Artemisinin susceptibility in the malaria parasite Plasmodium falciparum: propellers, adaptor proteins and the need for cellular healing

Abstract

Studies of the susceptibility of Plasmodium falciparum to the artemisinin family of antimalarial drugs provide a complex picture of partial resistance (tolerance) associated with increased parasite survival in vitro and in vivo. We present an overview of the genetic loci that, in mutant form, can independently elicit parasite tolerance. These encode Kelch propeller domain protein PfK13, ubiquitin hydrolase UBP-1, actin filament-organising protein Coronin, also carrying a propeller domain, and the trafficking adaptor subunit AP-2μ. Detailed studies of these proteins and the functional basis of artemisinin tolerance in blood-stage parasites are enabling a new synthesis of our understanding to date. To guide further experimental work, we present two major conclusions. First, we propose a dual-component model of artemisinin tolerance in P. falciparum comprising suppression of artemisinin activation in early ring stage by reducing endocytic haemoglobin capture from host cytosol, coupled with enhancement of cellular healing mechanisms in surviving cells. Second, these two independent requirements limit the likelihood of development of complete artemisinin resistance by P. falciparum, favouring deployment of existing drugs in new schedules designed to exploit these biological limits, thus extending the useful life of current combination therapies.

Keywords: artemisinin resistance; endocytosis; haemoglobin; malaria parasites; proteasome; protein recycling.

Figure 1.

RSA for artemisinin susceptibility testing is based upon variation in drug killing across the in vitro cell cycle. (A) Schematic of timing of drug exposure and wash-off when carrying out the RSA of Witkowski et al. (2013). A 4-h pulse is shown, but this can vary in literature reports. (B) Data from Dogovski et al. (2015) redrawn to directly compare the DHA dose response across the first 30 h of the cell cycle in vitro of four parasite isolates from Pailin, Cambodia, that differ in pfk13 genotype as indicated. This illustrates that the tolerance phenotype is only seen in the first few hours of intra-erythrocytic growth. Schizonts from all four lines were highly susceptible (LD₅₀ ∼20 nM). WT—wild-type propeller domain sequence in pfk13.

Figure 2.

Two-component model for parasite survival of artemisinin treatment in vivo. Conceptual model in which two components of P. falciparum cellular biology (endocytosis—top panel; cellular healing capacity—middle panel) contribute to survival across 48 h of intra-erythrocytic schizogony following exposure to artemisinin treatment in vivo within an approximate ‘artemisinin exposure window’ during which suppression of Hb endocytosis can be advantageous (bottom panel). This simplistic representation sees both components as binary (wild-type/reduced; basal/enhanced), but this is unlikely to hold true. Only parasites encountering artemisinin during this ‘window’ (the first few hours of intra-erythrocytic development) may survive; later stages remain susceptible and are unlikely to survive whatever their phenotype (not shown). Horizontal axis: hours post-invasion. Red vertical axis: relative efficiency of Hb endocytosis up to mature trophozoite stage. The alternate state of ‘reduced Hb endocytosis’ can be generated by the presence of Kelch 13 mutations (Ariey et al. ; Straimer et al. 2015) that reduce the endocytic uptake (from host cytosol) of haemoglobin entering through cytostomes or other endocytic structures (Yang et al.2019). The arrow shows direction of benefit, which is only exerted in ring stages. The same alternate state can be reached by expression of variants of PfCoronin (Demas et al. 2018), AP-2m, UBP-1 (Henrici et al. 2020), disruption of falcipain-2a (Klonis et al. ; Xie et al. 2016) and hypo- or hyper-thermal pulses at early ring stage (Henrici et al. 2019a). Blue vertical axis: capacity for cellular healing through adaptive modification of the proteasome, the unfolded protein response (UPR), chaperone-mediated protein re-folding, autophagy and DNA repair; the arrow shows direction of benefit. These phenotypes have not been adequately characterised to allow a precise prediction of impact, and any particular parasite may have enhanced capacity in only some, but not all, of these functions. It is assumed these functions are most active prior to schizogony and segregation of merozoites. Cellular healing capacity is encoded independently of ring-stage tolerance, is thought to be multigenic and may also require epigenetic adaptations. Therefore, it is likely to be rarely achieved, and may come at a cost to overall reproductive fitness in the absence of intense drug pressure. Green vertical axis: relative survival capability in a treated host of parasites with each of the four combined states possible in a simplistic binary model for each of the two components—haemoglobin endocytosis and capacity for cellular healing, respectively. In contrast to in vitro assays, damaged parasites surviving the artemisinin exposure window may be unlikely to survive host clearance mechanisms without an enhanced cellular healing phenotype. Only the effects of an artemisinin encounter at early ring stage are shown.

Figure 3.

Two-component model of in vivo artemisinin tolerance and recovery. Plasmodium falciparum survival of artemisinin exposure in vivo requires two distinct adaptive components, as even surviving cells are likely to have sustained oxidative damage that would increase the chances of clearance by host defences (O'Flaherty et al. 2019). Left hemisphere: Recent in vitro studies of parasite lineages displaying ring-stage artemisinin-tolerant phenotypes have revealed a common mechanism involving dysregulation of haemoglobin endocytosis via the cytostome. Two complexes seem to be involved: K13-labelled cytostome collar complex (blue) and AP-2-labelled cytostome cargo complex (yellow), which may be linked by KIC7 (green ellipse), a common interacting protein. Two-colour super-resolution SIM imaging (left-most panel; Henrici 2018) of mature schizonts by the authors demonstrates the spatial relationship between these two complexes and recapitulates the ring-shaped K13-defined structure of the cytostome collar with AP-2µ inside this ring in several optical planes. Manipulation of this common pathway leads to a decrease in haemoglobin-derived haem in the DV, and thus a corresponding diminution of activated DHA from supra-lethal concentrations to sub-lethal concentrations (purple and light green bars). This decrease in ring-stage haemoglobin uptake and artemisinin activation allows a fraction of cells to survive. Right hemisphere: Surviving cells may have adaptations that mitigate oxidative damage via a variety of as-yet poorly defined pathways. Such enhanced cellular healing phenotypes synergise with the ring-stage artemisinin tolerance phenotype to increase chances of survival in the treated immunocompetent host and enable emergence of artemisinin tolerance in circulating parasite lineages.