Background

Inflammatory bowel disease (IBD) has recently been viewed as a gastrointestinal disorder with an increasing burden worldwide [1]. The disease is characterized by chronic inflammation of the gastrointestinal tract due to exaggerated inflammatory responses against intestinal microbiota in genetically susceptible individuals [2]. In most cases, two clinical phenotypes of IBD are recognized: ulcerative colitis (UC) and Crohn’s disease (CD) [3]. In UC, the inflammation occurs in the colonic mucosa starting from the rectum and extending proximally in a continuous manner to involve part of, or the entire, colon. CD is characterized by transmural inflammation involving any part of the gastrointestinal tract [4]. The exact etiology of IBD is still not precisely defined and its pathogenesis not well elucidated. Nevertheless, it is suggested that interactions between intestinal microbes, environmental factors and host genetic predisposition are required to trigger irregular immune responses that are participated in provoking chronic relapsing and remitting inflammation [2].

There is overwhelming evidence drawn from population-based studies that genetic factors play an important role in pathogenesis of IBD [5]. The risk of contracting the disease increases in blood-relatives of patients, and the disease concordance rate is higher among monozygotic twins than among dizygotic twins [6]. Further, genome-wide association studies have described the importance of multiple genetic loci in increasing the risk of IBD, particularly those involved in modulating the inflammatory response [7]. Studies in patients of European and non-European ancestry have identified more than 200 IBD susceptibility loci that are proposed to have a functional role in the intestinal inflammatory immune response [8, 9]. Among these loci are genes located in a region of human chromosome 6 (6p21) termed inflammatory bowel disease 3 (IBD3). The region is of interest at the genome level because it encompasses genes of the classical and non-classical human leukocyte antigen (HLA) system, as well as HLA-class III genes, such as tumor necrosis factor (TNF)-α gene (TNFA) [10]. HLA gene products play a major functional role in immune responses. Further, HLA alleles have shown associations with susceptibility to different diseases; especially those related to immune dysregulation [11]. Besides, patients with IBD are at a greater risk of developing autoimmune and inflammatory diseases that have established HLA-associations; for instance, psoriasis, ankylosing spondylitis, type 1diabetes, and multiple sclerosis [12]. HLA-G is one such locus, the products of which play a major role in modulating immunity due to their tolerogenic properties [13].

HLA-G belongs to the non-classical class I HLA gene family, and is located within the major histocompatibility complex (MHC) in the chromosomal region 6p21.3 [13]. HLA-G was first described to play a major role in maintaining tolerance in the maternal-fetal interface [14]. Subsequently, HLA-G was shown to be involved in modulating innate and adaptive immune responses in a variety of pathological conditions; for instance, viral infections, malignancies, autoimmune diseases, transplant outcomes, and inflammatory diseases. These results highlighted that HLA-G encodes for molecules important for immune system functions [15]. Seven isoforms of HLA-G have been identified; four of them are membrane-bound (G1–G4), and three are soluble (G5–G7) [16].

The magnitude of HLA-G gene expression is controlled by variants in the promoter (5′-upstream regulatory region; 5′ URR) and in the 3′ untranslated region 3 (3′UTR) [13]. However, the HLA-G gene is described as a polymorphic locus with low number of alleles compared to other HLA loci. To date (14 October 2020), 82 alleles have been recognized at the DNA level (https://hla.alleles.org). A further polymorphism is exhibited by the HLA-G gene; it is 14 base-pairs (14-bp) insertion (Ins)/deletion (Del) at the 3′ UTR of exon 8 [17]. Numerous studies have examined the association between HLA-G variants and various diseases, especially inflammatory and immune-mediated diseases. Despite inconsistent observations made, it is suggested that HLA-G alleles are important biomarkers for monitoring the disease predisposition and progression [15].

With respect to the role of HLA-G molecules in etiology of IBD or their influence on pathogenicity of disease, few studies have been conducted with inconsistent findings. HLA-G expression in intestinal biopsies of UC and CD patients was first analyzed in 2004. The results depicted that expression of these molecules only occurred in the biopsies of UC patients, while CD samples showed no expression [18]. In a subsequent in vitro study, spontaneous secretion of sHLA-G was found in supernatant of cultured peripheral blood mononuclear cells (PBMCs) obtained from CD patients, while PBMCs obtained from UC patients or healthy control showed no secretion of sHLA-G [19]. A more recent immunohistochemical study demonstrated increased HLA-G expression in plasma cells/lymphocytes infiltrating the lamina propria in biopsies of UC and CD patients compared to controls, but greater cell staining was found in UC cells compared to CD cells [20]. In a study by our group, serum level of sHLA-G was significantly elevated in UC and CD patients compared to controls [21].

HLA-G polymorphisms have not been well investigated in IBD. Two previous studies suggested the importance of HLA-G 14-bp Ins/Del polymorphism in susceptibility to IBD [22, 23]. A third study by our group also favored a similar suggestion [21]. To the best knowledge of investigators, other HLA-G variants have not been investigated in IBD. In this study, the genetic association between three alleles of HLA-G gene (G*01:03, G*01:04 and G*01:05N) and susceptibility to IBD (UC and CD) was analyzed in Iraqi patients.

Methods

Populations investigated

A case-control study was conducted during January–June 2019 to investigate the genetic association of HLA-G alleles (G*01:03, G*01:04 G*01:05N) with two phenotypes of IBD; UC and CD. One hundred patients diagnosed with IBD (50 UC and 50 CD) were recruited from outpatient clinics in three hospitals in Baghdad (Al-Kindy Teaching Hospital, Baghdad Teaching Hospital and Gastroenterology and Hepatology Teaching Hospital). Standard clinical, radiological, endoscopic and histopathological criteria were followed in the diagnosis of UC and CD [24]. Patients presented with indeterminate colitis or other gastrointestinal autoimmune diseases were excluded. Data related to age, gender, cigarette-smoking, disease duration, family history of IBD, laboratory findings (hemoglobin; Hb, white blood cell count; WBC, erythrocyte sedimentation rate; ESR and sHLA-G), disease extension (UC: ulcerative proctitis, left-sided colitis and extensive colitis; CD: ileocecal colitis and ileocecal + colon), symptoms (abdominal/colon pain, diarrhea, and fever), and extra-intestinal manifestations (aphthous ulcer, arthralgia, skin rash, appendectomy, bowel stricture, colostomy, fistula, and hemorrhoid) were recorded for each patient (Table 1). A control sample of 100 healthy individuals (HC) matched to patients for age and gender was also included. Written consent to participate in the study was obtained from all participants. The approval of Ethics Committee in the target hospitals to conduct the study was obtained (Approval number: N264 on 13 January 2019).

Table 1 Baseline characteristics of inflammatory bowel disease (ulcerative colitis and Crohn’s disease) patients and controls
Full size table

Determination of HLA-G alleles

From each participating subject, 3 mL of peripheral blood was drawn in ethylene-diamine-tetra-acetic-acid (EDTA) tube. The blood was processed to isolate genomic DNA using DNA purification kit (Geneaid, Taiwan), and instructions of manufacturer were followed. Three alleles, HLA-G*01:03, HLA-G*01:04, and HLA-G*01:05N, were determined using sequence-specific polymerase chain reaction (PCR) assay followed by product digestion with restriction endonucleases (Hinf-I, BseR-I, and PpuM-I, respectively), as presented in Table 2 and previously described [25]. Briefly, the reaction mix (25 μL) consisted of 5 μL DNA (60 ng/mL), 12.5 μL 1× Green Master mix (Bioneer, Korea), 1 μL of each forward and reverse primer (10 pmol/μL) and 5.5 μL nuclease-free water. The optimized thermocycling conditions were initial denaturation at 94 °C (5 min), followed by 35 cycles of denaturation at 94 °C (30 s), annealing at 60 °C (30 s), and extension at 72 °C (30 s), and a final extension cycle at 72 °C (5 min). The amplified PCR products were digested with the restriction endonucleases Hinf-I (G*01:03), BseR-I (G*01:04), and PpuM-I (G*01:05N). The digestion products were electrophoresed in 3% agarose gels, and migrating bands were visualized with UV trans-illuminator after staining with ethidium bromide. Assignments of HLA-G alleles were based on migrating DNA fragments. For G*01:03, the two bands defining the allele were not observed (79 and 125 bp), and instead bands of 106 and 175 bp were found after digestion with Hinf-I, indicating absence of G*01:03. G*01:04 was defined by a band of 276 bp after digestion with BseR-I. Finally, G*01:05N was also defined by a band of 276bp but after digestion with PpuM-I (Table 2 and Fig. 1).

Table 2 Assignment of HLA-G*01:03, HLA-G*01:04, and HLA-G*01:05N alleles [25]
Full size table
Fig. 1
figure 1

Agarose gel electrophoresis of HLA-G gene PCR products in 3% agarose at 5 V/cm2. a PCR products digested with Hinf-I showing two bands (106 and 175 bp); Lanes 1–14: absence of HLA-G*01:03 allele. b PCR products digested with BseR-I showing three bands (40, 236, and 276 bp bands); lanes 1 and 8: homozygous HLA-G*01:04 allele; lanes 7 and 11: heterozygous HLA-G*01:04 allele; lanes 2–6, 9–10, and 12–13: absence of HLA-G*01:04 allele. c PCR products digested with PpuM-I showing three bands (108, 168, and 276 bp bands); lanes 1-3: homozygous HLA-G*01:05N allele; lanes 7–8 and 10: Heterozygous HLA-G*01:05N allele; lanes 4–6, 9, and 11–13: absence of HLA-G*01:05N allele. Lane M: 100 bp DNA ladder. The bands were visualized using UV trans-illuminator after staining with ethidium bromide

Full size image

Statistical analysis

Baseline data were given as mean ± standard deviation (continuous variables) or number and percentage (categorical variables). Least significant difference (LSD) was used to assess significant difference between continuous variables, while two-tailed Fisher’s exact or Pearson’s chi-square test was used for categorical variables. HLA-G alleles (G*01:03, G*01:04 and G*01:05N) were given as number and percentage. Logistic regression analysis was used to determine genetic association of HLA-G allele with UC or CD. The association was expressed as odds ratio (OR) and 95% confidence interval (CI). Two-tailed Fisher’s exact probability (p) was used to assess statistical significance of association. Bonferroni correction was applied to adjust the p value due to multiple comparisons. A corrected p (pc) ≤ 0.05 was considered significant. The statistical package IBM SPSS Statistics 25.0 (Armonk, NY: IBM Corp.) was employed to carry out these analyses. Power of sample size was estimated using G*Power software (version 3.1.9.4).

Results

Power of sample size (PSS)

At 0.05 α error of probability and 0.3 effect size, the PSS (1–β error probability) of 50 cases was 0.71, which was below the marginal limit (0.80). When UC and CD cases were combined into a single group (i.e. IBD), the PSS was raised to 0.93. Thus, statistical verification of sample size was proposed.

Baseline data of patients and controls

UC and CD patients matched HC for age and gender distributions, and no significant differences were recorded. In addition, no significant variations between UC and CD patients were observed regarding family history of IBD, disease duration, symptoms or extra-intestinal manifestations. For cigarette-smoking, 80 and 70% of UC and CD patients were smokers, respectively compared to 39% among HC (p < 0.001). In the case of laboratory findings, Hb, WBC, ESR, and sHLA-G means showed no significant variations between UC and CD patients; however, sHLA-G mean was significantly elevated in both IBD phenotypes compared to HC (Table 1).

HLA-G alleles

The G*01:03 allele was not detected in IBD patients (UC and CD) or HC, while G*01:04 and G*01:05N alleles showed polymorphic frequencies, and their distribution in IBD patients (total, UC, or CD) showed significant variations compared to HC. Whereas, the comparison between UC and CD patients revealed no significant differences in the distribution of G*01:04 and G*01:05N allele frequencies. The allele G*01:04 was significantly associated with susceptibility to UC (OR = 2.55; 95% CI = 1.27–5.13; pc = 0.018) and CD (OR = 4.45; 95% CI = 2.11–9.41; pc < 0.001), but the risk was higher in CD than in UC. Frequency of G*01:05N allele also showed a significantly elevated frequency in UC (18 vs. 5%; OR = 4.17; 95% CI = 1.32–13.21; pc = 0.032) and CD (20 vs. 5%; OR = 4.75; 95% CI = 1.53–14.78; pc = 0.014) patients compared to HC. These associations were more pronounced in IBD group (UC + CD), and a significantly increased risk for IBD was found with the alleles G*01:04 (OR = 3.32; 95% CI = 1.86–5.95; pc < 0.001) and G*01:05N (OR = 4.46; 95% CI = 1.59–12.47; pc = 0.008) (Tables 3 and 4).

Table 3 Frequency of HLA-G alleles in inflammatory bowel disease patients (ulcerative colitis and Crohn’s disease) and controls
Full size table
Table 4 Logistic regression analysis of HLA-G*01:04 and HLA-G*01:05N alleles in inflammatory bowel disease patients (ulcerative colitis and Crohn’s disease)
Full size table

Stratification of HLA-G alleles according to characteristics of patients

G*01:04 and G*01:05N alleles were stratified in total IBD patients according to the demographic and clinical characteristics given in Table 1. The frequency distribution of the two alleles showed no significant differences between the subgroups of patients in each stratum (Tables 5 and 6).

Table 5 HLA-G*01:04 and HLA-G*01:05N frequencies in inflammatory bowel disease patients classified according to demographic and clinical characteristics
Full size table
Table 6 HLA-G*01:04 and HLA-G*01:05N frequencies in inflammatory bowel disease patients (ulcerative colitis and Crohn’s disease) classified according to disease extension
Full size table

Impact of HLA-G alleles on sHLA-G

The data on sHLA-G serum level were obtained from an article previously published by our group in this journal [21]. The level did not show significant differences between participants (patients or controls) who had the G*01:04 or G*01:05N allele and those who did not have the allele. UC was an exception, and the presence of G*01:04 allele was associated with a significantly higher mean of sHLA-G compared to patients without the allele (189.6 ± 24.0 vs. 168.6 ± 27.2 ng/mL; p = 0.033) (Fig. 2).

Fig. 2
figure 2

Means of soluble HLA-G level distributed according to absence and presence of HLA-G*01:04 and HLA-G*01:05N allele in ulcerative colitis (UC) and Crohn’s disease (CD) patients and healthy controls (HC). p: Least significant difference probability (significant p value is indicated in bold)

Full size image

Discussion

In a previous study, we showed that serum level of sHLA-G was significantly elevated in current samples of UC and CD patients compared to HC. Further, the genetic association with an exon 8 polymorphism of HLA-G gene (14-bp Ins/Del polymorphism) was also analyzed, and the results indicated the susceptibility potential of this polymorphism to IBD [21]. These findings motivated us to extend the analysis of HLA-G gene in IBD to include three alleles (G*01:03, G*01:04 and G*01:05N). The G*01:03 allele was not evident in IBD patients or controls, while G*01:04 or G*01:05N were determined with polymorphic frequencies. With respect to G*01:03, the global estimated allele frequency was 6.3% (range 0.0% in Iberian populations from Spain and Han Chinese from South China to 11.9% in people of African ancestry from the southwestern USA) [17]. Prakash and colleagues also reported that G*01:03 was not recorded in populations from India, South Korea, Poland, Spain, Ghana, and Zambia, but in Denmark, a remarkably high frequency was found (43%) [26]. For G*01:04 and HLA-G*01:05N, their allele frequencies were higher than the estimated global frequencies (39 and 5% vs. 17.3 and 3.3%, respectively) [17]. In Iraq, frequency of the three alleles was only available for females; it was 5.2, 17.0, and 9.4%, respectively [27]. Inbreeding could theoretically lead to elevated frequencies of G*01:04 and G*01:05N alleles, because Iraqis generally obey some traditional practices (i.e., high prevalence of consanguineous marriage), which may influence HLA allele frequencies [28].

The present study sought to investigate the role of G*01:04 and G*01:05N alleles in genetic susceptibility to IBD in UC and CD patients. Suggestive evidence for the association of both alleles with susceptibility to UC and CD was presented by the study. It was observed that G*01:04 and G*01:05N alleles were associated respectively with 2.55- and 4.17-fold higher risk of susceptibility to UC. Similar observations were made in CD patients, and the corresponding ORs were 4.45 and 4.75, respectively. Further, no significant differences between UC and CD patients were found regarding frequencies of the two alleles. Therefore, it is possible to suggest that G*01:04 and G*01:05N alleles are involved in a common pathogenic mechanism between UC and CD. To highlight this suggestion, data were reanalyzed considering UC and CD in one group (i.e. IBD). This time, the association between the two alleles and susceptibility to IBD was sustained, and the disease risk increased by 3.32 and 4.46 times, respectively. Accordingly, G*01:04 and G*01:05N alleles may represent novel biomarkers for both IBD phenotypes (UC and CD).

G*01:04 has not been well screened in inflammatory diseases, but it has been reported to be coincided with Del allele of the 14-bp Ins/Del polymorphism at the 3′ UTR of exon 8 in HLA-G gene, and is associated with high serum level of sHLA-G [29]. The present study demonstrated that serum level of sHLA was significantly increased in UC patients carrier for G*01:04 allele compared to patients without the allele. The Del allele has also been associated with susceptibility to IBD, particularly CD, in Iraqi patients [21]. Thus, both alleles (G*01:04 and Del) could act synergistically in the context of susceptibility to inflammatory diseases due to their association with immunosuppressive functions of natural killer (NK) cells. Recently, G*01:04 has been shown to be a powerful catalyst in decidual NK cell activation. It also exerts protective effects against NK cell-mediated lysis. These findings underlie the exceptional role of G*01:04 as a mediator of immune tolerance [30]. In an earlier study, it was found that women with HIV-1 infection and bacterial vaginosis and carrying G*01:04:04 allele expressed the highest levels of genital HLA-G molecules [31]. In lung inflammatory diseases, such as cystic fibrosis, G*01:04 was associated with lower survival rates and higher frequency of chronic rejection after lung transplantation [32]. It was also reviewed that G*01:04 showed elevated frequency in couples with unexplained recurrent miscarriages, while it exhibited a protective role in one of the chronic inflammatory diseases (acute renal rejection and end-stage kidney disease) [33]. It has also been concluded that HLA-G haplotypes that include G*01:04 allele may be a good candidate marker for inflammation in asthma patients [34]. Together, these findings enforce the susceptibility role of G*01:04 in inflammatory diseases including IBD.

The null allele G*01:05N was also significantly associated with risk of UC and CD in current samples of Iraqi patients. The allele has not been investigated in IBD, but its susceptibility role in etiology of other human ailments has been suggested; for instance, Behçet’s disease, autistic spectrum disorders, recurrent pregnancy loss, and preeclampsia [35,36,37]. The G*01:05N allele represents a single base deletion causing a frame-shift reading that leads to a premature stop codon at the beginning of exon 4. As a consequence, this allele is associated with incomplete formation of the HLA-G isoforms G1, G4, and G5 that possess the α3 domain, while G2, G3, and G7 isoforms (lack the α3 domain) show normal expression and sustain the immune tolerogenic function of HLA-G [15, 38]. The allele is also associated with low serum level of sHLA-G, and unlike G*01:04, it coincided with Ins allele of the 14-bp Ins/Del polymorphism [39]. The Ins allele has been considered a risk factor for celiac disease (IBD-related disease) [40]. Both conditions are characterized by chronic intestinal inflammation, and may share similar etiology and immunopathogenesis [41]. Another study reported a positive association between G*01:05N and an infectious disease (HIV infection) in adult Caucasian females, and the authors suggested that this allele could be considered a marker of susceptibility or it might have a direct functional effect on the development of HIV infection [42].

Collectively, these data indicate that HLA-G plays an important role in regulating the inflammatory response in UC and CD. It has been indicated that HLA molecules exert immune-suppressive effects on different immune cells, including CD4+ and CD8+ T cells, NK cells, antigen-presenting cells, and B cells. In addition, they participate in inducing T regulatory cells and IL-10 producing dendritic cells [43]. In this context, it has been demonstrated that CD4+CD25+ T cells were elevated in UC patients with primary sclerosing cholangitis [44]. Two SNPs of HLA-G (rs66554220 and rs1063320) were also associated with low frequency of regulatory CD8+CD28- T cells [45]. Further studies indicated that HLA-G locus have been associated with higher susceptibility to UC and CD or higher severity of the two diseases [21, 23, 46]. HLA-G has also been indicated as predictive marker of response to therapy in inflammatory and infectious diseases, as well as cancer [47]. Thus, it is possible that these immunological and genetic signatures of HLA-G contribute to the susceptibility and persistence of UC and CD. However, this study might have provided preliminary evidence to suggest that HLA-G polymorphisms play a role in susceptibility to both phenotypes of IBD. Further studies are warranted to confirm or refute these results and to evaluate the role of G*01:04 and G*01:05N alleles in etiology and pathogenesis of IBD. It should be noted that although statistical validation of sample size was established, the study was still limited due to lower sample size and better evaluation of HLA-G alleles in IBD should be based on a larger sample size.

Conclusions

The results of this study indicated that HLA-G*01:04 and HLA-G*01:05N alleles are associated with susceptibility to UC and CD in Iraqi patients.