Genetic epidemiology of visceral leishmaniasis in northeastern Brazil

Familial clustering of disease, racial differences in asymptomatic:disease ratios, and studies of mice all point to a genetic component for disease susceptibility in visceral leishmaniasis. Analysis of 87 multi-case pedigrees (824 individuals; 138 nuclear families) from a region of northeastern Brazil endemic for Leishmania chagasi demonstrates a high relative risk ratio (lambda(2S) = 34) to further siblings of affected sibling pairs. Complex segregation analysis using POINTER and COMDS show that all single locus models, as well as polygenic and multifactorial models, provide a significantly (P < 0.001) better fit to the data than a sporadic model. Of the genetic models, the general single locus model was not significantly different from additive or dominant single locus models, all of which gave a gene frequency for the putative disease susceptibility allele of approximately 0.002. The general single locus model was strongly favored (P < 0.001) over a recessive single gene model. Using POINTER, polygenic and multifactorial models were clearly rejected (P < 0.001 in all cases) in favor of the general single locus model. Using COMDS, the analysis was extended to consider two locus models. Results under a general two-locus model did not differ significantly from the dominant, additive, or general single locus models. Under this model, one locus was estimated at a gene frequency of 0.0017, i.e., in the same range as the disease susceptibility locus for the most favored single gene models, with the second locus at a much lower frequency of 0.0002. Hence, the data support the hypothesis that a single major gene may be important in determining disease susceptibility in this population. To identify the gene(s) involved, a genome scan with replication using two subsets of these larger pedigrees with power to detect linkage is in progress.