It has been recognised for decades that serial blood passage of Plasmodium through rodents, primates, or humans universally increases parasite virulence. In 1917, induced malaria was first used as pyretic therapy for neurosyphilis, with Plasmodium vivax routinely inoculated to elicit a mild form of disease. Yet passage through the human host elevated parasitaemia and exacerbated disease, increasing the requirement for chemotherapeutic intervention [1]. Blood passage of Plasmodium knowlesi or Plasmodium cynomolgi, whether through human volunteers or nonhuman primates, similarly elevated parasite densities and disease severity [2–4]. And serial blood passage of every rodent malaria parasite species increased parasitaemia and pathogenicity [5–8]. On the other hand, it has been assumed for decades that mosquito transmission resets Plasmodium virulence [8]. At the Horton Mental Hospital, a pioneering centre for malaria therapy, Plasmodium strains were maintained by mosquito transmission to preserve their clinical and parasitological features [9]. Nevertheless, direct evidence that mosquito transmission resets Plasmodium virulence, and a mechanism to explain this phenomenon, have been missing [6,10,11].

We have recently shown that mosquito transmission modifies gene expression in blood-stage malaria parasites and in this way resets Plasmodium virulence [12]. Whereas serial blood passage of Plasmodium chabaudi leads to hyperparasitaemia and severe disease in laboratory mice, mosquito transmission of serially blood-passaged parasites leads to a low-grade, chronic, recrudescing infection with minimal pathology. Attenuation of virulence is not parasite clone- or dose-dependent and therefore cannot be explained by bottlenecking during mosquito transmission [13]. Instead, attenuation of the blood-stage parasite is dependent upon host genotype [14] and an intact host immune response [12] and associates with increased expression of Plasmodium subtelomeric multigene families, including those that code for VSA (Box 1). Mosquito transmission therefore modifies expression of parasite virulence genes and transforms host immunity in the pathogenic blood-stage of infection. As such, the mosquito vector both transmits malaria and regulates disease severity.

By recognising this key function of the mosquito, new research avenues open that can accelerate our understanding of the pathogenesis of human malaria. It is first important to delineate where, when, and how mosquito transmission modifies expression of Plasmodium virulence genes. This is likely to be a consequence, at least in part, of necessary changes in gene expression for progression through each step of the life cycle in both vector and host [15]. However, epigenetic reprogramming [16] of Plasmodium provides a mechanism by which expression of virulence genes could be reset within the vector. Heritable chromatin modifications control transcription of subtelomeric multigene families in the blood-stage parasite [15] and can thus promote adaptation of malaria parasites to their host [17]. Nevertheless, global erasure of epigenetic marks following gamete fusion in the mosquito could reset expression of multigene families and thus prepare Plasmodium for entry into a new unknown host (Fig 1).

Resetting Plasmodium gene expression could be particularly important when transmission is seasonal, and parasites undergo an extended period of host adaptation in a chronically infected individual before their return to the mosquito. In this context, it is important to know whether parasite virulence increases in the chronic phase of infection, as has been observed in human volunteers [3]. Thus, serial blood passage per se may not increase Plasmodium virulence; an alternative explanation is that virulence increases with time elapsed from the mosquito. This will be observed only in a new host and when mosquito transmission is bypassed.

By resetting Plasmodium gene expression, the mosquito can also control how blood-stage parasites elicit the systemic immune response in a new host. Mosquito transmission attenuates P. chabaudi virulence because merozoites that emerge from the liver induce an immune response that can rapidly control parasite growth without collateral damage [12]. This contrasts with the host response to serially blood-passaged parasites that causes severe immunopathology. Does increasing expression of Plasmodium VSA explain how the mosquito can transform the elicited host immune response? Or does mosquito transmission change the context in which blood-stage parasites initiate host immunity (e.g., by modifying invasion, cytoadherence, or sequestration)? Furthermore, it remains possible that immune priming and/or regulation during the pre-erythrocytic stages of infection can subsequently modify the systemic immune response to the blood-stage parasite. In all scenarios, the early immune response, elicited in the context of a mosquito bite, can shape malaria disease severity. In turn, the developing immune response is likely to influence expression of Plasmodium virulence genes and could therefore also directly regulate parasite pathogenicity.

Mosquitoes reset malaria parasites and can be used to strengthen the relevance of mouse models to human malaria. We should therefore aim to initiate experimental infections by the natural route of transmission wherever possible. We should also strive to study combinations of vector, parasite, and host that exist in nature to validate or improve our current experimental systems. Mouse models are important for interrogating the pathogenesis of malaria because they can answer research questions that cannot be addressed directly in humans. Moreover, relevant mouse models can act as a bridge between human studies. For example, vector regulation of Plasmodium virulence was first observed in human volunteers and subsequently reproduced and delineated in mice; the molecular mechanisms that operate within the mosquito to regulate Plasmodium virulence can now be dissected with human malaria parasites.

To this end, inoculation of human volunteers with Plasmodium is a powerful experimental model [18,19]. In this setting, it is possible to look for evidence of epigenetic reprogramming of P. vivax in laboratory-reared anopheline mosquitoes fed on infected volunteers. Interrogating expression and regulation of subtelomeric multigene families in gametocytes as they circulate, transmit, and then pass through each developmental checkpoint of sporogony is a priority. So, too, is examining how route of transmission influences the systemic host response to blood-stage infection. For this, the immune response to P. falciparum can be compared in peripheral blood obtained from human volunteers infected via mosquito bite versus direct inoculation of blood-stage parasites (isolated just 6–8 days after liver egress [20]). Nevertheless, mice are absolutely required to observe the interactions between parasites and the immune system that shape disease severity because these interactions occur in tissues, such as spleen. We should therefore aim to identify mouse models that share a common immune signature of infection in whole blood with human malaria and use these models to delineate the immune response to Plasmodium in relevant tissues.

A mosquito is not simply a flying syringe. Mosquitoes reset malaria parasites in preparation for entry into a new unknown host and thereby regulate Plasmodium virulence. Furthermore, they are a mixing pot for the generation of new recombinant parasites and can thus transmit previously unseen virulent strains. By studying events within the mosquito, we will accelerate our understanding of malaria disease severity.