Analysis of Killer Cell Immunoglobulin-Like Receptor Genes and Their HLA Ligands in Inflammatory Bowel Diseases

Genetic studies have illustrated that killer cell immunoglobulin-like receptor (KIR) genes could participate in various autoimmune disorders. We aimed to clarify the role of KIR genes, HLA ligands, HLA-KIR interactions, and their genotypes in inflammatory bowel disease (IBD) susceptibility. The study population was composed of 183 IBD subjects, comprising 100 ulcerative colitis (UC) patients, 83 Crohn's disease (CD) patients, and 274 healthy subjects. Polymerase chain reaction with sequence-specific primers (PCR-SSP) was used to evaluate the absence or presence of the 15 KIR genes, 5 HLA class I ligands, and 2 pseudogenes. We did not find any significant difference in allele frequency of KIRs and pseudogenes between IBD patients and healthy controls. In the case of HLA genes, there was a significant difference in HLA-B-Bw4Thr80 frequency between UC patients and healthy controls (P = 0.03, OR = 0.06, 95%CI = 0.008-0.4). Furthermore, we found a significant difference in HLA-C1Asn80 frequency between CD patients and healthy controls (P = 0.04, OR = 0.49, 95% CI = 0.3-0.8). In the full-array combination of KIR genes, there was no significant frequency difference between UC patients and healthy controls, while two KIR genotypes showed a significant susceptible association with CD. Our data do not support a strong role of NK cells in IBD susceptibility, but it does not rule out a role for KIR variability in IBD patients. However, there are some protective associations such as Bw4 alleles; these associations may be due to the interaction of the alleles to TCRs rather than KIRs.


Introduction
Inflammatory bowel diseases (IBDs) are categorized as chronic autoimmune diseases, which mostly target the gastrointestinal tract (GI). Ulcerative colitis (UC) and Crohn's disease (CD) are the two primary forms of IBDs [1][2][3]. The main parts involved in the UC are the rectum and colon, whereas the most part involved in CD is the ileum [4]. Although the etiology of IBDs is not completely understood, it has been shown that there are complex interactions between susceptible genes and environmental factors, which lead to defective inflammatory responses [5]. Environmental triggers such as microbiota in a genetically susceptible person may lead to the dysregulation of the immune system, and afterward, their interactions with immune and nonimmune cells can initiate the diseases [6].
The natural killer (NK) cells play important roles in the innate immune system and could be implicated in early responses against infected and transformed cells through cytokine production and direct toxicity [7,8]. Signaling through inhibitory and activating receptors determines the activity and function of NK cells. The HLA class I molecules act as ligands for killer cell immunoglobulin-like receptors (KIRs), which are presented on the surface of NK cells [7]. KIRs are a group of transmembrane proteins and are mainly presented by NK cells and some groups of T cells. Interaction between KIRs and their corresponding HLA class I ligands could play a critical role in the NK cell function. Furthermore, it has been documented that these genetic combinations could be implicated in cancers and autoimmune and infectious diseases [9,10]. The KIR receptors are encoded by a family of genes located on chromosome 19 which consists of 15 genes and two pseudogenes [11]. Depending on the intracellular length (long or short) and extracellular domains (2D or 3D), the KIR genes were categorized into 3 groups: activating KIR (aKIR) genes (KIR2DS1, 2DS2, 2DS3, 2DS4, 2DS5, and 3DS1), inhibitory KIR (iKIR) genes (KIR2DL1, 2DL2, 2DL3, 2DL5a, 2DL5b, 3DL1, 3DL2, and 3DL3), and pseudogenes (2DP1 with no protein expression and 3DP1 with a putative protein product). However, KIR2DL4 has the ability to act as both an inhibitory and activating receptor [12]. Interaction of KIR receptors with their appropriate HLA ligands (HLA-B and HLA-C alleles) leads to either activating or inhibitory signals according to the presence of intracellular immunoregulatory tyrosine-based activating or inhibitory motifs [11].
Two types of KIR haplotypes exist: A and B. Haplotype A only has one activating KIR, 2DS4, while haplotype B has various combinations of 2DS1, 2DS2, 2DS3, 2DS4, 2DS5, and 3DS1. HLA-C molecules are the main ligand for inhibitory KIRs. HLA-C has various variants; however, in the case of KIR, recognition could be reduced into two groups based on the amino acid that is located in the extracellular domain at position 80 [9,10]. Variants belonging to group 1 of HLA-C alleles are those with asparagine at this position and presented as ligands for KIR2DL2 and 2DL3, while variants belonging to group 2 of HLA-C alleles are those with lysine at this position and presented as ligands for KIR2DL1. There is no known KIR for HLA-B-Bw6, but Bw4, another allelic form of HLA-B, is recognized by KIR3DL1. Inhibitory and activating KIRs are almost identical in the sequence of HLA-binding domains; however, there is a low affinity between activating KIRs and HLA-C. This issue proposes that these class I molecules could not be used as a common ligand for aKIRs. There is an important structural difference in the cytoplasmic region of inhibitory and activating KIRs; the iKIRs have a long immunoreceptor tyrosine-based inhibitory motif in their tails, whereas the potential aKIRs need to transfer their signaling via adapter molecules that possess ITAM structures. Therefore, the aKIRs need an adaptor protein to activate NK and NKT cells.
Studies documented that frequency of activating and inhibitory KIRs and their haplotypes is different in various people [10,[13][14][15]. The body of evidence proposed that KIR genes have an essential role in various diseases. Some combinations of KIR-HLA could modulate the immune response of NK and NKT [16,17]. Genetic association stud-ies have documented that there is a link between aKIRs and various autoimmune and inflammatory diseases [13][14][15]. Alteration in NK cell activity in the gut through KIR-HLA interactions may propose a regulatory mechanism for inflammation. Jones et al. showed that there are differences in KIR frequencies, presenting that KIR and their interactions with cognate HLA could be implicated in the UC susceptibility [13].
In our study, the frequency of 15 KIR genes, 5 HLA ligands, and 2 pseudogenes was evaluated in the UC, CD, and control groups. We aimed to clarify the role of KIR and HLA genes and their genotypes in IBD susceptibility. To our knowledge, no study has evaluated the frequency of KIR genes, HLA ligands, and also their interactions in the Iranian population with IBDs.

Study Population.
The study population was made up of 183 IBD patients (100 UC patients (41 men and 59 women) and 83 CD patients (28 men and 55 women)) aged 44:3 ± 10:16 years (mean ± SD), who were attending the IBD outpatient clinic at Imam Khomeini Hospital, Tehran, Iran. The control group consisted of 274 unrelated randomly selected healthy controls (106 men and 168 women) aged 41:5 ± 12:11 years. All the subjects had Iranian racial background, and IBD patients were age-, sex-, and ethnicity-matched with the healthy control subjects. The diagnosis of UC [18] and CD [19] was according to the diagnostic guidelines and combination of histopathological, radiological, endoscopic, and biochemical tests. Controls had neither family history nor clinical manifestations of any type of autoimmune disease. Patients who had other types of colitis were ruled out of our study.
This study was performed based on the Declaration of Helsinki guidelines. Moreover, methods were carried out under the relevant regulations and guidelines by the mentioned institution. All the participants were informed of blood collection, DNA isolation, and genetic evaluation. In addition, the written informed consent was signed by all participants and/or their legal guardian/s. The ethics committee of the Tehran University of Medical Sciences approved this study.

Statistical Methods.
Statistical analyses were carried out by SPSS version 22 (IBM Corp., Armonk, NY, USA). In order to compare the frequency of KIR genes and their HLA ligands between the case and control groups, Pearson's chisquared test with continuity correction was applied. When the expected difference between the two groups was small, Fisher's exact test was applied. For estimations of strength of association, odds ratios (ORs) and 95% confidence intervals (CIs) were used. The Benjamini-Hochberg method was performed to control the false discovery rate (FDR) [29]. After correction for comparison, P values less than 0.05 (<0.05) were considered to be statistically significant. Pearson's chi-squared test with one degree of freedom was applied to confirm the Hardy-Weinberg equilibrium constant.
A geometric series was used in order to determine the specific genotype and its frequency. Genotype ID shows an exclusive number generated by the summation of members of a geometric series comprising products of each gene code (1 for positive and 0 for negative) multiplied by consecutive powers of 2: ðfirst gene × 20Þ + ðsecond gene × 21Þ, +ðthird gene × 22Þ, etc.

Implication of KIR and HLA Genes in IBD Susceptibility.
Primer sets (specific and internal) of the PCR-SSP assay for combined KIR-HLA genotyping are depicted in Supplementary file 1. Associations of KIR and HLA genes with IBDs (UC and CD) are depicted in Table 1. There were no significant differences in the frequency of inhibitory KIRs and pseudogenes between IBD patients and healthy subjects. Concerning aKIRs, the frequency of KIR2DS4 (full) was significantly dif-ferent between the UC and control groups (P = 0:02, OR = 0:53, 95% CI 0.3-0.9), and also, the frequency of KIR2DS3 was significantly different between the CD and control groups (P = 0:048, OR = 1:64, 95% CI 1.0-2.7). These differences were no longer significant after the FDR correction. Indeed, there were no significant differences in the frequency of KIR genes between IBD patients and healthy controls. In the case of HLA genes, the frequency of HLA-B-Bw4 Thr80 was significantly different between the UC and control subjects, and this association was significant after FDR correction (P = 0:03, OR = 0:06, 95% CI 0.008-0.4). Furthermore, the frequency of HLA-C1 Asn80 and HLA-B-Bw4 Thr80 was significantly different between the CD and control groups, but after FDR correction, just the frequency of HLA-C1 Asn80 remained to be significant (P = 0:04, OR = 0:49, 95% CI 0.3-0.8) ( Table 1).

Implication of KIR and/or HLA Genotypes on IBD
Susceptibility. The association of KIR genotypes with IBDs is illustrated in Table 2. There was no significant difference between the UC and control groups in the full-array combination of the frequency of KIR genes. In the case of CD patients and healthy controls, KIR genotype no. 5 and 6 showed a significant susceptible association with CD (P = 0:01, OR = 8:7, 95% CI 1.7-45.8 and P = 0:001, OR = 33:2, 95% CI 4.1-266.3, respectively).
The association of HLA gene combinations in IBDs is illustrated in Table 3. We did not find any significant difference in the frequency of the gene combinations neither between the UC and control groups nor between CD and control groups. In the evaluation of aKIR and iKIR genotype frequencies, there were no significant differences between IBD patients and healthy controls (Tables 4 and 5).

Receptor-Ligand
Interaction. Seventeen pair sets of KIR/HLA interaction in 2 possible conditions (absence and presence of interaction) were analyzed to clarify the role of these interactions in IBD susceptibility. Among all confirmed KIR/HLA pairs, there was no significant association between the UC and control groups.

Discussion
Investigations on IBD patients have shown that the innate immune system plays critical roles against microbial infections through the initiation of an inflammatory response in the bowel [30]. One of these innate immune cells with an important role in GI is NK cell. NK cells are regulated through KIR-HLA interactions. Studies showed that these interactions were implicated in various autoimmune rheumatic diseases [31][32][33]. Nonetheless, results obtained about KIR genes and IBD susceptibility from different studies are controversial [34].
Determining the role of KIRs and HLA ligands in IBD susceptibility needs further investigations since NK cells have 3 Journal of Immunology Research a vast repertoire of surface receptors, which are participating in NK cell activity [31].
During the past few years, many authors have been interested in the role of NK cells in IBDs [35][36][37]. Johansson et al. discussed the controversial roles of NK cells in the disease. In addition, they suggested that future investigations should be focused on the anatomical localization of NK cells and the cytokine environment [38]. Giacomelli et al. indicated that NK cells have a low level of killing activity when cocultured with PBMC of IBD patients. They proposed that this decreased activity could not be because of primary defects in NK cells since these NK cells were normal in killing activity. They suggested that inhibitory serum factors, likely from lymphocytes, could be produced in IBD patients, thereby leading to a low level of NK cell activity [39]. Another study has also shown that the NK cell cytotoxicity of UC patients is reduced [40].
There are reports that showed associations between the expression of surface receptors of NK cells and their cognate HLA ligands and autoimmune and infectious diseases [8]. For instance, ankylosing spondylitis is one of these rheumatic diseases in which the frequency of two KIR genes (KIR2DL3 and KIR2DL5) and two HLA ligands (HLA-B27 and HLA-C2 Lys80 ) was significantly different between AS patients and healthy controls [41]. Moreover, HLA-C w4 and KIR2DS4 were associated with rheumatoid arthritis [42]. Our recent study did not show any significant differences in the frequency of KIR genes between BD patients and healthy controls. In the case of HLA genes, HLA-B5, HLA-B51, HLA-B-Bw4 Ile80 , and HLA-C2 Lys80 showed a susceptible   A P value less than 0.05 was statistically significant. Genotype frequency less than 2% is not included. KIR: killer cell immunoglobulin-like receptor; CD: Crohn's disease; UC: ulcerative colitis. 5 Journal of Immunology Research association, while HLA-C1 Asn80 showed a protective effect against BD [26]. Jones et al. documented that KIR2DS2 and KIR2DL2 have a susceptible association with UC. In addition, they showed that the KIR2DL3 in the presence of its ligand, HLA-C1, has a protective effect against UC [13]. Hollenbach et al. reported that the KIR2DL2/3 has a susceptible association, while the C2 ligand has a protective effect on CD [43].
We evaluated the genetic diversity of KIRs and their corresponding ligands in the Iranian population with CD and UC. In the case of KIR genes, we reported that these genes are not associated with susceptibility to IBDs. Indeed, our results were in line with the studies of Saito et al. [44] and Wilson et al. [11], showing no significant association between KIR genes and CD patients. In the case of UC patients, our results were different from the mentioned studies. Indeed, each study tells a different story about the role of KIR genes in susceptibility to UC. The same data were obtained about the role of HLA class I with the disease susceptibility.
In our study, the HLA-B-Bw4 Thr80 showed a higher frequency in controls in comparison to UC patients and showed a protective role for the HLA-B-Bw4 Thr80 in UC patients. Furthermore, the HLA-C1 Asn80 also showed a higher frequency in controls in comparison to CD patients and showed a protective role for the HLA-C1 Asn80 in CD patients (Table 1). Therefore, we can only assume that there is a potential relationship between HLA genes and IBD diseases. However, none of these HLA genotype combinations were significant between the control and patient groups (Table 3). In the case of aKIR and iKIR genotypes, we did not show any significant association between these genotypes and IBD susceptibility (Tables 4 and 5). Gene-gene interaction analysis showed that the interaction between KIR2DL3 and HLA.C1 Asn has a susceptible role in CD. A P value less than 0.05 was statistically significant. Genotype frequency less than 2% is not included. A P value less than 0.05 was statistically significant. Genotype frequency less than 2% is not included. 6 Journal of Immunology Research

Conclusions
Collectively, our data do not support a strong role of NK cells in IBD susceptibility, but it does not rule out a role for KIR variability in IBD patients. However, there are some protective associations such as Bw4 alleles (present in 1% of UC patients and 4% of CD patients versus 14% of controls). These associations may be due to the interaction of these alleles to TCRs rather than KIRs.

Data Availability
All data generated or analyzed during this study are available upon request.

Ethical Approval
This study was performed based on the Declaration of Helsinki guidelines. The ethics committee of the Tehran University of Medical Sciences approved this study.

Consent
The written informed consent was signed by all participants and/or their legal guardian/s before enrolling in the study.

Conflicts of Interest
The authors declare that there is no conflict of interest regarding the publication of this paper.

Authors' Contributions
Fereshteh Beigmohammadi is responsible for the acquisition of data, interpretation of data, drafting of the article, and final approval of the article. Mahdi Mahmoudi helped in the conception and design of the study, analysis, interpretation of data, critical revision of the article, and final approval of the article. Jafar Karami is involved in the analysis, interpretation of data, drafting of the article, and final approval of the article. Nooshin Ahmadzadeh helped in the acquisition of data, interpretation of data, drafting of the article, and final approval of the article. Nasser Ebrahimi-Daryani helped in the conception and design of the study, interpretation of data, critical revision of the article, and final approval of the article. Nima Rezaei helped in the conception and design of the study, interpretation of data, critical revision of the article, and final approval of the article. A P value less than 0.05 was statistically significant. Genotype frequency less than 2% is not included. 7 Journal of Immunology Research