Influence of Common and Specific HLA-DRB1/DQB1 Haplotypes on Genetic Susceptibilities of Three Distinct Arab Populations to Type 1 Diabetes

Abstract

The contribution of HLA DRB-DQB to type 1 diabetes (T1D) in Bahrainis, Lebanese, and Tunisians was investigated. DRB1*030101-DQB1*0201 was a locus that conferred susceptibility in three populations, while DRB1*040101-DQB1*0302 was a locus that conferred susceptibility only in Tunisians and Bahrainis. The DRB1*100101-DQB1*050101 (Bahrainis) and DRB1*150101-DQB1*060101 (Lebanese) loci were largely protective. The contribution of HLA to T1D must be evaluated with regard to ethnic background.

Type 1 (insulin-dependent) diabetes (T1D) is caused by the autoimmune destruction of pancreatic β-islet cells (17). Many T1D susceptibility loci have been described and have been mapped into the IDDM1 to IDDM18 loci, of which the HLA region (identified as IDDM1) accounts for more than half of the genetic susceptibility to T1D (2, 16), whereas determination of a strong linkage disequilibrium between the DRB1 and the DQB1 alleles confirmed the contribution of specific HLA DRB1 and DQB1 alleles and DRB1-DQB1 haplotypes to the presence of T1D in many ethnic groups (7, 12, 16). This was highlighted by the association of DR3- and DR4-containing haplotypes with T1D in Caucasians (8, 10) but not Asians, including Japanese and Koreans (3, 9), in whom different haplotypes contribute to disease susceptibility (9). In view of the geographical/racial heterogeneity of Arabs (1), this study investigated the association of HLA class II (DRB1 and DQB1) haplotypes with T1D in Bahrain (Arabian Peninsula), Lebanon (eastern Mediterranean), and Tunisia (North Africa), in particular with regard to the identification of the specific haplotypes that confer susceptibility to and protection from T1D in each community.

Study subjects comprised unrelated Arab subjects from Tunisia (50 patients, 50 controls), Bahrain (126 patients, 126 controls), and Lebanon (78 patients, 111 controls). Non-Arabs (Berbers and Europeans in Tunisia, Armenians in Lebanon, and Iranians in Bahrain) or recently naturalized subjects, identified through personal interviews, were not included. A diagnosis of T1D was made according to clinical features and laboratory findings, and patients with other forms of diabetes (type 2 diabetes, latent autoimmune diabetes in adults, maturity-onset diabetes of the young, etc.) were excluded. The control subjects had normal fasting/random glucose levels and no family history of T1D or other autoimmune diseases. All subjects (patients and controls) were asked to sign a consent form according to the study protocol, and all institutional ethics requirements were met.

Total genomic DNA was extracted from EDTA-anticoagulated blood by the phenol-chloroform method. HLA-DRB1 and -DQB1 alleles were analyzed by a PCR sequence-specific priming (PCR-SSP) technique with an SSP2L HLA class II genotyping kit (One Lambda, Thousand Oaks, CA), according to the manufacturer's specifications. The frequencies of the most frequent haplotypes were determined by the maximum-likelihood method with Arlequin (version 2.000) population genetics data analysis software. Data were expressed as P values, odds ratios (ORs), and 95% confidence intervals (CIs) between the patients and the controls in each population.

Significant DRB1 and DQB1 allelic differences were seen between Lebanese, Bahraini, and Tunisian T1D patients and controls. Among the Lebanese subjects, DRB1*030101 (Pc < 0.033), DRB1*130701 (Pc = 0.008), and DQB1*0201 (Pc < 0.001) were positively associated with T1D and DRB1*110101 (Pc = 0.002), DQB1*030101 (Pc < 0.001), and DQB1*050101 (Pc = 0.006) were negatively associated with T1D, after the application of the Bonferroni correction (Pc). Among the Bahraini subjects, DRB1*030101 (Pc < 0.001), DRB1*040101 (Pc < 0.001), DQB1*0201 (Pc < 0.001), and DQB1*0302 (Pc < 0.001) were positively associated with T1D, while DRB1*110101 (Pc = 0.001), DQB1*030101 (Pc = 0.003), and DQB1*050101 (Pc < 0.001) were negatively associated with T1D. Among the Tunisian subjects, only DRB1*040101 (Pc = 0.017) was positively associated with T1D, while DRB1*110101 (Pc = 0.011), DQB1*030101 (Pc = 0.029), and DQB1*060101 (Pc = 0.009) were negatively associated with T1D.

The results of DRB1-DQB1 haplotype analysis are summarized in Table 1. DRB1*030101-DQB1*0201 served a T1D-susceptible role among all three populations. DRB1*040101-DQB1*0302 displayed a mixed association: it served a susceptible role among the Bahrainis and Tunisians and was protective in Lebanese subjects. DRB1*070101-DQB1*0201 and DRB1*110101-DQB1*030101 were negatively associated with T1D in the Bahrainis and Lebanese but not the Tunisians, while DRB1*100101-DQB1*050101 served a T1D-protective role only in the Bahrainis. Furthermore, DRB1*010101-DQB1*050101, DRB1*040101-DQB1*050101, and DRB1*150101-DQB1*060101 were negatively associated with T1D only in the Lebanese subjects. No T1D-protective haplotypes were identified among the Tunisian subjects.

TABLE 1.

DRB1*-DQB1* haplotype distribution^a

Haplotype	Bahrainis		Lebanese		Tunisians
	P	OR (95% CI)	P	OR (95% CI)	P	OR (95% CI)
DRB1*010101-DQB1*050101	0.683	0.51 (0.11-2.55)	<0.001	0.008 (0.00-0.189)4	0.609	0.47 (0.12-2.47)
DRB1*030101-DQB1*0201	<0.001	16.00 (8.29-29.04)	<0.001	81.00 (24.58-209.03)	0.028	2.15 (1.13-4.04)
DRB1*040101-DQB1*0302	<0.001	41.33 (12.35-109.72)	<0.001	0.049 (0.02-0.19)	0.007	3.75 (1.45-8.90)
DRB1*040101-DQB1*030201	1.000	1.44 (0.28-6.96)	0.187	1.29 (0.79-1.42)	0.490	0.58 (0.21-1.79)
DRB1*040101-DQB1*050101	0.467	0.623 (0.25-1.62)	<0.001	0.007 (0.00-0.07)	0.862	0.60 (0.14-3.20)
DRB1*070101-DQB1*0201	<0.001	0.13 (0.05-0.40)	0.028	0.132 (0.03-0.72)	0.733	0.78 (0.34-1.85)
DRB1*100101-DQB1*050101	0.002	0.10 (0.03-0.43)	0.144	0.28 (0.08-1.32)	0.317	0.99 (0.01-3.30)
DRB1*110101-DQB1*030101	<0.001	0.11 (0.04-0.32)	<0.001	0.055 (0.03-0.13)	0.170	0.34 (0.11-1.31)
DRB1*150101-DQB1*060101	1.000	0.84 (0.27-2.63)	0.011	0.049 (0.01-0.50)	0.419	0.39 (0.10-1.99)
DRB1*160101-DQB1*0201	<0.001	14.44 (3.30-47.92)	0.883	0.71 (0.20-2.83)	0.317	0.99 (0.01-3.30)

^aDRB1* and DQB1* alleles were assessed by PCR-SSP, and haplotype frequencies were determined by the maximum-likelihood method. Bahrainis comprised 126 patients and 126 controls, Lebanese comprised 78 patients and 111 controls, and Tunisians comprised 50 patients and 50 controls. Boldface indicates a significant difference between patients with T1D and age- and gender-matched controls in each community. P values were determined by Fisher's exact test.

Arabs comprise diverse ethnic populations that extend from the Arabian (Persian) Gulf to the Atlantic Ocean (2, 17). In view of the strong association of HLA DRB-DQB haplotypes with T1D (6, 9, 16), coupled with their various distributions in different racial/ethnic populations (7, 12, 15, 16), we investigated the association of HLA DRB-DQB haplotypes with T1D in individuals from three distinct Arab communities: Lebanon (Eastern Mediterranean), Tunisia (North Africa), and Bahrain (Arabian Peninsula). To avoid epidemiologic bias, only Arab subjects were included in the study; non-Arab nationals of the study communities, including Armenians (Lebanon), Berbers and Europeans (Tunisia), and Iranians (Bahrain), were excluded.

The differential association of the DRB1 and DQB1 alleles with T1D was noted in the three populations studied and was highlighted by the enrichment of DRB1*040101 in Tunisian and Bahraini but not Lebanese T1D patients and the strong association of DRB1*030101 and DQB1*0201 with T1D in Lebanese and Bahraini subjects. DQB1*0302, a traditional T1D-susceptible allele among Caucasians (5, 6, 8, 10), was not individually associated with T1D among the Lebanese and Tunisians but was associated with T1D susceptibility among the Tunisians and protection among the Lebanese. It is possible that the disease association conferred by DQB1*0302-containing haplotypes is linked to the nature of the DRB1 allele contained in the haplotype. Alternatively, it may be due to racial considerations, highlighted by the pathogenic gradient-like nature of DQB1*0302, where DQB1*0302 is linked with T1D in Northern European countries, including Finland (5, 6), but becomes neutral in Southern European and Mediterranean countries (12, 14).

Common and unique DRB-DQB haplotypes conferring susceptibility to and protection against T1D were identified in each community. The former was highlighted by the DRB1*030101-DQB1*0201 haplotype, which conferred disease susceptibility in all three populations, as has also been shown for other Caucasian populations (4, 10). A number of DRB-DQB haplotypes appeared to be population specific, as they were associated with T1D in some but not all populations (DRB1*100101-DQB1*050101 in Bahrainis, DRB1*040101-DQB1*050101 in Lebanese) or had an opposing association, depending on the population (DRB1*040101-DQB1*0302 conferred susceptibility in Bahrainis and protection in Lebanese). Similar to other Caucasian populations (5, 6), DRB1*040101-DQB1*0302 conferred T1D susceptibility in Bahraini and Tunisian patients. However, it was found to be largely protective among the Lebanese, as well as other Mediterranean communities (11, 12). This was the consequence of the high prevalence of DRB1*04 among Lebanese and other Mediterranean populations (11, 13).

Our data indicate that the association of HLA DRB-DQB haplotypes is variable throughout Arabian communities, with both common and unique haplotypes conferring disease susceptibility and protection. This is most likely due to genetic drift from Middle East ancestors (Arabian Peninsula) and the admixture with other populations, including Africans and Berbers (Tunisians), Europeans (Lebanese), and Iranians and Indians (Bahrainis). Further studies are required to identify common (and unique) HLA DRB-DQB haplotypes that confer susceptibility to and protection from T1D in related and distant Arab communities.