British journal of Haernatoloqy, 1992, 82, 757-763
Rheological properties of rosettes formed by red blood cells parasitized by Plasmodium fakiparum B . N A S H ,BRIAN M . C O O K E , JOHAN CARLSON' A N D MATS WAHLGREN* Department of i%ernatology, The Medical School, University of Birmingham, Birmingham, U.K., and *Department of Immunology, Stockholm University, Stockholm, Sweden GERARD
Rrlceived 3 Iune 1992; acceptedfor publication 9 July 7 9 9 2
Summary. A proportion of red blood cells parasitized by Plismodium falciparum form rosettes with non-parasitized red cells. Although these rosettes are thought to impair microcirculatory flow, their rheological characteristics have not been fully described. Using dual-micropipette manipulation to pull apart individual rosettes, we found that the forces binding rosettes together were strong (average force for removal of a cell was 4.4 x lo-"' N, approximately 5 times that required to detach a parasitized cell adhered to cultured endothelium). If disrupted rosettes were re-formed, cells rosetted immediately on contact, but the strength of attachment increased over minutes, and did not apparently reach its maximal level
for hours. All non-parasitized cells tested could adhere to rosette-forming parasitized cells. Rosettes could withstand arterial flow stresses (1.4-1'6 Pa) for minutes without disintegration. To test the effects of rosetting on flow resistance, the time required for entry into a 4.3 pm pipette was measured. Entry times depended strongly on the number of cells in the rosette, and averaged 3 5 times longer than for non-parasitized cells. Our studies show that the cell-cell attachments within rosettes are strong, and suggest that rosettes might survive both the arterial circulation and passage through microvessels and could contribute to the ischaemic complications of falciparum malaria.
assess their role in circulatory pathology, it is necessary to investigate the rheological characteristics of rosettes. It is known that rosettes are resistant to shear stress such as that imposed by pipetting of cell suspensions (Howard & Gilladoga, 1989). and that they increase the peripheral resistance when perfused into isolated rat vasculature (Kaul et al. 1991). Here we report detailed information on the rheological characteristics of an in vitro selected, stable rosetting strain of P. falciparum (R+PAl).
Ischaemic complications of falciparum malaria, particularly cerebral malaria, may be life threatening. The microvasculature of those who have died from severe malaria has been found to be filled, and presumably blocked, with cells containing mature, pigmented parasites (MacPherson et al, 1985; Pongponratn et al, 1991). It has formerly been thought that this sequestration arises from adhesion of parasitized cells to vascular, and particularly venular, endothelium (Luse&Miller,1971: Udeinyaet al, 1981).However, recent studies have shown that certain strains of P. falciparum, including those newly isolated from patients in endemic areas, can induce parasitized cells to act as the focus for rosettes with non-parasitized cells (for review see Wahlgren, 1992). The percentage of rosetting parasites varies widely between clinical isolates (Carlson et a/, 1990a: Hasler et al, 1990; Wahlgren et a/. 1990: Ho et al, 1991a), and some (Carlson et al, 1990a; Treutiger et al, 1992) but not all (Ho et a!. 1991a) studies have found a relation between this percentage and the severity of the clinical pathology. Rosettes could add to microvascular occlusion, either by hindering flow through capillaries or by forming on to Parasitized cells already adhered to venular endothelium. To
METHODS
Malarial culture. Cryopreserved samples of Plasmodium falciparum (strain R + PA1, cloned by limiting dilution: Udomangspetch et al, 1989) were thawed and cultured in human red blood cells (blood group A) suspended in RPMI 1640 (Flow Laboratories Ltd, Irvine, U.K.) supplemented with HEPES (37.5 mM). glucose (11 mM), glutarnine (4 mM), gentamycin sulphate (28 mg/l) and human AB serum (10%). Culture was carried out using the candle-jar method of Trager & Jensen (1976). Parasitaemia and stage of parasite maturation were assessed by examination of blood films stained with Giemsa. The majority of experiments were carried out on the day (average 20 h) or 3 d after thawing of isolates, so that synchronous cultures were utilized. Other-
Correspondence: Dr G. Nash, Department of Haematology, The Medical School, University of Birmingham, Birmingham B15 2TT.
75 7
758
Gerard B. Nash et a1
Fig 1. Video image of a rosette composed of five red cells being held between two micropipettes with internal diameters of approximately 2 and 1 . 2 pm.
wise. synchronization of cultures was carried out by incubation of cells in 5% w/v sorbitol on the day before analysis, which caused selective lysis of trophozoites and schizonts. On average. the cultures tested had a total parasitaemia of 2.4%. with pigmented trophozoites making up over 90% of the parasites. The percentage of all parasitized cells that were contained within rosettes (rather than existing as single cells in suspension) was defined as the rosetting frequency. This was assessed by light microscopy after cultures had been mixed 1 : 1 with 2% glutaraldehyde in phosphate buffered saline. A 4 0 x water-immersion objective lens (NA =0.75, Carl Zeiss (Oberkochen) Ltd. Germany) allowed clear identification of pigmented parasites in single cells or rosettes (see, e.g.. Figs 1. 2 and 6). A total of 1 0 0 parasitized cells were counted in each sample. and the percentage within rosettes was determined. In this study, rosetting frequency averaging 68 & 8% (mean i SD. seven cultures). Micropipette analysis of rosettes. Samples of malarial cultures were diluted 100-fold in culture medium with pH adjusted to 7.4. The suspension was placed in a micropipette chamber which was approximately 1 m m deep and made up of a microscope slide and coverslip. separated by two parallel strips of multiply-folded parafilm. The chamber was placed on the stage of a microscope, and two micropipettes were introduced into it from opposite directions. The pipettes were connected to separate hydrostatic reservoirs whose heights were adjusted by micrometers. so that suction pressures were independently. precisely controlled in each pipette. The pipette tips were viewed using a 4 0 x water-immersion lens and a video-microscope system which gave a final magnification on the video monitor of approximately x 6000. In general, a larger pipette (pipette A, i.d. about 2 pn) was used to hold rosettes firmly with a n applied pressure of 2-4 cmHLO (200-400 Pa), while the other (pipette B, i.d.=Dp= 1.2-1'5 pm) was used to pick u p desired cells or rosettes. Disruption and reforrnation of rosettes. A rosette was picked up using pipette B and passed to the larger, stationary pipette
A. The pressure in pipette B was adjusted to zero and then suction was increased in steps. At each step, a red cell from the stationary rosette was aspirated into the smaller pipette B and this pipette was withdrawn (Fig 1).The cell either pulled free from pipette B or became detached from the rosette. The pressure (P)required to detach the cell from the rosette was taken as the average of the final pressure (which detached the cell) and the previous pressure which had not detached the cell. The pressure was used to calculate the force (F)required to detach the cell from the rosette, using the equation:
F=P*zD:
(Nash et al, 1992)
This procedure is essentially the same as that previously used to measure the force required to remove parasitized cells after they had adhered to endothelial cells or melanoma (C32) cells grown on microbarrier beads (Nash et al. 1992). In some experiments. after all the non-parasitized cells had been removed from the rosetting, parasitized cell, the last cell to be removed was brought back into contact with the parasitized cell (Fig 2). After a timed delay, the force to detach the non-parasitized cell was re-measured. as described above. This measurement was repeated: delay times of 5, 60, 1 2 0 and 300 s were tested. The time actually taken to vary the pressure and to remove the cell averaged 3 0 s. Thus, the measured force for detachment was considered to be that which was acting at the midpoint of the removal procedure, i.e. 20, 75, 1 3 5 and 3 1 5 s. In other experiments, cells were picked u p a t random from those settled in the chamber using pipette B and a low suction pressure (about 20-40 Pa), and brought into contact with the parasitized cell held in pipette A. The pressure in pipette B was maintained and the pipette was withdrawn within a few seconds of cell-cell contact. It there was a significant interaction with the parasitized cell, then the newly adhered cells either pulled free from pipette B or showed deformation and elongation before pulling free from the parasitized cell. To avoid re-testing of cells that had just been removed from the
Rheology of Rosettes Induced by P. fakiparum
759
Fig 2. Series of video images showing a non-parasitized cell being brought into contact with a rosetting. parasitized cell (A-C). and then being reaspirated and pulled away (D-F). The arrow in A indicates the parasitized cell held stationary by the larger pipette.
rosette, the parasitized cell was first moved to a new region of the chamber. Kesistance of rosettes to micropipette aspiration. The effect of rosetting on flow resistance was assessed by measuring the time required for rosettes of different size to enter a micropipette (i.d. = 4 . 3 pm). The method is essentially the same as that previously used to study the deformability of individual parasitized cells (Nash et al, 1989). The micropipette apparatus was as described above, except that a single micropipette was introduced into a micropipette chamber made up with a U-shaped gasket. A fixed suction pressure (30 Pa) was applied, and the pipette tip brought close to the cell or rosette to be tested. The time required for entry was measured electronically, using Ag :AgC12 electrodes: one within the pipette and one within the chamber. A constant current was parsed between the electrodes and the cell entry was sensed as a voltage pulse, which was analysed by microcomputer after analogue-to-digital conversion. The time between the signal first crossing a preset threshold, and its falling to 90% of its peak value was taken to be the entry time. Measure-
ments using simultaneous electronic and video-playback analysis verified that this procedure accurately measured the entry time. Exposure of rosettes to shear stress. Malarial cultures were diluted to 5 x lo8 cells/ml in culture medium with pH adjusted to 7.4. Aliquots of these suspensions were then mixed with dextran-containing media as follows: (A) 1: 1 with 22% w/v dextran 500 (MW = 500 000), to give a final viscosity of about 30 mPa.s at room temperature; (B) 1: 1 with 20% w/v dextran 40 (MW = 40 000), to give a final viscosity of about 7 mPa.s at room temperature: (C) 1 :2 with 20% w/v dextran 40, to give a final viscosity of about 7 mPa.s at 37OC. The dextrans (Sigma Chemical Co. Ltd, Poole. U.K.) were dissolved in water with 20 mM HEPES (pH = 7.4) and the osmolarity adjusted to 290 mOsmol/kg by addition of NaCI. The suspension under analysis was placed on the plate of a cone-plate viscometer (model LVT, Brookfield Engineering Laboratories Inc., Stoughton, Mass., U.S.A.) and exposed to a uniform shear rate of 4 6 s-' for suspension A. or 2 30 s for suspensions B or C. Suspensions A and B were tested at
760
Gerard B. Nash et a1 Table I. Force required t o detach individual non-parasitiaed cells from rosettes. O 5
1
FORCE
Median 90th percentile .\rean +SI) ( r i = 1 5 t Mean rtS1)ofsaniplemeans I r r = h t
3. 3
(RELATIVE)
i.X
4.4zt2.2 4.1il.3
Force for detachment from endothelium: W a n +Sl)ofsarnplc mcitns(ii= 5 ) 0.9rtO.1
I
0O 2I!
i1
1:orce for detachment from melanoma cells: Mean rSIlofsample means (rt=4) O.hrt0.2
* Individual cells were detached from rosettes by iiiicromaiiipulation. The force in Newtons (N\ was calculated from the aspiration pressure (in Pascals) required to remove each cell. and the pipette diameter (see Methods). I t i ) = number of cells or number of cultures tested. f Data from Sash (11 1992 1. obtained by micropipette manipulation olparasitized cells (strain ITO?) into comiii't with endothelial or melanoma cells. and detachment as for rosettes. c3f
room temperature, while suspension C was tested at 3 7°C. The actual shear stress to which the cells were exposed was measured by the viscometer. and averaged 1.46&0.14 Pa (mean iSD, 10 experiments). After 1 min of shearing. the viscometer was halted and 100 PI of cell suspension removed and fixed by mixing 1 : 1 with glutaraldehyde ( 2 % in phosphate buffered saline). This process took 5-10 s. The suspension was sheared for a further 4 min. and another 100 p1 aliquot fixed. Rosetting frequency in the fixed, sheared samples was counted as described earlier. Results were compared to those obtained for aliquots of cells suspended in dextran but not sheared (samples were fixed before and after the period of shearing). and for aliquots of the original culture. RESCLTS Values for the force required to remove individual nonparasitized cells from rosettes which had been cultured. undisturbed overnight, are summarized in Table I. The distribution of the strengths of attachment was positively skewed (e.g. mean > median: 90th percentile > twice median) and varied widely from cell to cell, and to a lesser extent. between samples. The average size of the rosettes tested in these experiments was 4.8 f 1.4 cells (mean +SD). Table I also shows the force required to remove individual parasitized cells which had adhered to either cultured human vascular endothelium or cultured amelanotic melanoma cells. These data were obtained by a similar micropipette method (Nash et nl, 1992): the force for either cell line was about one-fifth of that required to remove a cell from a rosette. Following rosette disruption, the strength of attachment of a newly re-formed rosette was tested as a function of the time after cell-cell contact. The force required to disrupt the rosette
O I
I
I
0
100
I
I
200
I
I
I
300
TIME (SEC)
Fig 3. Variation in the force required to remove non-parasitized red cells from a rosetting parasitized cell as a function of the time after the cells were brought into contact. The force is expressed relative to the force required to remove the non-parasitized cell from the original rosette. Data are mean &SD of means from three experiments: five cells were tested in each experiment.
FF(EQuENcy
0
0
1
2 3 TIME (MIN)
4
5
Fig 4. Variation in rosetting frequency of Parasitized cells as a function of the time of exposure to a shear stress averaging 1.46 Pa. Data are mean ltSE for experiments carried out using suspending medium containing dextran 500 at room temperature (m.n = 4 ) . dextran 40 at room temperature ( 0 .n = 3 ) or dextran 40 at 37°C ( 0 . I f = 31.
increased with time, and after 5 min, reached 40% of the force required to disrupt the original rosette (Fig 3 ) . After rosette disruption, new red cells were chosen from those settled in the micropipette chamber, and brought into contact with the parasitized cell. All of 110 non-parasitized cells tested in this way adhered to the parasitized cell (11cells from three cultures). In addition, non-rosetting parasitized cells containing pigment also adhered (24/24, tested with nine rosettors). but would not adhere to non-parasitized cells. Rosetting parasitized cells would adhere to each other (three pairs tested).
Rheology of Rosettes lnnduced by P. falciparum 25
761
-
20 -
ENFlY TIME
(SEC) 15 -
0
2
4
6
>7
NUMBER OF CELLS Fig 6. Variation in the time required for entry of a rosette into a micropipette as a function of the number of cells in the rosette. Data are mean +SE for between three and 25 rosettes in each sixe category: a total of 8 5 rosettes were tested. The value at one cell ( 0 )is the mean entry time for 110 non-parasitized cells.
Fig 5. Series of video images showing a rosette being aspirated into a micropipette with internal diameter of 4.3 pm.
If rosettes were sheared in a cone-plate viscometer in a medium containing dextran 500, rosetting frequency was reduced by a quarter after 1 min exposure to 1.4 Pa, and by two-thirds after 5 min (Fig 4). However, it was noted that rosetting frequency reduced from 66%in the original culture to 43% in the dextran 500, even before stress was applied. In a second series of experiments, dextran 40 was used to increase the medium viscosity. In these experiments, up to 5 min at 1.4-1.7 Pa had no systematic effect on rosetting frequency, at room temperature or at 3 7°C (Fig 4). Exposure to dextran T40 for 15 min had no effect on rosetting per se
(rosetting frequency 59 f 12% before versus 60 f7% after suspension in dextran). The time required for entry into a micropipette (internal diameter 4.3 pm) at a constant aspiration pressure (30 Pa) was compared for 85 rosettes (size 4.95 f 1.3 cells, mean fSD) and I 10 non-parasitized cells from four cultures. Fig 5 shows a series of pictures of a rosette entering the micropipette. Generally, the first (non-parasitized) cells entered rapidly, the parasitized cell entered next, and then the remainder of the rosette was dragged slowly into the pipette. If the .rosette was ejected from the pipette after complete entry, the rosette remained intact, although occasionally a cell became detached during ejection. The variation in pipette entry time with rosette size is shown in Fig 6. Entry time increased with rosette size in a strongly non-linear manner. If the entry time for each rosette was expressed relative to the mean value for non-parasitized cells from the same culture, relative rosette entry time had a median value of 16 f5 and a mean of 3 5 f 1 4 (mean &SD of four sample medians or means). A few rosettes (6/85) would not enter the pipette at the applied pressure. In absolute terms, non-parasitized cells had a mean entry time of 0.12 5 f 0.01 s (mean f SD for four sample means). DISCUSSION
The present study investigated the rheological properties of a stable rosetting strain of P. fakiparum (K + PA1) (Udomangspetch et al, 1989).We found that the strength of attachment between non-parasitized and parasitized cells in pre-formed rosettes was at least 5 times higher than that between parasitized cells and either cultured vascular endothelial cells or melanoma cells (Nash et aJ, 1992). When new rosettes were formed by micro-manipulation, adhesion between cells
762
Gerard B. Nash et al
occurred immediately after contact, but the strength of the interaction increased quite slowly thereafter: Fig 3 suggests that re-development of the full interaction between disrupted rosettes might take hours. In bulk suspension, rosettes were not disrupted by a shear stress of about 1 ' 5 Pa, which is of the order of the stress in arterial flow (Chien, 1987). In addition. rosettes remained intact after aspiration into and ejection from a micropipette with internal diameter of 4 . 3 pm. Nevertheless. resistance to flow into the micropipette was increased dramatically by rosetting. This pipette diameter is of the order of the size of the entrance of a capillary. and the results show the potential for rosettes to impair capillary perfusion. An additional, fortuitous finding was that dextran could reverse rosetting. When rosettes were suspended in buffer which contained dextran with MW 5 0 0 0 0 0 , there was a consistent reduction in rosetting frequency. However, this was not the case when dextran with MW 40 000 was used. even at a much higher molarity. Dextrans bind to the red cell membrane, and dextran 500, but not dextran 40. induces marked red cell-red cell aggregation (Brooks et aJ. 1980).It has previously been shown that rosetting is inhibitable by heparin (Carlson et a/, 1990b). Heparin and other sulphated polysaccharides have recently been shown also to inhibit the adhesive interaction between rolling neutrophils and endothelium in vivo (Tangelder & Arfors. 1991 ). It appears that the ability of surface modifying agents and polysaccharides to inhibit the adhesive interactions of parasitized cells is worthy of further investigation. The pathology of falciparum malaria may be influenced by the rheological abnormalities of invaded red blood cells. Parasitized cells become increasingly less deformable. due to an increase in membrane rigidity and to the presence of the large rigid parasite itself (Nash et a / , 1989). At the pigmented trophozoite stage of development. the parasitized cells develop the ability to adhere to microvascular endothelium and disappear from the peripheral circulation (Luse & Miller. 1971: Udeinya et a/. 1981). Rosetting is an additional. recently discovered alteration in the properties of parasitized red cells, which also has the potential to inhibit perfusion of the microcirculation (Ldomangspetch et al. 1989: Howard & Gilladoga. 1989; Kaul et a / , 1991). If present in the arterial circulation. rosettes might inhibit flow into capillaries. Emerging down stream, they might actually inhibit adhesion to endothelium. Rosetting has been shown to reduce cytoadhesion of parasitized cells in a static assay (Howard & Gilladoga, 1989 I. However, observations in the micropipette (e.g. see Fig 6 ) suggest that when flowing through a narrow vessel, the parasitized cell within a rosette would be forced into contact with the vessel wall. Under these conditions, the non-parasitized cells would not necessarily inhibit the parasitized cell from adhering to the endothelium. Our findings also show that the inter-cellular interaction in rosettes is at least as strong as that which binds parasitized cells to the endothelial adhesion receptors ICAM-1 and CD36. Thus. if rosettes are disrupted during flow through capillaries. as suggested by Kaul et a1 ( 199 1), they might re-form on to parasitized cells which had already adhered to venular walls. This would worsen vascular occlusion. Endothelial adhesion
of individual parasitized cells alone, should not necessarily block microvessels, since even a fully lined venule (typically ofdiameter 20-50 pm) should have a lumen sufficiently large to allow continued flow, albeit with increased resistance. Margination of neutrophils, which are much larger cells, can be demonstrated in exposed microvessels and does not stop flow of red cells in the central region (Bagge & Brinemark, 1977: Tangelder & Arfors, 1991). Overall, rosetting appears capable of contributing to ischaemic complications of malaria, particularly in conjunction with cytoadhesion. Post-mortem studies of those dying with severe malaria have indicated that rosettes may indeed be found along with parasitized cells adhered to endothelium in the microcirculation (Pongponratn et al, 1991). Recent studies have shown that the rosetting frequency in isolates derived from peripheral blood drawn from children with cerebral malaria is higher than in isolates from those with uncomplicated malaria (Carlson et al, 1990a: Wahlgren et al, 1990). Other studies have examined the correlation between cytoadhesion and the clinical severity of malaria (Marsh et al, 1988: Ho et al, 1991b: Ockenhouse et al, 1991): these have not shown a connection between adhesiveness of parasitized cells and cerebral malaria, but have indicated that increased adhesion is related to impairment of the function of other organs. Rheological studies of rosetting and cytoadhesion in clinical isolates of P. fakiparum might clarify the role of these phenomena in pathology. ACKNOWLEDGMENTS This investigation received the financial support of the UNDP/World Bank/WHO Special Programme for Research and Training in Tropical Diseases (TDR). We gratefully acknowledge the assistance of Mr Don Ritchie who designed and constructed the apparatus for electronic analysis of cell entry into micropipettes. REFERENCES Bagge. U. 8: Brinemark. P.-I. (1977) White blood cell rheology. An intravital study in man. Advances in Microcirculation, 7, 1-1 7 . Brooks. D.E.. Grieg, R.G. & Jansen, J. (1980) Mechanisms of erythrocyte aggregation. Erythrocyte Mechanics and Blood Flow (ed. by G. R. Cokelet. H. J. Meiselman and D. E. Brooks), pp. 119-140. Alan R. Liss. New York.
Carlson.1.. Helmby. H., Hill. A.V.S.. Brewster. D.. Greenwood, B.M. & Wahlgren, M. ( 1990a)Human cerebral malaria: associationwith erythrocyte rosetting and lack of anti-rosettingantibodies. Lancet, 3 36, 145 7-1 460.
Carlson.J.C.. Holmquist. G.. Taylor, D.W.. Perlmann, P. & Wahlgren, M. (1990b) Antibodies to a histidine-rich protein (Pf-HRPI) disrupt spontaneously formed Plasmodium Julciparurn erythrocyte rosettes. Proceedings ofthe National Academy ofsciences ofthe United States ofAmerica. 87, 2511-2515. Chien. S. (198 7) Physiological and pathophysiological significance of hemorheology. Clinical Hemorheology (ed. by S. Chien. J. Dormandy, E. Ernst and A. Matrai), pp. 125-164. Martinus Nijhoff, Dordrecht. Hasler. T.. Handunnetti. S.M.. Aguiar. J.C., vanschranvendijk, M.R., Greenwood, B.M.. Lallinger, G.. Cegielski.P. &Howard,R.J.(1990) In vitro rosetting. cytoadherence.and microagglutination proper-
Rheology of Rosettes Induced by P. falciparum ties of Plasmodium falciparum-infected erythrocytes from Gambian and Tanzanian patients. Blood, 76, 1845-1 852. Ho. M.. Davis, T.M.E., Silamut. K.. Bunnag, D. & White, N.J. (1991a) Rosette formation of Plasmodium falciparum-infected erythrocytes from patients with acute malaria. Infection and Immunity, 59, 2 1 35-2 139. Ho. M.. Singh, B., Looareesuwan, S., Davis, T.M.E., Bunnag, D. & White. N.J. (1991b) Clinical correlates of in vitro P. falciparum cytoadherence. Infection and Immunity, 59, 873-878. Howard. R.J. & Gilladoga. A.D. (1989) Molecular studies related to the pathogenesis of cerebral malaria. Blood. 74. 2603-2618. Kaul. D.K.. Roth. E.F.. Nagel. R.L., Howard. R.J. &Handunnetti,S.M. ( 1991) Rosetting of Plasmodiumfalciparum-infected red blood cells with uninfected red blood cells enhances microvascular obstruction under flow conditions. Blood. 78, 812-819. Luse, S.A. & Miller, L.H. (1971) Plasmodium falciparum malaria: ultrastructures of parasitized erythrocytes in cardiac vessels. American journal of Tropical Medicine and Hygiene, 20, 655-660. MacPherson, G.G.. Warrel, M.J.. White, N.J., Looareesuwan. S. & Warrel. D.A. (1985) A quantitative ultrastructural analysis of parasitized erythrocyte sequestration. American Journal of Pathology. 119, 385-401. Marsh, K.. Marsh, V.M.. Brown, J., Whittle, H.C. & Greenwood, B.M. ( 1988) Plasmodiumfalciparum: the behaviour of clinical isolates in an in vitro model of infected red blood cell sequestration. Experimental Parasitology, 6 5 , 202-208. Nash, G.B.. Cooke, B.M.. Marsh, K., Berendt. A., Newbold, C. & Stuart, J. (1992) Rheological properties of the adhesive interactions of red blood cells parasitized by Plasmodiumfalciparum. Blood, 79, 798-807. Nash. G.B., O’Brien. E., Gordon-Smith, E.C.&Dormandy,J.A. (1989) Abnormalities in the mechanical properties of red blood cells caused by Plasmodium falciparum. Blood. 74, 855-86 1. Ockenhouse, C.F.. Ho, M., Tandon. N.N.. Van Seventer. G.A., Shaw. S., White, N.J.. lamieson, G.A.. Chulay, 1.0. & Webster. H.B.
763
(1991) Molecular basis of sequestration in severe and uncomplicated Plasmodium falciparum malaria: differential adhesion of infected erythrocytes to CD36 and ICAM-I ( 0 5 4 ) . jourrinl of Infectious Diseases, 164, 163-169. Pongponratn. E.. Riganti, M.. Punpoowong, B. & Aikawa. M. ( 1 991) Microvascular sequestration of parasitized erythrocytes in human falciparum malaria: a pathological study. Amprican Joirrnaf o/ Tropical Medicine and Hygiene, 44. 168-1 75. Tangelder, G.J. & Arfors, K.-E. (1991) Inhibition of leukocyte rolling in venules by protamine and sulfated polysaccharides. Blood. 77, 1565-1571. Trager. W. & Jensen, J.B. (1976) Human malaria parasites in continuous culture. Science, 193, 673-675. Treutiger, C.-J.. Hedlund, I.. Helmby. H., Carlson, 1.. Jepson. A,, Twusami, P.. Kwiatkowski. D.. Greenwood, B.M. & Wahlgren. M. (1992) Rosetting of Plasmodium falciparum isolates and antirosetting activity of sera from Gambians with cerebral and uncomplicated malaria. American lournu1 of Tropical Medicine and Hygiene. in press. Udeinya, I.]., Schmidt, J.A.. Aikawa. M.. Miller, L.H. & Green, I. (198 1) Falciparum malaria-infected erythrocytes specifically bind to cultured human endothelial cells. Science. 21 3, 555-557. Udomangspetch, R.. Wlhlin, B.. Carlson. J., Berxins, K., Torii. M.. Aikawa. M., Perlmann, P. & Wahlgren, M. (1989) Plasmodium falciparum-infected erythrocytes form spontaneous erythrocyte rosettes. Journal of Experimental Medicine, 169, 18 3 5-1 840. Wahlgren. M. ( 1 992) Erythrocyte rosetting. sequestration and severe Plasmodium falcipurum malaria. Experimental Parasitology, in press. Wahlgren. M., Carlson. J.. Ruangjirachuporn. W.. Conway. D.. Helmby. H.. Martinez, A,. Patarroyo. M.E. & Riley, E. (1990) Geographical distribution of Plasmodium falciparum rosetting and frequency of rosetting antibodies in human serum. Arneriran Journal of Tropical Medicine and Hygiene, 4 3 , 333-338.
Rheological properties of rosettes formed by red blood cells parasitized by Plasmodium fakiparum B . N A S H ,BRIAN M . C O O K E , JOHAN CARLSON' A N D MATS WAHLGREN* Department of i%ernatology, The Medical School, University of Birmingham, Birmingham, U.K., and *Department of Immunology, Stockholm University, Stockholm, Sweden GERARD
Rrlceived 3 Iune 1992; acceptedfor publication 9 July 7 9 9 2
Summary. A proportion of red blood cells parasitized by Plismodium falciparum form rosettes with non-parasitized red cells. Although these rosettes are thought to impair microcirculatory flow, their rheological characteristics have not been fully described. Using dual-micropipette manipulation to pull apart individual rosettes, we found that the forces binding rosettes together were strong (average force for removal of a cell was 4.4 x lo-"' N, approximately 5 times that required to detach a parasitized cell adhered to cultured endothelium). If disrupted rosettes were re-formed, cells rosetted immediately on contact, but the strength of attachment increased over minutes, and did not apparently reach its maximal level
for hours. All non-parasitized cells tested could adhere to rosette-forming parasitized cells. Rosettes could withstand arterial flow stresses (1.4-1'6 Pa) for minutes without disintegration. To test the effects of rosetting on flow resistance, the time required for entry into a 4.3 pm pipette was measured. Entry times depended strongly on the number of cells in the rosette, and averaged 3 5 times longer than for non-parasitized cells. Our studies show that the cell-cell attachments within rosettes are strong, and suggest that rosettes might survive both the arterial circulation and passage through microvessels and could contribute to the ischaemic complications of falciparum malaria.
assess their role in circulatory pathology, it is necessary to investigate the rheological characteristics of rosettes. It is known that rosettes are resistant to shear stress such as that imposed by pipetting of cell suspensions (Howard & Gilladoga, 1989). and that they increase the peripheral resistance when perfused into isolated rat vasculature (Kaul et al. 1991). Here we report detailed information on the rheological characteristics of an in vitro selected, stable rosetting strain of P. falciparum (R+PAl).
Ischaemic complications of falciparum malaria, particularly cerebral malaria, may be life threatening. The microvasculature of those who have died from severe malaria has been found to be filled, and presumably blocked, with cells containing mature, pigmented parasites (MacPherson et al, 1985; Pongponratn et al, 1991). It has formerly been thought that this sequestration arises from adhesion of parasitized cells to vascular, and particularly venular, endothelium (Luse&Miller,1971: Udeinyaet al, 1981).However, recent studies have shown that certain strains of P. falciparum, including those newly isolated from patients in endemic areas, can induce parasitized cells to act as the focus for rosettes with non-parasitized cells (for review see Wahlgren, 1992). The percentage of rosetting parasites varies widely between clinical isolates (Carlson et a/, 1990a: Hasler et al, 1990; Wahlgren et a/. 1990: Ho et al, 1991a), and some (Carlson et al, 1990a; Treutiger et al, 1992) but not all (Ho et a!. 1991a) studies have found a relation between this percentage and the severity of the clinical pathology. Rosettes could add to microvascular occlusion, either by hindering flow through capillaries or by forming on to Parasitized cells already adhered to venular endothelium. To
METHODS
Malarial culture. Cryopreserved samples of Plasmodium falciparum (strain R + PA1, cloned by limiting dilution: Udomangspetch et al, 1989) were thawed and cultured in human red blood cells (blood group A) suspended in RPMI 1640 (Flow Laboratories Ltd, Irvine, U.K.) supplemented with HEPES (37.5 mM). glucose (11 mM), glutarnine (4 mM), gentamycin sulphate (28 mg/l) and human AB serum (10%). Culture was carried out using the candle-jar method of Trager & Jensen (1976). Parasitaemia and stage of parasite maturation were assessed by examination of blood films stained with Giemsa. The majority of experiments were carried out on the day (average 20 h) or 3 d after thawing of isolates, so that synchronous cultures were utilized. Other-
Correspondence: Dr G. Nash, Department of Haematology, The Medical School, University of Birmingham, Birmingham B15 2TT.
75 7
758
Gerard B. Nash et a1
Fig 1. Video image of a rosette composed of five red cells being held between two micropipettes with internal diameters of approximately 2 and 1 . 2 pm.
wise. synchronization of cultures was carried out by incubation of cells in 5% w/v sorbitol on the day before analysis, which caused selective lysis of trophozoites and schizonts. On average. the cultures tested had a total parasitaemia of 2.4%. with pigmented trophozoites making up over 90% of the parasites. The percentage of all parasitized cells that were contained within rosettes (rather than existing as single cells in suspension) was defined as the rosetting frequency. This was assessed by light microscopy after cultures had been mixed 1 : 1 with 2% glutaraldehyde in phosphate buffered saline. A 4 0 x water-immersion objective lens (NA =0.75, Carl Zeiss (Oberkochen) Ltd. Germany) allowed clear identification of pigmented parasites in single cells or rosettes (see, e.g.. Figs 1. 2 and 6). A total of 1 0 0 parasitized cells were counted in each sample. and the percentage within rosettes was determined. In this study, rosetting frequency averaging 68 & 8% (mean i SD. seven cultures). Micropipette analysis of rosettes. Samples of malarial cultures were diluted 100-fold in culture medium with pH adjusted to 7.4. The suspension was placed in a micropipette chamber which was approximately 1 m m deep and made up of a microscope slide and coverslip. separated by two parallel strips of multiply-folded parafilm. The chamber was placed on the stage of a microscope, and two micropipettes were introduced into it from opposite directions. The pipettes were connected to separate hydrostatic reservoirs whose heights were adjusted by micrometers. so that suction pressures were independently. precisely controlled in each pipette. The pipette tips were viewed using a 4 0 x water-immersion lens and a video-microscope system which gave a final magnification on the video monitor of approximately x 6000. In general, a larger pipette (pipette A, i.d. about 2 pn) was used to hold rosettes firmly with a n applied pressure of 2-4 cmHLO (200-400 Pa), while the other (pipette B, i.d.=Dp= 1.2-1'5 pm) was used to pick u p desired cells or rosettes. Disruption and reforrnation of rosettes. A rosette was picked up using pipette B and passed to the larger, stationary pipette
A. The pressure in pipette B was adjusted to zero and then suction was increased in steps. At each step, a red cell from the stationary rosette was aspirated into the smaller pipette B and this pipette was withdrawn (Fig 1).The cell either pulled free from pipette B or became detached from the rosette. The pressure (P)required to detach the cell from the rosette was taken as the average of the final pressure (which detached the cell) and the previous pressure which had not detached the cell. The pressure was used to calculate the force (F)required to detach the cell from the rosette, using the equation:
F=P*zD:
(Nash et al, 1992)
This procedure is essentially the same as that previously used to measure the force required to remove parasitized cells after they had adhered to endothelial cells or melanoma (C32) cells grown on microbarrier beads (Nash et al. 1992). In some experiments. after all the non-parasitized cells had been removed from the rosetting, parasitized cell, the last cell to be removed was brought back into contact with the parasitized cell (Fig 2). After a timed delay, the force to detach the non-parasitized cell was re-measured. as described above. This measurement was repeated: delay times of 5, 60, 1 2 0 and 300 s were tested. The time actually taken to vary the pressure and to remove the cell averaged 3 0 s. Thus, the measured force for detachment was considered to be that which was acting at the midpoint of the removal procedure, i.e. 20, 75, 1 3 5 and 3 1 5 s. In other experiments, cells were picked u p a t random from those settled in the chamber using pipette B and a low suction pressure (about 20-40 Pa), and brought into contact with the parasitized cell held in pipette A. The pressure in pipette B was maintained and the pipette was withdrawn within a few seconds of cell-cell contact. It there was a significant interaction with the parasitized cell, then the newly adhered cells either pulled free from pipette B or showed deformation and elongation before pulling free from the parasitized cell. To avoid re-testing of cells that had just been removed from the
Rheology of Rosettes Induced by P. fakiparum
759
Fig 2. Series of video images showing a non-parasitized cell being brought into contact with a rosetting. parasitized cell (A-C). and then being reaspirated and pulled away (D-F). The arrow in A indicates the parasitized cell held stationary by the larger pipette.
rosette, the parasitized cell was first moved to a new region of the chamber. Kesistance of rosettes to micropipette aspiration. The effect of rosetting on flow resistance was assessed by measuring the time required for rosettes of different size to enter a micropipette (i.d. = 4 . 3 pm). The method is essentially the same as that previously used to study the deformability of individual parasitized cells (Nash et al, 1989). The micropipette apparatus was as described above, except that a single micropipette was introduced into a micropipette chamber made up with a U-shaped gasket. A fixed suction pressure (30 Pa) was applied, and the pipette tip brought close to the cell or rosette to be tested. The time required for entry was measured electronically, using Ag :AgC12 electrodes: one within the pipette and one within the chamber. A constant current was parsed between the electrodes and the cell entry was sensed as a voltage pulse, which was analysed by microcomputer after analogue-to-digital conversion. The time between the signal first crossing a preset threshold, and its falling to 90% of its peak value was taken to be the entry time. Measure-
ments using simultaneous electronic and video-playback analysis verified that this procedure accurately measured the entry time. Exposure of rosettes to shear stress. Malarial cultures were diluted to 5 x lo8 cells/ml in culture medium with pH adjusted to 7.4. Aliquots of these suspensions were then mixed with dextran-containing media as follows: (A) 1: 1 with 22% w/v dextran 500 (MW = 500 000), to give a final viscosity of about 30 mPa.s at room temperature; (B) 1: 1 with 20% w/v dextran 40 (MW = 40 000), to give a final viscosity of about 7 mPa.s at room temperature: (C) 1 :2 with 20% w/v dextran 40, to give a final viscosity of about 7 mPa.s at 37OC. The dextrans (Sigma Chemical Co. Ltd, Poole. U.K.) were dissolved in water with 20 mM HEPES (pH = 7.4) and the osmolarity adjusted to 290 mOsmol/kg by addition of NaCI. The suspension under analysis was placed on the plate of a cone-plate viscometer (model LVT, Brookfield Engineering Laboratories Inc., Stoughton, Mass., U.S.A.) and exposed to a uniform shear rate of 4 6 s-' for suspension A. or 2 30 s for suspensions B or C. Suspensions A and B were tested at
760
Gerard B. Nash et a1 Table I. Force required t o detach individual non-parasitiaed cells from rosettes. O 5
1
FORCE
Median 90th percentile .\rean +SI) ( r i = 1 5 t Mean rtS1)ofsaniplemeans I r r = h t
3. 3
(RELATIVE)
i.X
4.4zt2.2 4.1il.3
Force for detachment from endothelium: W a n +Sl)ofsarnplc mcitns(ii= 5 ) 0.9rtO.1
I
0O 2I!
i1
1:orce for detachment from melanoma cells: Mean rSIlofsample means (rt=4) O.hrt0.2
* Individual cells were detached from rosettes by iiiicromaiiipulation. The force in Newtons (N\ was calculated from the aspiration pressure (in Pascals) required to remove each cell. and the pipette diameter (see Methods). I t i ) = number of cells or number of cultures tested. f Data from Sash (11 1992 1. obtained by micropipette manipulation olparasitized cells (strain ITO?) into comiii't with endothelial or melanoma cells. and detachment as for rosettes. c3f
room temperature, while suspension C was tested at 3 7°C. The actual shear stress to which the cells were exposed was measured by the viscometer. and averaged 1.46&0.14 Pa (mean iSD, 10 experiments). After 1 min of shearing. the viscometer was halted and 100 PI of cell suspension removed and fixed by mixing 1 : 1 with glutaraldehyde ( 2 % in phosphate buffered saline). This process took 5-10 s. The suspension was sheared for a further 4 min. and another 100 p1 aliquot fixed. Rosetting frequency in the fixed, sheared samples was counted as described earlier. Results were compared to those obtained for aliquots of cells suspended in dextran but not sheared (samples were fixed before and after the period of shearing). and for aliquots of the original culture. RESCLTS Values for the force required to remove individual nonparasitized cells from rosettes which had been cultured. undisturbed overnight, are summarized in Table I. The distribution of the strengths of attachment was positively skewed (e.g. mean > median: 90th percentile > twice median) and varied widely from cell to cell, and to a lesser extent. between samples. The average size of the rosettes tested in these experiments was 4.8 f 1.4 cells (mean +SD). Table I also shows the force required to remove individual parasitized cells which had adhered to either cultured human vascular endothelium or cultured amelanotic melanoma cells. These data were obtained by a similar micropipette method (Nash et nl, 1992): the force for either cell line was about one-fifth of that required to remove a cell from a rosette. Following rosette disruption, the strength of attachment of a newly re-formed rosette was tested as a function of the time after cell-cell contact. The force required to disrupt the rosette
O I
I
I
0
100
I
I
200
I
I
I
300
TIME (SEC)
Fig 3. Variation in the force required to remove non-parasitized red cells from a rosetting parasitized cell as a function of the time after the cells were brought into contact. The force is expressed relative to the force required to remove the non-parasitized cell from the original rosette. Data are mean &SD of means from three experiments: five cells were tested in each experiment.
FF(EQuENcy
0
0
1
2 3 TIME (MIN)
4
5
Fig 4. Variation in rosetting frequency of Parasitized cells as a function of the time of exposure to a shear stress averaging 1.46 Pa. Data are mean ltSE for experiments carried out using suspending medium containing dextran 500 at room temperature (m.n = 4 ) . dextran 40 at room temperature ( 0 .n = 3 ) or dextran 40 at 37°C ( 0 . I f = 31.
increased with time, and after 5 min, reached 40% of the force required to disrupt the original rosette (Fig 3 ) . After rosette disruption, new red cells were chosen from those settled in the micropipette chamber, and brought into contact with the parasitized cell. All of 110 non-parasitized cells tested in this way adhered to the parasitized cell (11cells from three cultures). In addition, non-rosetting parasitized cells containing pigment also adhered (24/24, tested with nine rosettors). but would not adhere to non-parasitized cells. Rosetting parasitized cells would adhere to each other (three pairs tested).
Rheology of Rosettes lnnduced by P. falciparum 25
761
-
20 -
ENFlY TIME
(SEC) 15 -
0
2
4
6
>7
NUMBER OF CELLS Fig 6. Variation in the time required for entry of a rosette into a micropipette as a function of the number of cells in the rosette. Data are mean +SE for between three and 25 rosettes in each sixe category: a total of 8 5 rosettes were tested. The value at one cell ( 0 )is the mean entry time for 110 non-parasitized cells.
Fig 5. Series of video images showing a rosette being aspirated into a micropipette with internal diameter of 4.3 pm.
If rosettes were sheared in a cone-plate viscometer in a medium containing dextran 500, rosetting frequency was reduced by a quarter after 1 min exposure to 1.4 Pa, and by two-thirds after 5 min (Fig 4). However, it was noted that rosetting frequency reduced from 66%in the original culture to 43% in the dextran 500, even before stress was applied. In a second series of experiments, dextran 40 was used to increase the medium viscosity. In these experiments, up to 5 min at 1.4-1.7 Pa had no systematic effect on rosetting frequency, at room temperature or at 3 7°C (Fig 4). Exposure to dextran T40 for 15 min had no effect on rosetting per se
(rosetting frequency 59 f 12% before versus 60 f7% after suspension in dextran). The time required for entry into a micropipette (internal diameter 4.3 pm) at a constant aspiration pressure (30 Pa) was compared for 85 rosettes (size 4.95 f 1.3 cells, mean fSD) and I 10 non-parasitized cells from four cultures. Fig 5 shows a series of pictures of a rosette entering the micropipette. Generally, the first (non-parasitized) cells entered rapidly, the parasitized cell entered next, and then the remainder of the rosette was dragged slowly into the pipette. If the .rosette was ejected from the pipette after complete entry, the rosette remained intact, although occasionally a cell became detached during ejection. The variation in pipette entry time with rosette size is shown in Fig 6. Entry time increased with rosette size in a strongly non-linear manner. If the entry time for each rosette was expressed relative to the mean value for non-parasitized cells from the same culture, relative rosette entry time had a median value of 16 f5 and a mean of 3 5 f 1 4 (mean &SD of four sample medians or means). A few rosettes (6/85) would not enter the pipette at the applied pressure. In absolute terms, non-parasitized cells had a mean entry time of 0.12 5 f 0.01 s (mean f SD for four sample means). DISCUSSION
The present study investigated the rheological properties of a stable rosetting strain of P. fakiparum (K + PA1) (Udomangspetch et al, 1989).We found that the strength of attachment between non-parasitized and parasitized cells in pre-formed rosettes was at least 5 times higher than that between parasitized cells and either cultured vascular endothelial cells or melanoma cells (Nash et aJ, 1992). When new rosettes were formed by micro-manipulation, adhesion between cells
762
Gerard B. Nash et al
occurred immediately after contact, but the strength of the interaction increased quite slowly thereafter: Fig 3 suggests that re-development of the full interaction between disrupted rosettes might take hours. In bulk suspension, rosettes were not disrupted by a shear stress of about 1 ' 5 Pa, which is of the order of the stress in arterial flow (Chien, 1987). In addition. rosettes remained intact after aspiration into and ejection from a micropipette with internal diameter of 4 . 3 pm. Nevertheless. resistance to flow into the micropipette was increased dramatically by rosetting. This pipette diameter is of the order of the size of the entrance of a capillary. and the results show the potential for rosettes to impair capillary perfusion. An additional, fortuitous finding was that dextran could reverse rosetting. When rosettes were suspended in buffer which contained dextran with MW 5 0 0 0 0 0 , there was a consistent reduction in rosetting frequency. However, this was not the case when dextran with MW 40 000 was used. even at a much higher molarity. Dextrans bind to the red cell membrane, and dextran 500, but not dextran 40. induces marked red cell-red cell aggregation (Brooks et aJ. 1980).It has previously been shown that rosetting is inhibitable by heparin (Carlson et a/, 1990b). Heparin and other sulphated polysaccharides have recently been shown also to inhibit the adhesive interaction between rolling neutrophils and endothelium in vivo (Tangelder & Arfors. 1991 ). It appears that the ability of surface modifying agents and polysaccharides to inhibit the adhesive interactions of parasitized cells is worthy of further investigation. The pathology of falciparum malaria may be influenced by the rheological abnormalities of invaded red blood cells. Parasitized cells become increasingly less deformable. due to an increase in membrane rigidity and to the presence of the large rigid parasite itself (Nash et a / , 1989). At the pigmented trophozoite stage of development. the parasitized cells develop the ability to adhere to microvascular endothelium and disappear from the peripheral circulation (Luse & Miller. 1971: Udeinya et a/. 1981). Rosetting is an additional. recently discovered alteration in the properties of parasitized red cells, which also has the potential to inhibit perfusion of the microcirculation (Ldomangspetch et al. 1989: Howard & Gilladoga. 1989; Kaul et a / , 1991). If present in the arterial circulation. rosettes might inhibit flow into capillaries. Emerging down stream, they might actually inhibit adhesion to endothelium. Rosetting has been shown to reduce cytoadhesion of parasitized cells in a static assay (Howard & Gilladoga, 1989 I. However, observations in the micropipette (e.g. see Fig 6 ) suggest that when flowing through a narrow vessel, the parasitized cell within a rosette would be forced into contact with the vessel wall. Under these conditions, the non-parasitized cells would not necessarily inhibit the parasitized cell from adhering to the endothelium. Our findings also show that the inter-cellular interaction in rosettes is at least as strong as that which binds parasitized cells to the endothelial adhesion receptors ICAM-1 and CD36. Thus. if rosettes are disrupted during flow through capillaries. as suggested by Kaul et a1 ( 199 1), they might re-form on to parasitized cells which had already adhered to venular walls. This would worsen vascular occlusion. Endothelial adhesion
of individual parasitized cells alone, should not necessarily block microvessels, since even a fully lined venule (typically ofdiameter 20-50 pm) should have a lumen sufficiently large to allow continued flow, albeit with increased resistance. Margination of neutrophils, which are much larger cells, can be demonstrated in exposed microvessels and does not stop flow of red cells in the central region (Bagge & Brinemark, 1977: Tangelder & Arfors, 1991). Overall, rosetting appears capable of contributing to ischaemic complications of malaria, particularly in conjunction with cytoadhesion. Post-mortem studies of those dying with severe malaria have indicated that rosettes may indeed be found along with parasitized cells adhered to endothelium in the microcirculation (Pongponratn et al, 1991). Recent studies have shown that the rosetting frequency in isolates derived from peripheral blood drawn from children with cerebral malaria is higher than in isolates from those with uncomplicated malaria (Carlson et al, 1990a: Wahlgren et al, 1990). Other studies have examined the correlation between cytoadhesion and the clinical severity of malaria (Marsh et al, 1988: Ho et al, 1991b: Ockenhouse et al, 1991): these have not shown a connection between adhesiveness of parasitized cells and cerebral malaria, but have indicated that increased adhesion is related to impairment of the function of other organs. Rheological studies of rosetting and cytoadhesion in clinical isolates of P. fakiparum might clarify the role of these phenomena in pathology. ACKNOWLEDGMENTS This investigation received the financial support of the UNDP/World Bank/WHO Special Programme for Research and Training in Tropical Diseases (TDR). We gratefully acknowledge the assistance of Mr Don Ritchie who designed and constructed the apparatus for electronic analysis of cell entry into micropipettes. REFERENCES Bagge. U. 8: Brinemark. P.-I. (1977) White blood cell rheology. An intravital study in man. Advances in Microcirculation, 7, 1-1 7 . Brooks. D.E.. Grieg, R.G. & Jansen, J. (1980) Mechanisms of erythrocyte aggregation. Erythrocyte Mechanics and Blood Flow (ed. by G. R. Cokelet. H. J. Meiselman and D. E. Brooks), pp. 119-140. Alan R. Liss. New York.
Carlson.1.. Helmby. H., Hill. A.V.S.. Brewster. D.. Greenwood, B.M. & Wahlgren, M. ( 1990a)Human cerebral malaria: associationwith erythrocyte rosetting and lack of anti-rosettingantibodies. Lancet, 3 36, 145 7-1 460.
Carlson.J.C.. Holmquist. G.. Taylor, D.W.. Perlmann, P. & Wahlgren, M. (1990b) Antibodies to a histidine-rich protein (Pf-HRPI) disrupt spontaneously formed Plasmodium Julciparurn erythrocyte rosettes. Proceedings ofthe National Academy ofsciences ofthe United States ofAmerica. 87, 2511-2515. Chien. S. (198 7) Physiological and pathophysiological significance of hemorheology. Clinical Hemorheology (ed. by S. Chien. J. Dormandy, E. Ernst and A. Matrai), pp. 125-164. Martinus Nijhoff, Dordrecht. Hasler. T.. Handunnetti. S.M.. Aguiar. J.C., vanschranvendijk, M.R., Greenwood, B.M.. Lallinger, G.. Cegielski.P. &Howard,R.J.(1990) In vitro rosetting. cytoadherence.and microagglutination proper-
Rheology of Rosettes Induced by P. falciparum ties of Plasmodium falciparum-infected erythrocytes from Gambian and Tanzanian patients. Blood, 76, 1845-1 852. Ho. M.. Davis, T.M.E., Silamut. K.. Bunnag, D. & White, N.J. (1991a) Rosette formation of Plasmodium falciparum-infected erythrocytes from patients with acute malaria. Infection and Immunity, 59, 2 1 35-2 139. Ho. M.. Singh, B., Looareesuwan, S., Davis, T.M.E., Bunnag, D. & White. N.J. (1991b) Clinical correlates of in vitro P. falciparum cytoadherence. Infection and Immunity, 59, 873-878. Howard. R.J. & Gilladoga. A.D. (1989) Molecular studies related to the pathogenesis of cerebral malaria. Blood. 74. 2603-2618. Kaul. D.K.. Roth. E.F.. Nagel. R.L., Howard. R.J. &Handunnetti,S.M. ( 1991) Rosetting of Plasmodiumfalciparum-infected red blood cells with uninfected red blood cells enhances microvascular obstruction under flow conditions. Blood. 78, 812-819. Luse, S.A. & Miller, L.H. (1971) Plasmodium falciparum malaria: ultrastructures of parasitized erythrocytes in cardiac vessels. American journal of Tropical Medicine and Hygiene, 20, 655-660. MacPherson, G.G.. Warrel, M.J.. White, N.J., Looareesuwan. S. & Warrel. D.A. (1985) A quantitative ultrastructural analysis of parasitized erythrocyte sequestration. American Journal of Pathology. 119, 385-401. Marsh, K.. Marsh, V.M.. Brown, J., Whittle, H.C. & Greenwood, B.M. ( 1988) Plasmodiumfalciparum: the behaviour of clinical isolates in an in vitro model of infected red blood cell sequestration. Experimental Parasitology, 6 5 , 202-208. Nash, G.B.. Cooke, B.M.. Marsh, K., Berendt. A., Newbold, C. & Stuart, J. (1992) Rheological properties of the adhesive interactions of red blood cells parasitized by Plasmodiumfalciparum. Blood, 79, 798-807. Nash. G.B., O’Brien. E., Gordon-Smith, E.C.&Dormandy,J.A. (1989) Abnormalities in the mechanical properties of red blood cells caused by Plasmodium falciparum. Blood. 74, 855-86 1. Ockenhouse, C.F.. Ho, M., Tandon. N.N.. Van Seventer. G.A., Shaw. S., White, N.J.. lamieson, G.A.. Chulay, 1.0. & Webster. H.B.
763
(1991) Molecular basis of sequestration in severe and uncomplicated Plasmodium falciparum malaria: differential adhesion of infected erythrocytes to CD36 and ICAM-I ( 0 5 4 ) . jourrinl of Infectious Diseases, 164, 163-169. Pongponratn. E.. Riganti, M.. Punpoowong, B. & Aikawa. M. ( 1 991) Microvascular sequestration of parasitized erythrocytes in human falciparum malaria: a pathological study. Amprican Joirrnaf o/ Tropical Medicine and Hygiene, 44. 168-1 75. Tangelder, G.J. & Arfors, K.-E. (1991) Inhibition of leukocyte rolling in venules by protamine and sulfated polysaccharides. Blood. 77, 1565-1571. Trager. W. & Jensen, J.B. (1976) Human malaria parasites in continuous culture. Science, 193, 673-675. Treutiger, C.-J.. Hedlund, I.. Helmby. H., Carlson, 1.. Jepson. A,, Twusami, P.. Kwiatkowski. D.. Greenwood, B.M. & Wahlgren. M. (1992) Rosetting of Plasmodium falciparum isolates and antirosetting activity of sera from Gambians with cerebral and uncomplicated malaria. American lournu1 of Tropical Medicine and Hygiene. in press. Udeinya, I.]., Schmidt, J.A.. Aikawa. M.. Miller, L.H. & Green, I. (198 1) Falciparum malaria-infected erythrocytes specifically bind to cultured human endothelial cells. Science. 21 3, 555-557. Udomangspetch, R.. Wlhlin, B.. Carlson. J., Berxins, K., Torii. M.. Aikawa. M., Perlmann, P. & Wahlgren, M. (1989) Plasmodium falciparum-infected erythrocytes form spontaneous erythrocyte rosettes. Journal of Experimental Medicine, 169, 18 3 5-1 840. Wahlgren. M. ( 1 992) Erythrocyte rosetting. sequestration and severe Plasmodium falcipurum malaria. Experimental Parasitology, in press. Wahlgren. M., Carlson. J.. Ruangjirachuporn. W.. Conway. D.. Helmby. H.. Martinez, A,. Patarroyo. M.E. & Riley, E. (1990) Geographical distribution of Plasmodium falciparum rosetting and frequency of rosetting antibodies in human serum. Arneriran Journal of Tropical Medicine and Hygiene, 4 3 , 333-338.
