HLA-DQ and susceptibility to celiac disease: Evidence for gender differences and parent-of-origin effects

Celiac disease (CD) is twice as frequent among female than male. Despite the large number of reports on the DQ2/DQ8 association, no systematic studies have investigated a possible different role of the HLA genes in the two genders. We performed case-control and family-based analyses of DR-DQ variants in a pediatric CD cohort with the aim of comparing female to male associations and to investigate the paternal/maternal inheritance of the disease-predisposing haplotypes. METHODS: A total of 281 female and 156 male pediatric celiac patients, 292 nuclear families, and 551 controls were genotyped for HLA-DRB1, DQA1, and DQB1 loci. Odds ratio, parental origin of the disease-associated haplotypes, and transmission ratio distortion were evaluated in-between male and female cases. RESULTS: DQ2/DQ8 were more frequent in female than in male patients (94% F, 85% M; P = 1.6 x 10(-3)) with a 99.1% and 90.5% calculated negative predictive value of the HLA test, respectively. Surprisingly, the majorit

BACKGROUND AND AIMS: Celiac disease (CD) is twice as frequent among female than male. Despite the large number of reports on the DQ2/DQ8 association, no systematic studies have investigated a possible different role of the HLA genes in the two genders. We performed case-control and family-based analyses of DR-DQ variants in a pediatric CD cohort with the aim of comparing female to male associations and to investigate the paternal/maternal inheritance of the disease-predisposing haplotypes. METHODS: A total of 281 female and 156 male pediatric celiac patients, 292 nuclear families, and 551 controls were genotyped for HLA-DRB1, DQA1, and DQB1 loci. Odds ratio, parental origin of the disease-associated haplotypes, and transmission ratio distortion were valuated in-between male and female cases. RESULTS: DQ2/DQ8 were more frequent in female than in male patients (94% F, 85% M; P = 1.6 x 10(-3)) with a 99.1% and 90.5% calculated negative predictive value of the HLA test, respectively. Surprisingly, the majority of the 39 DQ2/DQ8 negative cases were male. The analysis of the DQ2 haplotype origin showed that 61% of female patients and 42% of male patients carried a paternal combination (P = 0.02). The transmission disequilibrium test (TDT) proved the major distortion in the DR3-DQ2 transmission from fathers to daughters. CONCLUSIONS: CD is confirmed to be more prevalent in female than in male (F:M = 1.8) but, in DQ2/DQ8 negative patients, we found an unexpected male excess (F:M = 0.7). Moreover, only the inheritance of a paternal DQ2 haplotype led to a daughters predominance. These data show a role of HLA genes on the disease sex bias and suggest a possible different effect of parent-specific epigenetic modifications in the two genders.