Three models of var gene switching in P. falciparum malaria infections

Introduction: Understanding naturally acquired immunity (NAI) and its relationship to immune evasion strategies in P. falciparum infections is important for malaria control. NAI is assumed to comprise specific antibodies to PfEMP1 proteins expressed on the surface of infected red blood cells. To escape from immune response, parasites switch among PfEMP1 variants. Using a simulation model which considers innate immunity, variant-specific and variant-transcending immunity, we explore different switching mechanisms and their implications for the immune mechanisms. Methods: Three models were implemented: Random switching, sequential switching and hierarchical switching (log-normal distributed target probabilities); assuming variant-specific growth rates. We modified the following parameters: maximum efficacy of variant-transcending immunity; threshold for variant-specific immunity; switching rate per cycle. The simulated curves were evaluated, using infection characteristics derived from 35 patients. Results: Random switching reproduced the characteristics for low switching rates (<0.1%) and high variant-specific immunity threshold. Although all variants were quickly expressed, one variant dominated the first peak of parasitaemia, followed by an increasing percentage of concurrent variants. Sequential switching needed a highly efficient variant-transcending immunity, otherwise the duration of infection became unrealistically long. A maximum of about 17% of the variants were concurrently active. Hierarchical switching reproduced the characteristics with less effective variant-transcending immunity. During early infection, many variants existed concurrently, but only few at a detectable level. Variants not present at the peak remained inactive. Conclusions: Random switching showed reasonable results only with questionably low switching rates. Sequential and hierarchical switching correspond better to published antibody thresholds and switching rates. The time lag between the activation of variants is important: sequential switching guarantees delays in activation, but there is no experimental evidence that switching occurs in a fixed order. In the hierarchical model, the time lag is produced by different target probabilities which could be caused by different promoters and the location on the chromosome.