Impact of changing drug treatment and malaria endemicity on the heritability of malaria phenotypes in a longitudinal family-based cohort study

Abstract

Despite considerable success of genome wide association (GWA) studies in identifying causal variants for many human diseases, their success in unraveling the genetic basis to complex diseases has been more mitigated. Pathogen population structure may impact upon the infectious phenotype, especially with the intense short-term selective pressure that drug treatment exerts on pathogens. Rigorous analysis that accounts for repeated measures and disentangles the influence of genetic and environmental factors must be performed. Attempts should be made to consider whether pathogen diversity will impact upon host genetic responses to infection.We analyzed the heritability of two Plasmodium falciparum phenotypes, the number of clinical malaria episodes (PFA) and the proportion of these episodes positive for gametocytes (Pfgam), in a family-based cohort followed for 19 years, during which time there were four successive drug treatment regimes, with documented appearance of drug resistance. Repeated measures and variance components analyses were performed with fixed environmental, additive genetic, intra-individual and maternal effects for each drug period. Whilst there was a significant additive genetic effect underlying PFA during the first drug period of study, this was lost in subsequent periods. There was no additive genetic effect for Pfgam. The intra-individual effect increased significantly in the chloroquine period.The loss of an additive genetic effect following novel drug treatment may result in significant loss of power to detect genes in a GWA study. Prior genetic analysis must be a pre-requisite for more detailed GWA studies. The temporal changes in the individual genetic and the intra-individual estimates are consistent with those expected if there were specific host-parasite interactions. The complex basis to the human response to malaria parasite infection likely includes dominance/epistatic genetic effects encompassed within the intra-individual variance component. Evaluating their role in influencing the outcome of infection through host genotype by parasite genotype interactions warrants research effort.

Figure 1. The incidence rate (mean and SEM) of clinical P. falciparum episodes per person-trimester (PFA) according to age classes (from left to right on the X-axis) <5 years, [5–15], [15–35] and ≥35 years that best describe the effect of age on PFA.

Figure 2. Proportion of variance in the number of clinical P. falciparum episodes per trimester explained by additive genetic (solid line), intra-individual (dotted line, squares) and house (thin dotted line, triangles) effects.

Figure 3. Comparative scatter plot of the Intra-individual estimates per individual per drug period for those individuals present throughout the study period.

Individuals with significant intra-individual estimates at any period are shown in color: red stars (significant in CQ1), green squares (significant in CQ2), blue triangles (significant in Fansidar) and yellow circles (significant in ACT).

Figure 4. Proportion of variance in the prevalence of P. falciparum gametocytes during clinical P. falciparum episodes (Pfgam) explained by additive genetic (solid line), intra-individual (dotted line, squares) and house (thin dotted line, triangles) effects.