Despite progress in malaria control, malaria remains one of the most important infectious diseases worldwide. In 2020, about 267 million malaria cases, including 409,000 deaths, were recorded [1]. Concerning all malaria parasites, the deadliest, Plasmodium falciparum, has a complex life cycle that alternates between Anopheles mosquitoes and humans. The asexual cycle that takes place in humans consists of the liver stage (multiplication of the parasite in hepatocytes) and the intraerythrocytic cycle (multiplication of the parasite in erythrocytes). Each intraerythrocytic cycle lasts approximately 48 h, during which the merozoite that invaded the erythrocyte develops through the ring stage to the trophozoite and finally the schizont. At the end of the intraerythrocytic cycle, the newly formed merozoites are released, ready to invade new erythrocytes. Some merozoites develop into gametocytes, which must be taken up by female Anopheles mosquitoes to complete their sexual development. The complexity of the parasite’s life cycle and its masterful ability to evade its elimination by the host immune system challenge efforts to combat the disease.

To survive in the human host, P. falciparum has evolved unique mechanisms, two of which, called sequestration and antigenic variation, rely mainly on a highly diverse protein family, the P. falciparum erythrocyte membrane protein 1 (PfEMP1). The PfEMP1s, which the parasite exposes on the surface of its host cell from the trophozoite stage onwards, have at least a dual function. First, they bind to various endothelial cell receptors (ECRs) on the walls of blood vessels (sequestration), thereby disappearing from the peripheral circulation and bypassing removal by the spleen. Unlike other plasmodial species, the deformability of P. falciparum infected erythrocytes (PfIEs) decreases as the parasite matures, so that circulating trophozoites and schizonts would be retained in the spleen and removed from circulation by resident macrophages ^{[2][3][4][5][6]}[2,3,4,5,6]. Second, PfEMP1 represents the main target of the humoral immune response [7], but due to the presence of numerous copies of var genes encoding PfEMP1, the parasite can sequentially present different PfEMP1 variants on the surface of its host cell and use them for sequestration.

2. The PfEMP1 Family

The PfEMP1 family is encoded by about 45–90 var genes per parasite genome ^[8][12]. Expression of the var genes is mutually exclusive in ring-stage parasites, such that only a single PfEMP1 variant is present on the surface of trophozoite- or schizont-stage PfIEs at any given time ^[9][10][13,14] for review ^[11][10]. Mutually exclusive expression relies on very complex mechanisms. These are based on both epigenetic regulation and cis-acting DNA elements and RNA transcripts involved in var gene activation and silencing (for review ^[11][10]). The var genes and their encoding PfEMP1s vary greatly from parasite to parasite, and recombination constantly generates new variants, so there is an enormous repertoire of var genes in nature ^{[9][10][12][13]}[13,14,15,16]. The molecular masses of PfEMP1s range from 150 to 400 kDa. These proteins consist of an intracellular acidic terminal segment (ATS domain), a transmembrane domain, and a variable, extracellularly exposed region responsible for receptor binding. This extracellular region contains a single N-terminal segment (NTS; main classes A, B, and pam) and a variable number of different Duffy binding-like domains (DBL; main classes α–ζ and pam) and cysteine-rich interdomain regions (CIDR; main classes α–δ and pam) ^{[14][15][16][17]}[17,18,19,20]. Approximately two-thirds of var genes localize in the subtelomeric regions of the chromosomes. Most of the subtelomeric and central localized var genes are located in regions of electron-dense heterochromatin at the nuclear periphery, with the active var gene shifting to the region of lower electron density ^[10][14]. Depending on the chromosomal localization, the upstream sequence, and the direction of transcription of the var genes, PfEMP1s can be classified as A, B, C, or E ^{[14][18][19][20]}[17,21,22,23]. A few conserved, strain-transcendent var variants have been described: var1, var2csa (group E), and var3. The var1 gene occurs in two variants in the parasite population, var1-3D7 and -IT, is often conserved as a pseudogene, and the encoded protein may not be presented on the erythrocyte surface ^[8][21][12,24]. VAR2CSA has an atypical domain architecture, mediates binding to chondroitin sulfate A (CSA) in the placenta, and is, thus, important in pregnancy-associated malaria ^[22][25]. VAR3 proteins are very short PfEMP1s with unknown receptor binding phenotypes ^[23][26]. Analysis of 399 different PfEMP1 sequences from seven P. falciparum genomes allowed the identification of 23 domain cassettes (DCs) that could be important for protein folding and binding to human ECRs, as well as for reflecting recombination breakpoints ^[14][17]. About 10% of PfEMP1s variants belong to group A and are usually longer proteins with a head structure that includes a DBLα1 and either a CIDRα1 domain (CIDRα1.4–7) that binds to the endothelial protein C receptor (EPCR) or a CIDRβ/γ/δ domain with unknown receptor binding phenotype ^[8][12]. Groups B and C make up the majority of PfEMP1s (at least 75%) and typically have DBLα0-CIDRα2–6 head structures that bind to CD36, followed by only two additional extracellular domains (DBLδ1, CIDRβ/γ). A subset of chimeric B-type proteins (group B/A, also known as DC8-containing proteins) has a DBLα2 domain (chimeric DBLα0/1 domain) and an EPCR-binding CIDRα1.1 or CIDRα1.8 domain typically attached to a complement component C1q receptor (C1qR)-binding DBLβ12 domain ^{[24][25][26][27][28][29][30][31]}[27,28,29,30,31,32,33,34]. Thus, the head structure confers mutually exclusive binding properties to either EPCR (14%), CD36 (72%), CSA (3%), or to one or more unknown ECRs via the CIDRβ/γ/δ domains (10%) or VAR3 (1%) ^[32][35].

3. Knobs—Anchor Point for PfEMP1s

PfEMP1s are concentrated in nanoscale, electron-dense protrusions of the plasma membrane of PfIEs, the so-called knobs. They are formed in erythrocytes about 16 h after parasite invasion and reach their highest density 20 h after infection ^[33][34][39,40]. Single knobs have a hemispherical ellipsoid shape with a minor axis of 20 nm and a major axis of 120 nm ^[35][41]. Knobs are composed of various submembrane structural proteins, including the major protein of this structure, knob-associated histidine-rich protein (KAHRP). These consist of PfEMP3, the ring-infected red cell antigen (RESA), the mature parasite-infected red cell surface antigen (MESA)/PfEMP2, and Pf332 ^[35][36][41,42]. The knobs consist of a highly organized skeleton made of a spiral structure located beneath specialized areas of the erythrocyte membrane ^[37][43]. The arrangement of PfEMP1s in a cluster near the top of the knobs is assumed to increase the binding capacity of PfIEs, especially under flow conditions (see below) ^{[38][39][40][41]}[44,45,46,47].

4. P. falciparum and CD36

Looking at the PfEMP1 family, the question arises why, depending on the parasite genome, between 75–85% of var genes encode PfEMP1s, which have a CIDRα2–6 domain for CD36 binding ^[14][24][28][17,27,31]. Interestingly, the CIDRα domains were shown to be present only in the P. falciparum-containing branch (clade B) of the Laverania subgenus. This could indicate that the binding to CD36 provides a selective advantage for P. falciparum ^[42][103]. What kind of selection advantage this was is yet unclear. What advantage does the parasite have in retaining this large number of CD36-binding PfEMP1 variants in its genome? Additionally, what is the difference between the individual variants or, more generally, between CD36 binding mediated by group B or C PfEMP1s?

5. CD36

CD36 is a pattern recognition receptor (PRR) that belongs to the class B scavenger receptor family. It is a glycoprotein present in many tissues and involved in several key processes. These include lipid processing and uptake, thrombostasis, glucose metabolism, immune function, angiogenesis, and fat taste (for review ^{[43][44][45][46][47]}[104,105,106,107,108]. CD36 is found on platelets, mononuclear phagocytes, adipocytes, hepatocytes, myocytes, some epithelia and, as mentioned above, expressed on the endothelia of liver, spleen, skin, lung, muscle, and adipose tissue ^[48][49][50][81,82,109]. On microvascular ECs, CD36 is a receptor for thrombospondin-1 and related proteins and functions as a negative regulator of angiogenesis. At least 60 variants have been described in the coding region of the CD36 gene. The mutations of CD36 caused by gene variants can also influence the adhesion of PfIEs and ECs. This could directly influence the severity of a malaria infection via the degree of cytoadhesion. There are several studies on this, but with contradictory results ^[51][110]).

6. CD36 Binding PfEMP1 Variants—Benefits for Parasite and Host

Several observations may help explain why a large number of CD36-binding PfEMP1 variants is not only beneficial for parasite development, but may also be an advantage for the infected host.

• The parasite targets a region of CD36 that is essential for its physiological role in fatty acid uptake because mutation of F153 disrupts the interaction of CD36 with CIDRα2–6 but also abolishes the binding of CD36 to oxidized LDL particles. This reduces the likelihood that the human host can escape from PfEMP1 binding by altering its CD36 ^[25][28].
• In contrast to the EPCR binding surface of CIDRα1 domains, which protrudes and is a structure that is likely to be well recognized by antibodies, the CD36 binding site is concave, and the conserved hydrophobic residues are hidden in a pocket, so maybe they are less easily recognized. In addition, the binding site is surrounded by a sequence-diverse protein surface containing a flexible loop that may make antibody recognition less likely. This unique interaction site of the parasite with CD36, which protects essential residues from exposure to the immune system, appears to allow the parasite to utilize an antigenically diverse set of CIDRα2–6 for cytoadhesion to CD36 to be protected from splenic clearance ^[25][28].
• CD36 is found in cells of the innate and adaptive immune system ^{[43][44][[47]}[104^45][46],105,106,107,108]. It has been shown that PfIEs can adhere to dendritic cells (DCs). This attachment inhibits maturation of these cells and their ability to stimulate T cells. Thus, the parasite can trigger dysregulation of the immune system. This favors the development of the parasite by impairing the host immune system’s ability to clear the infection ^{[47][52][53][54][55]}[108,111,112,113,114]. However, there is also an observation that the mechanism of DC inhibition by PfIEs may be independent of PfEMP1 and CD36 ^[56][115].
• The previously determined hierarchy of var expression upon parasite entry into human blood begins with group B and suggests that most parasites bind to CD36, as they all encode a CD36-binding phenotype. Most infected individuals, including those who are not immune, do not develop severe malaria, and cytoadhesion of PfIEs occurs in extensive microvascular beds in tissues other than the brain (skin, muscle, adipose tissue). Therefore, cytoadhesion in such non-vital tissues could promote survival and transmission of the parasite while minimizing host damage and death ^{[57][58][59][60]}[87,88,89,90].
• Antibody-induced selective binding and internalization of CD36 do not result in proinflammatory cytokine production by human macrophages. Interestingly, CD36-mediated phagocytosis of PfIEs also did not result in cytokine secretion by primary macrophages ^[61][116]. However, CD36-mediated binding of PfIEs increases the likelihood of phagocytosis by macrophages. This can lead to a reduction in parasitemia, but also allows the parasite to maintain a viable infection without causing too much damage to the host through high parasitemia ^{[47][55][62][63]}[108,114,117,118].
• DCs react to P. falciparum very early during infection and can, thus, influence the development of immunity. Internalization of PfIEs by DCs and subsequent pro-inflammatory cytokine production of DCs, NK, and T cells depends on CD36. Notably, plasmacytoid DCs regulate innate and adaptive immunity to malaria via the production of proinflammatory cytokines. As this effect is particularly evident at low levels of parasitemia, the role of CD36 for malaria immunity appears to take place early during infection and to promote the development of protective immunity against malaria ^[63][64][118,119].

7. Binding Phenotypes of PfIEs

Cytoadhesion of PfIEs is divided into the three phases: “tethering”, “rolling”, and “immobilization”, comparable to leukocyte diapedesis ^[65][66][67][120,121,122]. However, the dynamics of cytoadhesion of PfIEs to the vascular endothelium is controversial. For example, some authors describe cytoadhesion to ICAM-1 as rolling, and to CD36 as stationary, or vice versa ^{[68][69][70][71][72]}[123,124,125,126,127]. However, there is increasing evidence that PfIEs are very likely to roll over CD36 ^{[71][72][73][74][75][76]}[126,127,128,129,130,131]. Recently, the binding phenotype for different ECRs was investigated using a laminar flow system with transgenic Chinese hamster ovary (CHO) cells carrying different ECRs on their surface ^[72][127]. Rolling was observed upon interaction with CD36, and the rolling behavior of disc-shaped PfIEs at the trophozoite stage (flipping) differed from the rolling behavior of round-shaped PfIEs at the schizont stage (continuous rolling). Moreover, PfIEs in the schizont stage roll more stably than PfIEs in the trophozoite stage at different shear stresses ^[72][127]. The rolling motion of PfIEs was also seen on transgenic mouse fibroblasts presenting CD36 ^[73][128] and on recombinant CD36 instead of transgenic eukaryotic cells ^[74][129]. As described above, the dermal endothelium has large amounts of CD36. Rolling movements of PfIEs have also been found on dermal ECs, as well as on human skin grafts, on which large amounts of CD36 are found ^{[71][73][75][76]}[126,128,130,131]. Additionally, last but not least, the rolling CD36 binding phenotype was also confirmed by in silico modeling ^[77][78][132,133]. However, depending on the experimental setup, the parasite isolates used, and the parasite stage, different velocities were measured at similar shear forces. For trophozoite-stage parasites confronted with recombinant CD36, average velocities between 140 µm/min to 680 µm/min were measured at a shear force of 1.6 Pa, depending on the isolate ^[74][129]. When transgenic CHO cells presenting CD36 on the surface were used instead of recombinant CD36 in a similar experimental setup, average velocities ranging from 11 µm/min to 33 µm/min, i.e., a 12–20 fold lower value, were measured, also depending on the parasite stage and isolate ^[72][127]. If PfIEs cytoadhere for approximately 30 h during their intraerythrocytic development, they travel distances between 25–122 cm or 2–6 cm, respectively, depending on the experimental setup ^[72][74][127,129]. In both cases, however, the probability of passing over the spleen and being removed accordingly is low. Further studies showed that initial contact of PfIEs to CD36 under flow conditions activates Scr-family kinases, leading to dephosphorylation of CD36 via p130CAS signaling. This increases the binding affinity of PfIEs to CD36 and, thus, leads to increased adhesion of the PfIEs. This mechanism also leads to actin cytoskeletal remodeling and subsequent CD36 clustering, which further increases PfIE adhesion ^[73][76][79][128,131,134]. It is postulated that a small number of strongly adherent PfIEs activate the endothelium, and thus enhance the cytoadhesion of most parasites ^[76][131]. However, the binding mode of PfIEs also seems to be strongly dependent on the respective ECR. For ICAM-1, CD9, P-selectin, as well as CSA, stationary binding, instead of rolling, was observed under flow conditions ^[72][127]. Stationary binding to ICAM-1 was also demonstrated in an earlier study ^[71][126]. However, while binding to CD36 occurred at shear forces below 4 dyn/cm², binding to ICAM-1, CD9, P-selectin, and CSA occurred mostly at lower shear forces (from 2 dyn/cm²) ^[72][127]. Of note, the origin and environment of the ECR studied (recombinant or presented on eukaryotic cells) also seems to be important for characterising the binding phenotype. Antia and colleagues observed a rolling binding type for PfIEs, with an average rolling velocity of about 10 µm/s at 1–2 kPa and of 1–3 µm/s when recombinant ICAM-1 or CD36 was used, respectively ^[68][123]. Interestingly, in the same study, stationary binding of PfIEs, as also described by Lubiana and colleagues ^[72][127], was observed on transgenic CHO cells presenting ICAM-1 ^[68][123]. However, the binding showed large variations. Thus, the PfIEs came to a standstill for a few seconds, but were then also able to detach from the CHO cells again ^[68][123].