Recent developments in vaccination against malaria: Metrizamide density gradients for separation of different developmental stages of malarial parasites*

Bulletin of the World MOMa.'t Organ.;asor., 57 (Suippi 1) 181-18? L1979) Metrizamide density gradients for separation of different developmental stages of malarial parasites* ELSIE M. EUGUI I & ANTHONY C. ALLISON 2 Density gradients with metrizamide, a tri-iodinated benzanmido derivative ofglucose, have been used to separate erythrocytes inifected wilh three species of murnne plasmodia. Uninfected erythrocytes separated wellfrom eryrhrocyIes containing parasites in different developmental stages. With metrizamide solutions, the densities required for isopyknic separation can be obtained without hypertonicity or high viscosity, and the viability and metabolism ofparasites and erythrocytes are not detectably modiJied by exposure to these solutions. This type of separation has many possible applications to biochemical and im- munological investigations. For many purposes it is useful to separate parasit- ized from non-parasitized erythrocytes and to separ- ate the various developmental stages of the parasite in poorly synchronized infections. Ideally the separation system should provide a means of obtaining popu- lations of erythrocytes containing each of the main developmental stages (early trophozoites. late tropho- zoites, and schizonts) in high purity under conditions that preserve the infectivity, antigenicity. and other properties of the parasitc and that do not change the biochemical reactions in the infected cells. V'arious density-gradient systems have been devised for this purpose, including continuous sucrose (1), con- tinuous bovine serum albumin (2), discontinuous phthalate (3), discontinuous Ficolla (4-5), anrd dis- continuous Stractan II (arabinogalactan)a (6). More recently, Plasmagelb has been used to concentrate schizont-infected red cells ofPlasmodiumfa(ciparum from human blood (7). Because red cells have a higher density than leuko- cytes, isopyknic separation of infected red cells re- quires gradient systems of correspondingly high den- sity. As discussed by Kreier (8), each of the systems so far used has limitations. In general, infected cells are separated from uninfected cells, but separation of the various stages of development is poor. Gradients of sucrose and other low molecular weight solutes are nmarkedly hypertonic and alter the density of the cells * From the Cell Patholog DIh ision, MRC Clinical Rescarcl Centre, Watford Road, Harrow, HAI 3UJ, England Research Fellow. Head of Division. Present address: Director General, Inter- national Laboratory for Rcmearlih on Animal Diseases, P.O Box 30709, Nairobi, Kenya This is the address to which rcquests for reprinla should be sent. as well as the properties of parasites and erytlirocytes. In addition, some of these substances, such as Ficoll and Stractan II, require a long preparation, with exhaustive dialysis against distilled water fol- lowed by concentration, before they are ready fol use. Often the viability of the parasiEes or cells aind their biochemical properties have not been determined. Hence there is a need for a density-gradient system that gives good separation of infected cells in various stages of development from one another and from uninfected cells, and that does not exert significant osmotic effects or impair the viability or metabolism of parasites and erythrocytes. Observations presented in this paper show that gradients of metrizamide used previously to separate different types of nucleated cells as well as subcelIular components and viruses (9) achieve this aim. It was also interesting to observe that changes in- duced in the density of red cells when they are infected with murine malarial parasites are different according to the species of Plasmodiuni used. MATERIALS AND METHODS Animiials and parasites P. yoelii, line YM, P. chabaudi, line AS, and P. inckeipetteri (kindly provided by Dr D. Walliker, IrnstituLe of Animal Genetics, Edinburgh, Scotland) originally passaged in C57BL mice. were maintained in BALB/C, adult female mice, obtained from the 37 Frorr. Sigma London C,iemical Co. Ltd, Poole, Dorset, England. b FTom Laboratoire Roger Belloni, Neuilly, Fiance. 3816 -181- 182 E. M. EUGUI & A. C. ALLISON Animal Breeding Unit of the Clinical Research Centre. In all 3 parasite strains. the infecting dose used was 1 x 106 parasitized red blood cells, suspended in phosphate-buffered saline (PBS), given intraperitoneallv. The strain of P. yoelii used is virulent, and at the dose given kills the animals on day 7, with a para- sitaemia averaging 70-80%. P. chabaudi and P. v. petteri both induce a mild and transient infection in the strain ofmice used, with parasitaemia averaging 50-60% on days 9-10, followed by recovery. Animals were usuatly bled on the day of peak parasitaemia; samples were taken at the same time to determine the percentage of parasitized red cells. Animals were always maintained with 0.5 g/litre (0.05%) of p-aminobenzoic acid in the drinking water. When fractions separated by density-gradient cen- trifugation were used later for reinjection or bio- chemical studies, cells after collection were immedi- ately kept in an ice bath, and all subsequent mnanipu- lations were performed at 4 C. Before loading on the gradient, cells were washed twice with the same buffer to be used for the preparation of the metrizamnide solution, but adding inactivated fetal calf serum (FCS) in a concentration of 20 g/litre (27o). Density gradients Solutions of metrizamide, a tri-iodiniated benza- mido derivative of glucose,c were prepared by adding different concentrations of the solute very slowly, with magnetic stirring, into a buffer solution. This buffer was the same as that described by Munthe-Kaas & Seglen (10) and contained 7.14 g of NaCl, 0-5 g of KCI, 0.18 g of CaCl.2H20, 2.4 g of HEPES, and 7.7 ml of I mol/litre NaOH in I litre of deionized dis- tilled water adjusted to pH 7.4. To prepare cell suspensions and wash the cells, 20 g of FCS per litre were incorporated as mentioned above. The metrizamide solutions were kept in the dark at 4 'C for no more than 1 week before use. Solu- tions prepaied under sterile conditions can be stored indefinitely at -20'C. Densities of solutions con- taining different concentrations of metrizamide were determined refractometrically and gravimetrically. Both of these measurements follow a linear relation- ship according to the concentration of metrizamnide (g/litre). Continuous gradients were prepared using a gradient-mixing apparatus, which allows the gradual mixing of two different concentrated solutions. Ac- cording to the requirements of each experiment, con- centrations were varied as described in the section on results. Usually 9 ml of each of the solutions were used, giving a final volume of 18 ml, the mixture I From Nyegaard & Co. A/S, Oslo, Norway. being placed in 25-ml polycarbonate MSE centrifuge tubes. Up to 100 pil of the pellet of washed red cells, suspended in 2 ml of the buffer plus FCS (20 g/litre) were loaded at the top of each gradient. Tubes were centrifuged at 5000 g for 30 min at 4-5 °C in an MSE 3 x 25 ml swing-out rotoi. In such an isopyknic gradient, cells produce bands when their density equilibrium is reached. After cen- trifugation, the tubes were photographed and the dif- ferent bands were collected, either from a hole in the bottom of the tube or with a Pasteur pipette from the top. The refractive index of each fraction was measured, and from this it was possible to calculate the density of the different cell fractions and con- sequently the metrizamide concentration required for optimal separation. A sample of cells recovered at each level of the gradient was washed, deposited in a cytocentrifuge, and the smears stained with Giemsa stain. Duplicates of each sample were screened and the parasites classified as young ring forms, tropho- zoites (intermediates), and schizonts (immature and mature), following accepted criteria as described by Carter & Diggs (11). RESULTS Gradient separation As far as we are aware, metrizamide has not been used before to separate red cells of any species, so we started by determining the density of normal red blood cells of BALB/c mice using this gradient sys- tem. Under the conditions described these cells equili- 30 1,150 t1 X 5~~~~~~~~~~~~~~~25 -T Refractive index 20 1,050 - I I i 10 15 20 25 30 Metrizam ide concentration 1% wlv) Fig. 1 The relationship between densities, refractive indices, and concentrations of metrizamide solutions. SEPARATION OF DEVELOPMENTAL STAGES brate between relative densities of 1.124 and 1.141, equivalernt to metrizamide concentrations ranging from 225 to 267 g/litre (22.5-26.7%), as can be seen in Fig. 1, where concentrations are referred to den- sities and refractive indices simultaneously. Since it is III*I1 _~~~~~~~~~ II I B C known that parasitized red blood cells have a lower density than normal erythrocytes, lower concen- trations than those required for normal erythrocytes were tried in different gradients to obtain the best separation. As the species of parasites used infect red blood cells in different stages of maturity, we have also adapted the gradient conditions to each of the three plasmodia used in this study. With P. yoelii, excellent separation can be obtained using a continuous metrizamide gradient between 230 and 175 g/lhtre (d = 1.126 to 1.095). In Fig. 2A, two tubes are shown after centrifugation, one with normal erythrocytes (right). in whiclh most of the cells have sedimented to [he bottom, and the other (left) con- taining erythrocytes parasitized with P. yoelii with many different fractions displayed from the bottom to the top. Up to 7 fractions containing parasitized cells in different stages ofmaturity havrc been collected from this gradient. At the bottom (d- 1.126) lie most of the normal red cells and also cells containing parasites at very early stages of maturity (nng forms, Fig. 3A). Between this and the next layer there is a band without cells. The next Jayers, corresponding lo metrizamide concentrations betweeni 221 and 198 g/litre, contain the intermiediate stages of maturicy of the parasite, from immature to mature trophozoites and immature schizonts. Cells collected at one of these intermediate fractions can be observed in Fig. 3B. In these layers, practically 100°o of the cells are parasitized. Cells collected at the tOp of the gradi- ent (d less than 1.095) contain a purified population of mature schizonis and merozoites (Fig 3C). The proportions of the different stages of parasite at 3 levels of the gradient are summarized in Table 1. To separate red cells infected with P. chabaudi AS or P. v. petteri (both mild infections) the conditions required were different. Appropriate gradients were formed by using metrizamide concentrations from 240 to 200 g/litre and blood cells obtned on the day of peak parasitaemia were used. ln Fig. 2B is a com- parison of tubes with gradients in which P. chabaudi- Fig. 2. Tubes showing continuous metTizamide gradients, after centnfugation, in which several layers of parasitized red cells can be observed. A P. yoelii-infected erythrocytes, separated by densities between 1 126 and 1.095, in which several bands can be ob- served (left). On the right, a comparable number of normal erythrocytes have sedimented to the bottom of the gradient. B. P. chabaud:-infected erythrocytes, separated bydensi- ties between 1 .132 and 1 119, showing several bands. Onthe left, a comparaole number of normal erythrocytes have sedi- mented to the bottom of the gradient. C. P vinckei penen-parasitized red cells, separated by densitiesbetween 1.132 and 1.092. On the left, 50ulof eryth- rocytes of an infected mouse have been separated into several layers. On the rigiit, a comparable number of normal red cells have sedimented to the bottom 183 I I 4 I I I I Iy I A I E. M. EUGUI & A. C. ALLISON Table 1. Differential counts of P. yoeli-parasitized cells, following density-gradient centrifugation in metrizamidea Differential counts 1%) Relative Fraon densty Unnfncted Ring Tropho- cells forms zoites Schizonts Bottom 1 126 71 4 20.3 8 3 0 Intermediate 1 109 1.3 65 843 79 Top <1 095 0 3 0 14.7 80O * Values are averages of two separate expenMents, in each of which a mean of 500 paratized red cells was calculated Table 2. Differential counts of P. chabaudi-parasitized red blood cells, following density-gradient centrifugation in metrizamidea Differential counts 1%) Relative Fracton denshy Uninfected Ring Tropho- ceia Faorns loiltes Schizonts Bottom 1 132 80 0 9.5 70 35 Iniernediate 1 126 388 4.6 420 146 Top 1 119 40 0 110 85.0 a Values are mean counts of 500 erythrocytes I *.O' .1.I .I. I I . ,. I.w nr 0 _ . 4 Fig. 3. Photomicrographs of P. yoe/ii-infected cells recovered from gradients, washed, deposited in a cytocentrifuge and stained with Giemsa stain. A. Lowest fraction showing uninfected erythrocytes and early developmental forms of the parasite. B. Intermediate fraction showing trophozoites of varying stages of maturity. C. Uppermost fraction showing schizonts. parasitized and normal cells have been centrifuged. Here again, several layers can be collected and cells in the bottomn fraction are mostly normal red cells with some ring forms. The fraction at the top has a con- siderable enrichment of mature schizonts, similar to that described for P. yoelii. The proportions of the different stages of parasite development in some frac- tions of the gradient are compared in Table 2. Blood of mice infected with P. v.petteri was used at different periods after inoculation, on the day of peak 184 SEPARATION OF DEVELOPMENTAL STAGES parasitaemia (day 9) as well as during recovery (day 12). The profile of the gradient obtained, using metri- zamide solutions of the same density as for P. cha- baudi, can be observed in Fig. 2C. The macroscopic appearance of the tubes was similar for both of the periods tested. However, when microscopic slides were observed, one interesting difference in the cells collected at the top of the gradient was apparent (Fig. 4A-C and Table 3). Blood collected at the peak day of infection shows a rather uniform composition in the uppermost fraction, with an average of 820/o of mature schizonts, a figure similar to that observed with P. chabaudi. However, when blood had been collected from recovering animals, fractions.at the top have a more heterogeneous composition, with a variable proportion of unparasitized cells. as well as other cells containing shrunken and degenerating parasites. Table 3. Differential counts of P. v. peteri-infected cells, ob- tained at the time of peak parasitaemia or during recovery, following centrifugation in metrizamide density gradients Differential cour.ts l5b% Relative Fraction density Unin- fected Ring Tropho Schiz- Degen- cells formLs zoites onts erates Peak parssta crn a Bottom 1132 864 123 14 0 0 lixernted:ate 1 116 596 10 7 297 0 0 Top 1 092 2 0 0 6 I5 6 81.8 0 Duirng recovery b Bottom 1.132 84 5 13 2 4.1 0 0 Intemiediate 1.116 77 0 12 2 7 9 0 5 2 4 Top 1.092 42 5 0 15 0 27 0 15 5 a Values are mean counts ot 00 cells b Values are averages of two separate expenments, in each of which a mean o 500 red cells was calculated. Jz I Fig. 4. Photornicrographs of P vincket petteri-infected cells recovered from gradients, washed, deposited in a cytocentri- fuge, and stained with Giemsa stain. The fractions were obtained at the time of recovery from #he infection A. Lowest fraction, containing uninfected erythrocytes and early forms of the parasite B. Intermediate layer with trophozoites at varous stages of maturrty. C. Uppermost fraction containing schizonts as well as cells with degenerate forms of the parasite and occasional unin- fected cells of unusually low density 185 E. M EUGUI & A C. ALLISON Viability ofparasites after gradient separation Erythrocytes parasitized with P. yoeldi were ob- tained directly from eluate fractions. The cells were washed once with buffer containing FCS (20 g/litre) to eliminate metrizamide. Serial 10-fold dilutions of three different fractions (top, initermediate, and bottom), beginning with the I x 101 infecting dose, were inoculated into 3 BALB/c mice per group. Sim- ultaneous controls received the same dilution of para- sitized erythrocytes from the original blood. At the highest dilution used (J04), both in test and control animals, parasites appear in peripheral blood one day later than in animals receiving higher concentrations. It was evident that parasites at early stages of develop- ment, collected from the bottom fraction, induced a parasitaemia that rose sooner and at a faster rate than in the controls. Also, these animals died one day before the others. In the other groups, no differences were observed compared with the controls; animals died between days 9 and II, according to the dose. These observations show that viability of the parasites is not altered by metrizamide separation. Permeability oferythrocyte plasma mnembranes Several studies (12-13) have shown an increased influx of glycine, D-glucose and other compounds into erythrocytes of mice infected with Plasmodium or Babesa parasites. Since this is an important meta- bolic change and is reversed by products of immune spleen cells (13), it would be of interest to relate it to different stages of parasite development. Preliminary observations carried out in collaboration with B. Elford and J. Christensen shlow that the influx of glycine into parasitized erythrocytes is not affected by centrifugation in metrizamide gradients. DISCUSSION Metrizarnide density gradients could be useful for the following purposes: (I) preparation of schizont-specific and other stage-specific antigens; (2) concentration of late merogonic stages for short-term cultures to allow liberation of merozoites; (3) biochemical analysis of parasitized and unpara- sitized erythrocytes from infected animals or erythro- cyte cultures, to ascertain in what way the properties of each are modified; (4) separation of sporozoites from accompanying materials, since metrizamide gradients are isotonic and not harmful to parasites. The particular advantages of metrizamide gradients are the following: (I) they are easy to prepare; (2) solutioins are stable and can be stored indefinitely at -20 °C; (3) it is possible to obtain solutions of the re- quired density to separate parasitized and unparasit- ized erythrocytes without hypertonicity or high vis- cosity; (4) there is no detectable effect on parasite metabolism or viability; and (5) there is no detectable effect on erythrocyte membrane permeability. The observations show that each species of Plas- modium changes the density of infected erythrocytes in a characteristic way. Hence, different gradients are required for each parasite. Once a system has been devised for any species ofPlasmodium, it is highly rc- producible when using preparations from different mice during the period of rising or peak parasitaemia. However, when the animals are recovering from P. chabaudi or P. v. petteri infections, the fraction with lowest density contains not only schizonts but also cells with degenerate parasites and others with no detectable infection. Since normal erythrocytes are not found in this fraction, the apparently uninfected erythrocytes may be cells from which degenerate parasites have been extruded, or the density of which has been altered in some other way during the infec- tion. The densest fractions contain early ring forms as well as normal erythrocytes and are suitable for estab- lishing short-term synchronized cultures. P.fali- parum cultures (14) tend to become asynchronous, and metrizamnide gradients may be suitable for the preparation of developmental stages. In contrast to the murine malaria infections here described, P. knowlesi infections in primates are synchronous, and then other methods are suitable for the prepar- ation of parasite antigens (15). However, even in these infections it would be convenient for some pur- poses to be able to separate infected from uninfected erythrocytes, and metrizamiide gradients may have some advantages over previously used methods. ACKNOWLEDGEMENTS This investigation received financial support from the UNDP/JWorld Bank/WHO Special Programme for Research and Training in Tropical Diseases. The authors would like to thank Nyegaard & Co. for providing samples of metnzamide. 186 SEPARAUTION OF DEVELOPMENTAL STAGES 187 RESUM2 EMPLO1 DE GRADIENTS DE DENSITt A BASE DE METRIZAMIDE POUR LA SEPARATION DES DIVERS STADES INTRACELLULAIRES DES PARASITES DU PALUDISME Des gradients continus de densite a base de m6trizamide, un trijodo-benzamido d&riv& du glucose, ont e employEs pour la separation isopycnique des erythrocytes selon Js stades de parasites muunns qu'ils contiennent. Les gradLents A la metrizamide sont stables et faciles A preparer et on peut obtenir les densit&s requises sans constater d'hypertonicite ou de forte viscosit:. Les pr6parations, fortement enrichies, de chacun des stades de parasites qui ont 6rc uti1is6es provenatent d'animaux chez qui l'infection est peu syn- chrone. La m6trizamide ne semble pas avoir eu d'effet notable sur la viabilitF et le m6mabolhsme des parasites, et la perm&abi1it6 des &rythrocyres infcct6s a egaIement 6tt p;&- servee. Les gradlents a Ia mntnzamide pourraient Etrc utilement employ6s pour les etudes blochimiques et immunologiques des parasites du paludisme, concernant notamment: a) la prtparation d'anfigEncs sp6cifiques du schizonte et d'autres stades; b) la concentration de stades m6rogoniques avances en vue de la liberation de m0rozoites en cultures de courte dur6e; c) 1'analyse biocliinique des 6rythrocytes parasit6s et non parasites provenant d'animaux infectes ou de cultures d'erythrocyies pour d6terminer en quoi leurs caracttris- tiques sont rmodifi6es; d) Ia separation des sporozortes des aulres materLaux prTsen:s. REFERENCES 1. WILLIAMSON, 3. & COVER, 0. Transactions of tre Royal Society of Tropical Medicine and Hygiene. 60: 425-427 (1966). 2. ROWLEY, P. T. ETAL. Journal of laboratory ..rnd clinical mnedicine, 70: 933-937 (1967). 3. MILLER. L. H. & CHIEN, S. EVpertrtnentalparastiology, 29:451-456 (1971). 4. LUNDE, M. M. & POMWERS, K. S. Annals of tropical medicine and parasitology, 70: 283-291 (1976). 5. ELING, W. Bulletin of the World Hfealth Organizan,ron, 55: 105-114(1977). 6. MCALLISTER, R. 0. & GORDON, D. M. Journal of parasitology, 62: 664-669 (1976). 7. PASVOL, G. ET AL. Annals of tropical medicine and parasitology, 72: 87-88 (1978). 8. KREIER, J. P. Bulletin of the Wordd Health Organiz- ation, 55: 317-331 (1977). 9. RicxNAO0D, D. & BuIR1E, C. D. FEBS Letters, 50: 102-110 (1975). 10. MUNT,HEF;-KAS, A. C & SEGLEN, P. 0 FEBS Letters, 43: 252-256 1974). 11. CARTER, R. & DiGGS, C. L. Piasmodia of rodents. in: Kreier, J. P., ed. Parasitic protozoa, New York, Academic Press, :vol. 3, 1977, pp 359-466. 12. NEAMc, K. D.& HOMIEWOOD. C. A. Internationaljournal for parasitology, 5: 537-540 (1975). 13. ELFORD, B. & CHRISTENSEN. G. Submitted for publi- cation. 14. TtAGER. W. & JENSEN, J. B. Bulletan of the World Health Organization, 55. 363-365 (1977). 15. MITCHELL, G. H. ET AL. International journal for parasitology, 3: 443-445 (1973).