TEXT

Top Abstract Text References

Under in vitro conditions, Plasmodium falciparum-infected erythrocytes (P. falciparum-IE) may show rosette formation, the adherence of two or more uninfected erythrocytes to an erythrocyte containing a mature-stage parasite. Rosette formation has been associated with severity of malarial disease in young children in some studies (4, 10, 12, 15, 18, 20), but not all (1, 17), and has been proposed to contribute to microvascular obstruction in organs such as the brain (9). It was reported to be rare in isolates from pregnant Cameroonian women (11).

In regions of malaria endemicity, pregnant women are more susceptible to malaria infection than their nonpregnant counterparts (3). Placental malaria infection is clearly of major importance in the pathogenesis of malaria in pregnancy, with mature-stage IE frequently found in the placental intervillous spaces. Only a proportion of placental IE appear to be adherent to the syncytiotrophoblast (23, 24), and many of the erythrocytes in the intervillous spaces are uninfected. We postulated that placental accumulation of IE may result in part from rosette formation, which could lead to disturbances in blood flow (22). We therefore compared rosette formation by parasites cultured from peripheral blood of children and pregnant women with that of P. falciparum isolates obtained from placental blood at delivery.

Peripheral and placental blood samples were collected from pregnant women and peripheral blood was collected from children at Queen Elizabeth Central Hospital, Blantyre, Malawi, after obtaining informed consent. Peripheral blood was processed and cultured as described previously, in medium supplemented with 10% human blood group AB serum (2). Placental blood was collected from the cut maternal surface of the placenta, diluted to 5% hematocrit in complete medium supplemented with 10% AB serum, and examined for rosetting after incubation at 37°C for a minimum of 60 min. Rosetting assays were performed by staining IE with acridine orange and by examination with combined light and fluorescence microscopy, as previously described (16). A rosette was defined as two or more uninfected erythrocytes adherent to an IE. Each isolate was tested in duplicate, and a minimum of either 200 trophozoite-containing IE or 100 high-power (40× objective) fields were examined, and samples with one or more rosettes identified were classed as rosetting isolates. Rosette formation is reported as a percentage of all trophozoite-containing IE in rosettes.

Pearson's ² test and the Kruskall-Wallis test were used to compare qualitative and quantitative variables, respectively, by using Stata version 5.0 (Stata Corp., College Park, Tex.). Ethical approval was obtained from the College of Medicine Research Committee, University of Malawi.

Twenty-three of 38 isolates cultured from the peripheral blood of pregnant women and 5 of 22 isolates from placental blood examined ex vivo formed rosettes, whereas 136 of 139 isolates from children examined over the same period formed rosettes (Table 1). The percent rosette formation (the percentage of IE from a given sample found to be in rosettes) was lower for placental blood isolates than for maternal peripheral blood isolates (P = 0.002) and lower for maternal peripheral or placental blood isolates than for isolates from children (P < 0.001) (Fig. 1). The prevalence of rosetting isolates (proportion of isolates with any rosettes detected) was also significantly lower for placental blood isolates than for peripheral maternal blood isolates (P = 0.013) and for placental or peripheral maternal blood isolates than for isolates from children (P < 0.001). Prevalence and degree of rosette formation in pregnant women did not differ with age and gravidity, and isolates from women in the Antenatal Clinic or in labor showed similar rosetting characteristics (Table 1). For 12 women, we were able to compare rosetting by isolates from matched peripheral and placental blood. For 7 women, there were no rosettes in cultured peripheral or placental blood, and for 5 women, there was a low degree of rosetting (range, 0.25 to 3%) in peripheral blood but no rosetting in placental blood. This suggests that circulating parasites may comprise different populations, including some that sequester in the placenta and do not rosette and others that sequester elsewhere and form rosettes to some degree. Some isolates were also tested for adherence to purified receptors (2). Of 10 peripheral maternal blood isolates, 4 bound to chondroitin sulfate A (CSA) and 2 others bound to CD36; of 9 placental isolates, 7 bound to CSA and 1 also bound to CD36. In most instances parasites expressed functional adherence proteins on the IE surface but did not rosette.

View larger version (9K): [in this window] [in a new window]   FIG. 1.   Rosette formation (%) by IE cultured from children's peripheral blood (1), cultured from maternal peripheral blood (2), or isolated from placental blood (3).

Lack of rosetting by placental isolates did not appear to be attributable to reversible inhibitors, such as antibodies, present in placental blood: in some instances, IE were washed repeatedly before incubation in culture medium, and in others, they were incubated in culture medium overnight, without changes in rosetting (data not shown). Some placental isolates were adapted to ongoing culture, yet rosetting levels did not change appreciably (data not shown).

Placental blood isolates differed from peripheral blood isolates in that they were not cultured. Culturing could result in the selection of subpopulations of parasites with different adhesive properties. However, our results suggest that the lack of rosetting in placental blood isolates is not a function of their different processing. Rosette formation (mostly at high [>20%] levels) has previously been noted in uncultured isolates from the peripheral blood of Thai adults with severe malaria whose circulating parasites included mature asexual stages (21).

Our findings suggest that rosetting is not important in the development of placental malaria, whereas it has been implicated in the pathogenesis of malaria disease in children (4, 16, 18). Our results are supported by the findings of Maubert et al., who examined parasites from the placentas of 23 Cameroonian women, none of whose parasites formed rosettes (11). In comparison, cryopreserved peripheral blood parasites (which may change phenotype after thawing [14]) from 1 of 12 women in labor and from 8 of 12 nonpregnant adults showed rosette formation. We could not test parasites from nonpregnant adults for rosetting, but children examined over the same period were almost always infected with rosetting parasites, and other studies suggest that P. falciparum rosetting is similar in adults and children (7, 8, 11, 15).

Placental infection may develop in part through adhesion to a ligand(s) such as CSA on the syncytiotrophoblast lining of the intervillous spaces, although not all isolates from pregnant women bind to CSA (2, 6). Our results suggest that parasites from the placenta have different adherence and antigenic characteristics from those infecting children (2). Both rosette formation and adhesion to CSA have been ascribed to P. falciparum erythrocyte membrane protein 1 (5, 13, 19). A possible explanation for the relative lack of rosetting in CSA binding parasites from pregnant women and the scarcity of CSA binding in isolates from children (17) would be that CSA binding and rosetting are mutually incompatible properties of parasitized erythrocytes. This possibility remains to be proven.

In conclusion, pregnant women are not commonly infected with parasites that form rosettes, and rosetting parasites are rarely found in the placenta, indicating that rosette formation is not a major factor in the pathogenesis of malarial infection of the placenta.