The aim of this study was to assess the distribution of human leukocyte antigen (HLA) groups in Turkish children with celiac disease (CD) and to investigate the association of HLA types and clinical manifestations of CD.

Seventy-five children with CD were evaluated in two groups: Group I consisted of 45 classical celiac patients (15 males, 6.7±3.8 years); Group II consisted of 30 atypical celiac patients (9 males, 9.3±4.3 years). The control group consisted of 100 healthy renal transplantation donors. HLA typing was made serologically using standard lymphocytotoxicity techniques.

HLA A29, B51, CW5, DR14, DR16, and DQ1 were the most common antigens in the control group. Frequency of HLA B13, CW7, B8, DR7, DR17 and DQ2 was higher in CD patients than in the control group (p<0.005, <0.05, <0.001, <0.001 and <0.001, respectively). The relative risks for HLA DQ2, B8, DR17 and B13 were 14.9, 13.6, 7.1 and 3.6, respectively. Frequency of HLA B35, DR11 and DQ7 was higher in classical CD than atypical CD, while a positive association was found between HLA B8 and atypical CD.

A positive association was found between HLA B13, CW7 and DR17 in Turkish celiac patients in addition to HLA B8, DR7 and DQ2. This study also suggested that a correlation may exist between genotype and clinical manifestations.

Genetic, environmental and immunological factors may play important roles in the pathogenesis of the disease[3,5]. A significant association between CD and human leukocyte antigen (HLA) types was shown both in family and population studies. CD was first found to be associated with HLA B8. Later, it was shown that the disease is associated with expression of HLA DQ2 and HLA DQ8[6-8]. The frequency of HLA groups varies significantly in different populations. Although association of HLA group and CD has been reported in numerous studies, the influence of HLA type on the clinical manifestations of CD has not yet been investigated. The aim of this study was to assess the distribution of HLA groups in Turkish children with CD and to investigate the association of HLA type and clinical manifestations of CD.

HLA Typing
Human leukocyte antigen class I and II typing was studied serologically by the standard microlymphocytotoxicity method[6].

Statistical Analysis
The distribution of HLA types in the groups of patients was compared using χ² test. When frequency of HLA types was below 2%, Fisher’s exact test was used. P<0.05 was considered significant. In order to estimate relative risk, the odds ratio method was also used.

Table I: The Frequency of HLA-A, B and C in Celiac Disease Patients and Control Group

Table II: The Frequency of HLA-DR and DQ in Celiac Disease Patients and Control Group

In odds ratio test, HLA DQ2, B8, DR17 and B13 were demonstrated to have higher risk for development of CD (14.9, 13.6, 7.1 and 3.6, respectively) (Table III).

Table III: HLA Types Associated with Celiac Disease

Classical CD was significantly associated with HLA B35 (24.4%) (p<0.05), DR11 (36.7%) (p<0.01) and DQ7 (31.8%) (p<0.05). In atypical CD, HLA B8 (58.6%) was the most common HLA type in HLA groups (p<0.05) (Tables IV, V).

Table IV: The Frequency of HLA-A, B and C in Classical and Atypical CD

Table V: The Frequency of HLA-DR and DQ in Classical and Atypical CD

Exposure to gluten in genetically susceptible individuals is the major factor of development of CD. A gluten molecule contains two fractions: gliadin and glutenin. Gliadin fraction is known to be the main cause of CD. The possible interactions of gliadin (and/or its peptide derivates) with intestinal epithelia and the mechanism through which it crosses the epithelial barrier to reach the submucosa are not yet completely explained. After passage of gliadin into the subepithelial compartment, toxic gliadin derivates are deaminated by tissue transglutaminase. Deaminated gliadin epitopes are taken up by antigen presenting cells and are presented by MHC class II molecules HLA DQ2 and/or DQ8 to CD4+ helper T lymphocytes. Activation of T cells results in release of proinflammatory mediators causing damage of enterocytes resulting in villous atrophy[9-11]. It is thought that class II HLAs are efficient to present peptides produced from digestion of gliadin in the lamina propria of the small bowel mucosa to T cells[11,12]. It is known that most celiac patients (>90%) carry a particular DQ heterodimer encoded by DQ2 (DQA1*0501/DQB 1*0201) alleles[12,13]. Association of HLA DQ alleles with CD can explain the immunological concept of the disease pathogenesis; however, it does not explain the wide range of disease manifestations.

Our results showed that HLA B8, B13, CW7, DR7, DR17 and DQ2 were found in higher frequencies in celiac patients. Firstly, frequency of HLA B8 in celiac patients has been reported as 45-88%[14]. Later, it was observed that frequency of HLA DR3 is also increased and that the association with HLA B8 simply reflected the linkage disequilibrium between the alleles that coded for these antigens[15]. In our study, frequency of HLA B8 in CD was 41.9%, which is similar to other reports. Our study confirmed the strong association of CD and HLA DQ2. However, HLA DQ8, which is known as one of the most common antigens in CD, was not found to be frequent among our celiac patients.

The association between HLA DR3/DR7 and CD is explained by the linkage disequilibrium of these alleles with DQ2 allele[16]. Although in our study, frequency of HLA DR7 in celiac patients was significantly increased, we did not demonstrate an association between DR3 and CD. Difference of HLA distribution may be associated with genetic heterogeneity and geographic region. While in northern Europe, frequency of HLA DR3 is 95%, in southern Europe, HLA DR3 and DR7 are the most common HLAs[17,18]. It was previously reported in Turkish children by Erkan et al.[19] and Altuntaş et al.[20] that HLA A2, A9, B35, B4, DR2, DR4, DR53, CW4 and DQ4 were the most common antigens among celiac patients. According to these data, we may conclude that the HLA groups that are found to be strongly associated with CD are not seen in high frequency in the general Turkish population.

Based on recent epidemiological data, prevalence of CD is similar in Europe and North America, but clinical presentation is different. Early onset classical form is more commonly seen in Europe than in North America. It is unclear why presentation of CD differs from country to country or person to person. It is suggested that the difference in clinical presentations may be due to HLA-related predisposing genes or gluten intake[21]. Genotype-phenotype association was reported by Vogelsang et al.[22], who found that HLA B8 and CW7 might lead to late onset of CD. To our knowledge, association of HLA types and clinical manifestations has not been studied previously. In our study, an association was found between HLA types and different forms of CD. While in classical CD patients with chronic diarrhea, weight loss, abdominal distention and malnutrition, HLA B35, DR11 and DQ7 were the most frequent antigens, HLA B8 was the most frequent antigen in atypical CD patients who presented with extraintestinal manifestations.

In conclusion, this study demonstrated that HLA types are important in the development of CD, and HLA B8, CW7, DR7, DQ2, B13 and DR17 were the most frequent antigens in Turkish children with CD. Our study also showed that a relation may exist between HLA types and clinical presentations.