Neutrophils play a critical role in the pathogenesis of experimental cerebral malaria

Abstract

The role of neutrophils in experimental cerebral malaria (ECM) is not well understood. In this study we used a MoAb, RB6-8C5, to deplete the peripheral neutrophils of ECM-susceptible CBA/NSlc mice 24 h before Plasmodium berghei ANKA (PbA) infection. We found that early neutrophil depletion prevented the development of ECM and dramatically decreased the sequestration of monocytes and microhaemorrhage in the brain. The depletion of neutrophils also down-regulated tumour necrosis factor-alpha, interferon-gamma and IL-2 mRNAs and abrogated IL-12p40 mRNA expression in the brain as examined by competitive reverse transcriptase-polymerase chain reaction. Although depletion of neutrophils decreased the expression of Th1 cytokines in both spleen and brain, our results did not show the shift of a Th1 to a Th2 immune response since there was no obvious augmentation of expression of Th2 cytokine mRNAs (IL-4 and IL-10). We conclude that neutrophils play a role in the pathogenesis of ECM via enhancement of the expression of Th1 cytokines in the brain.

INTRODUCTION

Malaria is a major health problem in the world. Hundreds of thousands of people, especially children in African, die of cerebral malaria every year. However, the pathogenesis of this deadly disease is incompletely understood. To investigate the pathogenesis of cerebral malaria (CM), an experimental model involving the infection of genetically susceptible CBA mice with Plasmodium berghei ANKA (PbA) blood stages was developed [1]. In this model, the mice develop a neurologic syndrome 6–14 days after infection which leads to a cumulative mortality of about 90%. Among the various parameters, including immunological, non-immunological and histopathological findings of experimental and human CM (summarized in [2]), the mouse model mimics most parts of human CM, except that in mice the main cells sequestered in cerebral blood vessels are monocytes, whereas in human CM the major cells sequestered are parasitized erythrocytes. This murine model of experimental cerebral malaria (ECM) is the one most widely used.

Excessive production of tumour necrosis factor-alpha (TNF-α) by monocytes/macrophages plays a key role in the pathogenesis of both human CM and ECM. Patients with CM have high serum TNF-α levels [3–5] and administration of neutralizing antibodies to TNF-α prevents ECM in mice [6]. The immunopathology of ECM is also dependent on both CD4⁺[7] and CD8⁺ cells [8]. Since neutrophils are not sequestered in the brain, their role in the pathogenesis of ECM has received little attention, in spite of the fact that our previous studies [9–12] and other reports [13,14] show that neutrophils contribute significantly to the immune response by modulating both cellular and humoral immunity, especially via the synthesis and release of immunoregulatory cytokines. Most recent studies also indicate that human and mouse neutrophils secrete IL-12 [15,16]. In the present study, we found that depletion of neutrophils 1 day before infection prevented ECM in susceptible CBA mice and changed the expression of various cytokine mRNAs in brain and spleen. The study suggests that neutrophils play a role in the pathogenesis of ECM by modulating Th1 cytokine expression.

MATERIALS and METHODS

Mice and parasites

ECM-susceptible CBA/NSlc and resistant BALB/c 7–10-week-old female mice were purchased from Japan SCL Inc. (Hamamatu, Japan) and housed in the Animal Centre of our University under specific pathogen-free conditions. Mice were infected by intraperitoneal (i.p.) injection of 10⁶ parasitized erythrocytes suspended in 0·2 ml PBS. PbA was kindly provided by Dr M. Suzuki (Gumma University, Japan). Blood was collected from the tail vein every other day for determination of parasitaemia and measurement of haematological parameters.

Antibodies and reagents

The hybridoma which secretes a MoAb which depletes mouse granulocytes, RB6-8C5, was a gift from Dr R. Coffman (DNAX Research Institute, Palo Alto, CA). The MoAb is a rat IgG2b that selectively binds to and depletes mature neutrophils and eosinophils, but not lymphocytes and macrophages [17–21]. RB6-8C5 was obtained from hybridoma culture supernatants, purified by ammonium sulphate precipitation, dialysed against PBS, assayed for protein concentration with a colourimetric kit (BioRad Labs, Richmond, CA), and stored at −80°C. To deplete mice of neutrophils, 0·25 mg of the antibody was administered intraperitoneally 1 day before or 5 days after initiating infection. Treatment with this dose of the antibody induced severe neutropenia for up to 5 days which is similar to what has been reported [17–21]. Control mice received an equivalent amount of commercial normal rat IgG (Caltag Labs, Burlingame, CA).

Histopathology and immunohistochemistry

For conventional histopathology, paraffin-embedded sections (3 μm thick) were stained with haematoxylin–eosin (H–E). For immunohistochemistry, after deparaffinization and dehydration sections were treated on ice for 30 min with freshly prepared 1% H₂O₂ in methanol to block endogenous peroxidase activity. Non-specific background was blocked by incubating sections in 5% skim milk in PBS. After washing three times in PBS, 5 μg/ml purified RB6-8CA, or normal rat IgG as a negative control, were applied to the sections. They were incubated at 4°C overnight in a humidified chamber, washed four times with PBS and then incubated for 30 min with secondary polyvalent biotinylated antibody (MBL, Nagoya, Japan). After washing in PBS, sections were incubated with streptavidin–horseradish peroxidase (MBL) in a humidified chamber at room temperature for 30 min. The sections were then washed in PBS before the addition of 0·02% 3,3′-diaminobenzidine tetrahydrochloride substrate in 1 m Tris–HCl pH 7·5 containing 0·01% H₂O₂. The reaction was terminated by rinsing the slides in running water. Finally, the sections were counterstained with haematoxylin, dehydrated and mounted.

RNA isolation and semiquantitative reverse transcriptase-polymerase chain reaction analysis

Total RNA for each time point was extracted from spleen and brain tissue samples taken from two or three mice 1 and 9 days after infection using ISOGEN (Nippon Gene, Fuji City, Japan) according to the manufacturer’s protocol. The total RNA was reverse transcribed by AMV reverse transcriptase XL (TaKaRa, Ohtsu, Japan) using random 9mers primers and reverse transcriptase (RT) conditions of 10 min at 30°C and 25 min at 55°C. The reaction was terminated by heating at 99°C for 5 min, and the cDNA samples were stored at −20°C. The levels of β-actin, interferon-gamma (IFN-γ), TNF-α, IL-2, transforming growth factor-beta (TGF-β), and IL-10 were determined by competitive polymerase chain reaction (PCR) using a multicompetitor pmCK3.1 [22], a kind gift from Dr R. L. Tarleton (University of Georgia, Athens, GA). Primers for IFN-γ amplify a 429-bp fragment and a 565-bp fragment, those for TNF-α a 275-bp fragment and a 373-bp fragment, those for IL-2 a 423-bp fragment and a 565-bp fragment, those for TGF-β a 451-bp fragment and a 563-bp fragment and those for IL-10 a 390-bp fragment and a 469-bp fragment of the cDNA and the multicompetitor, respectively. IL-4 and IL-12p40 were determined by self-constructed competitors using a Competitive DNA Construction Kit (TaKaRa). The primer pairs used for amplification of IL-4 were previously published by Kopf et al. [23]. The primer pairs used for amplification of IL-12p40 were: 5′-AACAGCGCACCCACTTCATCAA and 5′-TTGAGATGATGCTTTGACA, and the primers for IL-4 and IL-12 amplify a 216- and 384-bp fragment of the cDNAs and a 340- and a 540-bp fragment of the competitors, respectively. After adjusting the cDNA samples to equal concentrations of β-actin, 4 μl of the diluted cDNA and specific primers were mixed with 0·5 μl of four-fold serial dilutions of the competitors (for IL-4 and IL-12p40, competitor was 10-fold serially diluted) in a total volume of 10 μl and used for a PCR reaction. PCRs were performed as follows: β-actin, IFN-γ, TNF-α and IL-12p40, 1 min at 94°C, 40 s at 55°C and 1 min at 72°C; IL-2, 1 min at 94°C, 40 s at 53°C and 1 min at 72°C; and IL-4, IL-10 and TGF-β, 1 min at 94°C, 40 s at 57°C and 1 min 72°C. After 35 cycles the final elongation was followed by a 5-min hold at 72°C for all cytokine cDNA amplifications. PCR was performed in a DNA Thermal Cycler (Perkin-Elmer Corp., Norwalk, CT). During the PCR, competitive reactions occur due to competition for the primers, resulting in fragments that differ in size by about 70–150 bp. The PCR products were resolved on a 2% agarose gel. The relative amount of the respective cytokine cDNA in a given sample was determined by comparing their relative ethidium bromide staining intensities.

Detection of TNF-α

Serum TNF-α levels 9 days after infection were determined by an indirect ELISA. In brief, microtitre plates were coated with hamster anti-TNF-α MoAb, and then incubated with 10-fold diluted sera. The bound TNF-α was probed with rabbit anti-TNF-α and detected with a peroxidase-conjugated goat anti-rabbit antibody. Recombinant mouse TNF-α was purchased from Genzyme (Cambridge, MA). The limit of detection was 20 pg/ml.

The spleens, brains and sera were obtained from CBA/NSlc mice which had neurological symptoms at day 9 after infection. They were used for the RT-PCR or ELISA assays.

Statistical analysis

Significance of survival curves was determined by Kaplan–Meier survival analysis. Analysis of differences of other data was performed using Student’s t-test. P < 0·05 was considered statistically significant.

RESULTS

Early neutrophil depletion prevents ECM

In a preliminary experiment we observed that one dose of RB6-8C5 5 days after infection did not prevent the development of ECM (data not shown), but when the MoAb was administered 1 day before infection, it prevented the appearance of neurological manifestations in >90% of susceptible CBA/NSlc mice (Fig. 1). Therefore, in further studies we used the MoAb 1 day before infection for depletion of neutrophils. The mice treated with MoAb eventually died with severe anaemia (< 10⁶ erythrocytes/μl) (Fig. 2a) and hyperparasitaemia (80–90%) 3–4 weeks after infection (Fig. 2b) but without neurological symptoms. In contrast, the mice treated with rat IgG developed a neurological syndrome which included paralysis (mono-, hemi-, para- or tetraplegia), ataxia, convulsions and coma 8–11 days after infection but had low parasitaemia (8–18%) (Fig. 2b). All of the control mice died within 48 h after the onset of neurological symptoms. There was no difference in the parasitaemia in MoAb-treated and control mice (P > 0·05) (Fig. 2b), indicating that neutrophils did not directly affect parasitaemia. There was also no significant difference in the number of peripheral blood erythrocytes in the two groups of mice (P > 0·05) (Fig. 2a). This study suggests that neutrophils play a role in the pathogenesis of ECM. The percentage of neutrophils in peripheral blood increased dramatically to about 50% in control CBA mice 1 and 2 days after infection, decreased to a normal level at 4 days and reached another peak 8 days after infection (Fig. 2c). This suggests that PbA infection stimulates neutrophils which may play some role in innate immunity to this infection. Resistant BALB/c mice did not present neurologic lesions and died of hyperparasitaemia and severe anaemia about 3–4 weeks after infection (data not shown).

Fig. 1.

Prolonged survival of Plasmodium berghei ANKA (PbA)-infected CBA/NSlc mice whose neutrophils were depleted by RB6-8C5 MoAb (○). Cerebral malaria (CM)-susceptible CBA/NSlc mice (n = 12) were injected intraperitoneally with 0·25 mg RB6-8C5 24 h before infection with 10⁶ blood form PbA. Control mice (n = 12) were injected intraperitoneally with 0·25 mg normal rat IgG 24 h before infection (□). Survival was measured daily. Kaplan–Meier survival analysis was performed with Stat-View software (Abacus Concepts, Inc, Berkeley, CA). P = 0·001. Similar results were obtained in another experiment.

Fig. 2.

No alteration of parasitaemia and anaemia in Plasmodium berghei ANKA (PbA)-infected CBA/NSlc mice by neutrophil depletion. (a) Peripheral blood erythrocyte counts. (b) Parasitaemia. (c) Percentage of neutrophils in peripheral blood. Results presented are the means ±s.e.m. for groups of 12 mice. Similar results were obtained in another experiment.

Neutrophil depletion changes the splenic and cerebral histopathology

We investigated the histopathology of PbA-infected mice. In control mice injected with rat IgG there were more neutrophils in the red pulp of the spleen 1 day after infection (Fig. 3A) than in normal mice (data not shown), but this was not observed in the mice treated with MoAb (Fig. 3B). There were much fewer microhaemorrhages (2·7 ± 0·3) on the brain surface regions including cerebellum, olfactory lobe/bulbs, frontal lobe, parietal lobe and mid-brain in the neutrophil-depleted mice than in the controls (20·0 ± 3·5) (P < 0·01). The sequestration of monocytes and erythrocytes in brain blood vessels was much more obvious in control mice (Fig. 3C) than in those treated with MoAb (Fig. 3D). However, petechial haemorrhages in the sections were only observed in the controls (Fig. 3E). We did not find neutrophil sequestration in the brain venules examined by immunohistochemistry (data not shown). No monocyte/erythrocyte sequestration was observed in the brain venules of BALB/c mice (data not shown).

Fig. 3.

Histopathology of spleen and brain of Plasmodium berghei ANKA (PbA)-infected CBA/NSlc mice treated or not treated with a MoAb that selectively depletes neutrophils. CBA/NSlc mice were injected intraperitoneally with 0·25 mg RB6-8C5 or control IgG 24 h before PbA infection. One day post-infection spleen sections were examined for the presence of neutrophils by immunohistochemistry. Nine days after infection brains were histologically examined after haematoxylin–eosin staining. (A) A spleen section from CBA/NSl mice treated with control IgG. Note neutrophils which are indicated by arrows in the red pulp. (B) A spleen section from CBA/NSlc mice treated with RB6-8C5. Note no neutrophils were observed. (C) A brain section from control CBA/NSlc mice infected with PbA exhibiting cerebral symptoms 9 days after infection. Intravascular accumulation and adherence of large numbers of monocytes to endothelial cells. (D) Brain from a RB6-8C5-treated CBA/NSlc mouse infected with PbA showing no experimental cerebral malaria (ECM) 9 days after infection. Note few monocytes sequestered in venule. (E) Brain of a mouse with conditions similar to (C). Note the microhaemorrhage indicated by an arrow. (A,B, immunohistochemistry; C–E, H–E staining; bars = 20 μm).

Neutrophil depletion alters expression of cytokine mRNAs

Since early depletion of neutrophils in CBA mice prevented ECM, we speculated that the immune response may have been shifted from one that was Th1 dominant to one that was Th2 dominant. Therefore, we used a semiquantitative competitive RT-PCR analysis to investigate the expression of cytokine mRNAs in both spleen and brain.

As shown in Fig. 4a and Table 1, the expression of IL-2, IL-10, IL-12p40, TNF-α and IFN-γ was induced in the spleen 1 day after PbA infection of control mice, whereas IFN-γ and IL-12p40 mRNAs were not detected in neutrophil-depleted mice at this time. Both control mice and those treated with MoAb expressed about the same levels of TNF-α mRNA, and the expression of IL-10 mRNA in MoAb-treated mice was four-fold less than in control mice at this time. Nine days after infection, the spleen cells of control mice still expressed about the same levels of these cytokines as 1 day after infection, except that IL-2 and IL-10 showed a 4–16-fold increase compared with 1 day after infection. The levels of TNF-α, IFN-γ and IL-10 were comparable in the two groups of mice, and those of IL-2, IL-4 and IL-12p40 in control mice expressed 4–10-fold higher than in neutrophil-depleted mice 9 days after infection.

Fig. 4.

(a) Semiquantitative competitive reverse transcriptase-polymerase chain reaction (RT-PCR) analysis of IL-2, IL-4, IL-10, IL-12p40, IFN-γ, and tumour necrosis factor-alpha (TNF-α) mRNAs in the spleens and brains of CBA/NSlc mice infected with Plasmodium berghei ANKA (PbA) whose neutrophils were or were not depleted. RB6-8C5 or normal rat IgG (control) treated CBA/NSlc mice were infected with 10⁶ parasitized erythrocytes. Spleens were obtained 1 and 9, and brains 9 days after infection. Total RNA were obtained from two or three mice for each time point. mRNAs were reverse transcribed and the cDNAs were adjusted to equal amounts of β-actin cDNA. The transcripts of IL-2, IL-4, IL-10, IL-12p40, IFN-γ, and TNF-α were amplified in the presence of competitor. For quantification of TNF-α, the competitor was serially diluted four-fold starting at 62·5 attomoles (day 1) or 4000 attomoles (day 9, for brain from 62·5 attomoles); of IFN-γ, the competitor was serially diluted four-fold starting at 62·5 attomoles (day 1) or 250 attomoles (day 9, for brain from 16 attomoles); of IL-2, the competitor was serially diluted four-fold starting at 62·5 attomoles (day 1) or 250 attomoles (day 9); of IL-10, the competitor was serially diluted four-fold starting at 16 attomoles (day 1) or 62·5 attomoles (day 9); of IL-4, the competitor was serially diluted 10-fold starting at 40 000 attomoles; and of IL-12p40, the competitor was serially diluted 10-fold starting at 60 000 attomoles. If there was no target cytokine cDNA amplified when the lowest dilution of competitor was used in competitive PCR, the negative result was confirmed by using the same PCR condition without adding competitor (data not shown). The amount of competitor which resulted in equal or approximately equal amounts of target cytokine cDNA in PCR is indicated above the amplified competitor. Very similar results were obtained in another experiment. (b). Semiquantitative competitive RT-PCR analysis of IL-2, IL-4, IL-10, IL-12p40, TNF-α and IFN-γ mRNAs in the spleens and brains of BALB/c mice infected with PbA. BALB/c mice were infected with 10⁶ parasitized erythrocytes. The spleens 1 and 9 days after infection and the brains 9 days after infection were obtained. Total RNA were obtained from two or three mice for each time point. mRNAs were reverse transcribed and the cDNAs were adjusted to amounts equal to the samples of the CBA/NSlc mice β-actin. For quantification of TNF-α, IL-4, IL-10 and IL-12p40, the same amounts of competitors were used as in CBA/NSlc mice; for quantification of IFN-γ, the competitor was serially diluted four-fold starting at 62·5 attomoles (both spleen and brain); and of IL-2, competitor was serially diluted four-fold starting at 16 attomoles (day 1) or 250 attomoles (day 9). If the lowest dilution of competitor was used in competitive PCR, there was no target cytokine cDNA amplified. The negative result was confirmed by using the same PCR condition without adding competitor (data not shown). Very similar results were obtained in another experiment. D1, Day 1; D9, day 9; N, normal; RB, RB6-8C5; IgG, normal rat IgG.

Table 1.

Semiquantitative reverse transcriptase-polymerase chain reaction (RT-PCR) analysis of cytokine mRNAs in the spleens and brains of neutrophil-depleted or undepleted CBA/NS1c and BALB/c mice infected with Plasmodium berghei ANKA (PbA)

	CBA/NS1c								BALB/c
	Spleen			Brain			Spleen			Brain
		RB6-8C5		IgG			RB6-8C5	IgG
Cytokines	N	D1	D9	D1	D9	N	D9	D9	N	D1	D9	N	D9
IL-12p40	ND	ND	60–600	600	600	ND	ND	600	ND	600	6·0	ND	600
TNF-α	ND	62·5	62·5	62·5	62·5	1·0	4·0	16·0	ND	4·0	0·25	16·0	250
IFN-γ	ND	ND	4·0	4·0	4·0	ND	0·06	1·0	ND	0·25	0·25	ND	0·25
IL-2	ND	4·0	4·0	4·0	16·0	ND	16·0	62·5	ND	4·0	16·0	ND	62·5
IL-10	ND	1·0	62·5	4·0	62·5	ND	ND	ND	ND	1·0	4·0	ND	ND
IL-4	ND	ND	4·0	ND	40·0	ND	ND	ND	ND	ND	4·0–40·0	ND	ND

N, Normal spleen or brain; ND, not detectable; D1, 1 day after infection; D9, 9 days after infection.

The original data are shown in Fig. 4a,b. Values indicate the attomoles of competitor which resulted in equal or approximately equal amounts of target cytokine cDNA in competitive PCR.

Nine days after infection the expression of TNF-α, IFN-γ and IL-2 mRNAs in the brains of control mice was 4–16-fold higher than in neutrophil-depleted mice. IL-12p40 mRNA was detectable in control but not in MoAb-treated mice. IL-4 and IL-10 transcripts were undetectable in the brains of both groups of mice (Table 1). In summary, PbA infection induced the expression of Th1 cytokines (IL-2, IL-12p40 and IFN-γ), TNF-α, and Th2 cytokines (IL-4 and IL-10) in spleen and IL-2, IL-12p40, IFN-γ and TNF-α in brain. Depletion of neutrophils decreased the expression of Th1 cytokines (IL-2, IL-12p40 and IFN-γ) in both spleen and brain, and TNF-α in the brain. Thus, the present study confirms the previous findings that PbA-induced ECM is associated with over-expression or production of TNF-α and IFN-γ and a predominant Th1 response [24–27], and that protection from ECM is associated with reduced production of Th1 cytokines and TNF-α without a concomitant augmented Th2 response [25].

In general in ECM-resistant BALB/c mice the levels of all cytokine transcripts (Fig. 4b and Table 1) in the spleens 1 and 9 days after infection were lower than in those of IgG-treated CBA mice. Nine days after infection TNF-α, IFN-γ, IL-2 and IL-12p40 mRNAs were also detected in brains. The levels of IL-2 and IL-12p40 transcripts were the same as in IgG-treated CBA mice, IFN-γ mRNA expression was four-fold lower than in IgG-treated CBA, but the level of TNF-α message was 16-fold higher than in CBA mice. TGF-β transcripts were not detectable in any of the samples from both CBA/NSlc and BALB/c mice (data not shown).

DISCUSSION

The present study demonstrates that early depletion of neutrophils prevents the occurrence of ECM, monocyte sequestration in brain blood vessels, and changes in the patterns of expression of various cytokine mRNAs in the brains of ECM-susceptible mice. However, the anti-neutrophil MoAb did not modify parasitaemia and the number of peripheral erythrocytes showed that neutrophils do not effect parasite growth. Senaldi et al. [28] previously reported that depletion of neutrophils prevents the mortality and vascular permeability in brain and lung. It also abolishes neutrophil sequestration in the lung and partially decreases monocyte sequestration in the brain and the lung. The major differences between our study and theirs are that: (i) the prevention of ECM occurred when neutrophils were depleted 5 days after infection in their work, while in our study the effect was observed only when neutrophils were depleted 1 day before infection; (ii) their report shows that depletion of neutrophils prevented ECM while increasing the occurrence of microhaemorrhages in the brain, which is contradictory to other observations in the study. The reason for this finding is difficult to find; (iii) they did not examine the effects of neutrophils on immunomodulation by examining the expression of cytokine mRNAs. We could not reproduce their results when the MoAb was used at day 5 after infection. The reason for the different results is unknown at present, but the discrepancy may be related to differences in the strain of PbA and neutrophil-depleting antibodies used.

An explanation of our results may involve the following mechanisms. Neutrophils have long been thought of as terminally differentiated cells and their role in immunoregulation has received little attention. However, recent convincing evidence suggests that neutrophils can contribute significantly to the initiation and amplification of both cell-mediated immunity and the humoral immune response [13,14]. Our previous studies revealed that neutrophils play a critical role in DTH [9,10]. Furthermore, we also found that neutrophils are involved in resistance to experimental Chagas’ disease and murine Leishmania major infection, and play an immunoregulatory role in these infections (Chen, unpublished results). Considering these results and the modulation of cytokine mRNA expression by neutrophil depletion found in the present study, a change in the profile of cytokine production may be responsible for the inhibition of ECM by neutrophil depletion. At present, we do not know exactly how neutrophils modulate the expression of cytokine mRNAs in brain and spleen. However, inasmuch as activated neutrophils themselves produce various kinds of cytokines such as IL-1, IL-6, IL-8, IL-10, IL-12, IFN-α, TNF-α, granulocyte-macrophage colony-stimulating factor (GM-CSF) and macrophage colony-stimulating factor (M-CSF) [13–16] and chemokines including macrophage inflammatory protein-1α (MIP-1α), MIP-1β, monokine-inducible by IFN-γ (MIG), IFN-inducible T cell α chemoattractant (I-TAC), and IFN-γ-inducible protein-10 (IP-10) [29–32], we speculate that these cytokines and chemokines produced by neutrophils regulate the functions of lymphocytes and monocytes, including their cytokine production. According to recent studies, not only IL-12, which has been identified as a strong inducer of Th1-type immune response, but also chemokines such as MIP-1α, MIP-1β, MIG, I-TAC and IP-10, can serve as chemoattractants for selective mobilization of Th1 cells [33–37]. Previous studies indicated that a cell-mediated or Th1 immune response plays a role in the pathogenesis of CM and that a critical imbalance in cytokine production is associated with the development of the neurovascular lesions that contribute to the pathology of this disease [24–26]. Our results confirm these observations, since neutrophil depletion that inhibited ECM also reduced the expression of Th1 cytokines. The administration of RB6-8C5 MoAb did not affect the course of infection in CM-susceptible CBA mice, while it did when the MoAb was used 1 day before initiating infection, suggesting that neutrophils at the very early stage of the infection contribute significantly to the development of a Th1-type immune response which is responsible for the pathogenesis of ECM.

It is well established that TNF-α is important for the development of ECM [3–6]. However, in the present study ECM was inhibited irrespective of the expression of TNF-α mRNA in the brains of CBA mice whose neutrophils were depleted, although the strength of expression was weaker than in control mice. Since there was no IL-12p40 detectable in the brains of susceptible CBA mice whose neutrophils were depleted, it seems possible that TNF-α without concomitant IL-12 production in the brain is not sufficient to cause development of ECM. It seems that IL-12 expressed in the brain may be a central effector molecule for development of ECM. Th2 cytokines (IL-4 and IL-10) and TGF-β, the natural antagonists of Th1 cytokines, were not detectable in the brains of ECM mice. These results support the idea that in ECM there is a lack of production of negative regulatory cytokines for TNF-α, IFN-γ, IL-2 and IL-12 over-expresson in the brain which promotes the development of CM.

It is interesting that the brains of ECM-sensitive and -resistant strains of mice constitutively express TNF-α mRNA, but at a much lower level than in PbA-infected mice. The expression of mRNA may not always be proportional to the production of the corresponding protein, as indicated by a study [38] which showed that TNF-α mRNA was detectable in the brain 3 days after infection, but TNF-α protein was not found at that time. It was unexpected that serum TNF-α was not detected in both ECM-susceptible CBA/NSlc and ECM-resistant BALB/c mice. However, it is not the systemic level of TNF-α which is important in CM, but rather the level produced by sequestered monocytes in the microcirculation or microglia and/or astrocytes in the brain [38–40]. In a clinical investigation [41] it was found that some malaria patients with high serum TNF-α levels do not develop CM, but some with low serum TNF-α levels do, and that serum TNF-α levels are related to the severity of malaria, but not necessarily to CM.

Our study shows that although both TNF-α and IL-12 are highly expressed in the brains of BALB/c mice, the mice do not develop CM. According to a recent study, brain microvascular endothelial cells (MVEC) from CBA and BALB/c mice exhibit different sensitivities to TNF-α. In response to TNF-α, CBA brain MVEC displayed a higher capacity to produce IL-6 and to up-regulate intercellular adhesion molecule-1 (ICAM-1), vascular cell adhesion molecule-1 (VCAM-1), and both p55 and p75 TNF-α receptors than BALB/c brain MVEC [42]. This may provide an explanation for our results, which show that the brains of BALB/c mice express even higher levels of TNF-α than those of CBA mice, but they do not develop CM.

As mentioned in the Introduction, PbA induces a cerebral malaria syndrome in susceptible mice that parallels the human disease in some respects. The major difference is that in malaria-infected mice, the focal arrest of monocytes/macrophages and lymphocytes, not erythrocytes, in cerebral blood vessels is the dominant feature. This may be caused in part by the binding of CD11a on the monocytes to ICAM-1 expressed on endothelial cells in the brain [28,42], whereas in human CM the major cells arrested are parasitized erythrocytes because of the binding of Plasmodium falciparum-infected erythrocyte membrane protein 1 to ICAM-1, VCAM, E-selectin and CD36 on the endothelial cells [43]. There is no evidence showing that neutrophils are sequestered in the brain vessels in either human or mouse CM. Thus, neutrophils may not be directly involved in the pathological changes in the brain, although they can affect other organs, since Senaldi and co-workers observed sequestration of neutrophils in the lung in PbA-infected mice and neutrophil depletion abolished this manifestation [28]. Neutrophils may indirectly influence cerebral sequestration of monocytes, because our previous studies indicated that neutrophil depletion markedly decreased the migration of monocytes in DTH reactions and inhibited macrophage infiltration into the peritoneal cavity induced by i.p. injection of a streptococcal preparation, OK-432 [7,44]. However, the precise mechanisms by which neutrophils affect monocyte function in ECM need to be studied further. In addition, the immunoregulatory roles of neutrophils in human cerebral malaria also need to be studied, since an increase in neutrophil elastase and malaria pigment-containing neutrophils was observed in patients with severe malaria, indicating that neutrophils are also involved in the pathogenesis of human CM [45,46].

In conclusion, the present study demonstrates the importance of neutrophils in the induction of ECM through the modulation of cytokine production. Among the cytokines, IL-12 expressed in the brain may be the most critical factor. Thus, these studies indicate a potential therapeutic role for interventions targeted at IL-12 produced in the brain.