polymorphisms of KIR3DL1/DS1 With HLA, the balance of KIR3DL1/KIR3DL2 licensed NK cells and the different capacities of Behçet’s Disease. these receptors to eliminate pathogens could be involved in the etiopathogenesis of BD. Front. Immunol. 10:2755. doi: 10.3389/fimmu.2019.02755 Keywords: Behçet’s disease, HLA, KIR, NK cells, functional polymorphisms Frontiers in Immunology | www.frontiersin.org 1 November 2019 | Volume 10 | Article 2755 Castaño-Núñez et al. KIR, HLA-Ligand, and Behçet’s Disease INTRODUCTION The KIR-ligands are HLA class I molecules and the interaction between the pair ligand-receptor is fundamental for the Behçet’s disease (BD) [OMIM #109650] is a rare, chronic regulation of NK cell activity. All the classical HLA-class I and systemic vasculitis characterized by recurrent oral and molecules (HLA-A, B and C) present peptides to the CD8+ T genital ulcers, although other clinical manifestations, such as cells, however, not all of them are ligands of the KIR molecules. In skin lesions, ocular, gastrointestinal, and neurological disorders this sense, all the HLA-C molecules are ligands of KIR and they are relatively common. This chronic disorder is an immune- can be grouped as C1 (Asn80) or C2 (Lys80) which are ligands mediated disease in which imbalances between the innate and of KIR2DL2 and KIR2DL1, respectively. Nevertheless, only some adaptive immune response triggered by infectious agents or HLA-A and B molecules are KIR ligands, specifically, those environmental factors in genetically predisposed individuals have bearing the Bw4 epitope are KIR3DL1 ligands and A3 and A11 been suggested as the underlying mechanisms of the disease have been described to be KIR3DL2 ligands (4). HLA-B51 and (1). Among the genetic factors, HLA-B51 has been associated other HLA molecules associated with BD have the Bw4 epitope with the disease in many populations (2); more recently, other and interactions between Bw4 and KIR3DL1 have been proposed genes related to the immune system such as IL23R, IL10, STAT4, as one of the possible underlying mechanisms to explain the CCR1, CCR3, KLRC4, ERAP1, TNFAIP3, and FUT2 have been relationship between HLA class I and BD (7). However, a large associated with this disease (3). number of common HLA-B alleles encoding molecules with the Natural killer (NK) cells are lymphocytes of the innate Bw4 epitope (e.g., B∗ 44, B∗ 49, etc.) have never been associated immune system with several activators and inhibitor receptors with this disease, even those with isoleucine at position 80 (Bw4- on their surface. Among them, the Killer-cell Immunoglobulin- 80I), which are the strongly interacting ligands. The Bw4 epitope like Receptor (KIR) gene family consisted of a set of 15 spans the residues 77–83 but it has been described that changes genes: KIR2DL1, KIR2DL2, KIR2DL3, KIR2DL4, KIR2DL5A at specific positions in the HLA-B molecule outside this epitope, KIR2DL5B, KIR2DS1, KIR2DS2, KIR2DS3, KIR2DS4, in particular, position 97, affect the interaction of Bw4 with KIR2DS5, KIR3DL1, KIR3DL2, KIR3DL3, and KIR3DS1 KIR3DL1 and, interestingly, the position 97 of HLA-B has been and 2 pseudogenes: KIR2DP1 and KIR3DP1 located near to found strongly associated with susceptibility to BD (8–10). the leukocyte receptor complex in the chromosome 19. The With this picture, great complexity and a relative lack of KIR gene content is variable among individuals being a major the knowledge of the polymorphism of the KIR system and contributor to the KIR diversity in the human population and its relationship with its ligands, the objective of this study the NK cell repertoires (4). At present, 625 genotypes (different was to investigate the contribution of certain KIR functional sets of KIR genes) named by numbers assigned sequentially polymorphisms to the susceptibility to BD. have been found in the human population1 . Based on their gene content, two kinds of KIR haplotype groups, A and B (that give rise to the genotypes AA and Bx), have been described. Although MATERIALS AND METHODS there is no single specific criterion to distinguish them, from a Study Population practical point of view, one individual is considered as Bx when The study included a total of 466 BD-unrelated patients its set of KIR genes included at least one of the following genes: (44.2% males) who fulfilled the 1990 International Study 2DL2, 2DL5, 3DS1, 2DS1, 2DS2, 2DS3, 2DS5, and it is assigned Group classification criteria for BD (11), and 444 unrelated as AA when all these genes are absent. The KIR genes encoded healthy individuals (50% males) included as the control activator or inhibitor molecules, in general, the molecules with group. All the subjects were Spanish European recruited long cytoplasmic tails (L) contains two immune tyrosine-based from 17 Spanish hospitals across the country. The study was inhibitory motifs (ITIM) which transduce inhibitory signals approved by all local ethical committees of the corresponding to the NK cell, whereas the molecules with short cytoplasmic hospitals, and all the study participants gave their written tails (S) possess a positively charged amino acid residue in their informed consent to participate. Clinical features of the patient transmembrane region which allows them to associate with group were the following: 100% had oral ulcers, 59.4% a DAP12 molecule generating an activation signal (4). The A genital ulcers, 53.9% uveitis, 42% arthritis, and 21% vascular, Haplotypes have fewer activator genes than haplotypes B, in 18.2% neurological, 16.3% positive pathergy test, and 15.4% fact, the only two activator genes present in AA individuals are gastrointestinal involvement. KIR2DS4 and KIR2DL4, but only KIR2DS4 binds HLA classical molecules. Besides, each KIR gene has a variable number of DNA Extraction alleles, some of them with functional significance. Indeed, a Peripheral blood collected in EDTA tubes was obtained from different expression in the cell membrane of diverse KIR3DL1 the healthy controls, and peripheral blood or saliva served as alleles, ranging from practically null (in KIR3DL1∗ 004, which starting material from patients. Genomic DNA was extracted encodes a misfolded receptor mostly retained inside the cell) to using the QIAamp DNA Mini Kit (Qiagen, Barcelona, Spain) high expression, or production of soluble forms in the case of according to the manufacturer’s recommendations and stored at KIR2DS4 (5, 6). −20◦ C. The purity of DNA was determined using a NanoDrop spectrophotometer (Thermo Fisher Scientific, Wilmington, DE, USA). Only DNA samples having a 260/280 absorbance ratio of 1 http://www.allelefrequencies.net/kir6001a.asp 1.7–2.0 and a final concentration of 10–20 ng/µl were considered Frontiers in Immunology | www.frontiersin.org 2 November 2019 | Volume 10 | Article 2755 Castaño-Núñez et al. KIR, HLA-Ligand, and Behçet’s Disease appropriate. A total of 14 DNA samples from saliva were association. The odds ratios (ORs) and 95% confidence intervals eliminated because they did not meet these quality criteria. (95% CI) were calculated using the web software OpenEpi (Open Source Epidemiologic Statistics for Public Health, Versión Genotyping 3.01. www.OpenEpi.com). The HLA class I (A, B, and C) genotyping was carried out using a PCR-SSOP Luminex method, LABType SSO (One Lambda Inc., Canoga Park, CA), according to the manufacturer’s instructions. RESULTS Briefly, target DNAs were PCR-amplified using group-specific HLA Molecules KIR-Ligands primers (HLA-A, -B, or -C) and the biotinylated-PCR products A total of 435 BD patients (96.2%) and 439 controls (98.9%) were were denatured and hybridized with specific probes bound fulfilled genotyped and included in the statistical analysis. The to colored-coded microspheres. Phycoerythrin conjugated to distribution of different HLA molecules KIR-ligands is displayed Streptoavidin was used to label and reveal reactions which were in Table 1. The strongest associations were: HLA-B51, conferring read in a flow analyzer, LABScanTM100, to identify fluorescent risk and A3/11, conferring protection. The frequency of A3/11 intensity on each microsphere. The software HLA fusion 2.0 was decreased in both patient groups, B51 positive and negative (One Lambda Inc.) was used to assign the HLA typing of each (18 and 21%, respectively, vs. 26 and 33% in their corresponding locus. This method allows a medium resolution genotyping of controls), although statistical significance was reached only in each class I HLA gene. After class I genotyping, the samples were the negative group. Bw4 and Bw4-80I confer risk but with a classified as A3/A11, Bw4, Bw4-80I, C1, and C2. lower OR than B51. Among the group B51 negative (250 patients The set of KIR genes presents in each sample was determined and 370 controls), the distribution of Bw4 and Bw4-80I was also using a PCR-SSOP Luminex method, KIR SSO Genotyping not significantly different in patients and controls (Bw4 76.8% Test (One Lambda Inc., Canoga Park, CA), according to in patients vs. 76.2 in controls; Bw4-80I 56.0% in patients vs. the manufacturer’s instructions. The principle of the method 56.2% in controls, P > 0.05 in both cases). No differences in the is the same as that employed for HLA-class I genotyping distribution of C1 and C2 were detected in any case. with the corresponding group-specific primers and bound- probes. Also, to distinguish the presence/absence of KIR genes, KIR Genes this method allows discriminating between some groups of The distribution of the KIR genes in our cohort of patients alleles, specifically, between the group of KIR2DS4 alleles with and controls is summarized in Table 2. All the KIR genes the full sequence (KIR2DS4∗ 001/011/014/015) which encode and pseudogenes were detected in this study and four of functional proteins and the group with a deletion of 22 base them, corresponding to the centromeric (3DL3 and 3DP1) and pairs (KIR2DS4∗ 003/004/006/007/008/009/010/012/013) which telomeric (2DL4 and 3DL2) framework genes, were found in all encode non-functional variants (6). the samples. The distribution of genes was similar in patients KIR3DL1/S1 individuals were genotyped in the rs149123986 and controls and statistically significant differences were not by real-time PCR using TaqMan SNP Genotyping Assays reached in any case. Regarding the distribution of the KIR (Applied Biosystems, Barcelona, Spain) in a LightCycler 480 haplotypes, Bx was the most common in both groups and (Roche, Barcelona, Spain). This single nucleotide polymorphism no significant differences between patients and controls were (SNP, A/G) allows to distinguish two groups of KIR3DL1 observed. Concerning the KIR genotypes, a total of 51 genotypes alleles with different expression patterns: the KIR3DL1∗ 004 were found in our cohort, but only 12 had a frequency upper 2% (rs149123986G) with a very low or null membrane expression in at least one of the groups. The genotype 2 was slightly less (3DL1Null) and the rest of alleles (rs149123986A), which have a normal although variable membrane expression (3DL1Exp)2 . TABLE 1 | Distribution of HLA molecules KIR-ligands in Spanish BD patients and Statistical Analysis controls. Phenotypic and genotypic frequencies were estimated by direct HLA BD (%) Controls (%) P OR (95% CI) counting and distributions were compared using the χ2 test. N = 435 N = 439 This test was applied to check the association of (a) different ligands: A3/11, B51, Bw4, Bw4-80I, C1, and C2; (b) the KIR B51 185 (42.5) 69 (15.7) <10−7 3.96 (2.89–5.48) genes, haplotypes and genotypes; (c) the KIR3DL1/DS1 variants A3/11a 87 (20.0) 140 (31.9) <10−4 0.53 (0.39–0.73) in the full group and the subgroups B51 Bw4, Bw4-80I; and (d) Bw4 377 (86.7) 351 (80.0) 0.008 1.63 (1.13–2.35) the KIR2DS4 variants in the total groups and among individuals Bw4-I80 325 (74.7) 277 (63.1) 0.0001 1.73 (1.29–2.31) AA and Bx and stratified according to the presence/absence C1 329 (75.6) 351 (79.9) >0.05 of the combination ligand and receptor. The P-values were C2 313 (71.9) 302 (68.8) >0.05 corrected by Bonferroni’s adjustment (Pc) considering the C1C1 122 (28.0) 137 (31.2) >0.05 number of tests in each case. Those comparisons with Pc- C1C2 207 (47.6) 214 (48.7) >0.05 values < 0.05 were considered associated whereas those with C2C2 106 (24.4) 88 (20.0) >0.05 P < 0.05 but with Pc > 0.05 were considered suggestive of a B51positive group: 34 patients and 18 controls A3/11 positive (p = 0.09); B51 negative 2 https://www.ebi.ac.uk/ipd/kir/align.html group: 53 patients and 122 controls A3/11 positive (p = 0.001). Frontiers in Immunology | www.frontiersin.org 3 November 2019 | Volume 10 | Article 2755 Castaño-Núñez et al. KIR, HLA-Ligand, and Behçet’s Disease TABLE 2 | Distribution of KIR genes in Spanish BD patients and healthy controls. The distribution of the individuals in the six possible groups taking into account both, the KIR3DL1/S1 and the rs149123986, KIR BD Controls P Pc was different in patients and controls (p = 0.002 in 2 × 6 Table) n = 435 (%) n = 439 (%) with down-representation of 3DL1Null/3DL1 (6.2 vs. 11.4%, p = Genes 0.007, Pc = 0.042; OR = 0.51; 95% CI 0.32–0.84) and a trend 2DL1 416 (95.6) 422 (96.1) to an over-representation of 3DL1Exp/3DL1Exp (38.1 vs. 31.9%, 2DL2 270 (62.1) 267 (60.8) p = 0.052) individuals among patients. The distribution of the 2DL3 373 (85.7) 382 (87.0) three allele groups: 3DL1Exp, 3DL1Nul, and 3DS1 was different 2DL4 435 (100) 439 (100) in patients and control (p = 0.01, in 2 × 3 contigency Table) 2DL5A/2DL5B 241 (55.4) 261 (59.4) having the group of patients a decreased frequency of 3DL1Null (0.15 vs. 0.20, P = 0.005, Pc = 0.015; OR = 0.70; 95% CI 0.54– 2DP1 416 (95.6) 423 (96.4) 0.90) and a slight increased frequency of 3DL1Exp (0.61 vs. 0.56, 2DS1 174 (40.0) 194 (44.2) P = 0.03, Pc > 0.05). Results were similar among B51 positive 2DS2 269 (61.8) 262 (59.7) and negative individuals with a suggestive down-representation 2DS3 156 (35.9) 153 (34.8) of 3DL1Null/3DLS1 in both groups (B51positive: 5.9% in patients 2DS4 408 (93.8) 421 (95.9) vs. 15.9% in controls P = 0.009 Pc > 0.05; OR = 0.34; 95% CI 2DS5 141 (32.4) 168 (38.3) 0.13–0.83 and B51 negative: 6.4% in patients vs. 10.5% in controls 3DL1 409 (94.0) 422 (96.1) P = 0.04 Pc > 0.05; OR = 0.58; 95% CI 0.3–1.05) (Table 4). 3DS1 181 (41.6) 192 (43.7) Regarding the deletion of 22 bp of 2DS4, differences in the 3DL2 435 (100) 439 (100) distribution of individuals with at least one copy of the 2DS4Full 3DL3 435 (100) 439 (100) between patients and controls were not detected (42.7 vs. 37.5%, 3DP1 435 (100) 439 (100) P = 0.13). Since 2DS4 is the only activator binding classical HLA Haplotypes n = 435 (%) n = 439 (%) molecules in AA individuals, patients and controls were stratified AA 109 (25.1) 100 (22.8) according to their haplotype group in AA or Bx but no differences Bx 326 (74.9) 339 (77.2) were detected (Data not showed). Genotypes n = 435 (%) n = 439 (%) 1 109 (25.1) 100 (22.8) 2 38 (8.7) 62 (14.1) 0.01 >0.05 DISCUSSION 3 23 (5.3) 21 (4.8) 4 65 (14.9) 63 (14.3) The main finding of this study is the association of the 5 39 (9.0) 45 (10.2) KIR3DL1∗ 004 with susceptibility to BD. This allele is protective 6 26 (6.0) 36 (8.2) in the development of the disease and, according to our results, 7 18 (4.1) 16 (3.6) its effect is independent of HLA-B51. 13 10 (2.3) 4 (0.9) Previous studies analyzed the relationship between the 70 9 (2.1) 5 (1.1) distribution of the KIR genes and the disease. Similarly to our 71 8 (1.84) 12 (2.7) results, no differences between patients and controls have been 72 14 (3.2) 11 (2.5) reported in these studies (12–15). Only one of these studies, 73 12 (2.8) 10 (2.3) performed in the Iranian population, investigated KIR genotypes. The resta 64 (14.7) 54 (12.3) That study reported a different distribution of various KIR a Genotypes genotypes, including genotype 2, although the direction of the with a frequency lower than 2%. association was contrary to that found in the present study (15). KIR3DL1/DS1 has focused on interest because of its function as the receptor of the Bw4 epitope. The distribution of KIR3DL1 frequently represented in patients than in controls (8.7 vs. 14.1%, and 3DS1 in BD was evaluated in a large cohort of Turkish. respectively, P = 0.01, Pc > 0.05; OR 0.58; 95% CI 0.38–0.89). Similarly to our results, no association was found in the whole cohort neither in subgroups with B51, Bw4, or Bw4- Functional KIR Polymorphyms 80I (16) but this study did not investigate variants on this Data regarding the polymorphism of KIR3DL1/S1 are displayed gene. Variants in KIR3DL1 have been reported as associated in Table 3. The distribution of individuals KIR3DL1/S1 was not with the disease in two high throughput studies (10, 17), significantly different in patients and controls (KIR3DL1+DS1+ although none of them was designed to investigate the possible 35.6 vs. 39.9%, KIR3DL1+DS1– 58.3 vs. 56.3%, KIR3DL1-DS1+ involvement of functional variants in this disease. Very recently, 6.0 vs. 3.9%, respectively). Nevertheless, the distribution of the during the submission of this paper, a study that addresses genotypes of rs149123986 was significantly different in both this question in a smaller cohort composed of ethnically- groups (p = 0.01 in 2 × 3 contigency Table) with an over- mixed patients and controls was published (18). The main representation of AA in patients (73.5% vs. 64.3 in controls, Pc = conclusion of that study is the same as ours, an association 0.006; OR = 1.54; 95% CI 1.16–2.07) and a down-representation of the functional alleles of KIR3DL1, independent of HLA- of AG (23.0 vs. 31.7%, Pc = 0.015, OR = 0.65; 95% CI 0.48–0.88). B51 with the disease. Nevertheless, they are some discrepancies Frontiers in Immunology | www.frontiersin.org 4 November 2019 | Volume 10 | Article 2755 Castaño-Núñez et al. KIR, HLA-Ligand, and Behçet’s Disease TABLE 3 | Frequency of the KIR3DL1/DS1 functional polymorphism in BD patients and controls. KIR rs149123986 BD Controls P Pc OR (95%CI) 3DL1 3DS1 n = 435 % n = 439 % + – AA 3DL1Exp /3DL1Exp 166 38.1 140 31.9 0.05 >0.05 + – AG 3DL1Exp /3DL1Null 74 17.0 89 20.3 + + AA 3DL1Exp /3DS1 128 29.4 125 28.5 + – GG 3DL1Null /3DL1Null 14 3.2 18 4.1 + + AG 3DL1Null /3DS1 27 6.2 50 11.4 0.007 0.042 0.51 (0.32–0.84) - + AA 3DS1/3DS1 26 6.0 17 3.9 AF AF 3DL1Exp 534 0.61 494 0.56 0.03 >0.05 3DL1Null 129 0.15 175 0.20 0.005 0.015 0.7 (0.54–0.90) 3DS1 207 0.24 209 0.24 >0.05 >0.05 AF, allelic frequency. Bold values are those statistically significant after correction. KIR3DL1∗ 004 encodes a misfolded receptor which is retained TABLE 4 | Frequency of the KIR3DL1 functional polymorphism in BD patients and controls stratified according to the HLA-B51. in the cell, although it has been described that a small quantity of folded protein can be exported to the membrane (22). One B51+ B51– possible explanation to the protective effect of KIR3DL1∗ 004 BD Controls BD Controls in these pathologies is that this misfolded molecule retains 185 69 250 370 intracellularly the corresponding class I-risk molecules. In this case, differences in the distribution of KIR3DL1∗ 004 n % n % n % n % among patients with and without the risk molecule would be expected, nevertheless, no differences were found in the 3DL1Exp /3DL1Exp 76 41.1 25 36.2 90 36.0 115 31.1 present study neither in other pathologies (20, 21, 23) and 3DL1Exp /3DL1Null 28 15.1 11 15.9 46 18.4 78 21.1 no evidence supporting the intracellular interaction between 3DL1Exp /3DS1 53 28.7 19 27.5 75 30.0 106 28.6 KIR3DL1 and MHC class I molecules have been found (5). 3DL1Null /3DL1Null 7 3.8 3 4.3 7 2.8 15 4.1 Interestingly, KIR3DL1∗ 004 is within a common telomeric 3DL1Null /3DS1 11 5.9 11 15.9 16 6.4 39 10.5 haplotype Tel-A1-V which, also, bears two alleles encoding 3DS1/3DS1 10 5.4 0 0.0 16 6.4 17 4.6 secreted receptors, KIR2DL4∗ 008 and KIR2DS4∗ 006 and another 3DL1Exp 233 0.63 80 0.58 301 0.60 414 0.56 that encodes a membrane receptor KIR3DL2∗ 003/005/011/020 3DL1Null 53 0.14 28 0.20 76 0.15 147 0.20 (24)3 . Therefore, this Tel-A1-V haplotype, protective in BD, 3DS1 84 0.23 30 0.22 123 0.25 179 0.24 carries non-functional alleles with the only exception of KIR3DL2. Consequently, KIR3DL1∗ 004 could be a marker of the Tel-A1-V haplotype and the protective association found in our in the associated genotypes between these two studies. When study could be explained by the fact that KIR3DL2 is the only the allele frequencies of the study by Petrushkin et al. are functional gene in this haplotype. deduced of their data, results are similar to ours regarding the KIR3DL2 is an inhibitor framework gene, found in 100% of allele KIR3DL1∗ 004, which is underrepresented in patients (with the population and it has many allelic forms encoding many similar OR), although the statistical significance is only reached different proteins (161 and 111, respectively, in the last database), in our cohort. Nevertheless, we have opposite results concerning the relative expression levels of these different proteins are KIR3DL1Exp and KIR3DS1. In our cohort, the alleles that unknown. The KIR3DL2 HLA-ligands are A3 and A11 and there encode KIR3DL1Exp are significantly overrepresented among is some evidence that KIR3DL2 can bind these HLA molecules patients whereas, in its cohort, it is KIR3DS1 which tends in vivo (25). The education of the NK cells is based on the to be overrepresented among patients. In our opinion, the interaction between the inhibitory receptors and the self-MHC- distribution of the allelic frequencies reinforces the idea of I molecules, only those individuals bearing the HLA-ligand association with the allele KIR3DL1∗ 004. Discrepancies could be molecules have licensed NK cells that express the corresponding due to different reasons, among them, methods used in KIR3DL1 receptor and therefore, only the individuals A3 and A11 have genotyping, approaches, stratification of cohorts with a relatively mature NK cells expressing KIR3DL2. Noticeably, similarly to small number of individuals and ethnical differences. Besides, the findings in the present cohort, different studies in several there are previous studies in other class-I pathologies such as populations have found a protective association of HLA-A3 with psoriatic disease (PD) and ankylosing spondylitis (AS), in which BD, independent of HLA-B51 (9, 10, 26). This protective effect a protective association of the KIR3DL1∗ 004 allele has been reported (19–21). 3 https://www.ebi.ac.uk/ipd/kir/sequenced_haplotypes.html Frontiers in Immunology | www.frontiersin.org 5 November 2019 | Volume 10 | Article 2755 Castaño-Núñez et al. KIR, HLA-Ligand, and Behçet’s Disease of HLA-A3, not yet explained, could be based on the presence of In conclusion, our results suggest a protective role of a licensed KIR3DL2 NK cells in these individuals. telomeric KIR haplotype with only one functional gene in BD. KIR molecules having 3 extracellular domains (KIR3DL1 and Therefore, the education of the NK cells depending on the HLA, KIR3DL2) are pattern recognition receptors (PRR), they sense the balance of licensed KIR3DL1/KIR3DL2 NK cells and the pathogen-associated molecular patterns (PAMPs), specifically different capacities of these receptors to eliminate pathogens single-stranded microbial DNA molecules containing CpG could be involved in the etiopathogenesis of BD. motifs (CpG-ODN). KIR3DL2 binds CpG-ODN by the D0 domain and the complexes are internalized and recognized DATA AVAILABILITY STATEMENT by TLR9 located in the endosomal compartment what results in cell activation. Nevertheless, in spite of KIR3DL1 binds The raw data supporting the conclusions of this article will be CpG-ODN, the complexes are not internalized (27). Thus, made available by the authors, without undue reservation, to any one possible explanation to the protective effect of the Tel- qualified researcher. A1-V haplotype is that the lack of expression of KIR3DL1 in these individuals permits a better efficiency to eliminate ETHICS STATEMENT pathogens by KIR3DL2 NK cells. In other words, NK cells of individuals with the KIR3DL1Exp would be lesser efficient The studies involving human participants were reviewed and in eliminating pathogens resulting in a perpetuation of the approved by CEI de los Hospitales Universitarios Vírgen release of inflammatory cytokines. In this sense, a protective Macarena-Virgen del Rocío. The patients/participants provided effect of KIR3DL1∗ 004 against the progression of the VIH their written informed consent to participate in this study. infection to AIDS has been described (28). This mechanism would explain why the association is independent of B51 AUTHOR CONTRIBUTIONS and Bw4 in our study but also psoriasis. Moreover, some HLA-Bw4 molecules, such as B27 (associated with ankylosis M-FG-E: experimental design. ÁC-N: experimental procedures spondylitis) and B51 could be more efficient to license KIR3DL1 performance. M-AM-C and J-RG-L: experimental support. NK cells and this could be related to its association with ÁC-N, M-AM-C and M-FG-E: data analysis and preparation of inflammatory diseases. In this sense, the activity of ERAP1, the manuscript. NO-C, F-JG-H, GE, GG-G, JS-B, M-RJ, RS, RB, an enzyme that has been related in an epistatic way with the A-CB-M, RG, PF, MR-C, LR-R, TC, SC, J-JA-S, and JM: samples HLA molecules involved in these pathologies and that can and clinical data providers. conditioner the repertoire of peptides binding to them, could modify the role of the diverse HLA-Bw4 molecules in these FUNDING diseases (29). Lastly, the Tel-A1-V haplotype includes KIR2DS4∗ 006, which This work was supported by Fondo de Investigaciones Sanitarias, has the deletion of 22 base pairs, but according to our results, this Instituto de Salud Carlos III (ISCIII, 13/01118 and 16/01373), deletion does not influence the disease susceptibility. Fondos FEDER and Plan Andaluz de Investigación (CTS-0197). REFERENCES 7. Petrushkin H, Hasan MS, Stanford MR, Fortune F, Wallace GR. Behçet’s disease: do natural killer cells play a significant role? Front Immunol. (2015) 1. Mendes D, Correia M, Barbedo M, Vaio T, Mota M, Gonçalves O, et al. 24:134. doi: 10.3389/fimmu.2015.00134 Behçet’s disease–a contemporary review. J Autoimmun. (2009) 32:178–88. 8. Sanjanwala B, Draghi M, Norman PJ, Guethlein LA, Parham P. doi: 10.1016/j.jaut.2009.02.011 Polymorphic sites away from the Bw4 epitope that affect interaction 2. de Menthon M, Lavalley MP, Maldini C, Guillevin L, Mahr A. HLA-B51/B5 of Bw4+ HLA-B with KIR3DL1. J Immunol. (2008) 181:6293–300. and the risk of Behçet’s disease: a systematic review and meta-analysis of doi: 10.4049/jimmunol.181.9.6293 case-control genetic association studies. Arthritis Rheum. (2009) 61:1287–96. 9. Ombrello MJ, Kirino Y, de Bakker PI, Gul A, Kastner DL, Remmers EF. doi: 10.1002/art.24642 Behcet disease-associated MHC class I residues implicate antigen binding 3. Leccese P, Alpsoy E. Behçet’s disease: an overview of etiopathogenesis. Front and regulation of cell-mediated cytotoxicity. Proc Natl Acad Sci USA. (2014) Immunol. (2019) 10:1067. doi: 10.3389/fimmu.2019.01067 111:8867–72. doi: 10.1073/pnas.1406575111 4. Middleton D, Gonzelez F. The extensive polymorphism of KIR 10. Ortiz-Fernández L, Carmona FD, Montes-Cano MA, García-Lozano JR, genes. Immunology. (2010) 129:8–19. doi: 10.1111/j.1365-2567.200 Conde-Jaldón M, Ortego-Centeno N, et al. Genetic analysis with the 9.03208.x immunochip platform in Behçet disease. Identification of residues associated 5. Taner SB, Pando MJ, Roberts A, Schellekens J, Marsh SG, Malmberg KJ, in the HLA class I region and new susceptibility Loci. PLoS ONE. (2016) et al. Interactions of NK cell receptor KIR3DL1∗ 004 with chaperones 11:e0161305. doi: 10.1371/journal.pone.0161305 and conformation-specific antibody reveal a functional folded state as 11. Criteria for diagnosis of Behcet’s disease. International Study Group for well as predominant intracellular retention. J Immunol. (2011) 186:62–72. Behcet’s Disease. Lancet. (1990) 335:1078–80. doi: 10.4049/jimmunol.0903657 12. Middleton D, Meenagh A, Sleator C, Gourraud PA, Ayna T, Tozkir H, et al. 6. Middleton D, Gonzalez A, Gilmore PM. Studies on the expression of the No association of KIR genes with Behcet’s disease. Tissue Antigens. (2007) deleted KIR2DS4∗ 003 gene product and distribution of KIR2DS4 deleted 70:435–8. doi: 10.1111/j.1399-0039.2007.00929.x and nondeleted versions in different populations. Hum Immunol. (2007) 13. Arayssi TK, El Hajj N, Shamseddine W, Ibrahim G, Nasr J, Sabbagh 68:128–34. doi: 10.1016/j.humimm.2006.12.007 AS, et al. Killer cell immunoglobulin-like receptor genotypes in Behçet’s Frontiers in Immunology | www.frontiersin.org 6 November 2019 | Volume 10 | Article 2755 Castaño-Núñez et al. KIR, HLA-Ligand, and Behçet’s Disease disease patients: any role for the 3DP1∗ 001/002 pseudogene? Genet Test Mol 24. Hou L, Chen M, Ng J, Hurley CK. Conserved KIR allele-level haplotypes Biomarkers. (2009) 13:319–24. doi: 10.1089/gtmb.2008.0108 are altered by microvariation in individuals with European ancestry. Genes 14. Moon SJ, Oh EJ, Kim Y, Kim KS, Kwok SK, Ju JH, et al. Diversity of killer cell Immun. (2012) 13:47–58. doi: 10.1038/gene.2011.52 immunoglobulin-like receptor genes in uveitis associated with autoimmune 25. Augusto DG, O’Connor GM, Lobo-Alves SC, Bass S, Martin MP, diseases: ankylosing spondylitis and Behçet disease. Ocul Immunol Inflamm. Carrington M, et al. Pemphigus is associated with KIR3DL2 expression (2013) 21:135–43. doi: 10.3109/09273948.2012.754905 levels and provides evidence that KIR3DL2 may bind HLA-A3 and 15. Mohammad-Ebrahim H, Kamali-Sarvestani E, Mahmoudi M, Beigy A11 in vivo. Eur J Immunol. (2015) 45:2052–60 doi: 10.1002/eji.2014 M, Karami J, Ahmadzadeh N, et al. Association of killer cell 45324 immunoglobulin-like receptor (KIR) genes and their HLA ligands with 26. Takeuchi M, Mizuki N, Meguro A, Ombrello MJ, Kirino Y, Satorius C, susceptibility to Behçet’s? disease. Scand J Rheumatol. (2018) 47:155–63. et al. Dense genotyping of immune-related loci implicates host responses doi: 10.1080/03009742.2017.1340510 to microbial exposure in Behçet’s disease susceptibility. Nat Genet. (2017) 16. Erer B, Takeuchi M, Ustek D, Tugal-Tutkun I, Seyahi E, Özyazgan Y, et al. 49:438–43. doi: 10.1038/ng.3786 Evaluation of KIR3DL1/KIR3DS1 polymorphism in Behçet’s disease. Genes 27. Sivori S, Falco M, Carlomagno S, Romeo E, Soldani C, Bensussan A, et al. A Immun. (2016) 17:396–9. doi: 10.1038/gene.2016.36 novel KIR-associated function: evidence that CpG DNA uptake and shuttling 17. Kim SJ, Lee S, Park C, Seo JS, Kim JI, Yu HG. Targeted resequencing of to early endosomes is mediated by KIR3DL2. Blood. (2010) 116:1637–47. candidate genes reveals novel variants associated with severe Behçet’s uveitis. doi: 10.1182/blood-2009-12-256586 Exp Mol Med. (2013) 45:e49. doi: 10.1038/emm.2013.101 28. Martin MP, Qi Y, Gao X, Yamada E, Martin JN, Pereyra F, et al. Innate 18. Petrushkin H, Norman PJ, Lougee E, Parham P, Wallace GR, Stanford partnership of HLA-B and KIR3DL1 subtypes against HIV-1. Nat Genet. MR, et al. KIR3DL1/S1 allotypes contribute differentially to the (2007) 39:733–40. doi: 10.1038/ng2035 development of Behçet disease. J Immunol. (2019) 203:1629–35. 29. Guasp P, Barnea E, González-Escribano MF, Jiménez-Reinoso A, Regueiro doi: 10.4049/jimmunol.1801178 JR, Admon A, et al. The Behçet’s disease-associated variant of the 19. Ahn RS, Moslehi H, Martin MP, Abad-Santos M, Bowcock AM, Carrington M, aminopeptidase ERAP1 shapes a low-affinity HLA-B∗ 51 peptidome by et al. Inhibitory KIR3DL1 alleles are associated with psoriasis. Br J Dermatol. differential subpeptidome processing. J Biol Chem. (2017) 292:9680–9. (2016) 174:449–51. doi: 10.1111/bjd.14081 doi: 10.1074/jbc.M117.789180 20. Berinstein J, Pollock R, Pellett F, Thavaneswaran A, Chandran V, Gladman DD. Association of variably expressed KIR3dl1 alleles with psoriatic disease. Conflict of Interest: The authors declare that the research was conducted in the Clin Rheumatol. (2017) 36:2261–6. doi: 10.1007/s10067-017-3784-5 absence of any commercial or financial relationships that could be construed as a 21. Díaz-Peña R, Vidal-Castiñeira JR, Alonso-Arias R, Suarez-Alvarez B, Vicario potential conflict of interest. JL, Solana R, et al. Association of the KIR3DS1∗ 013 and KIR3DL1∗ 004 alleles with susceptibility to ankylosing spondylitis. Arthritis Rheum. (2010) Copyright © 2019 Castaño-Núñez, Montes-Cano, García-Lozano, Ortego-Centeno, 62:1000–6. doi: 10.1002/art.27332 García-Hernández, Espinosa, Graña-Gil, Sánchez-Bursón, Juliá, Solans, Blanco, 22. Pando MJ, Gardiner CM, Gleimer M, McQueen KL, Parham P. The protein Barnosi-Marín, Gómez de la Torre, Fanlo, Rodríguez-Carballeira, Rodríguez- made from a common allele of KIR3DL1 (3DL1∗ 004) is poorly expressed Rodríguez, Camps, Castañeda, Alegre-Sancho, Martín and González-Escribano. at cell surfaces due to substitution at positions 86 in Ig domain 0 and 182 This is an open-access article distributed under the terms of the Creative Commons in Ig domain 1. J Immunol. (2003) 171:6640–9. doi: 10.4049/jimmunol.17 Attribution License (CC BY). The use, distribution or reproduction in other forums 1.12.6640 is permitted, provided the original author(s) and the copyright owner(s) are credited 23. Zvyagin IV, Mamedov IZ, Britanova OV, Staroverov DB, Nasonov EL, and that the original publication in this journal is cited, in accordance with accepted Bochkova AG, et al. Contribution of functional KIR3DL1 to ankylosing academic practice. No use, distribution or reproduction is permitted which does not spondylitis. Cell Mol Immunol. (2010) 7:471–6. doi: 10.1038/cmi.2010.42 comply with these terms. Frontiers in Immunology | www.frontiersin.org 7 November 2019 | Volume 10 | Article 2755