CYTOADHERENCE CHARACTERISTICS OF PLASMODIUM FALCIPARUM INFECTED ERYTHROCYTES FROM MALAWIAN CHILDREN WITH SEVERE AND UNCOMPLICATED MALARIA

Transcription

1 Am. J. Trop. Med. Hyg., 61(3), 1999, pp Copyright 1999 by The American Society of Tropical Medicine and Hygiene CYTOADHERENCE CHARACTERISTICS OF PLASMODIUM FALCIPARUM INFECTED ERYTHROCYTES FROM MALAWIAN CHILDREN WITH SEVERE AND UNCOMPLICATED MALARIA STEPHEN J. ROGERSON, RODNEY TEMBENU, CARLOTA DOBAÑO, SABRINA PLITT, TERRIE E. TAYLOR, AND MALCOLM E. MOLYNEUX Malaria Project and Wellcome Trust Centre, College of Medicine, Queen Elizabeth Central Hospital, Blantyre, Malawi; Wellcome Trust Centre for Clinical Tropical Medicine, University of Liverpool, Liverpool, United Kingdom; Institute of Cell, Animal and Population Biology, University of Edinburgh, Edinburgh, Scotland; Department of Public Health Sciences, Faculty of Medicine, University of Alberta, Edmonton, Alberta, Canada; Department of Internal Medicine, College of Osteopathic Medicine, Michigan State University, East Lansing, Michigan Abstract. Cytoadherence of Plasmodium falciparum-infected erythrocytes to the microvascular endothelium is believed to be a key factor in the development of cerebral malaria. Erythrocyte rosette formation has been correlated with malaria severity in studies from east and west Africa. We cultured fresh isolates from Malawian children with severe (n 76) or uncomplicated (n 79) malaria to pigmented trophozoite stage and examined rosette formation and adherence to CD36, intercellular adhesion molecule-1 (ICAM-1), chondroitin sulfate A (CSA), and thrombomodulin (TM). Most (126 of 148) isolates bound to CD36, and 76 of 136 bound to ICAM-1. Fewer bound to CSA (4 of 148) or TM (23 of 148). After controlling for parasitemia, there was an inverse association between binding to CD36 (P.4) or ICAM-1 (P.1) and disease severity. Parasites from children with severe malaria anemia bound least to CD36, whereas ICAM-1 binding was lowest in children with cerebral malaria. There was no difference in rosette formation between any of the groups. In Malawian children, there was no evidence of a positive association between adherence to any of the receptors examined and disease severity. The negative association found raises the possibility that adherence to certain receptors could instead be an indicator of a less pathogenic infection. Erythrocytes infected with mature stages of Plasmodium falciparum sequester in the deep circulation and are rarely found in the peripheral blood. At post-mortem examination, parasitized red blood cells (PRBC) containing trophozoites or schizonts are found in capillaries and postcapillary venules of the brain and other organs of individuals who die of cerebral malaria. 1 This sequestration is believed to be necessary, but not sufficient, for the development of cerebral malaria. 2,3 In vitro systems have shown that erythrocytes parasitized with mature stage parasites may adhere to receptors found on the endothelial cell surface. Such receptors include CD36, 4 intercellular adhesion molecule-1 (ICAM-1), 5 chondroitin sulfate A (CSA), 6 E-selectin and vascular cell adhesion molecule-1 (VCAM-1), 7 thrombomodulin (TM), 8 platelet endothelial cell adhesion molecule-1 (PECAM-1), 9 and P-selectin. 1 A key question is whether the ability of PRBC to adhere to these receptors is a risk factor for severe disease. Previous studies found no association between cerebral malaria and parasite adherence, but many have had inadequate sample sizes In one of these studies, there was an association between adherence to C32 melanoma cells and severe (but not cerebral) malaria in Thai adults. 12 A recent large study from Kenya, using cryopreserved and thawed isolates, showed a trend to more binding to ICAM-1 by isolates from children with cerebral malaria than from children with uncomplicated or asymptomatic disease. 15 In contrast, formation of erythrocyte rosettes (a trophozoite-infected PRBC with two or more uninfected RBC adherent to it) is the only parasite adherence property to have been repeatedly associated with disease. 16,17 We have examined the adherence profiles of PRBC cultured from the blood of Malawian children with severe or uncomplicated malaria. We have measured binding to CSA and TM (not previously evaluated as a correlate of disease 467 severity in children), CD36, and ICAM-1 and formation of erythrocyte rosettes in 155 children. METHODS The study took place at the Malaria Research Project and Wellcome Trust Centre (MRP), Queen Elizabeth Central Hospital in Blantyre, Malawi between January and June Venous blood samples collected into tubes containing lithium heparin or EDTA were obtained from children with severe malaria defined according to World Health Organization criteria, 18 and from children with uncomplicated disease. Informed consent for venisection was obtained from the parents or guardians of all children enrolled in the study. The study was approved by the Malawi National Health Sciences Research Committee. Cerebral malaria was defined as a Blantyre coma score 2 19 in the absence of other apparent cause, and severe anemia was defined as a hematocrit 15%. Patients with uncomplicated disease were from several sources: 1) ambulant children screened for enrollment in studies of novel antimalarial therapy at Ndirande Health Centre, Blantyre, 2) ambulant controls attending the MRP for blood examination for malaria parasites, 3) patients discharged from the MRP who were significantly parasitemic but without any malarial complications at a follow-up visit 2 weeks to 4 months later, and 4) patients admitted to the MRP ward with a final diagnosis of uncomplicated malaria. Controls had no history of convulsions or coma and a Blantyre coma score of 5 of 5. The numbers of isolates successfully grown and tested from children from each group are shown in Table 1. Clinical data, including basic demography, conscious state, history of convulsions, history of prior drug treatment, and age, hematocrit, and outcome of infection were recorded for all patients. Plasma was removed from the sample, and RBC were

2 468 ROGERSON AND OTHERS Age (months) Parasitemia (%) for binding Prior drug treatment Deaths Sequelae Death or sequelae Severe malaria (n 76) (19.7%) 9 (11.8%) 7 (9.2%) 16 (21%) TABLE 1 Patient characteristics* CM SMA (n 54) (16.7%) 7 (13.%) 7 (13.%) 14 (25.9%) CM only (n 38) (2.1%) 6 (15.8%) 5 (13.2%) 11 (28.9%) SMA only (n 22) (27.3%) 2 (9.1%) 2 (9.1%) Controls (n 79) (2.5%) * Results are expressed as mean SD or number (%). CM SMA cerebral malaria with or without severe anemia; CM cerebral malaria; SMA severe malaria anemia. n values are number of children whose parasites were tested for adhesion to one or more receptors. Uncomplicated malaria controls were recruited from the following sources: antimalarial drug trail (39 children), ambulant hospital controls (9 children), follow-up patients (21 children), and uncomplicated malaria resulting in hospital admission (1 children). P.1, SMA versus CM or SMA versus controls, by one-way analysis of variance. washed three times in phosphate-buffered saline (PBS) with buffy coat depletion. When parasitemia was at least 1%, an aliquot of washed RBC was cultured in RPMI 164 medium (Life Technologies, Ltd., Paisley, United Kingdom) supplemented with 25 mm HEPES, gentamicin (1 g/ml) (Life Technologies, Ltd.), NaHCO3 (2 mg/ml), and 1% human AB serum from non-malaria exposed Australian blood donors. Parasites were cultured in a gas mixture of 1% O 2,5% CO 2, 94% N 2, for hr until parasites were at least.5% healthy mature stages as judged by thin film examination. We counted at least 5 erythrocytes and recorded stages (rings, early mid or late trophozoites, and schizonts) to assess binding parasitemia. Assays were performed when 5% of the parasites were at the trophozoite stage. Purified receptors CD36, ICAM-1, CSA, and TM (8 l each) were spotted onto triplicate 35-mm petri dishes (Nunc, Roskilde, Denmark) and coated at least overnight in a humidified atmosphere. Dishes were blocked with 1% bovine serum albumin in PBS for 3 6 min and washed with RPMI 164 medium-hepes. Red blood cells were centrifuged and resuspended in 5% AB serum in RPMI 164 medium-hepes, ph 6.8, at a 1% hematocrit, and 1.7 ml was added to each dish. Assays were incubated at 37 C for 6 min with gentle swirling every 15 min. Dishes were gently washed with RPMI 164 medium-hepes until no non-adherent RBC were visible by inverted microscopy, and bound cells were fixed with 2% glutaraldehyde (Agar Scientific, Stansted, United Kingdom) in PBS, stained with Giemsa (Merck, Ltd., Poole, United Kingdom), and counted by light microscopy. Numbers of mature stage parasites adherent to each receptor were determined by counting 25 fields using a1 eyepiece and a 1 objective under oil immersion in a standardized manner. Numbers of parasites bound were standardized for a parasitemia of 1% by dividing PRBC bound by parasitemia. Significant binding was defined as 5 PRBC/mm 2 and was calculated both at binding parasitemia and after parasitemia had been adjusted to 1%. Rosette formation was measured on triplicate aliquots of parasite culture (15 l) added to 7 1 l of acridine orange on a glass slide. After placing a mm cover slip over the slide, the sample was examined by fluorescence and direct light microscopy, and the proportion of trophozoite-infected cells in rosettes was measured for each sample as previously described. 2 Two or three hundred trophozoite infected PRBC were counted for each replicate. Data were expressed as the mean of triplicate measurements for most data points. A shortage of ICAM-1 meant that only duplicates were performed in some assays, and technical problems with fixing in single replicates led to those being excluded in a small proportion of cases. Data were entered into Microsoft (Redmond, WA) Access version 7. and analyzed using a combination of Epi- Info version 6. (Centers for Disease Control and Prevention, Atlanta GA), Microsoft Excel, and SPSS version 7.5 (SPSS Institute, Chicago, IL). Because the data were not normally distributed, the Mann-Whitney U test and the Kruskal-Wallis test were used for comparing means, unless otherwise indicated. A P value.5 was considered significant for this analysis. RESULTS Parasites from 76 children with severe malaria and 79 children with uncomplicated disease were grown successfully to the trophozoite stage and binding was measured in at least one assay. Children with severe malaria had cerebral malaria (38), severe malaria anemia (22), or both (16) and were similar to controls in age (mean SD versus months) and sex distribution (48.7% versus 48.4% males). Two of the controls had a history of antimalarial treatment (which was always sulfadoxine-pyrimethamine) in the past week, whereas 15 children with severe malaria had this treatment. Children with severe malaria anemia were significantly younger (P.1) than children with cerebral malaria or controls, consistent with previous observations. The characteristics of these patients and their parasitemias are shown in Table 1. Table 2 shows the comparison of mean (range) parasitemia-adjusted binding (PRBC/mm 2 ) for isolates from all children with severe malaria, subsets of severe malaria cases, and the isolates from children with uncomplicated malaria. Figure 1 shows the percentage of isolates binding at significant levels ( 5 PRBC/mm 2 ) to each receptor. There was no evidence of a higher prevalence of significant binding or of greater numbers of PRBC bound when isolates from patients with severe malaria were compared with those with uncomplicated disease. Indeed, for all receptors, the trend was in the opposite direction (Figure 2), and reached statistical significance for both CD36 (P.4) and ICAM-1 (P.1). The binding of ICAM-1 was significantly lower (P.11) for isolates from children with severe malaria even when raw (non-parasitemia adjusted) data were used.

3 P. FALCIPARUM CYTOADHERENCE AND ROSETTE FORMATION 469 TABLE 2 Binding of clinical isolates of Plasmodium falciparum to cytoadherence receptors and to uninfected erythrocytes* Severe malaria (n 76) CD ( 454.1; 73) ICAM-1 5 ( 45.1; 69) CSA 1.4 ( 2.1; 73) TM 1.1 ( 19.2; 73) Rosette formation (64) CM SMA (n 54) 9.8 ( 454.1; 51) 4.1 ( 45.4; 48) 1.1 (-13.5; 51) 1.1 ( 19.2; 51) (46) CM only (n 38) 95.7 ( 454.1; 36) 4.9 ( 45.4; 333) 1.2 ( 13.5; 36) 1.4 ( 19.2; 36) (32) SMA only (n 22) 26.1 ( 167.7; 22) 7.2 ( 35.4; 21) 1.9 ( 2.1; 22).7 ( 3.9; 22) (18) Controls (n 79) ( 2,12.7; 74) 12.7 ( 117.7; 67) 6 ( 41.3; 74) 1.6 ( 4.6; 74) (62) * Results are given as mean (range; number of isolates tested) binding of parasitized red blood cells (PRBC)/mm 2 adjusted for parasitemia, except for rosette formation, expresssed as the mean SEM% PRBC in rosettes (number tested). CM SMA cerebral malaria with or without severe anemia; CM cerebral malaria; SMA severe malaria anemia; ICAM-1 intercellular adhesion molecule-1; CSA chondroitin sulfate A; TM thrombomodulin. Few isolates bound to CSA or TM (which contained a CSA chain). These results were not significantly different when follow-up samples from patients previously admitted were excluded. There was consistently lower binding to all receptors in the group of children with severe anemia compared with controls. An inverse relationship between hemoglobin concentration and receptor binding was demonstrated by comparing the binding levels of CD36 and ICAM-1 and patient hematocrit measurements of the total study population (Figure 3). The binding of isolates to CD36 was significantly different at each hemoglobin level; however, the ICAM- I binding was only significantly different between the 25% and 25% categories. There was no significant difference in mean rates of rosette formation between PRBC from children with severe malaria, subsets of severe malaria, and uncomplicated malaria (Table 1 and Figure 4). Only three isolates failed to form rosettes, two from children with severe malaria and one from a child with mild disease. Correlations between binding of ICAM-1 and CD36 and rosette formation were examined. A significant correlation between CD36 and ICAM-1 binding was found (P.1); however, no correlation with rosette formation and isolate binding was noted. All children with uncomplicated malaria survived. Of children with severe disease, nine died and seven had detectable sequelae at discharge. Binding of isolates from children with fatal malaria or sequelae was not significantly different from those from children who made a full recovery or children with mild disease. When rosette formation was analyzed according to disease outcome (died or survived), there was also no difference (Table 3). There was no difference between binding of isolates or rosette formation from children with or without previous drug therapy (although numbers in the former group were small). DISCUSSION We have examined cytoadherence and rosette formation in 155 patient isolates cultured to the trophozoite stage. We found no correlation between rosette formation and disease severity, and no evidence of a predisposition for isolates from children with more severe or complicated malaria to bind at higher rates to any of the purified receptors used. After standardizing for parasitemia, parasite isolates from children with uncomplicated disease bound at higher levels to all receptors, significantly so for ICAM-1 and CD36. There was a significant association between hematocrit and parasite adhesion, with lower binding to CD36 and ICAM- 1 by isolates from children with severe anemia than children with mild or no anemia. FIGURE 1. Percentage of isolates tested showing significant binding to CD36, intercellular adhesion molecule-1 (ICAM-1), chondroitin sulfate A (CSA), and thrombomodulin (TM) by patient group. Significant binding was defined as 5 parasitized red blood cells/mm 2 using raw data (not adjusted for parasitemia). CM cerebral malaria; SMA severe malaria anemia. FIGURE 2. Mean SEM binding to CD36, ICAM-1, CSA, and TM for isolates from each patient group. Results are expressed as PRBC bound/mm 2 after adjusting for parasitemia. Numbers of isolates tested are indicated in Table 2. For definitions of abbreviations, see Figure 1.

4 47 ROGERSON AND OTHERS FIGURE 3. Relationship between patient hematocrit at presentation and binding to CD36 (solid line) or intercellular adhesion molecule-1 (ICAM-1) (dotted line). Data include all patients for whom initial hematocrit was available, irrespective of disease severity. Bars show the mean SD binding for each hematocrit group. The current study is most like that of isolates from Kilifi, Kenya, 15 but with some important differences in methodology and results. The Kilifi study was performed on cryopreserved and thawed isolates, whereas the current study used fresh patient isolates. Individual patient isolates may be composed of multiple clones, 21 and within one clonal population there may be multiple variant antigenic types (VATs) with different adherence profiles. 22 Genotyping by the polymerase chain reaction and phenotyping with antibodies to merozoite surface protein-1 (MSP-1) and MSP-2 2 showed that almost all infections in these children are composed of multiple clones (Dobaño C, unpublished data). To avoid a possible selective loss of some VATs over others with cryopreservation, we performed studies on site. Changes in dominant antigenic types in a sample following cryopreservation are reported, as reflected by the agglutination profile obtained using semi-immune adult sera. 23 Examining patient isolates placed directly into culture minimizes but does not eliminate the risk of selecting some variant antigenic types from a mixed population for binding. Due to the difficulty of getting fresh erythrocytes from malaria-free and antimalarial-free blood donors in Malawi, we did not dilute high-binding parasitemia to a standard level. We may therefore have underestimated binding from some isolates. 15 However, we found no difference in the relationship between groups in mean binding when we compared only isolates with a binding parasitemia 1%, which has previously been found to be in the range giving linear binding with parasitemia (Beeson J, unpublished data). Similar to findings from Kilifi, 15 binding to CD36 was lowest in children with severe malaria anemia; unlike their results, we found lower levels of binding to ICAM-1 in severe disease in general, rather than specifically severe malaria anemia. Since there are potential problems examining parasites from children with severe malaria anemia or cerebral malaria together, we have analyzed the results both together and separately. Their pathogenesis may be quite different, 24 and our finding of a correlation between patient hematocrit and parasite binding (Figure 3) is consistent with this hypothesis (although cytoadherence may be implicated in the development of both cerebral malaria and severe malaria anemia). Rosette formation was very frequent in our series: all but three isolates tested formed rosettes to some degree. Studies from east and west Africa have found a high prevalence of FIGURE 4. Rosette formation by isolates from children with severe or uncomplicated malaria. Mean percentage rosette formation for individual isolates is plotted for children with cerebral malaria (CM), severe malaria anemia (SMA) without CM, CM with or without SMA, and children with uncomplicated malaria (control). non-rosette-forming parasite isolates from individuals with uncomplicated malaria (from 3% to 45%) with higher prevalences of rosette formation in children with severe malaria. 16,17,25 In contrast, in a large study from Papua New Guinea, 26 all isolates tested formed rosettes. The reasons for this difference may include host genetic differences in the prevalence of rosette formation receptors, such as complement receptor 1 polymorphisms, 27 differences in heparan sulfate molecule type or density on the uninfected erythrocyte surface, or differences in prevalence of other blood group determinants. 28 We examined a larger number of trophozoiteinfected erythrocytes per sample than other studies reported. In the Papua New Guinean study, there was no difference in rosette formation between isolates from different patient groups, a finding borne out in the current study. It will be important to examine the prevalence of rosette formation receptors in different populations in relation to the frequency and intensity of rosette formation and in relation to the associations between rosette formation and disease severity in those populations. Our study did not set out to examine the associations between PRBC binding and prior drug treatment or disease outcome. Examining the results in relation to these variables, we found no evidence for impaired cytoadherence by isolates from children with prior treatment within the previous week with pyrimethamine-sulfadoxine, as has been described in in vitro and in vivo drug exposed parasites for quinine and artemisinins. 29 We also found no association between disease outcome (death or sequelae versus recovery) TABLE 3 Binding to receptors according to outcome for severe malaria* CD36 ICAM-1 CSA TM Rosette formation Severe malaria plus sequelae or death (n 16) Severe malaria, full recovery (n 6) * Results are expressed as the mean SD parasitized red blood cells (PRBC) bound/ mm 2 after adjustments for parasitemia; rosette formation is expressed as the mean SD% PRBC in rosettes. For definitions of abbreviations, see Table 2.

5 P. FALCIPARUM CYTOADHERENCE AND ROSETTE FORMATION 471 and binding (Table 3). Again, the trends were in the opposite direction. What might be the explanation for the negative association between cytoadherence and disease severity? It is unlikely, given the level of significance, that this is due to chance alone, and the same findings were observed on several preliminary examinations of the data. A negative association between CD36 or ICAM-1 binding and disease severity could suggest that PRBC binding to those receptors is in fact more prone to cause less severe disease. Such parasites could sequester in less harmful parts of the body. The population in the patient s peripheral blood represents the progeny of PRBC that were sequestered throughout the body on the previous parasite cycle, only a proportion of them in the brain. Almost all of these will have the same adherence type and VAT as their parents, and so reflect a population previously sequestered in a heterogeneous range of sites, including notably organs such as the gut, brain, skin, and lung. Endothelium differs between tissues, and it is highly likely that different adherence mechanisms may predominate in different tissues. Thus the circulating population may contain a relatively small proportion of PRBC capable of cerebral sequestration (and thus facilitating the development of cerebral malaria) and a larger proportion that were previously sequestered in other non-cerebral sites, whose adherence profile obscures that of the more pathogenic minority. Alternatively, if an infection contains two highly synchronous broods of parasites, it is possible that the brood sequestered in the brain and initiating the events leading to cerebral malaria differs in adherence behaviour or genotype from the brood circulating at the time of admission, again obscuring a pathogenic parasite characteristic. It is known that populations of parasites can change rapidly in asymptomatic individuals in endemic areas, 3 and similar dynamic processes may be occurring in these symptomatic children. In the subset of children who died, we are comparing MSP- 1 and MSP-2 phenotypes of circulating and sequestered parasites using monoclonal antibody typing (Dobaño C, unpublished data). Several of our isolates showed no binding to any of the receptors tested, although all isolates with no binding still formed rosettes to some extent. These parasites must be able to sequester in the host, and their failure to do so in vitro could reflect a failure of our system. We did not examine binding to PECAM-1, P-selectin, VCAM- 1, or E-selectin. Given the findings from Kilifi of very low levels of adherence of cultured parasites to PECAM-1, VCAM-1 and E- selectin, 15 it suggests either that there are other as yet unidentified receptors that are playing a critical role in the sequestration process, or that cooperation between receptors is essential. Our findings could have implications, for example, for development of vaccines aimed at blocking adherence to CD36. Such a vaccine, if successful, could select positively for more pathogenic variants that use alternative sequestration receptors expressed on cerebral endothelium. Further evaluation of vaccine candidates derived from sequestrin 31 or P. falciparum erythrocyte membrane protein 1 32 will need to take this possibility into account. There is clearly a great diversity of cytoadherence profiles from patient isolates from many places in the world 11 15,23,33 that is explained in part by the variety of different assay systems used. For results to be comparable between settings, standardization of techniques should be encouraged as much as possible. It is clear that most or all PRBC bind to CD36, and that binding to ICAM-1 is also common, whereas a minority of isolates bind to other receptors identified. Examination of adherence of PRBC to isolated purified receptors in static in vitro systems may be a poor reflection of what happens in vivo. The proportion of PRBC introduced into the assay that remain attached to the receptor is far lower than the proportion of PRBC of similar stages that sequester in vivo, and how representative the bound population is of the total population is unknown. Flow systems have been proposed as a more physiologic method. 1,34 As well as the effects of flow, cooperation between receptors or a multiple-step process of sequestration such as that described for leukocytes 35 may be of importance in malaria. 36 It will be important to develop systems that more closely resemble in vivo conditions that incorporate endothelial cell lines and parasites under physiologic flow. Acknowledgments: We thank the patients and guardians for essential cooperation with this study, the clinicians and nurses of the Malaria Project for sample collection, and the Department of Paediatrics, College of Medicine, University of Malawi (Blantyre, Malawi) for provision of laboratory space. Kathy Davern provided ICAM-1, Dr. Michael Berndt provided CD36, and Dr. Brian Grinnell provided TM. Financial support: Stephen J. Rogerson is a Wellcome Trust Career Development Fellow, Carlota Dobaño is a Wellcome Trust Postgraduate Student, and Malcolm E. Molyneux is a Wellcome Trust Research Leave Fellow in Clinical Tropical Medicine. Terrie E. Taylor is supported by NIH grant RO1 AI Sabrina Plitt was a World University Service of Canada student funded by the Canadian International Development Agency. Authors addresses: Stephen J. Rogerson and Malcolm E. Molyneux, Malaria Project and Wellcome Trust Centre, College of Medicine, Queen Elizabeth Central Hospital, Blantyre, Malawi and Wellcome Trust Centre for Clinical Tropical Medicine, University of Liverpool, Liverpool L69 3GF, United Kingdom. Rodney Tembenu, Malaria Project and Wellcome Trust Malaria Centre, College of Medicine, Queen Elizabeth Central Hospital, Blantyre, Malawi. Carlota Dobaño, Institute of Cell, Animal and Population Biology, University of Edinburgh, Edinburgh EH9 3JT, Scotland. Sabrina Plitt, Department of Public Health Sciences, Faculty of Medicine, University of Alberta, Edmonton T6E 1Y8, Alberta, Canada. Terrie E. Taylor, Department of Internal Medicine, College of Osteopathic Medicine, Michigan State University, East Lansing, MI REFERENCES 1. MacPherson GG, Warrell MJ, White NJ, Looareesuwan S, Warrell DA, Human cerebral malaria. A quantitative ultrastructural analysis of parasitized erythrocyte sequestration. Am J Pathol 119: Berendt AR, Turner GDH, Newbold CI, Cerebral malaria: the sequestration hypothesis. Parasitol Today 1: Clark IA, Rockett KA, How important is mechanical blockage of blood vessels in human cerebral malaria? Parasitol Today 1: Ockenhouse CF, Chulay JD, Plasmodium falciparum sequestration: OKM5 antigen (CD36) mediates cytoadherence of parasitized erythrocytes to a myelomonocytic cell line. J Infect Dis 157: Berendt AR, Simmons DL, Tansey J, Newbold CI, Marsh K, Intercellular adhesion molecule-1 is an endothelial cell

6 472 ROGERSON AND OTHERS adhesion molecule for Plasmodium falciparum. Nature 341: Rogerson SJ, Chaiyaroj SC, Ng K, Reeder JC, Brown GV, Chondroitin sulfate A is a cell surface receptor for Plasmodium falciparum-infected erythrocytes. J Exp Med 182: Ockenhouse CF, Tegoshi T, Maeno Y, Benjamin C, Ho M, Kan, KE, Thway Y, Win K, Aikawa M, Lobb RR, Human vascular endothelial cell adhesion receptors for Plasmodium falciparum-infected erythrocytes: roles for endothelial leukocyte adhesion molecule 1 and vascular cell adhesion molecule 1. J Exp Med 176: Rogerson SJ, Novakovic S, Cooke BM, Brown GV, Plasmodium falciparum: infected erythrocytes adhere to the proteoglycan thrombomodulin in static and flow based systems. Exp Parasitol 86: Treutiger CJ, Heddini A, Fernandez V, Muller WA, Wahlgren M, PECAM-1/CD31, an endothelial receptor for binding Plasmodium falciparum-infected erythrocytes. Nat Med 3: Udomsangpetch R, Reinhardt PH, Schollaardt T, Elliott JF, Kubes P, Ho M, Promiscuity of clinical Plasmodium falciparum isolates for multiple adhesion molecules under flow conditions. J Immunol 158: Marsh K, Marsh VM, Brown J, Whittle HC, Greenwood BM, Plasmodium falciparum: the behavior of clinical isolates in an in vitro model of infected red blood cell sequestration. Exp Parasitol 65: Ho M, Singh B, Looareesuwan S, Davis T, Bunnag D, White NJ, Clinical correlates of in vitro Plasmodium falciparum cytoadherence. Infect Immun 59: Ockenhouse CF, Ho M, Tandon NN, Van-Seventer GA, Shaw S, White NJ, Jamieson GA Chulay JD, Webster HK, Molecular basis of sequestration in severe and uncomplicated Plasmodium falciparum malaria: differential adhesion of infected erythrocytes to CD36 and ICAM-1. J Infect Dis 164: Goldring JD, Molyneux ME, Taylor T, Wirima J, Hommel M, Plasmodium falciparum: diversity of isolates from Malawi in their cytoadherence to melanoma cells and monocytes in vitro. Br J Haematol 81: Newbold CI, Warn P, Black G, Berendt AR, Craig A, Snow R, Msobo K, Peshu N, Marsh K, Receptor-specific adhesion and clinical disease in Plasmodium falciparum. Am J Trop Med Hyg 57: Carlson J, Helmby H, Hill AVS, Brewster D, Greenwood BM, Wahlgren M, 199. Human cerebral malaria: association with erythrocyte rosetting and lack of anti-rosetting antibodies. Lancet 336: Rowe A, Obeiro J, Newbold CI, Marsh K, Plasmodium falciparum rosetting associated with malaria severity in Kenya. Infect Immun 63: Warrell DA, Molyneux NE, Beales PF, 199. Severe and complicated malaria. Trans R Soc Trop Med Hyg 84: Molyneux ME, Taylor TE, Wirima JJ, Borgstein A, Clinical features and prognostic indicators in paediatric cerebral malaria: a study of 131 comatose Malawian children. Q J Med 71: Rogerson SJ, Chaiyaroj SC, Reeder JC, Al-Yaman F, Brown GV, Sulfated glycoconjugates as disrupters of Plasmodium falciparum erythrocyte rosettes. Am J Trop Med Hyg 51: Conway DJ, McBride JS, Population genetics of Plasmodium falciparum within a malaria hyperendemic area. Parasitology 13: Roberts DJ, Craig AG, Berendt AR, Pinches R, Nash G, Marsh K, Newbold CI, Rapid switching to multiple antigenic and adhesive phenotypes in malaria. Nature 357: Reeder JC, Rogerson SJ, Al-Yaman F, Anders RF, Coppel RL, Novakovic S, Alpers ND, Brown GV, Diversity of agglutinating phenotype, cytoadherence and rosette-forming characteristics of Plasmodium falciparum isolates from Papua New Guinean children. Am J Trop Med Hyg 51: Kurtzhals JAL, Adabayeri V, Goka BQ, Akanmori BD, Oliver- Commey JO, Nkhrumah FK, Behr C, Hviid L, Low plasma concentrations of interleukin 1 in severe malarial anaemia compared with cerebral and uncomplicated malaria. Lancet 351: Wahlgren M, Carlson J, Ruangjirachupom W, Conway D, Helmby H, Martinez A, Patarroyo ME, Riley E, 199. Geographical distribution of Plasmodium falciparum erythrocyte rosetting and frequency of rosetting antibodies in human sera. Am J Trop Med Hyg 43: Al-Yaman F, Genton B, Mokela D, Raiko A, Kati S, Rogerson S, Reeder J, Alpers M, Human cerebral malaria: lack of significant association between erythrocyte rosetting and disease severity. Trans R Soc Trop Med Hyg 89: Rowe JA, Moulds JM, Newbold CI, Miller LH, P. falciparum rosetting mediated by a parasite-variant erythrocyte membrane protein and complement-receptor 1. Nature 388: Carlson J, Wahlgren M, Plasmodium falciparum erythrocyte rosetting is mediated by promiscuous lectin-like interactions. J Exp Med 176: Udomsangpetch R, Pipitaporn B, Krishna S, Angus B, Pukrittayakmee S, Bates I, Suputtarnongkol Y, Kyle DE, White NJ, Antimalarial drugs reduce cytoadherence and rosetting Plasmodium falciparum. J Infect Dis 173: Daubersies P, Sallenave-Sales S, Magne S, Trape J-P, Contamin H, Fandeur T, Rogier C, Mercereau-Puijalon O, Druilhe P, Rapid turnover of Plasmodium falciparum populations in asymptomatic individuals living in a high transmission area. Am J Trop Med Hyg 54: Ockenhouse CF, Klotz FW, Tandon NN, Jamieson GA, Sequestrin, a CD36 recognition protein on Plasmodium falciparum malaria-infected erythrocytes identified by anti-idiotype antibodies. Proc Natl Acad Sci USA 88: Baruch DI, Ma XC, Singh H, Bi X, Pasloske BL, Howard RJ, Identification of a region of PfEMP1 that mediates adherence of Plasmodium falciparum infected erythrocytes to CD36: conserved function with variant sequence. Blood 9: Chaiyaroj SC, Angkasekwinai P, Buranakiti A, Looareesuwan S, Rogerson SJ, Brown GV, Cytoadherence characteristics of Plasmodium falciparum isolates from Thailand: evidence for chondroitin sulfate A as a cytoadherence receptor. Am J Trop Med Hyg 55: Cooke BK Coppel RL, Cytoadhesion and falciparum malaria: going with the flow. Parasitol Today 11: Springer TA, Traffic signals for lymphocyte recirculation and leukocyte emigration: the multistep paradigm. Cell 76: McCormick CJ, Craig AC, Roberts D, Newbold CI, Berendt AR, Intercellular adhesion molecules and CD36 synergize to mediate adherence of Plasmodium falciparum-infected erythrocytes to human microvascular endothelial cells. J Clin Invest 1: