Placental malaria is a distinct form of the disease. Unlike IE infecting non-pregnant individuals, IE isolated from the placenta generally have a highly conserved phenotype (adhesion to CSA), and predominantly transcribe one var gene, termed var2csa [6, 11, 12, 14, 21]. Antibodies that develop naturally in pregnant women cross-react in blocking adhesion of placental isolates from different regions to CSA, suggesting conserved epitopes may mediate protection . Var2csa appears to be the most important gene encoding adhesion to CSA and HA; targeted disruption of var2csa ablated adhesion to CSA, and after repeated panning little or no CSA adhesion was regained [16, 17].
PfEMP1 family members other than VAR2CSA, or different proteins, might also mediate placental adhesion, and IE might accumulate in the placenta by adhering to other ligands expressed on the SCT. To investigate this possibility, CS2KO IE were selected by repeated panning on the BeWo trophoblast cell line, which has been shown to be a useful in vitro model of IE adhesion to SCT [18–20, 31]. CS2KO IE selected on BeWo gradually increased adhesion to BeWo, which reached a plateau after 7 cycles of panning. Adhesion of selected CS2KO IE remained lower than that of wild type CS2 IE despite repeated cycles of selection, and adhesion was not CSA-dependent. Instead, CS2KO9 IE adhered to ICAM-1, which is known to be expressed on SCT and BeWo cells and upregulated in malaria infection . Importantly, in clinical studies placental IE do not adhere to ICAM-1 [6, 21]. Many parasite isolates from children adhere to ICAM-1 [36, 37], and adults frequently have antibodies to VSA expressed by such isolates , so host immunity may prevent them from establishing infection in pregnant women. ICAM-1 did not appear to be the principal ligand for CS2KO9 IE on BeWo as parasite line E8B, which binds to ICAM-1 and CD36 but not to CSA or HA, adhered to BeWo at significantly lower levels than CS2KO9, and a blocking antibody to ICAM-1 did not significantly decrease adhesion of CS2KO9 to BeWo.
On unfixed snap-frozen placental tissues (Figure 3) CS2KO9 bound at low levels in the intervillous space and to the SCT compared to CS2, and also bound to villous tissue (to which IE have no access in vivo), probably using CD36 as a ligand. Adhesion of CS2KO9 was not inhibited by a blocking antibody to ICAM-1. Normal, uninfected placental tissue was used, and it is possible that placental malaria results in induction of expression of other parasite receptors on villi or secreted into the intervillous space. Nevertheless, the present findings suggest that CS2KO IE selected on BeWo bind at low levels to available placental ligands, and that VAR2CSA mediated adhesion is the major driving force behind placental sequestration.
Recently, Viebig et al used the FCR3 parasite line with disrupted var2csa [16, 17] to study adhesion of FCR3KO parasites to BeWo cells, in experiments similar to those reported here . The overall conclusion was similar, that an intact var2csa gene appears essential for IE to show a pregnancy malaria phenotype, but there were some differences in findings. Like CS2KO, the FCR3KO parasites showed some ICAM-1 adhesion; but unlike CS2KO, adhesion to BeWO was partly blocked by anti-ICAM-1. This may be because several var genes (at least two abundantly transcribed) were detected in the FCR3KO, whereas by reverse transcription and by Northern blot, we detected a single, different var in CS2KO, termed var19. FCR3KO parasites selected on BeWo were not recognized in a parity-dependent manner by immune sera, whereas one of our sera sets showed parity dependent recognition of CS2KO, and we saw low levels of adhesion of CS2KO. Neither isolate showed significant specific adhesion to SCT. Together these data suggest that IE panned on BeWo may bind to multiple receptors, but that these are not necessarily relevant to the pathogenesis of placental sequestration or pregnancy malaria.
Gender and parity dependent antibody responses develop in pregnant women, which are directed against VSA expressed by placental or CSA binding IE . These responses have been associated with protection from anaemia, premature birth and low birth weight babies in some studies [40, 41]. If IE with a phenotype similar to CS2KO9 are an important cause of placental malaria, then a similar pattern of antibody responses would be expected. By flow cytometry, typical gender and parity dependent responses to CS2 were observed in samples from Malawian and Papuan New Guinean men and women, as expected. CS2KO9 IE were frequently recognized by malaria exposed adults. In one series, similar gender and parity-dependent trends were observed for responses to CS2KO9, but there was no evidence of these in the other series. It is possible that the gender and parity dependent recognition we observed for CS2KO9 reflects the ability of similar variants to elicit pregnancy-specific immunity under certain circumstances. These data suggest that variants expressing similar VSA to It4var19 could be a minor cause of malaria in pregnancy in some cases, but they contribute little to placental sequestration. It4var19 may determine the phenotype of CS2KO9. The receptor for CS2KO9 IE adhesion to BeWo remains unknown.