Background
Adhesion of
Plasmodium-infected red blood cells to endothelial cells lining the microvasculature of tissues of infected individuals (sequestration) has been implicated in the development of several severe malaria syndromes [
1,
2]. For the human malaria parasite
Plasmodium falciparum, molecules known to interact with the endothelium are members of the
PfEMP1 family, encoded by the
var multigene family [
3]. These have been well characterized, and indeed multiple endothelial cell receptors for
PfEMP1 have been identified (reviewed by [
4,
5]). For the other human malaria species lacking
PfEMP1, or the rodent malaria species, little is known about the mechanisms of cytoadherence, with respect to either the parasite molecules or the host proteins with which they interact. The genomes of the human and rodent parasite species that lack the
var family encode several other multigene families, some of which are not represented in
P. falciparum or Plasmodium reichenowi, and thus rodent malaria can be used as a model to dissect the mechanisms involved [
6,
7]. For
PfEMP1 interaction with the host molecules, CD36 and ICAM-1 were identified early [
8], subsequently supported by both data from field studies and experimental models (reviewed by [
9]). In addition to the restriction of blood flow by occlusion of the microvessels, adherent parasites may also be the focus of an inflammatory response.
CD36 is a class B scavenger receptor expressed on the surface of microvascular endothelium and many other cells and has a role in inflammation, angiogenesis and lipid metabolism (reviewed by [
10]). It can bind to clinical isolates of
P. falciparum-infected red blood cells (iRBC) in vitro, however most clinical studies show an association of CD36 binding with uncomplicated malaria rather than severe disease (reviewed by [
9,
11]). CD36 is also involved in internalization and phagocytosis of
P. falciparum iRBC, suggesting a role in parasite clearance [
12]. In the experimental model of
Plasmodium berghei ANKA in mice, sequestration of iRBCs in lungs and adipose tissue in vivo, is partially dependent on CD36 [
13], and this process is thought to be a major contributor to lung pathology in the
P. berghei ANKA model [
14].
The cell-surface receptor, ICAM-1, has been shown both by in vitro binding assays and by field studies to be important for cytoadherence of
P. falciparum although the relationship with disease outcome is unclear. In some studies ICAM-1 expression has been associated with severe falciparum malaria [
15] and more severe infections of children [
16], whereas a negative correlation with disease was found in a Malawian childhood malaria study [
17]. In addition, in vitro binding of
Plasmodium vivax iRBCs to ICAM-1 was demonstrated in static and flow assays [
18,
19]. ICAM-1 expression on vascular endothelium is upregulated in experimental models of cerebral malaria
(P. berghei ANKA
) and infections of mice deficient in ICAM-1 expression (
icam1−
/−) show improved survival [
20,
21] and reduced sequestration in the alveolar capillaries of the lung [
21]. Upregulation of ICAM-1 expression has also been observed on macrophages in
P. berghei ANKA infections, and it has been suggested that binding of
P. berghei-iRBC to macrophages could restrict venous blood flow thereby contributing to the cerebral symptoms of this model of experimental cerebral malaria [
22]. In the
Plasmodium chabaudi model an increased expression of ICAM-1 has been associated with a high parasite liver burden [
23].
Sequestration has been described for other species of
Plasmodium including
P. vivax and the rodent-infecting species, which do not have the
var genes encoding
PfEMP1 [
13,
24‐
26]. However, in these cases the parasite ligands responsible for cytoadherence are not known.
In the rodent malaria parasite,
P. berghei ANKA, a large proteomic screen of parasite membrane proteins identified a 17kD schizont membrane-associated protein (SMAC) [
27], which was exported to the iRBC cytoplasm.
Plasmodium berghei ANKA parasites lacking SMAC (Δ
smac) exhibited a different cytoadherence pattern from that of wild-type parasites, with a high number of mature schizonts in the peripheral blood and reduced accumulation in adipose tissue, suggesting that SMAC, although not on the iRBC surface, is implicated in some way in sequestration. In the absence of CD36, both wild type (wt) and Δ
smac mutant parasites exhibited similar distribution patterns, indicating that Δ
smac might be involved in the CD36-mediated sequestration [
27].
In order to determine whether adherence to ICAM-1 or to CD36 and the involvement of SMAC was applicable to other rodent malarias, the effect of the lack of these molecules on a
P. chabaudi blood-stage infection of C57BL/6 mice was investigated. The advantage of the
P. chabaudi infection is that it maintains a naturally synchronous asexual developmental cycle [
28], and sequestration, mainly in liver and lung, can be observed both by removal of schizonts from peripheral blood and by sequestration in tissues at a defined time of the day [
29]. These data show that the CD36 receptor in mice and the lack of SMAC in the parasite, unlike for
P. berghei, have limited impact on sequestration and pathology. By contrast, the lack of the adhesion molecule ICAM-1 reduces parasite burden in liver and spleen, and reduced the clinical signs of an acute
P. chabaudi infection.
Therefore it appears that with the rodent malarias there may be multiple host and parasite molecules involved in sequestration. This may well be the case for other human malarias, such as P. vivax.
Discussion
ICAM-1 and CD36 are strong candidate receptors for interactions with iRBC containing mature
Plasmodium parasites within the microvasculature at the time of sequestration, in both human infections and rodent model infections, and for the associated development of severe malaria pathologies. As species-specific differences in human malaria manifest themselves in distinctions in the nature and severity of the pathological syndromes developed, such differences are also seen in experimental models of malaria where the different host and parasite combinations exert similar influences [
9].
The data presented here suggest that the ICAM-1 molecule is an important cytoadherence receptor for
P. chabaudi in the spleen and liver, and further supports the hypothesis that cytoadherence is important for the development of pathology. Infected mice lacking the ICAM-1 receptor are less anaemic and lose less weight than mice with control infections, despite developing a higher peripheral blood parasitaemia which remained high for a longer time. The increase in parasitaemia was mirrored by a decrease in accumulation of mature parasites in spleen and liver. The lack of ICAM-1 has a marked effect on tissue sequestration in spleen and liver. It has previously been shown that sequestration in the livers of wild-type animals and the associated liver damage and increased liver pathology are at a maximum at day 9, when there is also a considerable host response [
29]. It is likely that local production of cytokines in the tissues in response to sequestering parasites would further enhance the expression of ICAM-1 in the wild-type host, contributing both to weight loss and the development of anaemia. Although clearly the lack of ICAM-1 modulates the infection in vivo, the capacity of
P. chabaudi iRBC to bind to cell lines expressing different recombinant receptors in vitro would give more direct evidence of the relevant host/parasite interactions. Such binding has been shown for the human malaria species
P. vivax and
P. falciparum [
18,
19].
Plasmodium berghei ANKA infections with mutant parasites show reduced sequestration in lungs and adipose tissue and this is accompanied by a higher splenic parasite load, which could lead to more effective clearance of the parasites [
27,
37]. These authors show that the parasite mutants either develop lower peripheral blood parasitaemia (SBP1, MAHRP) or can persist to the same degree as wild-type parasites (SMAC) in splenectomized mice. Here, in infections of
icam1−
/− mice, reduced parasite sequestration in the liver occurs concomitantly with a reduced load in the spleen, thus peripheral blood parasitaemia increases and parasite clearance is delayed. Thus, the impaired ability to sequester rather than high parasitaemia per se has the greater impact on the health status of the mouse. As it has previously been shown that parasite sequestration in the liver is accompanied by tissue damage and reduced liver function [
29], it is possible that such damage is abrogated when sequestration is reduced. Although the possibility that uptake of parasitized cells by monocytes and macrophages of the reticuloendothelial system could contribute to the total measured luciferase activity in the organs here cannot be completely ruled out, this approach was validated previously using light and electron microscopy and parasite quantitation at the time of schizogony.
Although ICAM-1 has been shown to impact sequestration of
P. berghei ANKA iRBC in alveolar capillaries [
21],
P. chabaudi iRBC sequestration in the lungs was unaffected by ICAM-1. The level of ICAM-1 mRNA is increased in the livers of
P. chabaudi AS-infected mice [
23] and in addition, liver sequestration is reduced in the absence of IFNγR signaling [
29]. It is possible that the IFNγR signaling effect is mediated partially via ICAM-1 as IFNγ upregulates the expression of ICAM-1 on endothelial cells [
38]. The observed reduced liver sequestration is also seen after transmission through the mosquito, where parasites in both the spleen and liver are reduced compared to controls. This finding of a role for ICAM-1 in sequestration of
P. chabaudi parasites is in line with the observation that
P. vivax-infected erythrocytes exhibit enhanced binding to ICAM-1 in vitro, compared to CD36 or to the untransfected control CHO cells [
18]. This group found that treatment with anti-VIR antibodies abrogated the binding of
P. vivax-iRBC to human lung endothelial cells. The predominance of ICAM-1 cytoadherence is also supported by a later study using
P. falciparum transgenic lines expressing
P. vivax VIR proteins showing binding of a specific VIR protein to multiple host receptors in vitro under static conditions yet binding to ICAM-1 alone was maintained under flow conditions [
19]. For
P. falciparum, domains of the variant
PfEMP1 proteins responsible for interacting with specific host receptors, notably CIDRα1 for EPCR and CIDRα2-6 for CD36 and DBL2β for ICAM-1 [
39‐
41] have been extensively studied. By contrast, the parasite ligand(s) involved in
P. vivax sequestration are unknown and the relationship to pathology is unclear [
42]. The differences in the relative importance of the receptors involved in sequestration, shown here for
P. chabaudi, compared to
P. falciparum, may be due to the absence of
var and the involvement of a different set of parasite ligands, encoded by one of the other multigene families, shared by
P. vivax and the rodent malaria parasites.
It has previously been shown that
P. chabaudi iRBC accumulate in the spleen early in infection and later sequester in liver and lung, and we observed damage to liver, lung and kidney [
29]. Experiments with transgenic
P. falciparum lines engineered to express VIR proteins (encoded by the
pir/vir multigene family) also suggest a stronger interaction with ICAM-1, as under physiological or flow conditions these lines adhere to ICAM-1 but not to CD36 or CSA [
19]. Mosquito transmission of
P. chabaudi AS alters parasite gene expression including upregulation of expression of many
pir/cir genes, [
43]. As the effect of lack of ICAM-1 on tissue accumulation (spleen and liver) is a little more pronounced after the parasite has completed the full developmental cycle, it is possible that expression of the parasite ligands involved in binding to ICAM-1 may be increased after transmission through the mosquito.
The lack of CD36 had no effect on sequestration or pathology in
P. chabaudi infections, although infected erythrocytes have previously been shown to bind to recombinant human CD36 in vitro [
44]. This is in contrast to the
P. berghei ANKA model, where lack of CD36 afforded no protection from cerebral pathology but sequestration was reduced in lungs and adipose tissue [
13] and lung pathology, as manifested by vascular leakage, was reduced [
14,
45]. However, a recent study of lung pathology in this model used mouse bone-marrow chimeras to show that endothelial damage was greatest when CD36 was expressed on endothelial cells, suggesting that the presence of CD36 on trafficking cells may result in enhanced parasite clearance and hence offer some protection from endothelial damage [
14]. This could explain why in this study infection in the complete
cd36−
/− mouse KO is no different from that of wild-type mice as the two roles of CD36 could exert a balancing effect.
Interactions of endothelial cell receptors with parasite ligand(s) may be highly specific, as for CSA [
46] which interacts with a very limited number of
PfEMP1 proteins, or more promiscuous, in the case of CD36 [
16]. Receptors may also interact synergistically with a co-receptor. It has been suggested that for
P. falciparum, ICAM-1 and EPCR may interact to promote binding to brain endothelial cells (where CD36 expression is much lower), while ICAM-1 and CD36 may act as co-receptors promoting binding in other vascular beds (e.g., lungs). So the relative importance of a particular receptor or receptor combination may be dependent on the vascular niche (e.g., brain
vs lung
vs placenta) [
47]. However, the lack of an effect of CD36 on sequestration of pathology of
P. chabaudi infections, in spite of a capacity to bind human CD36, [
44] may indicate receptor redundancy rather than synergy.
Significantly more mature schizonts were observed in the peripheral blood at the time of schizogony (day 7 post-infection) in infections with
P. chabaudi Δ
smac parasites. However this was not mirrored by any changes in parasite accumulation in the spleen or sequestration in the lungs, liver or other tissues. This result contrasts with that of mice infected with
P. berghei ANKA Δ
smac parasites, which exhibited an increased parasite load in the spleen of mice and reduced load in lungs and fat of the same mice [
27]. However,
P. berghei ANKA Δ
smac parasites also demonstrated reduced CD36-mediated cytoadherence, whereas we could find no evidence of CD36-mediated cytoadherence or any associated pathology in the
P. chabaudi model. Thus, the lack of a strong effect of
P. chabaudi Δsmac parasites on tissue sequestration would suggest that the effect is specific to the particular receptor involved in the cytoadherence process. Two further
P. berghei ANKA gene deletion mutants (SBP1 and MAHRP1) that exhibit reduced sequestration phenotypes have recently been generated and these also show reduced binding to CD36 [
37]. It would be of interest to determine whether these molecules could play a role in CD36 independent sequestration, as seen in this
P. chabaudi infection.
Authors’ contributions
DAC, JWL and JL designed and conceived the project. DAC and JWL performed the experiments in all transgenic mouse and parasite lines. TB, IT and GK contributed to experiments in CD36-null and ICAM-1 null mice. WJ performed differential counts. JR performed the PFG. BFF designed and constructed the Δsmac plasmid and advised on the design of ΔsmacEFluc plasmid. DAC, JWL and JL analysed, discussed the data and wrote the manuscript. All authors read and approved the final manuscript.