Susceptibility role of soluble HLA-G and HLA-G 14-bp insertion/deletion polymorphism in inflammatory bowel disease

Inflammatory bowel disease (IBD) is a group of chronic inflammatory disorders of the gastrointestinal tract. It is fundamentally related to a dysregulated immune response in the intestinal mucosa against microbiota in genetically predisposed individuals. Among the genetic and immunological factors that are suggested to have role in etiology and pathogenesis of IBD are human leukocyte antigen (HLA)-G molecules. Therefore, soluble HLA-G (sHLA-G) serum level and genetic association with HLA-G 14-bp insertion (Ins)/deletion (Del) polymorphism was analyzed in 100 IBD patients; 50 ulcerative colitis (UC) and 50 Crohn’s disease (CD), and 100 controls. sHLA-G level was significantly elevated in IBD patients compared to controls (174.7 ± 27.1 vs. 126.8 ± 15.1; corrected probability [pc] < 0.001). The level was also elevated in UC patients compared to CD patients but the difference was not significant (180.5 ± 27.1 vs. 168.9 ± 26.3; p = 0.059). Receiver operating characteristic analysis confirmed the significance of sHLA-G in total IBD, UC, and CD patients (area under curve = 0.944, 0.961, and 0.927, respectively). The genetic association was analyzed under five genetic models (allele, recessive, dominant, overdominant, and codominant). At the allele level, Del allele frequency was significantly increased in total IBD patients (Odds ratio [OR] = 1.93; 95% confidence interval [CI] = 1.27–2.94; pc = 0.018) and CD patients (OR = 2.08; 95% CI = 1.23–3.54; pc = 0.042) compared to controls. Among UC patients, a similar increased frequency was observed, but the pc value was not significance (OR = 1.79; 95% CI = 1.07–3.00; p = 0.031). At the genotypic level, Del/Del genotype was associated with a significantly increased IBD-risk in total patients under codominant model (OR = 4.06; 95% CI = 1.56–10.56; pc = 0.024). sHLA-G level was not influenced by the Ins/Del polymorphism. This study demonstrated a significant increase in serum level of sHLA-G in UC and CD patients. Further, HLA-G 14-bp Ins/Del polymorphism may be associated with susceptibility to IBD, particularly CD.


Background
Inflammatory bowel disease (IBD) is a group of chronic inflammatory disorders of the gastrointestinal tract. It is fundamentally related to a dysregulated immune response in the intestinal mucosa against microbiota in genetically predisposed individuals [1]. Two major phenotypes of IBD are clinically recognized: ulcerative colitis (UC) and Crohn's disease (CD). The inflammation in UC is restricted to the colon and rectum, while in CD, it involves any part of the gastrointestinal tract in a noncontinuous fashion [2]. Their clinical presentations are similar, but diarrhea and bleeding are most likely presented in UC, whereas watery diarrhea and vague symptoms are encountered in CD [3]. Etiology of both phenotypes is incompletely understood, but complex interactions between genetic, environmental, immunological, and gut microbiomial factors are suggested to be involved. Their interactions orchestrate a cascade of inflammatory responses in the intestinal mucosa [4]. Among the genetic and immunological factors that are proposed to have a role in etiology and pathogenesis of IBD are human leukocyte antigen (HLA)-G molecules, which are encoded by the non-classical major histocompatibility (MHC) class I genes that are mapped to the short arm of chromosome 6 (6p21.3) [5][6][7].
HLA-G molecules are characterized by restricted tissue expression. Under non-pathological conditions, they are principally expressed on the extra-villous cytotrophoblast cells and tissues at the feto-maternal interface [8]. Functionally, HLA-G molecules create a state of tolerance to protect the fetus from the maternal immune response against the fetal semi-allograft [9]. The tolerogenic properties of HLA-G molecules are augmented through interactions with cell-surface receptors present on natural killer (NK) cells, T and B cells and antigenpresenting cells [10]. However, upregulation of HLA-G has also been demonstrated under pathological conditions, especially in diseases characterized by inflammation. Therefore, it has been suggested that these molecules enhance immune surveillance as a possible mechanism of tissue protection against inflammatory responses [11][12][13]. Seven isoforms of HLA-G molecules are recognized (HLA-G1-G7). The first four isoforms (G1-G4) are cellsurface bound proteins, while G5-G7 are soluble proteins (sHLA-G) [8].
The HLA-G gene is presented with low allelic polymorphism compared other classical HLA-class I molecules. Up to April 2020, 78 alleles have been recognized at the DNA level and 21 alleles at the protein level, as well as, four null alleles have also been registered [14]. In addition, HLA-G gene exhibits a further polymorphism, which is 14 base-pairs (14-bp) insertion (Ins)/deletion (Del) at the 3′ untranslated region (3′ UTR) of exon 8. Investigations have linked this polymorphism to post-transcriptional regulation of HLA-G molecules, and thus its effect on the stability of mRNA has been proposed [15,16].
The role of HLA-G molecules in pathogenesis of IBD or their effect on susceptibility to disease has not been wellelaborated. In 2004, the expression of HLA-G in intestinal biopsies was analyzed in UC and CD patients. It was found that HLA-G was only expressed in biopsies of UC patients, whereas CD samples showed no expression [17]. In a further study, the level of sHLA-G was determined in supernatants of cultured peripheral blood mononuclear cells (PBMCs) obtained from 18 UC and 10 CD patients. The results showed different expression of sHLA-G between the two phenotypes of IBD, and a spontaneous secretion of sHLA-G was depicted in CD patients but not in UC or healthy subjects [18]. In a more recent immunohistochemical study, expression of HLA-G was noticed in UC and CD biopsies; however, the expression of HLA-G5 isoform was differentially correlated with levels of tumor necrosis factor-α (TNF-α) and interleukin-10 (IL-10) [7]. With regard to HLA-G 14-bp Ins/Del polymorphism, two previous studies suggested the significance of this polymorphism in IBD risk [19,20].
Due to the growing importance of HLA-G molecules in etiology and pathogenesis of IBD (UC and CD), this study analyzed serum level of sHLA-G in UC and CD of Iraqi Arab patients categorized according to gender, cigarette-smoking status, disease duration, family history, disease extension and medication. The genetic association of HLA-G 14-bp Ins/Del polymorphism with both phenotypes of IBD was also analyzed.

Populations studied
During January-June 2019, a case-control study was conducted on 100 Iraqi Arab IBD patients (50 UC and 50 CD). Ethnicity matched 100 healthy subjects were taken as controls. The patients attended the outpatient gastrointestinal clinics at Al-Kindy Teaching Hospital, Baghdad Teaching Hospital and Gastroenterology and Hepatology Teaching Hospital in Baghdad for diagnosis and treatment. The diagnosis was based on standard clinical, radiological, endoscopic, and histopathological criteria [3]. Patients with indeterminate colitis or other related autoimmune diseases were excluded. The patients were stratified according the following parameters: age, gender, current cigarette-smoking status, disease duration, family history, laboratory findings (hemoglobin; Hb, white blood cell count; WBC and erythrocyte sedimentation rate; ESR), disease extension, symptoms, extra-intestinal manifestations, and medication ( Table 1). The control sample included blood donors who were healthy and there serum profile for anti-pathogen antibodies was negative (Central Blood Bank, Baghdad). All participants gave  written consent to participate in the study. The approval of Ethics Committees at the target hospitals was obtained to conduct the study (N264 on 13/01/2019). All participants (50 UC and 50 CD and 100 controls) were genotyped for HLA-G 14-bp Ins/Del polymorphism, while serum level of sHLA-G was determined in 90 randomly selected participants (30 UC, 30 CD, and 30 controls).

Determination sHLA-G
Serum level of sHLA-G was determined using human HLA-G enzyme-linked immunosorbent assay kit (catalog no.: E-EL-H1663) according to instructions of manufacturer (Elabscience, China).

Statistical analysis
Data were statistically analyzed using the statistical package IBM SPSS Statistics 25.0 (Armonk, NY: IBM Corp.). Categorical variables were given as numbers and percentages. Continuous variables were tested for normality distribution (Kolmogorov-Smirnov and Shapiro-Wilk tests). All continuous variables were normally distributed; therefore, they were presented as mean ± standard deviation (SD). Pearson's chi-squared and Fisher's exact tests were used to compare categorical variables. Oneway analysis of variance (ANOVA) was used to compare continuous variables followed by the post hoc test LSD (least significant difference). Receiver operating characteristic (ROC) was used to determine area under curve (AUC), sensitivity, specificity, and cut-off value. Direct gene counting method was used to estimate allele frequencies of Del/Ins 14-bp polymorphism. Assumption of Hardy-Weinberg equilibrium (HWE) was tested using Pearson's chi-squared goodness-of-fit test. Logistic regression analysis was employed to calculate odds ratio (OR) and 95% confidence interval (CI) under five genetic models (allele, recessive, dominant, overdominant and codominant). A probability (p) value ≤ 0.05 was considered significant after applying Bonferroni correction (corrected p; pc). G*Power software (version 3.1.9.4) was used to estimate power of sample size.

Power of sample size
Power of sample size (1-β error probability) was estimated for UC and CD. At an α error of probability of 0.05 and an effect size of 0.3, the power for a sample size of 50 cases was 0.71, which was below the marginal power of 0.8. However, merging UC and CD cases in one group (IBD) elevated the power to 0.93.

Baseline characteristics of populations studied
Mean age and gender distribution showed no significant variation between IBD patients (UC and CD) and control. However, most of UC and CD patients were cigarette-smokers (80.0 and 70.0%, respectively), while among control, smokers accounted for 39.0%. Such difference was statistically significant (p < 0.001). There were no significant differences between UC and CD patients regarding disease duration (≤ 3 and > 3 years), positive family history, laboratory findings (Hb, WBC, and ESR), symptoms (abdominal pain, diarrhea and fever), and extra-intestinal manifestations (aphthous ulcer, arthralgia, skin ulcer, appendectomy, bowel stricture, colostomy, fistula, and hemorrhoids). With regard to disease extension, UC patients were classified as ulcerative proctitis (40%), left-sided colitis (30%), and extensive colitis (30%), while most of CD patients were classified under ileocecal colitis (86%). For medication, four protocols were used: antiinflammatory and immunosuppressive, anti-inflammatory and antibiotics, anti-inflammatory, antibiotics and immunosuppressive, and immunosuppressive only. Patients who received the immunosuppressive medicines adalimumab or infliximab were divided into four groups according to the number of doses received: 0, 1-10, 11-20, and 21-30 doses (Table 1).

Serum level of sHLA-G
Serum level of sHLA-G was significantly elevated in IBD patients compared to controls (174.7 ± 27.1 vs. 126.8 ± 15.1; pc < 0.001). The level was also elevated in UC patients compared to CD patients but the difference was not significant (180.5 ± 27.1 vs. 168.9 ± 26.3; p = 0.059) ( Table 2). ROC analysis confirmed the significance of sHLA-G antigens in total IBD patients, as well as UC and CD patients (AUC = 0.944, 0.961, and 0.927, respectively) (Fig. 1).  Table 3).

Frequency of HLA-G 14-bp Ins/Del polymorphism in controls
Genotype frequencies of HLA-G 14-bp Ins/Del polymorphism in controls of this study were in a good agreement with HWE equilibrium. The allele Del was more frequently observed than Ins allele (59 vs. 41%). However, allele and genotype frequencies of the HLA-G 14bp Ins/Del polymorphism showed significant variation compared with other world populations (Table 4).

HLA-G 14-bp Ins/Del polymorphism-disease association
Five genetic models (allele, recessive, dominant, overdominant, and codominant) were adopted to assess the genetic association of HLA-G 14-bp Ins/Del polymorphism with IBD, UC and CD. At the allele level (Del vs. Ins), frequency of Del allele frequency was significantly increased in total IBD patients (OR = 1.93; 95% CI = 1.27-2.94; pc = 0.018) and CD patients (OR = 2.08; 95% CI = 1.23-3.54; pc = 0.042) compared to controls. In UC, a similar increased frequency was observed, but the significance was lost when the p-value was corrected (OR = 1.79; 95% CI = 1.07-3.00; p = 0.031). At the genotypic level, none of the adopted genetic models revealed variation with a corrected significant p value between patients and controls. The codominant model was an exception in total IBD patients. The Del/Del genotype vs. Ins/Ins maintained a significantly corrected p value (OR = 4.06; 95% CI = 1.56-10.56; pc = 0.024) ( Table 5).

Impact of HLA-G 14-bp Ins/Del polymorphism on sHLA-G level
Serum level of sHLA-G antigens was inspected in the three genotypes of HLA-G 14-bp Ins/Del polymorphism of IBD patients (UC and CD) and controls, and no significant impact of genotypes was recorded (Table 6).

Discussion
Among the characteristics that showed significant variation in UC and CD compared to controls was cigarettesmoking. Most of the patients were smokers (80 and 70%, respectively). This may underline a casual relation between IBD and cigarette-smoking; however, the literature disclosed that smoking has a dichotomous effect in UC and CD [38]. In CD, it has been established that smoking is a prominent environmental risk factor, and has a significant impact on disease development among smokers. Moreover, increased risk of relapse, steroid requirements, and hospitalization were associated with tobacco exposure. Conversely, cigarette-smoking is considered an environmental protective factor in UC, and a reduced risk of disease has been reported in active smokers. However, a reduced risk of developing UC has also been identified in patients who have never smoked. Therefore, the association between cigarette-smoking and UC is not yet fully understood [39].
The principal aim of present study was to understand the role HLA-G molecules in etiology and pathogenesis of UC and CD. Two approaches were adopted. In the first, serum level of sHLA-G was determined. In both IBD phenotypes, the level was significantly elevated in patients compared to controls. Further, ROC analysis depicted that sHLA-G occupied a significant AUC in IBD, but was higher in UC than in CD (0.961 vs. 0.927). Accordingly, it is suggested that sHLA-G molecules are risk factors with detrimental effect on the pathological course of UC and CD. There is no direct evidence to support these findings and previous studies investigating sHLA-G role in pathogenesis of UC or CD have not been well-elaborated. Consistent with this study, a significantly elevated level of sHLA-G was reported in CD patients compared to controls [20]. However, in an earlier study, the mucosal expression of HLA-G was detected in UC samples, while biopsies of CD patients did not show the expression of HLA-G molecules [17], whereas, a recent study demonstrated their gene and protein expression in biopsies of UC and CD patients [7]. In an in vitro study, the supernatant of cultured PBMCs obtained from UC and CD patients was examined for sHLA-G molecules using ELISA. This time, UC patients showed no detectable level of sHLA-G, while  the supernatant of CD patients showed spontaneous production of sHLA-G [18]. Further accomplishing data suggested that immunosuppressant therapies are able to normalize the production of sHLA-G in CD, while in UC patients such therapy initiates their release. This effect on sHLA-G release was not influenced by age and gender of patients, as well as, disease activity, localization, or duration [40]. In this study, medication showed no significant effect on sHLA-G level. However, there was a tendency for sHLA-G to have a decreased level in UC patients who received immunosuppressive therapy, and the decrease paralleled the number of therapy doses. In CD patients, the response to medication was different. Patients treated with protocol C (anti-inflammatory, antibiotics and immunosuppressive) showed the highest level of sHLA-G among the other medication groups. Further, the increased level of sHLA-G paralleled the number of immunotherapy doses (the level was Significant p is bold-marked SD standard deviation, UC ulcerative colitis, CD Crohn's disease, protocol A anti-inflammatory and immunosuppressive, protocol B anti-inflammatory and antibiotics, protocol C anti-inflammatory, immunosuppressive, and antibiotics, protocol D immunosuppressive only, p LSD (least significant) or one-way ANOVA (analysis of variance) probability, pc Bonferroni-corrected probability, NS not significant (pc > 0.05)   increased in patients who received 21-30 doses). These findings indicted that UC and CD patients responded differently to the administrated therapy. Previous in vitro study also suggested that immunosuppressant therapies may act differently in UC and CD patients [40]. Irrespective of these inconsistencies, this study points to the role for sHLA-G molecules in pathogenesis of UC and CD. Both phenotypes presented with upregulated inflammatory responses, and sHLA-G molecules may encounter the exaggerated inflammatory response due to their immunomodulating effects [41]. The HLA-G molecules may protect local tissue by inhibiting the activity of NK cells and may also control the balance between T helper 1 and 2 cells in mucosa tissues [5].
In the second approach, allele and genotype frequencies of HLA-G 14-bp Ins/Del polymorphism were analyzed in IBD patients (UC and CD) and controls. As in other world populations, both alleles (Ins and Del) showed polymorphic frequencies in the Iraqi sample of controls (41 and 59%, respectively). In most of the populations investigated (Table 4), Ins was the allele of minor frequency and had a range of 23.6% in Koreans [31] to 49.2% in Saudi Arabians [34] and Turkish [37]; however, Egyptians [25] and Tunisians [36] showed the opposite profile, and the Ins allele frequency exceeded this range (60 and 51.2%, respectively). Therefore, the Iraqi Ins allele frequency fits well the presented range. However, the comparison between patients and controls depicted variations in allele and genotype frequencies of HLA-G 14-bp Ins/Del polymorphism. In total IBD patients, as well as CD patients, logistic regression analysis suggested a risk potential of the allele Del. In UC patients, a similar conclusion was reached but the difference was significant before correction of p value. In Caucasian German patients, allele and genotype frequencies of HLA-G 14bp Ins/Del polymorphism showed no significant variations between UC or CD patients and controls, but there were significant differences between UC and CD patients. The heterozygous genotype and the Del+ phenotype were significantly elevated in UC patients compared to CD patients [19]. In a Tunisian study, neither allele nor genotype frequencies of HLA-G 14-bp Ins/Del polymorphism showed significant differences between CD patients and controls. However, the distribution of patients as young-and adult-onset showed significantly increased frequencies of Ins allele and Ins/ Ins genotype in young-onset patients but not adult-onset patients [20]. To the best knowledge of investigators, no further study was performed to explore the association of HLA-G 14-bp Ins/Del polymorphism with UC or CD. Although conflicting results were obtained, a possible risk effect of the HLA-G gene in both phenotypes of IBD is suggested.
This study also addressed that the upregulated serum expression of sHLA-G in UC and CD patients might be influenced by HLA-G 14-bp Ins/Del variant. The analysis revealed that sHLA-G level was similarly distributed in the three genotypes of patients or controls, and no significant influence was observed. However, a previous study suggested inconsistent functional relevance, and the exon 8 polymorphism (14-bp Ins/Del) may influence the transcription activity of HLA-G gene [15]. The authors observed that HLA-G mRNAs harboring the 92base deletion showed more stability than complete forms of mRNA. Accordingly, it was suggested that regulation of HLA-G post-transcription is controlled by this region. However, it has been reviewed that presence of Del allele in homozygous and heterozygous genotypes is correlated with higher production of sHLA-G (high-producer genotypes). A similar observation was made in CD patients, and Del/Del and Del/Ins genotypes were associated with an increased plasma level of sHLA-G, but no statistical significance was attended [20].

Conclusions
This study demonstrated a significant increase in serum level of sHLA-G in UC and CD patients. Further, HLA-G 14-bp Ins/Del polymorphism may be associated with susceptibility to IBD, particularly CD. However, the study was limited by low number of UC and CD patients, and increasing the sample size will certainly contribute to a further understanding of HLA-G role in immunopathogenesis of both IBD phenotypes.