Molecular mechanisms of cytoadherence in malaria

May Ho, Nicholas J. White


Microbial pathogens subvert host adhesion molecules to disseminate or to enter host cells to promote their own survival. One such subversion is the cytoadherence of Plasmodium falciparum-infected erythrocytes (IRBC) to vascular endothelium, which protects the parasite from being removed by the spleen. The process results in microcirculatory obstruction and subsequent hypoxia, metabolic disturbances, and multiorgan failure, which are detrimental to the host. Understanding the molecular events involved in these adhesive interactions is therefore critical both in terms of pathogenesis and implications for therapeutic intervention. Under physiological flow conditions, cytoadherence occurs in a stepwise fashion through parasite ligands expressed on the surface of IRBC and the endothelial receptors CD36, intercellular adhesion molecule-1 (ICAM-1), P-selectin, and vascular adhesion molecule-1. Moreover, rolling on ICAM-1 and P-selectin increases subsequent adhesion to CD36, indicating that receptors can act synergistically. Cytoadherence may activate intracellular signaling pathways in both endothelial cells and IRBC, leading to gene expression of mediators such as cytokines, which could modify the outcome of the infection.

  • Plasmodium falciparum
  • erythrocyte
  • adhesion molecules
  • pathogenesis

adhesion molecules of the immune system are surface receptors that facilitate cell-cell interaction. They are central to the recruitment and trafficking of immune cells and thus in the generation of an immune response. Many of the adhesive molecules are constitutively expressed, whereas others may be induced or upregulated by cytokines and microbial products. In the past decade, it has become increasingly recognized that some microbial pathogens subvert these molecules to promote their dissemination or entry into host cells and thus defend against host immune responses. Examples of such subversion have been found in three major classes of microbial pathogens, i.e., viruses, bacteria, and parasites, and involve the integrins, selectins, and members of the immunoglobulin superfamily. Understanding the molecular events involved in these adhesive interactions is important in terms of both pathophysiology and implications for therapeutic intervention. In this paper, the molecular basis of the adhesive interactions between Plasmodium falciparum-infected erythrocytes (IRBC) and vascular endothelium is reviewed in the context of the clinical disease.


Malaria is the most important parasitic infection, affecting an estimated 350–500 million people worldwide and resulting in between 0.5 and 2 million deaths annually (110). Human malaria is caused by one of four Plasmodiumspecies, namely, P. falciparum, P. vivax, P. ovale, and P. malariae. The life cycle of malaria parasites is complex, with asexual reproduction occurring in the mammalian host and sexual reproduction in the anopheline mosquito vectors (Fig. 1) (107). The parasites are transmitted to humans in the form of sporozoites through the bite of a female anopheline mosquito. The sporozoites circulate for up to 45 min before entering hepatocytes, in which they undergo asexual reproduction to form a large intracellular schizont. Hepatic schizonts contain thousands of merozoites when they are mature (within 5–15 days of sporozoite inoculation). The swollen hepatocyte eventually bursts, discharging merozoites into the bloodstream, where they rapidly invade erythrocytes to initiate the erythrocytic cycle. InP. vivax and P. ovale infections, dormant liver forms also occur that are capable of reactivating and forming a hepatic schizont weeks or months later. Merozoite attachment to the red blood cell is mediated via a specific erythrocyte surface receptor. Inside the erythrocyte, the parasite develops within a membrane-bound parasitophorous vacuole first as a trophozoite and then, during multiple nuclear division known as schizogony, as a schizont. When the schizont matures, the host red blood cell ruptures, liberating merozoites that rapidly invade fresh erythrocytes in the general circulation, thus continuing the life cycle. Some merozoites develop into sexual forms (gametocytes), which are taken into the mosquito gut with a blood meal. These may fuse forming a zygote and then undergo meiosis to form first an ookinete and later an oocyst in the gut wall. This phase of multiplication in the mosquito is known as sporogony. The oocyst also bursts, liberating large numbers of sporozoites that migrate to the salivary glands to await injection into the human host during the next blood meal.

Fig. 1.

The malaria transmission life cycle. [From White and Breman (107), with permission of The McGraw-Hill Companies.]

P. falciparum is the causative agent of the most severe of the four infections and accounts for most of the mortality. Most deaths occur in children under five years of age (110). Falciparum malaria is an acute febrile illness characterized by fever, chills, headache, anemia, and splenomegaly, which responds promptly to appropriate antimalarial therapy. Untreated, the patient will either die in the acute attack (<5%) or survive with almost no development of immunity against the next infection unless he or she has been exposed to years of repeated challenge. Approximately 5–10 million infected individuals per year develop complications during the acute infection, manifested as coma (cerebral malaria), metabolic acidosis, hypoglycemia, severe anemia, and, in adults, renal failure and pulmonary edema (108). The overall mortality from severe malaria varies from 15 to 30%, with the highest mortality resulting from cerebral malaria, metabolic acidosis, and pulmonary edema.


The pathogenicity of P. falciparumresults from its potential to multiply to high parasite burdens and the unique ability to adhere to capillary and postcapillary venular endothelium during the second half of the 48-h life cycle, a process that is called cytoadherence (52, 53). The resulting sequestration of infected erythrocytes (IRBC) leads to alterations in microcirculatory blood flow, metabolic dysfunction, and, as a consequence, many of the manifestations of severe falciparum malaria (Fig.2). Cytoadherence confers at least two survival advantages for the parasites: the microaerophilic venous environment is better suited for their maturation, and adhesion to endothelium allows them to escape clearance by the spleen, which recognizes their loss of deformability (26, 49) and opsonization with antibodies and/or complement components (40).

Fig. 2.

Cerebral venules packed with infected erythrocytes (arrows) in a fatal case of cerebral malaria.

As a result of cytoadherence, patients who die in the acute phase of falciparum malaria have intense sequestration of erythrocytes containing mature forms of the parasite in the microvasculature of vital organs (1, 53, 68). The organ distribution of sequestration varies and tends to reflect the clinical features of the preceding clinical illness; for example, patients with coma (cerebral malaria) show increased cerebral sequestration compared with that in other organs (76, 81). Even within the brain, variation in the degree of sequestration is seen between cerebral and cerebellar vessels, and white and gray matter (89). At the microvascular level, there is considerable heterogeneity among the individual vessels (N. J. White and K. Silamut, unpublished observations). Some are packed with IRBC containing fully developed schizonts, others with mature trophozoites that have yet to undergo schizogony, and others contain no parasites. This synchronous clustering suggests that, once the erythrocytes have adhered, “unsticking” and recirculation do not occur. The majority of parasites in fatal cases are at the mature trophozoite stage, but this probably represents antemortem drug effects that preferentially arrest development at this stage. In the acute phase of cerebral malaria there is remarkably little extravascular pathology, and, although occasional fibrin strands may be seen, platelets are also notable by their absence (53). Inflammatory cells are more prominent in patients who die many days after starting treatment and in whom parasites have largely cleared, and phagocytic cells are seen to be ingesting parasite pigment that has been released by ruptured IRBC or that remains in cytoadherent erythrocyte ghosts. There is no pathological evidence of acute vasculitis.

At the ultrastructural level, electron-dense, knoblike protrusions of the erythrocytic membrane are seen at the points of contact between the IRBC and endothelial cells (Fig. 3) (42,53, 68). These knobs are made up of parasite-encoded proteins that have been exported to the surface of the infected erythrocyte. They are essential for firm cytoadherence by facilitating the initial attachment of the infected erythrocyte to the endothelial cell and by concentrating the parasite ligands at a particular site. Although knobless IRBC can adhere to target cells in static binding assays in vitro (16, 102), ultrastructural studies of human tissues from fatal malaria cases have not shown cytoadherence independent of knobs. Furthermore, targeted disruption of the gene encoding one of the knob proteins, knob-associated histidine-rich protein (KAHRP or PfHRP1), results in the failure of knob formation and an inability of the IRBC to adhere under flow conditions in vitro (24). Thus these knobs appear to serve the same function as microvilli on the surface of leukocytes, where ligands for interaction with endothelium are presented (58). Parasitism would continuously favor the selection of knob-positive organisms able to form a more stable union with host cells and thus to allow the parasites to evade splenic clearance, thereby increasing the probability of survival and transmission.

Fig. 3.

Cytoadherence between infected erythrocytes (IRBC) and endothelial cells (EN) showing knobs (arrows) at the points of attachment. [From E. Pongponratn and D. Ferguson, with permisssion (unpublished observations).]


The stage and host cell specificity of cytoadherence suggests that the process involves specific parasite or host ligands expressed on the surface of IRBC and vascular endothelium. A number of endothelial receptor molecules have been identified, based on their ability to support the adhesion of laboratory-selected parasite lines and clones in static adhesion assays in vitro. The first of these molecules described was thrombospondin (TSP) (82). IRBC adhere to immobilized TSP in a dose-dependent manner but not to other adhesive proteins such as fibronectin or von Willebrand factor. Subsequent investigations showed that, although TSP may contribute to cytoadherence, it is not sufficient to mediate the process by itself. IRBC do not adhere to every cell line that secretes TSP (72), and anti-TSP antibodies do not inhibit cytoadherence to C32 melanoma cells that express CD36 and intercellular adhesion molecule-1 (ICAM-1) in addition to TSP (66).

The second receptor molecule to be implicated in cytoadherence was CD36 (7, 8, 66). CD36, or platelet glycoprotein IV, is found on monocytes, endothelial cells, platelets, and erythroblasts. Its natural ligands include collagen (98), TSP (5), and both the natural and oxidized forms of high-density lipoproteins, low-density lipoproteins, and very-low-density lipoproteins (21). A monoclonal antibody to CD36, OKM5, inhibits and reverses the cytoadherence of IRBC to a number of target cells in vitro, including dermal microvascular endothelial cells and C32 melanoma cells, as well as purified CD36 immobilized on plastic. Furthermore, although IRBC binding to TSP and CD36 is highly correlated, IRBC can adhere directly to CD36-transfected COS cells in the absence of TSP (69). With the use of peptides generated from a random CD36 domain library, it has been shown that IRBC interact with sites on the CD36 molecule that are distinct from the binding sites of TSP (6). The fact that OKM5 blocks the adhesion of all parasite isolates tested so far suggests that all IRBC bind to a common region on the molecule. The interaction of IRBC with CD36 expressed on monocytes leads to a respiratory burst (64), with the production of oxidative metabolites that are toxic to intraerythrocytic parasites (65).

A third endothelial molecule that acts as a receptor for IRBC selected on human umbilical vein endothelium (HUVEC) is ICAM-1 (13), which belongs to the immunoglobulin superfamily of adhesion molecules. ICAM-1 is a glycoprotein that acts as a ligand for the leukocyte integrin lymphocyte function-associated antigen-1 (LFA-1) and plays a central role in the generation of an immune response (86). It is distributed widely on venular endothelium, where it has been shown to be crucial for neutrophil adhesion before the transmigration of these cells into an inflammatory focus (91). ICAM-1 expression on endothelial cells can be upregulated by the proinflammatory cytokines tumor necrosis factor-α (TNF-α), interleukin-1 (IL-1), and interferon-γ (IFN-γ) (75). ICAM-1 is also the receptor for human rhinoviruses (92). Molecular studies have shown that IRBC binds to the first immunoglobulin domain of ICAM-1 at a site that is 180° from the binding sites of both LFA-1 and rhinovirus (12, 61).

Two other endothelial receptors have been shown to mediate the cytoadherence of a parasite clone that was derived from a clinical parasite isolate by sequential panning of IRBC on purified CD36, ICAM-1, E-selectin, and vascular adhesion molecule-1 (VCAM-1) proteins (67). VCAM-1 is not expressed constitutively on endothelial cells but can be induced by TNF-α, IL-1, and IL-4 on endothelium with kinetics similar to ICAM-1 (70, 71). Its interaction with its ligand, the β1-integrin very late antigen-4 (VLA-4; α4β1), provides a second lymphocyte-endothelium adhesion mechanism distinct from that of LFA-1 and ICAM-1. There is increasing evidence that VCAM-1 can mediate all three phases of leukocyte-endothelial cell interaction i.e., tethering, rolling, and adhesion (3, 14, 44). Neutrophils (80) and sickle reticulocytes (95) also adhere to VCAM-1 via VLA-4.

IRBC from the above parasite clone also adhere to E-selectin. The selectins are a family of differentially expressed adhesive molecules that recognize fucosylated and sialylated carbohydrate ligands (45). Both P-selectin and E-selectin are intimately involved in the initial interaction of neutrophils at the onset of acute inflammation. E-selectin is not constitutively expressed on endothelial cells but is induced (peak 4–6 h) after exposure to IL-1 and TNF-α (75). It supports the rolling of neutrophils on vascular endothelium in vitro (47) but probably at a later stage of inflammation than P-selectin, which can be quickly mobilized from an intracellular store in Weibel-Palade bodies (36).

More recently, chondroitin-4-sulfate (CSA), a glycosaminoglycan expressed throughout the microvasculature in association with various proteoglycans such as thrombomodulin, has been shown to mediate the cytoadherence of IRBC selected on Chinese hamster ovary cells (85). CSA is now recognized as the principal molecule mediating cytoadherence in the human placenta, where it is expressed on syncytiotrophoblasts (31). The unique interaction between IRBC and CSA in the placenta is discussed in cytoadherence in the placenta.


There are some inherent problems in extrapolating from results obtained in vitro with selected parasite lines and clones directly to malaria parasites causing clinical infections. The cytoadherent phenotype is known to be influenced by long-term culture in vitro (28, 100) and by the selection and cloning process itself (15). To determine the relative importance of the above receptor molecules for wild-type parasites, the cytoadherence of IRBC taken directly from the peripheral blood of patients with acute falciparum malaria has been examined in a number of studies using static binding assays (22, 35, 39, 60, 62,105). These heterogeneous populations of IRBC are cultured in vitro for 24–36 h (i.e., less than one asexual cycle) until they reach the cytoadherent trophozoite/schizont stages. Greater than 90% of clinical parasite isolates tested adhere to CD36, whereas ∼10% adhere to ICAM-1. For isolates that adhere to both molecules, the degree of adhesion to CD36 is at least 10-fold higher than adhesion to ICAM-1. Minimal or no adhesion to E-selectin, VCAM-1, or CSA is seen with most of the isolates. The cytoadherence to TSP, C32 melanoma cells, and purified CD36 is found to be directly proportional to parasitemia (35,39), expressed as the percent of infected erythrocytes. When the degree of cytoadherence to CD36 is compared at a fixed parasitemia, a range of intrinsic cytoadherent capabilities among different isolates becomes evident and in some instances correlates positively with the clinical severity of the infection (39, 62).

Immunopathological studies of postmortem tissues have also been performed to assess IRBC-endothelium interactions. In at least one study, colocalization of sequestration with ICAM-1 expression was noted, particularly in the brain (99). This finding may be interpreted as suggesting that ICAM-1 is the principal receptor for cytoadherence in the cerebral circulation. However, postmortem studies need to be interpreted with caution, as adhesion molecules could be expressed as a consequence of the circulatory disturbance and metabolic abnormalities resulting from cytoadherence rather than being its cause. For example, oxidant damage by adherent sickle cells has been shown to induce the expression of ICAM-1, E-selectin, and VCAM-1 on vascular endothelium (94).

From a population perspective, results from static binding assays suggest that ∼30% of IRBC in a given parasite isolate adhere to CD36, whereas 2–3% of IRBC are adherent to ICAM-1 (105). The percentage of IRBC adherent to E-selectin and VCAM-1 is negligible. This means that >60% of IRBC do not bind to any of the receptor molecules in vitro. There are several possible explanations for the in vitro observations. On the parasite side, not all IRBC may express the parasite proteins that are involved in cytoadherence. So-called “null” cells can appear after long-term culture in vitro (28), but their presence in overnight culture of wild-type parasites has not been shown. A high percentage of null cells is also unlikely in vivo, in that mature parasites are seldom seen in the peripheral blood, yet multiplication is remarkably efficient even in the face of splenic clearance. There may be other receptors for cytoadherence that have not been identified. Recent studies have suggested that platelet/endothelial cell adhesion molecule-1 may also have a role in mediating cytoadherence of clinical isolates (60). However, the minor degree of adhesion to this molecule would not account for the cytoadherence of the majority of the IRBC in a given isolate. A more likely scenario is that IRBC may need to interact with a number of endothelial receptor molecules for optimal adhesion. The interaction with some of the adhesion molecules such as E-selectin and VCAM-1, and possibly P-selectin, may be revealed only under flow conditions. To ensure its own survival, IRBC will likely use any readily available adhesion molecule to adhere to the vascular endothelium in vivo.


To address the above possibilities, studies have been performed using a parallel plate flow microscopy system that allows the interaction between IRBC and receptor molecules to be visualized directly under physiological flow conditions (104). IRBC from clinical parasite isolates were observed to interact with endothelial receptors in a stepwise process that involves tethering, rolling, and firm adhesion (Fig. 4). IRBC initially tether and then roll on CD36, ICAM-1, P-selectin, and VCAM-1. However, the strength of the rolling interaction with each receptor molecule varies, as reflected in differences in rolling velocity, and significant adhesion under shear is almost exclusively to CD36. Some IRBC bypass the rolling event and are arrested on CD36 immediately after tethering. There is no interaction with E-selectin under flow conditions.

Fig. 4.

Schematic diagram of the different phases of IRBC interaction with endothelium observed under physiological flow conditions.

The interaction between IRBC and P-selectin has been further characterized using this methodology (37). P-selectin is stored in Weibel-Palade bodies of endothelial cells and is rapidly translocated to the cell surface in response to various mediators including oxidants, histamine, thrombin, and cysteinyl leukotrienes (36). P-selectin may be important in mediating the capture and fast rolling of IRBC followed by slower rolling and attachment to CD36. The IRBC-P-selectin interaction is Ca2+ dependent and involves a sialic acid residue on IRBC. However, the monoclonal antibody G1, which inhibits the interaction between neutrophil and P-selectin, has no effect on IRBC rolling on P-selectin. This finding indicates that the primary binding site for IRBC on P-selectin is different from the epitope responsible for the binding of its natural counterreceptor P-selectin glycoprotein ligand-1 (58).

On C32 melanoma cells, which coexpress CD36 and ICAM-1, inhibition of rolling by an anti-ICAM-1 antibody reduces the subsequent adhesion of some parasite isolates to CD36. Similarly, inhibition of rolling by an anti-P-selectin antibody reduces IRBC adhesion to CD36 on activated platelets. The rolling interactions with molecules such as ICAM-1 and P-selectin appear to facilitate adhesion to CD36, even if they individually are of much lower avidity than that required to allow attachment. This is the first demonstration of synergism among receptor molecules for cytoadherence under flow conditions. There is also synergism between CD36 and ICAM-1 in mediating cytoadherence to human dermal microvascular endothelial cells under static conditions (56). Collectively, these results indicate that cytoadherence under physiological flow conditions may be mediated by multiple IRBC ligands that interact with different adhesion molecules in a cooperative fashion. The synergism may be particularly important in view of the fact that CD36 expression on microvascular endothelium does not appear to be upregulatable (74). In other words, the degree of cytoadherence of P. falciparum on vascular endothelium may be regulated at the level of expression of adhesion molecules such as P-selectin and ICAM-1 rather than that of CD36.

On the other hand, there are also parasite isolates that appear to be able to interact exclusively with CD36 on either C32 cells or platelets. These results highlight the diversity of the adhesive properties of clinical P. falciparumisolates relative to the more homogeneous adhesive mechanisms used by leukocytes. There are other striking differences between the adhesive interactions of IRBC and leukocytes with endothelial receptors. First, adhesion of IRBC to CD36 can occur without prior interaction with another molecule, whereas leukocytes do not adhere directly. Second, ICAM-1 acts as a receptor for firm adhesion of neutrophils but cannot tether or support rolling of these cells except at extremely low shear. In contrast, ICAM-1 mediates the tethering and rolling of IRBC. Third, leukocytes require activation after the initial interaction with endothelial receptors, which permits subsequent firm adhesion. However, IRBC are able to roll and adhere firmly to endothelial receptors without notable activation or the obvious requirement for chemotactic molecules, because the adhesive interactions can occur on formalin-fixed transfectants expressing the adhesion molecules (105) or endothelial cells (23).


Falciparum malaria adversely affects pregnancy (17, 18). In areas of intense transmission, where symptomatic disease in adults is rare, these adverse effects are largely confined to the fetus of the first pregnancy (57). There is intrauterine growth retardation and a reduction in birth weight (18). Infections with P. falciparum in nonimmune pregnant women tend to be severe (50). Pregnant women with cerebral malaria have a case mortality rate over twice that of nonpregnant adults. Premature labor and fetal death are common.

The placenta is a site of preferential IRBC accumulation (19), but on microscopic examination this sequestration appears to occur in the absence of cytoadherence, because IRBC are seen free in the intervillous space and not adherent to endothelial surfaces (54). Experimental evidence has shown that, although CD36 appears to be the main receptor for IRBC on venular endothelium, parasites obtained directly from human placentas adhere to CSA expressed on syncytiotrophoblasts that line the placental intervillous space (31). The CSA may be presented on proteoglycans that extend well beyond the cell surface. Moreover, immune sera from multigravid women but not males or primigravid women inhibit cytoadherence to CSA (32). The presence of anti-CSA antibodies may well explain why the falciparum-malaria-associated reduction in birth weight is lower or absent in women with multiple pregnancies.


At least five parasite-derived proteins are associated with the cell membrane of an infected erythrocyte at various stages of the developmental cycle (Fig. 5) (42). Three of the malarial proteins are associated with knobs [P. falciparum erythrocyte membrane proteins 1 and 2 (PfEMP1 and PfEMP2) and PfHRP1 or KAHRP]. Of these, PfEMP1 is the only protein that extends beyond the cell surface to mediate cytoadherence (55), whereas PfEMP2 and PfHRP1 or KAHRP remain on the internal face of the erythrocyte membrane in association with electron-dense material. The proteins appear to be exported from the intracellular parasite through the erythrocyte cytoplasm to the surface membrane via a complex system of vesicle trafficking pathways (77).

Fig. 5.

Schematic diagram of topological distribution of P. falciparum proteins in surface membrane of infected erythrocytes. Lipid bilayer of red blood cell membrane (RBCM) is indicated, together with the cytoskeleton and electron-dense material (EDM) under knobs. PfEMP, P. falciparum erythrocyte membrane protein; PfHRP,P. falciparum histidine-rich protein [also known as knob-associated histidine-rich protein (KAHRP)]; RESA, ring-infected erythrocyte surface antigen. [From Howard (42), reproduced with permission of S. Karger AG, Basel.]

PfEMP1 consists of a family of diverse-size (200–350 kDa) surface proteins that are linked to the erythrocyte cytoskeleton. The protein is encoded by a large family of vargenes. There are ∼50–150 copies of thevar gene per genome and a different set of var genes in eachP. falciparum clone (92). Several members of the family of var genes that encode PfEMP1 have been cloned. The protein has four Duffy-binding-like (DBL) domains, each containing five consensus motifs rich in cysteine residues (11, 93). Three cysteine-rich motifs, denoted CRM-1 to CRM-3, are located after DBL-1, DBL-2, and DBL-4, respectively. The presence of an RGD motif (amino acids 1212–1214) and an LDV motif (amino acids 142–144) in the extracellular domain suggests that PfEMP1 may participate in ligand-receptor interactions involving the integrin family. There is a single transmembrane domain, followed by a presumed intracellular domain encoded by the 3′ exon. Tryptic fragments of PfEMP1 cleaved from the surface of IRBC bind to CD36, ICAM-1, and TSP (9). The epitope on PfEMP1 that interacts with CD36 has been mapped to the rC1–2 region of CRM-1, which corresponds to amino acids 576–808 (10). Recently, CSA-adherent parasites have been shown to transcibe a particularvar gene with the production of a distinct variant PfEMP1, and inhibition studies with antibodies raised against different domains of the protein suggest that the binding site may be located on DBL-3 (79).

In addition to variations in molecular size, PfEMP1 undergoes a high rate (∼2% per cycle) of antigenic variation in vitro (83), with resultant changes in the cytoadherent phenotype. This considerable antigenic variation could contribute to immune evasion and thus to the survival of the blood stage infection. Whether this degree of antigenic plasticity occurs in vivo has yet to be firmly established. Studies of clinical parasite isolates would suggest that CD36 exerts a strong selective pressure on the IRBC cytoadherent phenotype of these heterogeneous parasite populations. Recent data in isogenicP. falciparum populations with defined adhesive phenotypes for CD36, ICAM-1, and CSA by selective panning indicate that multiple var genes may be transcribed early in the parasite cycle. However, a silencing mechanism appears to come into play as the parasite matures, so that only a specific var gene is transcribed in trophozoites (85).

A molecule similar to PfEMP1, of molecular mass 270 kDa, called sequestrin, has been demonstrated on the surface of infected erythrocytes using anti-idiotypic antibodies raised against OKM8, a monoclonal antibody for CD36 (63). This finding further strengthens the hypothesis that CD36 is the major receptor on vascular endothelium for the parasite cytoadherent ligand.

Two other parasite proteins, ring-infected erythrocyte surface antigen (RESA) and PfHRP2, are associated with the erythrocyte membrane but are not localized specifically to the knobs (Fig. 5). PfHRP2 is secreted into the circulation and is currently being exploited for diagnostic assays. The 155-kDa RESA antigen is transferred from the merozoite to the erythrocyte membrane during invasion (2). Once transferred, it is thought to be entirely submembranous. RESA has been shown to have cross-reactive epitopes with band 3 protein (41), the human erythrocyte anion transporter, and this too has been implicated as a ligand for cytoadherence mediated by CD36 (26). Presumably, changes in the erythrocyte cytoskeleton that occur as a result of parasitization could expose previously cryptic host molecules (neoantigens) on the cell surface.

A few other cytoadherent ligands remain to be defined. The sialylated P-selectin ligand is trypsin sensitive, as is PfEMP1. The major ligand for VCAM-1 is VLA-4 (α4β1). VLA-4 has been shown to mediate the adhesion of sickle reticulocytes to endothelial cells (95). IRBC may make use of the same mechanism. VLA-4 is expressed by late erythroid progenitors and erythroblasts but is not detectable on mature erythrocytes (73). It is conceivable that VLA-4 may be reexposed on the surface of the erythrocyte due to topographical changes induced by the maturing intracellular parasite or that IRBC may express a VLA-4-like ligand that is of parasite origin. An LFA-1-like ligand has been demonstrated on one clinical P. falciparum isolate, and an anti-LFA-1 monoclonal antibody partially inhibited its adhesion to cytokine-stimulated cerebral microvascular endothelium (96).

Regardless of the eventual identity of the cytoadherent ligand(s), a conserved component must be present, since all P. falciparum parasite isolates causing natural infections cytoadhere. In addition, there must be a strain-variable component, since inhibition or reversal of cytoadherence by immune sera occurs in a strain-specific manner (101). The constant and variant components could be either closely associated molecules or different epitopes on the same molecule.


The functional consequences of the adhesion process have been little studied. It is now well recognized that adhesion molecules are not just innocent bystander molecules that function only to mediate leukocyte interactions with endothelium but that they also initiate signal(s) that can activate and regulate adhesion-dependent leukocyte function, β2-integrin activation, O2 production, and T cell receptor stimulation (27, 51). A functional consequence of cytoadherence might be the activation of intracellular signaling pathways in the endothelial cell, leading to changes in the status of the adhesion molecules and/or gene expression of immunoregulatory molecules such as cytokines and nitric oxide synthase, which would modify the outcome of the infection. The triggering of downstream events by cytoadherence would be consistent with the observation of a respiratory burst following adhesion of IRBC to CD36 on monocytes (64). In this connection, it is interesting to note that CD36 is present in a subcompartment in the plasma membrane known as caveolae, which are present on endothelial cells. Caveolae are flask-shaped invaginations that have been proposed to represent a site of signal transduction. Although the exact mechanism by which the signal is activated is unknown, the src family kinases have been shown to colocalize in caveolae (48) and have been shown to coprecipitate with CD36 (29).

Signal transduction might also be induced in erythrocytes, even though they are often considered biologically inactive cells. Signal-dependent translation of a number of proteins has been described in activated human platelets, which, like erythrocytes, lack nuclei, cannot synthesize mRNA, and are considered incapable of regulating protein synthesis (106). More importantly, signal transduction might involve the metabolically active parasite via a putative parasitophorous duct pathway that allows for the transport of macromolecules directly from the environment into the parasite (78). A glycosylphosphatidylinositol (GPI) anchor moiety of P. falciparumhas been shown to induce nitric oxide synthase expression in HUVEC by a protein tyrosine kinase-dependent and protein kinase C-dependent signaling pathway (97). The toxin also upregulates ICAM-1, VCAM-1, and E-selectin expression in HUVEC via tyrosine kinase-dependent signal transduction (88).


The ultimate goal of any research on cytoadherence is the development of safe and effective methods of inhibiting or reversing the process. Soluble mediators such as cytokines, chemokines, nitric oxide, and certain drugs are known to regulate adhesion molecule expression and thus may modulate cytoadherence. Furthermore, these agents may have direct effects on IRBC or endothelial cells.

The effect of cytokines on cytoadherence is well documented. TNF-α, IL-1, and IFN-γ enhance the adhesion of IRBC from laboratory-adapted parasites to endothelial cells under static conditions (13, 43). These cytokines are known to be elevated in patients with P. falciparum malaria (34, 46). However, the effect of the cytokines on the rolling and adhesion of clinical parasite isolates to microvascular endothelium under shear stress has not been determined. Furthermore, the effect of other cytokines, either singly or in combination with the proinflammatory cytokines, is unknown. The effect of IL-10, an immunosuppressive cytokine produced by Th2 cells as well as monocytes and B cells, is of particular interest, because it inhibits TNF-α, IL-1, and IL-6 production by peripheral blood mononuclear cells in response to malarial antigens in vitro (38). It has also been shown to downregulate ICAM-1 expression induced by IFN-γ on monocytes (109). In fatal falciparum malaria, there appears to be a defect in the IL-10 control of proinflammatory cytokine production (N. P. Day and M. Ho, unpublished observations). IL-4 is also produced by some patients in response to P. falciparum, and it may contribute to cytoadherence by upregulating VCAM-1 expression (71).

Chemokines are small proinflammatory peptides that have a crucial role in the mobilization of the cells of the immune system through their leukocyte chemoattractant activity. The cytoadherence of IRBC to formalin-fixed target cells suggests that the process can occur in the absence of any chemotactic agents in vitro. However, this does not exclude a role for chemokines in enhancing cytoadherence in vivo. Human erythrocytes as well as postcapillary venular endothelial cells express the promiscuous Duffy antigen receptor (DARC) that binds to chemokines of both the C-X-C (e.g., IL-8, MGSA/gro, NAP-2) and C-C (e.g., RANTES, MCP-1) classes with high affinity (59). The presence of DARC on erythrocytes suggests that IRBC may respond to a chemotactic gradient. Plasma chemokine levels, including IL-8 and MIP-1, are elevated in falciparum malaria (20, 33).

Nitric oxide has been postulated to have both pathogenic and protective roles in severe falciparum malaria. Nitric oxide could derive from different synthetic enzymes in neuronal tissue or vascular endothelium or from an inducible form found in phagocytic cells. In acute malaria, increased nitric oxide production would be expected as a result of the proinflammatory response and microcirculatory ischemia. Moreover, nitric oxide synthase expression in endothelial cells can be induced by a GPI toxin from P. falciparum (97). Nitric oxide derivatives have been shown to inhibit the growth of P. falciparum in vitro (84). Nitric oxide may exert an inhibitory effect on IRBC rolling and adhesion, just as it inhibits neutrophil rolling on endothelial cells in models of ischemia reperfusion (30) and platelet aggregation (90). In patients with malaria, an inverse relationship between the severity of the infection and nitric oxide production/nitric oxide synthase expression has been demonstrated by some (4) but not others. The precise role of this evanescent molecule in host protection against malaria, vasoregulation, and the pathogenesis of cerebral malaria remains to be determined.

Last, antimalarials have been shown to inhibit cytoadherence at concentrations that inhibit overall nucleic acid and protein synthesis (103). Drugs that affect cytoadherence specifically have not been described. We have observed that, with suboptimal doses of the iron chelator desferrioxamine, the degree of inhibition of adhesion to CD36 consistently exceeds overall growth inhibition, suggesting that an additional mechanism may be involved in the activity of this drug against cytoadherence. Desferrioxamine has been shown to inhibit the transcription of several mitochondrial enzymes of P. falciparum (S. R. Meshnick, unpublished observations). The expression of the parasite cytoadherent ligand PfEMP1 may be similarly affected by environmental conditions such as low intraerythrocytic iron.


In human P. falciparum malaria infection, IRBC either sequester or are removed from the circulation primarily by the spleen. The balance between splenic clearance and sequestration, which allows the parasite to survive to initiate a new life cycle, is a major determinant of the rate of increase and magnitude of the infecting parasite burden. Within this paradigm, pathogenicity is proportional to the size of the sequestered parasite burden and the pattern of vital organ sequestration. In the past decade, detailed molecular studies have provided exciting new insight into the process of cytoadherence. The next challenge lies in translating the advances in our understanding of pathogenesis into improved treatment for the many millions who are affected by falciparum malaria.


We are grateful to Dr. Paul Kubes for reviewing the manuscript.


  • Address for reprint requests and other correspondence: M. Ho, Dept. of Microbiology and Infectious Diseases, 3330 Hospital Drive NW, Calgary, Alberta, Canada T2N 4N1 (E-mail:mho{at}

  • The cited research by the authors is supported by the World Health Organization, the Medical Research Council, Canada, the Alberta Heritage Foundation for Medical Research, Alberta, Canada, and the Wellcome, Mahidol University-Oxford Tropical Medicine Research Programme funded by the Wellcome Trust of Great Britain.

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.


View Abstract