Recent developments in the assessment of the immune response to malaria, especially as related to vaccination: Malaria infections in different strains of mice and their correlation with natural killer activity*

Bulletin of the World Health Organization, 57 (Suppi 1) 23l-23B (I 979) Malaria infections in different strains of mice and their correlation with natural killer activity* ELSIE M. EuGuiI1 & ANTHONY C. ALLISON ^ Different mouse strains shoiK inarked variation ini suiscepribility to murine malaria parasites. Of those studied, strain A mice showed the highest suisceptibility, most animals dying from Plasmodium chabaudi infections. In contrast, C57BI and CBA mice all re- covered from P. chabaudi infectionrs and were then resistanr to challenge with a virulent strain of P. yoelii. BALB/c mice recovered front P. chabaudi infections but were not uniformly resistant to P. voelii challenge. BIO.A mice, which share haplorypes (H-2IJ2) ivith strain A nzice, were resistant to P. chabaudi infections. Of all the correlates of resistance, one appears to be of special interest: natural killer (1NKi activity. This wvas high in C57BI mice and in malaria-infected mice considerable increases in spleern cell numbers and in NK activity were observed. In contrast, strain A mice shfowed lowNK activity and in mnalaria- infected animals spleen cell numbers andNK activity showed little or no increase. It is sug- gested thatNK cells, recruited and activated by Tlymphocyte-mnediated immnune responses to parasite antigens, recruit and activate NK cells and macrophages, and thlat a product of the reacting cells inay produce death ofparasites wirhin circulating erythrocytes. During the course of investigations aimed at immu- nization of mice against malaria, we have observed marked differences between strains of mice infected with various murine malaria parasites. These extend previous observations on the subject reviewed by Miller & Carter (1). For example, Nadel et al. (2) found that the survival time of mice infected with Plasmodium berghei varied, although the infection was uniformly lethal. Mice of the C57BI strain were the most resistant while mice of the strain DBA were the most susceptible. A comparable grading of re- sponses of mice has been found in our experiments. As detailed below, mice of the A strain have a very high parasitaemia when infected with P. chabaudi line AS and nearly always succumb to the infection. In contrast, mice of the C57BI strain soon recover from the infection with P. chabaudi and prove uniformly resistant to challenge with the virulent P. yoehii YM strain used in our laboratory. BALB/c mice are inter- mediate: they all recover from infection with P. chabaudi but when challenged with P. yoelii not all * From the Cell Pathology Division, LMRC Clinical Research Centre, Watford Road, Harrow, HAIl 3UJ, England This investi- gaton received financial xupport from the UNDP/World Bank/ WHO Special Programme for Research and Training in Tropical DLseases. Research Fellow. Head of Division. Present address: Director General. Inter- nabonal Laboratory for Research on Animal Diseases, P.O. Box 30709, Nairobi, Kenya. This is the address to which requests for re- prints should be sent. are resistant. Mice of the A strain also fail to recover from Babesia microi infections; they develop a patent parasitaemia that persists indefinitely.a Major histocompatibility determinants do not seem to be responsible for the variable susceptibility, since mice of the strain BlO.A recover from P. chabaudi and B. microti infections. The persistent B. microti and lethal P. chabaudi in- fections in strain A mice resemble the Babesia and malaria infections previously observed in this labora- tory in nu/nu mice which have a congenital deficiency of mature T lymphocytes (3). Since in mice of strain A the numbers and functions of T lymphocytes are normal, the reason for their susceptibility to baemo- protozoan infections must be different. A property of A strain mice that has emerged recently is the low capacity of their spleen cells to kill sensitive target cells (natural killer or NK activity) as compared with other strains (4, 5). If NK activity is important in resistance to haemoprotozoa it might be expected that this activity would increase about the time when the host is recovering from infection. We have found that NK activity, expressed as a constant ratio of spleen cells to target cells, is considerably increased at the period of peak parasitaemia. SinCe at this time the total number of spleen cells is also considerably in- creased, Ithe overall NK activity is much greater than in nor-mal animals. a E. MI. Eigui & A C. Allison, unpublished results. 3893 -231- E. M. EUGUJ & A. C. ALLISON MATERIALS AND METHODS Animals C57BV/I0, CBA, and BALB/c mice were supplied by the animal breeding unit of the Clinical Research Centre. BlO.A and A strains were obtained from OLAC 1976 Ltd., Shaw's Farm, Blackthorn, Bicester, Oxfordshire, England. Unless otherwise stated, animals were 6-8 weeks old at the time the experi- ments were initiated, and were sex-matched as indi- cated. The drinking water of inoculated animals con- tained 4-aminobenzoic acid in a concentration of 0.5 g/litre. Parasites P.yoelil line YM and P. chabaudi line AS, kindly provided by Dr D. Walliker, Institute of Animal Genetics, Edinburgh, Scotland, were maintained by blood-passage in C57B1 female mice when the parasit- aemia was rising. A standard dose of 1 x 10' parasit- ized red blood cells, given by intraperitoneal inocu- lation, was always used. The course of the infection was monitored, at first daily, when the parasitaemia was rising, and then every second day during recovery, by examination of Giemsa-stained smears of tail blood. Infections were expressed as the percentage of 200 erythrocytes parasitized. Spleen cellpreparation After removal, spleens were weighed and immedi- ately placed in Eagle's Minimal Essential Medium (EMEM), supplemented with antibiotics. Spleens were disrupted gently and cell suspensions collected after any debris had sedimented. The total number of nucleated cells per spleen was then determined by haemocytometer counting. Erythrocytes were lysed by treatment for 10 min at 37 IC with Tris-buffered armmonium chloride, pH 7.2. After washing, ceLs were resuspended in RPMI 1640 culture medium, con- taining 10% heat-inactivated fetal calf serum (FCS), penicillin (100 lU/mil) and streptomycin (50,g/ml). This complete medium was used for incubation in all the cytotoxicity experiments and also to maintain the YAC-1 tumour cell line. Viability of cells, calculated by trypan blue dye exclusion, always ranged between 75 and 85%o, and concentrationls were adjusted as required. Tumour cells The YAC-1 cell line was established in vitro from a Moloney virus-induced Iymphoma in A/Sn mice (6). Cytotoxicity assay NK activity was evaluated in a 1Cr release assay, using YAC-1 tumour cells as target, based on methods described by Kiessling et al. (7) with some modifi- cations. Target cell suspensions containing 5 x 106 to 10 x l06 cells were labelled byincubation for 30minat 37°C in 0.5 ml of RPMI 1640 with 2%o FCS and 3.7 x 105 Bq (100 QCi) of "I Cr as sodium chromate. The cells were washed twice and left in suspension for 30 min at room temperature before the third wash. Cells were fmally suspended in complete medium, counted, and adjusted to the desired concentration. Lymphocytes used as effector cells were obtained from spleens as described. Macrocytotoxicity tests were performed in triplicate in plastic tubes. Different proportions of effector cells were added to a standard number of 4 x 104 labelled target cells to a total vol- ume of 1 ml. The ratios of effector to target cells nor- mally used were 100:1 and 50:1, respectively. Appro- pnate controls, to which no lymphocytes were added, were included in order to estimate spontaneous release. This ranged between 10 and 25/o under the test conditions. Tubes were incubated for 8 h at 37 °C in air with 5%e CO2 . Maximum release was determined by treating the target cells with detergent, this releasing about 80%o of the total "I Cr (measured on counts per min, cpm). Results were expressed as percentage of specific lysis, according to the formula: ?70 lysis = test cpm - spontaneous cpm maximum release cpm - spontaneous cpm xIOO RESULTS The course of P. chabaudiAS infection in different strains of mice Groups of 5-6 mice of different strains to be com- pared were injected simultaneously with I x 106 para- sitized erythrocytes and the course of the infection followed according to the description presented under materials and methods. To assess whether genes map- ping within the H-2 complex are involved in the response to murine malaria parasites, strains with different H-2 haplotypes, as well as some other strains having thesame H-2 complex, were studied. In Table I the general susceptibilities of different mouse strains to P. chabaudi are compared. This parasite produces transient infection in most of the strains, but is lethal for A strain mice. As shown in Fig. 1 (A and B) the parasitaemia rose earlier and reached higher levels in strain A mice than in any of the other strains so far studied, and in strain A mice 10Oo of males and 80"lo of females succumbed to the infection (Fig. 2, A and B). A similar degree of resistance to that of C57BI and BALB/c mice was observed in the BlO.A and CBA strains. Since mice of strains A and BlO.A have the same H-2a haplotype but an entirely different sus- 232 INFECTIONS IN MICE AND NATURAL KILLER ACTIVITY ceptibility to P. chabaudi. these determinants do not seem to have a major effect on susceptibility. Table 1. Differences in the susceptibility of various mouse strains to haemopfotozoan infections Permstent Lethal B m,crot: P chabafdo Mouse strain Haploype infection irfection A,iHeCrc h-2a + - B10.A,'Ola F1. 2a - - CBA!CaCrc H-2k - - C57BI/1OScSACrc H-2b - - BALBkcCrc H -2d - - 100- s0- 60- L- 20 0 P cliatiaudi inc eALBic aA 0 2 4 5 8 10 1 - lo Days alter anfecoLDn 100-I 80- 70- 60- 50- 9-, 40 .c 30- 10- 20- la L. P.chabaud, it 9 m oA ,L 40 20- 0- @ X ~~~~~~~~~~PCh?Ntudi in 9 oA I I I tI~~~~ho c I I I I 0 2 4 a i 10 Days after infKctDn 4 6 8 1' 12 Days after inlEction 14 16 18 1i il. 16 Fig. 2. Percentage of survivors in groups of 5 mice of dif- ferent strains, after inoculation with 1Or erythrocytes para- sitized with P. chabaudiAS. (A) BALB/c males all recovered, compared to A strain all dying, and (B) C57BI and BALB/c females recovered, compared to A mice of which 80% die. 70- 6-5\ P chabaudi inm ~~oA 60 - BAL61c 50 40- 20- ,'r- it' 4 6 8 10 :2 14 It 13 Days alter .nfec't3n Fig 1. Course of parasitaemia in IA) strain C57B131O and strain A females and (B) strain BALB/c and strain A males, following intraperitoneal inoculation with 106 erythrocytes parasitized with P. chabaudi AS. In each groLp, 5 animals 6-12 weeks old were used. Cross immunity between P. chabaudi and P.yoelii in different inouse strains P. yoelii YM is a virulent strain that kills mice of all the strains studied in about 7-8 days at the dose used. However, animals that haye recovered from a pre- vious infection with P. chabaudi are often resistant to Table 2. Response to P. yoehi YM in mice t-vo months after recovery from P. cheaaaudi AS Mouse strair. Recovc ed animalsa Controla C-57Bei'ScSACrc 11;1 0,6 CBA. CaCrc 8;8 0/6 BA LB/i;Crc G6;10 017 Numerator number o'recovered ani.ma s; denominator total number at anirnals 233 E. M. EUGUI & A C ALLISON P ylisinl9QC5?BL t t6 * 2 mor o fram _ | ~~~oConlr iths alter P chabaud rol I O 4 6 8 10 Days after infeclion P yol inQCBA t6 * 2 months af tror. P chab o Control _ 8 I I 1 4 6 8 10 Days alter inlection challenge with the virulent P.yoelii. Further exper- iments were performed to ascertain whether strains of mice having different haplotypes develop a similar degree of immunity. At suitable intervals after P. cha- baudi parasites had been cleared from peripheral blood (21 days and 2 months), animals from different recovering strains were challenged with I x 10. P.yoelll-para- sitized erythrocytes. As shown in Table 2, at the longest period after recovery that we studied, all the C57B1 and CBA mice recovered from the challenge, but 40%70 of the BALB/c mice succumbed. In Fig. 3 the course of parasitaemia in these strains of mice 2 months after recovery is shown; C57BI and CBA mice 12 14 had a very low and transient parasitaemia, while in BALB/c mice parasitaemia was higher and some of the animals died. In general, these experiments reveal a wide range in susceptibility to murine malaria para- sites in different strains of mice, since BALB/c are more resistant to P. chabaudi than strain A, but at the same time they are more susceptible to P. yoelii than some other strains. Iter recovering baudi 1I P yoelinIn BALB#c L-j7 ©, * 2 months after recoverina Iroe P. chabatd o Control 6/10 Days alter Iniectlion Fig. 3. Parasitaemia in female mice challenged with P. yoe(ii YM 2 months after recovery from P. chabaudi and in non- pretreated controls. Numbers indicate animals used in each group; all controls died and all pretreated C57BI (A) and CBA (B) strains recovered. In BALB/c mice IC) only 6/10 re- covered. Changes in the spleen at different intervals qfter infection Because the spleen has a primary role in haemo- protozoan infections, as discussed elsewhere (8), we have measured the weights of the spleens and esti- mated the total number of nucleated cells per spleen of individual animals, as well as the NK activity. A significant increase in the spleen weight and total cell number was observed in C57B1 mice during the time of rising parasitaemnia, with a still greater increase at the time of recovery (Fig. 4). Nine days after infec- tion, while the parasitaemia was still rising, a rise of about 3- to 4-fold in the total cell number was ob- served. When the parasitaemia was falling, the in- crease in cell number was 5- to 6-fold, and by day 12, when the parasitaemia was nearly cleared, this rose to about 10 times the average in age-matched uninfected controls (Fig. 4). The increase in spleen weight goes in paraUlel with the increase in cell number (Table 3). However, at the same period after infection (9 days), strain A mice showed very little increase in spleen weight or total cell number, even when the parasit- aemia was at its peak and they were close to dying (Fig. 4-C and Table 4). Thus, in strain A mice the spleen cell response to haemoprotozoan infections differs markedly from that in other strains. Vatural killer activity in the spleens ofmice infected with malaria One of the characteristics of strain A mice that has become apparent during the last few years is the very low level cf NK activity in their spleens as compared with other strains. For this reason, among others, we 50 - 40 - c 30- x L 2 - 10 - 60 - 50 - 40- 30 - 20- I. E x CL 10 - s 5, 80 - 70 - 60 - 50 - 40 - 30 - 20 - 10 - 0 :-: - I u ~- . 234 I a INFECTIONS IN MICE AND NATURAL KILLER ACTIvTr Table 3. NK activity and some other parameters measured in female C5781/10 mice infected with P chabaudia Average Nucleated Group Days para- Spleen cells Specific and e!tter srtaemiab weight per spleen lysis c number infechon (%) (mg) x 10' 1%i Uninlected (7) - - 90.5 a 1 14 3 (±13 81 (=0.71 (=6 01 Rising parsa 9 27 5 518 0 18 1 32 3 sitaemnia (2) 1±.52 3) (=7 71 <=0 91 Falling para- 9 175 5385 267 281 sitaernia (2) .±50 21 =- 7it l-=' 0) Parasnaemia 12 1 0 851 0 37 6 35 0 nearly cleared (3) .67 71 1=15 71 1=1 53 8 Tnese results were obla.red in two separate exoienmerts 6 in penpheral blood c Effector to target cell ratio is 1 0 1 have measured NK activity in the spleens of mouse strains susceptible and resistant to malaria parasites. Fig. 5 shows the NK activity in infected C57B1 mice tested individually on day 9 (in some the parasitaemia was rising, while in others it was beginning to fa]l) and on day 12 (when the parasitaemia had been practically Table 4. NK activity and some other parameters measured in female A strain mice Infected with P. chabaudi nine days before Average Nucleated Group pars- Splsen cells Specific and sitaemia weight per spleen lysisb number 4%) 4mg) x 10 %) Unrifected 131 - 83 3 12.3 8.2(±15 5) (=1 41 1+1 61 Ristn.q parasilemi. 44) 64 7 174 5 11 0 12 9(.19 21 (.2.0) L± 4) a In peripheral nlood b Tns effector to target cal ratio wiVs 1C 1 cleared). The increased NK activity observed becomes still more signlficant when it is related to the total number of nucleated cells per spleen. It is also im- portant to recall that by the time ofpeak parasitaemia, and to an even greater extent during recovery, the normal ratio between ruclcated cell types in the spleen is completely altered. Because of the compensatory erythropoiesis in the recovery period from 'murine malaria infections, normoblasts predominate among nucleated cells (9). As effector cells. we used the same 9) C5781/10 60-, I9 C57B1I10 40 - __ 30- K C> x *= 20- c la- 10- (4) N (2) -L7 (2) I J 111"' 50.- N 40-Ci x L- = p30 -- c2 'J 2C- lOu j3 o0- [31 iM Paras;'aerria (3) 30 - ci x 20- C- 10-0 ' 0 - 1% Parasitaerri2 N 64.7% Parasltaeinia Fig. 4. Total nucleated cell number per spleen at different intervais after noculation .NwiTh P. cnabaudi AS: (A) 9 days after inoculation with parasitaemia rising orfalling in C57B1 females; (B) the same group of enimals 12 days after inoculation when parasitaemia is nearly cleared; (C female A strain, 9 days after similar inocuiation. N = controls. In brackets, the nuimber of animals per group. )'A' © 14)(3) I 235 C E. M. EUGUI & A. C. ALLISON effector: target 100:1 9 C57BII10 effector target 50:1 40 - (2) .J_ 30 - 20 - 0- N 27.51'4 17.5%$ Parasitaemia Fig. 5. NK activity against YAC-1 cells in the spleen of C57BI females at different intervals after P. chabeudi AS inoculation: (A) at 9 days, when some animals have rising parasrtaemia and others falling; (8) at 12 days when the parasitaemia is nearly cleared. N = controls. In brackets, the number of animals individually tested. Results are from two separate expeiiments. proportion of nucleated cells in normal animals as in parasitized ones. The ratio of lymphocytes to normoblasts is 70:20 in normal spleen, but about 20:70 at peak parasitaemia and at the beginning of the recovery penod (9). Since the number of lymphoid cells in our experiments was diluted some sevenfold byerythrocyte precursors, NK activity must havebeen tremendously increased in these animals. On the other hand, 9 days after infection, when the parasitaemia was at its highest and the animals were moribund, strain A mice had very low NK activity, practically identical with that of controls (Fig. 6). The difference between A and C57B1 mice is greater when it is con- sidered that the total splenic cell number and the spleen weight remained unaltered in the A strain. A general comparison of different measurements in A mice is given in Table 4. DISCUSSION The observations reported here and unpublished results show that mice of strain A seldom recover from malaria and Babesia infections. In contrast, mice of strain BO.A, which have the same hap]otype, 40- 30 'I 0 m- 0 10 Q(Al efftor targel 100 1 t4 13) T u N 64.7% Parusilmemis Fig. 6. NK activity against YAC-1 cells in the spleen of A strain females, 9 days after inoculation with P. chabaudiAS, when animals have peak parasitaemia and are close to dying. N = controls. In brackets, the numberof animals individually tested. 12) (3) I.. 1: :. 9 40 30 '- %f, '4 E- 20 US G - 10 0 C57BI110 13)El (3) N 1% Parasitaemia 236 INFECrIONS IN MCIE AND NATURAL KILLER ACTIVITY recover from both infections. At the other extreme, C57BlI 10 and CBA mice readily recover from P. cha- baudi infections and are then uniformly resistant to challenge with the virulent P.yoelii YM strain. BALB/c mice that have recovered from P. chabaudi infections are sometimes, but not always, resistant to challenge with P.yoeiii. Thus, there appears to be a spectrum of resistance to haemoprotozoan infections, from the highly susceptible strain A, through the intermediate BALB/c to the highly resistant C57Bl and CBA strains of mice. Among the various possible correlates of resistance, our attention has been focused on NK activity. This is known to be high in C57BI and CBA, intermediate in other strains of mice, and low in A mice, as reported by Kiessling et al. (7) and confirmed in our laboratory. IfNK activity is important in recovery from malaria infections, it would be expected to rise during the period of peak parasitaeniia. We have demonstrated that not only is the total number of cells in the spleen of C57B1 mice considerably increased at this time, but the NK activity per nucleated spleen cell, assayed against a standard target cell, is significantly augmented. Hence, total NK activity in the spleen is much higher than in uninfected animals. If this activity is con- centrated in certain areas, such as the splenic sinusoids or red pulp, it could be very effective in acting on parasitized erythrocytes. which pass slowly through these regions. In contrast, mice of the strain A at the period of high P. chabaudi parasitaemia show no sig- nifi'cant increase of the spleen weight or cell number and NK activity is only slightly augmented. It is of interest that intravenous or intraperitoneal injections of a BCG strain of Mycobacterium bovis or killed Corynebacterium parvum, which increase the resist- ance of mice to haemoprotozoan infections (10, 11), considerably augment the NK activity of spleen cells (12, 13). In contrast, corticosteroids, which increase susceptibility to haemoprotozoan infections and eliminate resistance produced by BCG (II), depress NK activity (14). Further studies are in progress using Fl and back-cross animals to ascertain whether NK activity and capacity to resist haemoprotozoan infec- tions segregate together or independently. The mechanism by which NK cells might produce degeneration of parasites within circulating erythro- cytes remains speculative. Immune spleen ceIls pro- duce factors that inhibit transport across the mem- branes ofBabesia-infected erythrocytes (8). Possibly, NK cells are the effectors in this reaction. In general, NK cells may represent one of the major effector mechanisms by which nonspecific imunnity is achieved. Their activity can be increased by specific immune responses, for example liberation of inter- feron by sensitized T lymphocytes or macrophages (8). Activation ofNK cells by interferon has been des- cribed byGidlund et al. (15). On the basis of these ob- servations a working hypothesis can be suggested (Fig. 7). Recovery from malaria infections depends on SPECIFIC IMMUNITY NONSPECIF1C IMMUIJITY Pargawl £nt'eu DCG. O)4.?fIa t.rflUI n P'.M.,Br',c'&. 4woin.. AntigrmcSFmula7ola tvypetion mI T ynzphqto mProdurn ; ofactlaiadparaspnOtesf.tl ytes: mCe dervd fmtreTlmpoye cannot reec(1n6er) t-enflecthval Psn Ic-orst Hrtper \ literidronl |Skiinu'ation marked T cell stimulation ( T lypctre (Sporandingoto antgenFiecwssnhibitse nqPrraeterdfeonpma weccll as producsta rcui n activt marpae i5stlred 2rytlvracvtes F(g. 7Postu1latedMn eraction of several celi typesin Im- rnunity to murine malarial parasites. a specific imaune response medrated by T sympho- cytes: mce deprived of mzaEure T lymphocytes cannot recover (1, 16). Mice infected with P.yoeiii show marked T cell stimuladon(p7). T lymphocytes res- Bondi to antigenic stimulation secrete interferon as well as products that recruit and activate macrophages (18, 19). Macrophages cadt be activated in} an im- munologically specific manner by products of T lymphocytes or nonspecifilcally by BCG, C parvum, Brucella abortus, and other agents (8). Among the products liberated by activated macrophages is inter- feron, which increases NK activity. It has long been known that mice infected with P. berghei have inter- feron in their plasma (20). Macrophages, NK cells, or cells ofsome other type secrete a factor or factors that inhibit the development of parasites in circulating erythrocytes, producing the degenerate intracellular forms that we consistently obseive in mice recovering from malaria and Babesia infections. The mechanism by which such factors exert their inhibitory effects is still a matter of speculation. However, their activity may be related to inhibition of transport systems in the membrane of the infected red cells, as recently dis- cussed for Babesia (8). ACKNOWLEDGEMENTS We are indebted to Miss A. Temple for expert advice on how to conduCt the cytotoxicity tests. 237 238 E. M. EUGUI & A. C ALLISON RtSUM1t LES INFECIlONS PALUDtENNES CHEZ DIFFARENTES SOUCHES DE SOURIS ET LEUR CORRfLATION AVEC L'ACTIVITt CYTOTOXIQUE NATURELLE Les travaux d6crits ont montrt que la sensibilit6 aux para- sites du paludisme infectant les murid6s itait trs vanlable selon la souche de souris. Chez les souris A, tous les mrnIes et la plupart des femelles ont succomb6 aprEs avoir 6tt infect6s avec une souche de Plasmodifum chabaud: d'une virulence relativement faible. Par contre, les souris des autres souches eprouv6es ont toutes gu&i et, lorsqu'elles ont ftt infectkes ult6rieurementavec une souche virulente deP.yoehl, on a pu constater qu'elles avaient acquis unc resistance contre cc parasite. Ce resultat a WtC constant pour les souris C57B1 et CBA. Chez les souris BALB/c toutefois, la rEsistance n'atait pas uniformae. Les divers facteurs pouvant etre associes A la rEsistance A ['infection palud&ennc ont W* analysCs. Les caract6ristiques g6n6tiques ne semblent pas exercer une in- fluence dtterminante puisque les souris B10.A, qui ont le meme g6notype H.2a que les souris A, ont toutes surmonte l'infection A P. chabaud,. Mais on a pu 6tablir une corr& lation entre la resistance au paludisme et I'activitt cyto- toxique naturelle (NK) des cellules spl&niques. Chez les souris C57BI r6sistantes, 1'activitt NK est normalement elevEe ec ellc augmente fortement, en meme temps que le nombre des cellules sp6eniques, lorsque ces souris sont infectees avec P. chabaud:. Chez les souris de la souche sensible A par contre, l'activit6 NK est normalement faible et, lorsqu'elles ont etW infectkes, le nombre des cellules spieniques et I'acti- vite NK ne s'accroissent que peu ou pas du tout. Lihypothese a ek 6mise que les cellules spl6ruques dot6os d'une activite cytotoxique et les macrophages pouvaient atre recrut6s et actives par la r6action immunitaire & m6diation cellulaire (lymphocytes T) contre les antigenes du parasite, et que ces populations de cellules lib&ent elles-memes des produIts capables de limiter le developpement des parasites dans les erythrocytes circulants. REFERENCES 1. MILLER, L. H. & CARTER, R. Experimental para- sitology, 40: 132-146 (1976). 2. NADEL, E. M. ET AL Genetics, 40: 620-626 (1953). 3. CLARK, 1. A. & ALLISON, A. C.Nature (London), 252: 328-329 (1974). 4. KIESSLING, R. ET AL. International journal of cancer, 15: 933-940 (1975). 5. HERBERMAN, R. B. ET AL. International journal of cancer, 16: 216-229 (1975). 6. CuCEs, M. ET AL. Journal of the Natronal Cancer Insti- tute, 50: 347-350 (1973). 7. KIESSLING, R. ET AL.Europeanjournalofimmunology, 5: 112-117 (1975). 8. ALLISON, A. C. ET AL In: Torrigiani, G., ed. The spleen and its role in parasitic infections, Basel, Karger, 1978. 9. FREEMAN, R. R. & PAR[SH, C. R. Immunology. 35: 479-484 (1978). 10. CLARK, I. A. ET AL Infection and immunity, 17: 430-438 (1977). 11. CLARK, 1. A. ET AL Parasitology, 74: 9-18 (1977). 12. HEREERMAN, R. B. ET AL International journal of cancer, 19: 555-564 (1977). 13. OJO, E. ETAl.. International journal of cancer, 21: 444-452 (1978). 14. HOCHM.AN, P. S. & CUDKOWICZ, 0. Journal of immu- nology, 119: 2013-2015 (1977). 15. GJDLUND, M. ET AL. NVature (London), 273: 759-761 (1978). 16. WEJNBAUM, F. l. ET AL. Journal of immunologv, 117: 1999-2005 (1976). 17. JAYAWARDENA, A. N. El AL Nature (London), 258: 149-151 (1975). 18. DAVLD, J. R. & DAVID, R. R. Progress in allergy, 16: 300449 (1972). 19. WYLER, D & CGALLEN, J. I. Journal of immunology. 118: 478.484 (1975). 20. HUANG, K. Y. Er AL. Science, 162: 123-124 (1968).