Competition, cooperation and immune selection of multi-strain Plasmodium falciparum malaria

Abstract Setup Malaria Plasmodium falciparum (Pf) species contains multiple strains with different immunogenic profiles, and expressed phenotypes. These strains circulate in host populations via mosquito transmission, and interact (compete, cooperate) on two levels: within - host (via cross-reactive immunity), and in host populations. Both factors, host immunity and transmission environment, play important part in evolution and selection. Conventional population-based models of malaria have limited capacity to accommodate parasite-immune dynamics within-host and strain diversity. Here we developed an in-host model for multi-strain malaria based on its genetic (immunogenic) makeup, which accounts for essential parasite-immune biology. The model allows efficient simulations of mixed-strain infections in individual hosts and in host ensembles over multiple transmission cycles. We use it to explore evolutionary implications (competition, selection) of malaria quasi-species, driven by host immunity and transmission intensity. Results The key ‘selectable’ trait within-host is strain transmissibility (TP), which measures cumulative odds of mosquito infection by a given strain over infection history. Here we adopt it to explore evolutionary implications of parasite-immune interactions on different time scales and transmission environments. Specifically, we explore (i) primary strain selection in naive host ensembles based on TP-fitness; (ii) evolution and selection of mixed multi-strain systems over multiple transmission cycles. On level (i) different strain mixtures competed in multiple hosts, to identify ‘most fit’ (highly transmissible) types. A key observation of (i) was fitness-cost of in-host competition, i.e. statistical TP-loss determined by multiplicity of infection (number of competing strains), and strain genotype (immunogenic profile). The most-fit strains maintained their high TP-values regardless of competing environment. We selected them for step (ii), to explore long-term evolution over multiple transmission cycles. Our analysis revealed peculiar features of evolution: success within-host (step (i)) did not guarantee strain survival over multiple cycles. Indeed, the latter was strongly associated with cooperative behavior, i.e. co-existence of a given strain in suitable mixtures, in multiple hosts over many generations. We examined the resulting population structure of evolving strains, in terms of their immune cross-reactivity. Overall, our results were consistent with predictions of strain theory [1–4], [5, 6]. Strain theory predicts that cross-reacting parasite strains in host population should organize themselves into ‘non-overlapping’ (immunogenically disjunct) clusters. In our case, no strict ‘immune separation’ arises, but cross-reactivity is lost over multiple cycles, and surviving clusters are ‘nearly disjunct’. Such weakly overlapping clusters (cooperating cliques) persisted over long (evolutionary) periods. Specifically, each clique was found to possess a core node -highly cooperative persistent strain, carrying a subordinate (transient) cluster. Our results shed new light on relative importance of competitive vs. cooperative behavior, and multi-level organization of genetically structured parasite system. They could have implications for malaria control and vaccine design.