Search

Freshly isolated human peripheral blood monocytes inhibited the growth of blood-stage asexual Plasmodium falciparum parasites in vitro. The monocytes contained intracellular parasite pigment and a few whole parasites, but the remaining parasites reinvaded fresh red cells successfully and were morphologically normal. Anti-parasitic activity of these macrophages was not significantly enhanced by treatment with recombinant tumour necrosis factor alpha, recombinant gamma-interferon or lymphoblastoid alpha-interferon. Catalase had no effect on this parasite inhibition, suggesting a hydrogen peroxide independent mechanism. Anti-parasitic activity was, however, enhanced by prior maturation of the monocytes. Monocytes matured for 6 days caused 100% killing of parasites. In contrast to identical concentrations of freshly isolated monocytes the parasites incubated with these matured macrophages showed intraerythrocytic death similar to the crisis forms seen in vivo. gamma-interferon present either during the assay or as a pretreatment had no significant enhancing effect on the killing, although cytotoxicity to tumour cell lines was enhanced. Conditioned medium from macrophages showed only moderate parasite inhibition.

The susceptibility of the human malaria parasite, Plasmodium falciparum, to killing in vitro by macrophage secretory products was investigated. The effect of O2 radicals and tumor necrosis factor on parasite viability was assessed both morphologically and by following the uptake of [3H]hypoxanthine. H2O2 produced by the interaction of glucose and glucose oxidase was found to reduce viability; this effect was reversed by the addition of exogenous catalase. Further studies indicated that the catalase level within the erythrocyte was not altered upon parasite invasion. O2 radicals produced during the xanthine-xanthine oxidase interaction also killed P. falciparum. The addition of various O2 radical scavengers (including catalase) did not reverse this effect; therefore, it was not possible to determine which of the O2 radicals were involved in the killing process. Samples from three different sources containing tumor necrosis factor, a nonspecific soluble mediator derived from Mycobacterium bovis BCG-activated macrophages treated with endotoxin, also killed the parasite. There was no evidence that tumor necrosis factor or the products of the xanthine-xanthine oxidase interaction caused damage to the erythrocyte membrane that could be implicated as an important aspect of the killing process. These findings all strongly suggest that such macrophage products play an important role in immunity to malaria.

The presentation and management of 10 cases of proven or presumptive fungal endophthalmitis in narcotic-drug abusers is described. Miconazole was found to be an effective agent in some patients when administered in a dosage of 2400 mg/day. Eight patients received treatment with a combination of miconazole and flucytosine (5-fluorocytosine). Regression of the infection with preservation of the eye was observed in each case. However, visual acuity in the affected eye improved only in four of the eight patients; it was unchanged in two and had deteriorated in the other two. In two patients, who received amphotericin B and flucytosine as initial treatment, control of the infection was achieved, but vision remained unchanged. Vitrectomy was performed in three patients to remove residual sites of infection. However, vision remained unchanged in two of these patients and worsened in the third. The selection of individual modalities of therapy and responses to treatment are discussed.

We show here that purified lipoarabinomannan (LAM) from Mycobacterium tuberculosis can cause the release of tumour necrosis factor (TNF) in vitro from human blood monocytes and activated mouse peritoneal macrophages, and the production of TNF in vivo in mice pretreated with Propionibacterium acnes, with a potency comparable to that of lipopolysaccharide (LPS) from Gram negative bacteria. Like LPS, LAM binds to polymyxin B. We confirmed that its activity was distinct from any contaminating LPS and was associated with the antigenic activity by affinity chromatography, using a monoclonal antibody specific for LAM. Treatment with dilute alkali greatly diminished the TNF-inducing activity, suggesting that omicron-acyl groups may be involved. When LAM was fractionated by electrophoresis on SDS-Page and blotted on nitrocellulose, most TNF-inducing capacity coincided with the bulk of the LAM, as estimated by molecular weight and antigenic activity. This modification of the Western blotting technique may be generally useful for the study of macrophage-triggering molecules. The ability of LAM to cause the release of TNF may be responsible for some of the characteristics of tuberculosis, such as fever, weight loss, raised acute phase reactants and necrosis that can be mediated by this cytokine.

Experiments with malaria in mice suggest that protective immunity depends not only on antibody but also on activation of macrophages. Activated macrophages may cause intra-erythrocytic death of parasites by releasing reactive oxygen intermediates and/or tumour necrosis factor. Macrophage activation for both types of product correlates well with the timing of recovery in a range of different malaria infections and in mice protected by vaccination.

The multiplication of Plasmodium falciparum in culture, as measured by [3H]hypoxanthine incorporation, was inhibited in a dose-dependent manner by rabbit tumor necrosis serum. The regimen by which tumor necrosis serum is produced caused significant increases in the levels of triglycerides and lipid peroxides, with the latter being indicated by the level of malondialdehyde in the serum. When tumor necrosis serum was depleted of lipoproteins by aerosil (fumed silica), no parasiticidal activity remained, and when it was separated by ultracentrifugation, more than 70% of the parasiticidal activity was found in the lipoprotein fraction. This suggests that lipid peroxides may account for most of the parasiticidal activity in tumor necrosis serum but that a nonlipid toxic factor may also be present.

Mice infected with lethal P. berghei or P. yoelii malaria develop anti-lymphocyte autoantibodies less than a week after infection. The autoantibodies are IgM, T-dependent and cytotoxic to a subpopulation of non-T lymphocytes at 37 degrees C in the presence of mouse or guinea-pig complement. Non-lethal P. yoelii and P. chabaudi do not induce these autoantibodies. X-irradiated P. berghei parasites induce autoantibodies in normal mice but not in mice infected with non-lethal P. yoelii suggesting an active suppressor mechanism in the non-lethal infection.

Cats infected with Brugia pahangi were immunosuppressed and the development of lymphatic pathology, as measured xeroradiographically by dilatation of the vessels, in these animals was compared with that of normally infected cats. Individual cats showed wide variations in response. Niridazole and antilymphocyte serum both reduced dilatation, niridazole being particularly effective in the early stages of infection. When silica, used as an inflammatory agent, was injected subcutaneously in conjunction with worm infection lymphatic dilatation developed, this being potentiated in degree and rate of development as compared with control animals. The possible role of a non-filarial inflammatory response in the development of lymphatic dilatation is discussed.

The mouse helper T-cell response to four plasmodial and babesial parasites was measured by using them as carriers for a standard hapten (TNP). Helper T cells appeared to recognize all the parasites, but not to be able to distinguish between them. Helper T-cell responses could be augmented by vaccination with formalin-fixed parasites. However vaccination did not always confer protection against infection. Conversely, mice resistant to infection because of prior recovery from a homologous or heterologous infection had normal or reduced helper T-cell responses. It is concluded that resistance to infection with these parasites, though dependent on T cells, may not only involve the helper T-cell subpopulation.

A single dose of a formalin-fixed malaria vaccine which normally protects mice against challenge with the live parasite, is ineffective in mice born to immune mothers. This inhibition of protection, which is due to maternally-derived IgG, can be overcome if a second dose of vaccine is given 10 or more days after the first. We show that this is related to the production of specific IgM antibody in response to the first dose of vaccine, which competitively blocks the inhibitory effect of the IgG. The implications of this finding in relation to immune regulation and immunization regimes are discussed.

Mice were protected against lethal Plasmodium yoelii malaria by vaccination with a Triton X-100 lysate of whole parasitized erythrocytes. For full effectiveness this vaccine required an adjuvant, and we have found that recombinant gamma-interferon has strong adjuvanticity in this model when given either intraperitoneally or subcutaneously. Specific immune responses that were enhanced included antibody, T cell help, and delayed hypersensitivity.

We have compared interferon-gamma (IFN-gamma) with saponin and interleukin-1 (IL-1) as adjuvants for a blood-stage malaria vaccine in mice with various immunological abnormalities. IFN-gamma was particularly effective in Biozzi low antibody responder mice, mice selectively bred to produce antibody of low affinity, and mice depleted of CD4+ T cells. IFN-gamma and other cytokines may be safe adjuvants for use in human immunodeficiency states.

Strain differences in ease of induction of autoantibody production were observed when mice were injected with rat RBC. Responsiveness was not linked to the H-2 locus. NZB and (NZB X BALB/c)F1 mice were hyper-responsive both in terms of the induction of autoantibody and in the production of agglutinating antibody to rat RBC. C57BL/c and (C57BL X BALB/c)F1 were poor responders. Injection of the rat RBC in FCA converted a poor responder into a good responder. Adult thymectomy and ALS treatment did not significantly enhance autoantibody production.

A cosmid vector carrying the Epstein-Barr virus origin of replication, the EBNA-1 gene, the hygromycin phosphotransferase (hph) gene and pBR322 sequences has been constructed. This cosmid can replicate autonomously in the nucleus of human tissue culture cells, even when it carries a 35-kb long insert. The cosmid can be rescued from the transfected cells by cloning it directly into ampicillin-sensitive Escherichia coli. A gene for human tumour necrosis factor (TNF) cloned into this cosmid vector was introduced in tissue culture cells, where it was transcribed into mature mRNA.

Normal lymphocytes labelled with 51Cr were injected into mice at various stages of lethal and non-lethal malaria infections. Marked alterations were seen in the uptake into spleen and liver, which correlated with the outcome of the infection. Non-lethal infections and lethal infections in mice protected by vaccination caused increased uptakes, especially in the liver. In lethal infections, particularly Plasmodium berghei, uptakes were below normal values at certain times: this was apparently due to destruction of lymphocytes, probably caused by autoantibody.

Plasmodium berghei XAT, an attenuated variant of lethal P. berghei, causes a resolving infection in Balb/c mice from which they recover in about 3 weeks. The parasitaemia displays an early peak at about 5 days, followed by a steep drop in parasite number associated with the appearance of degenerating forms inside mature erythrocytes; the parasites remaining are inside reticulocytes. By contrast, no degenerating parasites were seen in infections caused by the virulent parent, which was mainly confined to mature erythrocytes. However, P. berghei XAT was no more sensitive to reactive O2 metabolites, generated by alloxan, or to tumour necrosis serum, than its virulent parent. Furthermore, its early drop in parasitaemia was unaffected by silica. The drop still occurred in the absence of T cells, although the infection was then ultimately lethal, and it was not mediated by NK cells since it occurred in nude mice treated with anti-asialo GM1 serum to abolish NK cell activity. However, it was absent in splenectomized mice, in which P. berghei XAT infection was lethal. Thus, the attenuation of P. berghei XAT infection is not due to increased susceptibility to some of the agents thought to cause parasite destruction, but to some other mechanism in which the spleen is involved.

Tumour necrosis serum (TNS), from animals primed with macrophage activating agents and challenged with endotoxin, causes necrosis of some tumours and can kill certain tumour lines in vitro and malarial parasites in vitro and in vivo. We have tested the possibility that the same factor is responsible for killing both tumour cells and malarial parasites. In competitive inhibition experiments, parasitized erythrocytes, but not normal erythrocytes, inhibited the cytotoxicity of TNS against a tumour cell line. Conversely, the tumour cells inhibited the killing of Plasmodium yoelii in vitro by TNS. When rabbit TNS was fractionated by ion exchange chromatography followed by gel filtration and the fractions at each step were pooled according to their ability to kill the tumour cells, in vitro parasite killing activity was found to correlate with tumour cell cytotoxicity, to a final sample which was purified more than 600-fold. Our results suggest that in terms of function, at least, the same component of TNS is responsible for the killing of both tumour cells and malarial parasites.

Mice vaccinated with a formalin-fixed preparation of either Plasmodium berghei or P. yoelli exhibited delayed type hypersensitivity (DTH) to the homologous antigen. This manifested itself in increased delayed thickening of antigen-challenged pinnae of the vaccinated mice as compared to the non-vaccinated controls. DTH was also evident in the vaccinated mice using the homing of radio-labelled bone marrow cells (BMC) to the delayed lesion as a criterion of reactivity. When P. yoelii vaccinated mice were given a live infection P. yoelii, a marked migration of BMC into the spleen occurred, with a peak at 48 hr, and it is suggested that this was a systemic response of DTH. The splenic T-cells of P. yoelii-vaccinated animals transformed in vitro with a soluble extract of the homologous parasite. The potential function of cell-mediated mechanisms in immunity to malarial infections is discussed.

The faster degradation of antigen by macrophages in Biozzi low (L) responder mice, compared to Biozzi high (H) responder mice, is thought to be responsible for their lower antibody response. We have measured four functions associated with macrophages to see whether macrophages from L mice were generally more active than those from H mice. Peritoneal macrophages obtained from normal mice were compared with those from groups of mice given Mycobacterium bovis BCG or Propionibacterium acnes. Cells from normal H mice gave a stronger oxidative burst when triggered with phorbol myristate acetate, and were more cytotoxic for tumour cells than cells from L mice. Cells from all mice injected with BCG or P. acnes gave a stronger oxidative burst, and were more cytotoxic for tumour cells; again, both responses were higher in H mice than in L mice. By contrast, when groups of mice that had received P. acnes were given endotoxin and bled, higher titres of tumour necrosis factor were found in the sera of L mice. Spleen cells from both lines of mice released similar levels of interleukin-1, both spontaneously and in response to lipopolysaccharide. Our results suggest that these various macrophage responses are expressed independently in H and L mice.

Depletion of C3-binding cells from spleen virtually abolishes the T cell-dependent response to sheep erythrocytes as assessed by transfer experiments. It is suggested that a functionally thymus-dependent B-cell subpopulation (B2) can be distinguished by surface receptors for C3.

Following injection of rat red cells, mice develop anti-red cell autoantibodies and subsequently suppressor T cells specific for these. Likewise, following recovery from non-lethal malaria, they develop suppressor T cells which suppress the anti-lymphocyte autoantibodies induced by lethal malaria parasites. Neither type of suppressor cell affected non-autoantibody components of the response, nor a response to sheep red cells. However, there was variable but significant cross-suppression of the respective autoantibody responses by both types of suppressor cell. Possible reasons for this unexpected cross-reaction are discussed.

Both nonlethal Plasmodium yoelii and lethal Plasmodium berghei were killed in vitro by hydrogen peroxide at concentrations as low as 10(-5) M. Higher concentrations were required in the presence of added normal erythrocytes. Injection of hydrogen peroxide in vivo significantly reduced P. yoelii parasitemia but had less effect on P. berghei.

A purified 230,000 mol wt protein antigen from the lethal mouse malaria parasite Plasmodium yoelii YM which had previously been shown to be highly effective as a vaccine, was tested for its ability to stimulate specific helper T cells and T cells responsible for delayed hypersensitivity. Strong stimulation was found in both assays, but larger doses were required for delayed hypersensitivity, correlating well with the requirements for protection. It is suggested that T stimulation may be a requirement for effective protection by purified antigens in malaria.

Heat-stable soluble products of rodent malarial parasites induce activated peritoneal macrophages to secrete tumour necrosis factor (TNF) in vitro. Since heat-stable parasite antigens are known to be present in the circulation of patients with malaria and it has been suggested that much of the pathology of malaria is due to TNF, we investigated the ability of such antigens to induce the production of TNF in vivo and to be toxic to mice. Injection of antigens obtained from Plasmodium yoelii or from P. berghei into mice which had previously received the macrophage-activating agent Propionibacterium acnes induced the release of TNF into the serum in amounts equivalent to the maximum release induced by bacterial lipopolysaccharide (LPS). Specific antiserum blocked the ability to the boiled soluble antigens, but not of LPS, to induce release of TNF. Similarly, vaccination specifically inhibited the release of TNF into the serum in response to subsequent stimulation with the antigens, but not with LPS. Mice made hypersensitive to the lethal action of TNF by pretreatment with D-galactosamine were killed in a dose-related fashion by administration of antigen preparations; addition of specific antiserum or prior vaccination with the antigens protected such mice, but not those given LPS, from death. We conclude that, in malaria, soluble antigens derived from the parasites may act like a toxin by stimulating the production of TNF, an important mediator of endotoxic shock, and that immunization with such antigens may diminish TNF secretion and consequently many of the clinical manifestations of the disease.

Mice were injected with silica or Corynebacterium parvum, which, respectively, inhibit and stimulate macrophages in vivo, in an attempt to study the role of macrophages in lethal Plasmodium yoelii infection and in mice protected by immunization. In the normal infection, macrophages were able to control parasitaemia for up to 1 week, whereas in immunized mice they appeared to inhibit the sterilizing immune response. A model is proposed in which this dual role of activated macrophages may account for the chronic non-sterilizing course of natural malaria infections.

The use of enzymes as markers of T or B cells in tissue sections has been studied in mouse lymphoid tissue and lymph nodes from the gerbil, rat and cat. Lymphocytes in the T-cell areas of murine lymph nodes and spleen contained discrete dots of non-specific esterase and N-acetyl-beta-D-glucosaminidase (beta-glucosaminidase) activity, with weak acid phosphatase activity. Lymphocytes in the B-cell areas lacked this discrete staining. Cortical thymocytes contained slight esterase activity while medullary thymocytes were strongly positive for both esterase and beta-glucosaminidase. Lymphocytes with a T-cell staining pattern were only occasionally seen in lymph nodes from Nude (nu/nu) mice. ATPase staining was restricted to lymphocytes in the B-cell areas; weak 5'-nucleotidase staining was only present in a frew lymphocytes in both T- and B-cell areas. Blast cells stimulated by in vivo injection of ConA or PHA in the mouse showed strong discrete enzyme activity for non-specific esterase and beta-glucosaminidase. Lipopolysaccharide-stimulated blast cells and cells within germinal centres lacked this discrete staining. Comparison of lymph nodes from the gerbil, rat and cat suggested at least on enzymes as a T-cell marker in each species although considerable variation in staining profiles was seen in the different species.

The affinity of anti-DNP antibodies produced by mice of various ages has been studied at the cellular level by the plaque inhibition technique. The affinity of PFC produced shortly after a single injection of antigen was found to increase during the first 6 weeks of life. The responses of 1- and 2-week-old animals also showed apparently restricted heterogeneity. The difference in the affinity of anti-DNP--PFC between young and adult mice could not be attributed to different (serum) levels of antigen or to differences in the rate of maturation of affinity during the immune response. Cell transfer experiments suggested that the age-dependent increase was due to a change in the population of antibody-forming (B) cell precursors and not to a progressive improvement of T-cell function. This findings is interpreted as favouring somatic mutation theories of antibody diversity.

As tumour necrosis serum kills malarial parasites in vitro and inhibits the multiplication of some species of Plasmodium in mice, we examined the effect of recombinant mouse tumour necrosis factor (rTNF) on P. yoelii both in vitro and in vivo. Parasites incubated overnight with rTNF showed no loss of viability, but repeated injection of rTNF into infected mice reduced parasitaemia and significantly prolonged survival of mice infected with a lethal variant of the parasite. We conclude that TNF acts on blood-stage malaria in vivo via a host cell and that another molecule in tumour necrosis serum is involved in killing the parasite in vitro.

Mice were immunized with lethal Plasmodium yoelii blood-stage malaria antigens that had been fractionated by isoelectric focusing using a variety of Ampholines over the range pH 3 to 10. Fractions were tested for their ability to protect against live challenge and to prime for parasite-specific T-cell help. Both activities exhibited three major peaks in the pH regions 4.5, 6.5, and 8, the pH 4.5 peak being the most consistently protective. There was a significant correlation between protection and T-helper-cell priming, particularly with antigens from the first peak, suggesting that T-cell priming represents an important component of the function of some protective malaria vaccines.

Early experiments in mice suggested that tumour necrosis factor (TNF) might be cytotoxic to asexual blood-stage malaria parasites. This was based on the striking activity of tumour necrosis serum (TNS) on the parasite both in vitro and in vivo, and the inability to separate by physical means the parasite-killing and tumour-killing components. However, recombinant TNF does not have this cytotoxic effect in vitro, while its antiparasitic activity in vivo, though significant, is not as strong as that of an equivalent amount of TNS. Thus it appears that TNS contains another cytotoxic molecule and that TNF itself may act indirectly in vivo, perhaps by activating an effector cell. An example of this has been found in murine schistosomiasis, where macrophage-derived TNF is able to activate eosinophils to attack the infecting worms. One mechanism of schistosomule damage is by eosinophil cationic proteins, and these have also been found to be cytotoxic to blood-stage malaria. There may therefore be a pathway of TNF activity common to both parasites. In a similar way, the crisis-forming factor (CFF) found in the serum of certain immune Sudanese adults is clearly distinct from TNF, since CFF-containing sera do not kill TNF-susceptible tumour cells and rTNF does not kill Plasmodium falciparum in vitro. This confirms that there are other cytotoxic molecules, still to be identified, with a role in immunity to malaria and perhaps other parasites. TNF is also active against intracellular Trypanosoma cruzi and against some viruses but in both cases this appears to be an interferon-like mediatory effect and not direct cytotoxicity. It is not yet clear whether these antiparasitic activities are part of the biological role of TNF.

The anti-lymphocyte autoantibody response to irradiated lethal Plasmodium berghei malaria parasites in normal mice was significantly reduced when recipients were pre-treated with splenic T cells from mice recovered from a non-lethal Plasmodium yoelii infection. Suppression was specific for the autoantibody and did not affect the antibody response to the parasite. Experiments involving sequential P. yoelii-P. berghei infections in situ revealed that recovery from P. berghei was possible when the interval between the two infections was 14 days or more. This ability to recover from P. berghei correlated with a progressive reduction of anti-lymphocyte autoantibody suggesting a useful role for the suppressor cell. The possible link between suppressor cells and anti-lymphocyte autoantibodies in malaria is discussed.

The isoelectric heterogeneity of both IgM and IgG murine anti-sheep red cell serum antibodies has been examined using an adaptation of published methods. It was found that the IgM spectra were restricted in a characteristic manner, implying oligoclonality. In cell transfer experiments, the T cell dependency of the IgM responses was confirmed. Further, helper cells appeared to switch on IgG antibody production and, simultaneously, recruit many novel IgG-forming clones into the response. The pattern of the IgM oligoclonality was attributed in part to the inheritance of genes at, or closely linked to, the Igl heavy chain locus. These findings are discussed in relation to current research on the regulation of diversity.

Tumour necrosis factor (TNF) has been implicated as a mediator of toxicity in a number of infectious diseases, including malaria. We have shown that human and rodent blood-stage parasites liberate heat-stable soluble antigens that induce the release of TNF by activated macrophages in vitro and in vivo, and are toxic to mice made hypersensitive to TNF by D-galactosamine. These antigens induce T-independent antibodies which specifically block their ability, but not that of bacterial lipopolysaccharide, to cause the secretion of TNF. Cytokine release in vitro may be a useful strategy for identifying potentially toxic molecules of infectious organisms and for investigating the nature of antibodies that can protect the host against their damaging effects.

Mice were vaccinated against lethal blood stage P. yoelii malaria with lysates of parasitised blood prepared in various detergents. Two injections, containing 25 micrograms of protein, of a Triton X-100 lysate gave optimal protection when injected intraperitoneally with saponin as adjuvant. Such lysates provide a starting point for the isolation of potential antigens for vaccination, selected for their protective activity in vivo.

Mice from Biozzi high (H) and low (L) antibody responder lines were infected either with non-lethal Plasmodium yoelii 17X, with a lethal variant of the same parasite or with P. chabaudi (non-lethal). H mice recovered from infection with both P. yoelii and its normally lethal variant, whereas L mice died from both. However, L mice showed better initial control of parasitaemia with all three infections. Both H and L mice infected with P. chabaudi showed a similar rapid reduction of parasitaemia once it had reached 40-50%, but L mice were unable to clear the infection completely. When infected with either non-lethal P. yoelii or P. chabaudi, H mice made substantially more antibody than L mice, especially around the time of recovery. The courses of parasitaemia suggest a role for non-antibody factors in early control of infection and for antibody in recovery from P. yoelii. Although antibody was apparently necessary for complete recovery from P. chabaudi, the first rapid fall in parasitaemia appeared to be unrelated to antibody levels.

A method is described for detecting the simultaneous presence of Fc and C3 receptors on mouse spleen cells. A proportion of both T cells and non-T cells bear both receptors. Both T-cell and non-T-cell Fc receptors were blocked by aggregated mouse IgG2 to the same degree. C3, but not Fc, receptors were blocked by factors present in the serum of irradiated mice or mice undergoing graft-versus-host reaction. Thymocytes activated by injection into irradiated F1 hybrid mice, and thymocytes regenerating after irradiation and bone marrow injection, appeared to have increased Fc receptors. A general role for Fc and C3 receptors in T cell-B cell co-operation is suggested.

Sera from mice infected with Plasmodium yoelii or Plasmodium berghei and given endotoxin contained nonspecific mediators which killed both species of parasite and tumor cells in vitro. The sera resembled tumor necrosis sera obtained from mice given macrophage-activating agents such as Propionibacterium acnes (formerly designated Corynebacterium acnes) or Mycobacterium bovis BCG and then endotoxin. Cytotoxicity developed parallel to parasite killing activity and indicated that macrophages were activated. Activation occurred sooner with P. berghei, which is lethal, and serum activity remained on a plateau until the mice died. In nonlethal P. yoelii infections, activation was related to the course of parasitemia. Endotoxin given to mice infected with P. yoelii caused an immediate decrease in parasitemia, presumably through the release of parasite killing factors. The extent of the decrease depended upon the time of administration. No immediate drop in the parasitemia caused by P. berghei was observed at any time. Early administration of endotoxin prolonged survival; late administration accelerated death. Passive transfer of rabbit tumor necrosis serum to infected mice decreased the parasitemia caused by P. yoelii but not that caused by P. berghei. Other components of the immune response appeared to act together with these soluble mediators to eliminate P. yoelii; they may be absent or suppressed in infections with P. berghei.

Mice can be protected against several species of lethal malaria infection by vaccination, and their recovery correlates well with increased anti-malarial antibody levels, particularly IgG (ref.2). However, there is also a good correlation between protection by vaccines and priming for delayed-type hypersensitivity in the skin, although there is no obvious explanation for this effect. We now report an apparent relationship between protection and a cell-mediated immune response involving the migration of various types of cell capable of killing malaria parasites in vitro to the liver. We suggest that the effect of vaccination is to bring together parasites, specific antibody and nonspecific cytotoxic cells, and that the liver may be a major site for their interaction.