Lipopolysaccharide Preconditioning Augments Phagocytosis of Malaria-Parasitized Red Blood Cells by Bone Marrow-Derived Macrophages in the Liver, Thereby Increasing the Murine Survival after Plasmodium yoelii Infection

ABSTRACT

Malaria remains a grave concern for humans, as effective medical countermeasures for Plasmodium infection have yet to be developed. Phagocytic clearance of parasitized red blood cells (pRBCs) by macrophages is an important front-line innate host defense against Plasmodium infection. We previously showed that repeated injections of low-dose lipopolysaccharide (LPS) prior to bacterial infection, called LPS preconditioning, strongly augmented phagocytic/bactericidal activity in murine macrophages. However, whether LPS preconditioning prevents murine Plasmodium infection is unclear. We investigated the protective effects of LPS preconditioning against lethal murine Plasmodium infection, focusing on CD11b^high F4/80^low liver macrophages, which are increased by LPS preconditioning. Mice were subjected to LPS preconditioning by intraperitoneal injections of low-dose LPS for 3 consecutive days, and 24 h later, they were intravenously infected with pRBCs of Plasmodium yoelii 17XL. LPS preconditioning markedly increased the murine survival and reduced parasitemia, while it did not reduce tumor necrosis factor (TNF) secretions, only delaying the peak of plasma gamma interferon (IFN-γ) after Plasmodium infection in mice. An in vitro phagocytic clearance assay of pRBCs showed that the CD11b^high F4/80^low liver macrophages, but not spleen macrophages, in the LPS-preconditioned mice had significantly augmented phagocytic activity against pRBCs. The adoptive transfer of CD11b^high F4/80^low liver macrophages from LPS-preconditioned mice to control mice significantly improved survival after Plasmodium infection. We conclude that LPS preconditioning stimulated CD11b^high F4/80^low liver macrophages to augment the phagocytic clearance of pRBCs, which may play a central role in resistance against Plasmodium infection.

INTRODUCTION

According to the “World Malaria Report 2019” from the World Health Organization (WHO), malaria is one of the three major infectious diseases, along with AIDS and tuberculosis, and efforts are under way to eliminate it (1). However, increases in drug-resistant parasites and insecticide-resistant mosquitoes have threatened to worsen the infection control of malaria, leading to a high mortality (2, 3). The development of an effective malaria vaccine remains challenging (4). Therefore, there is an urgent need to establish effective medical countermeasures, including preventative and mitigation efforts, against severe Plasmodium infection that do not rely on antimalarial drugs.

The augmentation of host defense against Plasmodium infections is important for improving patient outcomes. Recent studies have shown that certain vaccines, such as Mycobacterium bovis BCG, an antituberculosis vaccine, can induce changes in the nonspecific innate immune system and exert off-target effects against other pathogens, including malaria (5, 6). This nonspecific protection is thought to be mediated by macrophages through training the innate immune system (7). If activation of innate immunity is effective for eliminating malaria parasites, a reduced malaria mortality can be expected, and instances of drug-resistant malaria parasites can be eliminated. However, previously reported methods of activating innate immunity may also induce an enhanced inflammatory response in the host, resulting in organ damage (8). Augmentation of the elimination of infected cells without enhancing the inflammatory response may be an ideal medical countermeasure against Plasmodium infection.

We recently reported that repeated low-dose lipopolysaccharide (LPS) injection, termed LPS preconditioning or LPS tolerance (9, 10), renders mice drastically resistant to bacterial infection, as such LPS preconditioning potently reduces the host’s inflammatory response to bacterial stimuli while markedly augmenting the host’s bactericidal activity (11). LPS preconditioning increases the population of monocyte-derived macrophages in the liver and enhances their phagocytosis and bactericidal activity.

The liver plays a crucial role in Plasmodium infection in the initial phase. We, therefore, attempted to use these attractive phenomena induced by LPS preconditioning as a novel treatment against Plasmodium infection and investigated whether or not LPS preconditioning can prevent severe Plasmodium infection using a murine Plasmodium infection model.

RESULTS

LPS preconditioning improved mouse survival after lethal Plasmodium yoelii 17XL infection.

LPS preconditioning was induced by intraperitoneal injections of 5, 50, or 500 μg of LPS/kg of body weight for 3 consecutive days in mice, and 24 h after the last LPS injection, they were intravenously infected with the green fluorescent protein (GFP)-expressing Plasmodium yoelii 17XL (PyLGFP) (5 × 10⁴ parasitized red blood cells [pRBCs]). This dose of PyLGFP infection was lethal for nontreated control mice (Fig. 1A to C). However, preconditioning with 5 μg/kg of LPS tended to prolong the murine survival time after PyLGFP infection (Fig. 1A). Interestingly, preconditioning with 50 μg/kg of LPS showed a 20% survival rate in PyLGFP-infected mice (Fig. 1B), and preconditioning with 500 μg/kg LPS resulted in more than half of the infected mice surviving (60% survival) (Fig. 1C), suggesting that LPS preconditioning significantly increased the survival of P. yoelii-infected mice.

FIG 1.

Survival after rodent P. yoelii 17XL infection in LPS-preconditioned mice. LPS preconditioning was induced in mice with i.p. injection of 5 (A), 50 (B), or 500 (C) μg/kg LPS for 3 consecutive days, and 24 h after the last LPS injection, the mice were i.v. infected with 5 × 10⁴ pRBCs of PyLGFP. n = 10 in each group (2 separate experiments with 5 mice per group). Significant differences in survival rate were analyzed by log rank (Kaplan-Meier) test. **, P < 0.01 and *, P < 0.05 versus control.

LPS preconditioning reduced the growth of P. yoelii in mice at 5 days after infection.

Next, we examined the effect of LPS preconditioning (5 or 500 μg/kg of LPS) on the growth of PyLGFP in mice. Although no significant reduction was noted at 3 days, LPS preconditioning with both 5 and 500 μg/kg LPS significantly reduced the parasitemia at 5 days after infection (Fig. 2A and B). We confirmed the significant reduction in parasitemia by LPS preconditioning using a flow cytometric analysis. LPS preconditioning with both 5 and 500 μg/kg LPS significantly reduced the proportion of GFP-positive red blood cells (RBCs), which indicated PyLGFP-parasitized RBCs, at 5 days after infection (Fig. 2C and D). This suggested that LPS preconditioning potently reduced growth of P. yoelii in mice during blood-stage infection (at 5 days), resulting in an improved survival after infection.

FIG 2.

The effects of LPS preconditioning on parasitemia after PyLGFP infection in mice. Mice were similarly subjected to LPS preconditioning with i.p. injection of 5 μg/kg or 500 μg/kg LPS, and 24 h later, they were i.v. infected with 5 × 10⁴ PyLGFP-parasitized RBCs. (A, B) Parasitemia at days 3 and 5 postinfection was measured by Giemsa stain. (C, D) The proportion of PyLGFP-parasitized RBCs was also evaluated at day 5 as the FITC intensity using a flow cytometer. Counting parasitemia and flowcytometric analyses were performed using the double-blind technique. Data are pooled from 2 separate experiments with 5 mice per group (in total, n = 10 in each group) and shown as mean ± SE in panels C and D (right). Representative data are shown with similar results in panels C and D (left). Statistics were calculated by nonparametric Mann-Whitney U test (A to D); *, P < 0.05 versus control.

LPS preconditioning augmented the phagocytic clearance of pRBCs by CD11b^high F4/80^low macrophages in the murine liver but not the spleen.

We examined the effect of LPS preconditioning on the phagocytic clearance of pRBCs in murine liver mononuclear cells (MNCs). Liver MNCs were obtained from the LPS-preconditioned mice 24 h after the last LPS injection to use for the following examinations. After coculture with pRBCs for 16 h, the liver MNCs of the LPS-preconditioned mice (with 500 μg/kg LPS) significantly reduced the number of pRBCs in the culture medium compared with those of the control mice (Fig. 3A), suggesting that LPS preconditioning augmented the phagocytic clearance of pRBCs by liver MNCs. We also examined the effect of LPS preconditioning on the phagocytic clearance of pRBCs in the splenocytes. However, LPS preconditioning did not augment phagocytic clearance of pRBCs by whole splenocytes (see Fig. S1A in the supplemental material).

FIG 3.

An in vitro phagocytosis assay. Liver MNCs (A) or sorted CD11b^high F4/80^low liver macrophages (B) were obtained from the LPS (500 μg/kg)-preconditioned mice and cocultured with PyLGFP-parasitized RBCs for 16 h. Thereafter, residual pRBCs were counted. Data are pooled from two separate experiments with three mice per group and shown as the mean ± SE. Statistical analyses were performed using Mann-Whitney U test (A) and one-way ANOVA (B). ***, P < 0.001; **, P < 0.01; *, P < 0.05.

Next, we examined the phagocytic clearance of pRBCs by the CD11b^high F4/80^low liver macrophages, which are markedly increased by LPS preconditioning (11). We sorted the CD11b^high F4/80^low liver macrophages and other liver MNCs of the LPS-preconditioned mice (with 500 μg/kg LPS) as shown in Fig. S2A in the supplemental material and cocultured these cells (5 × 10⁵ cells) with pRBCs for 16 h. The CD11b^high F4/80^low liver macrophages of the LPS-preconditioned mice significantly reduced the number of residual pRBCs in the culture medium compared with those of other MNC subsets in the LPS-preconditioned mice, although the non-CD11b^high F4/80^low subset showed a significant reduction in residual pRBCs compared to those of the whole-liver MNCs of the control mice (Fig. 3B). We also sorted the CD11b^high F4/80^low spleen macrophages and other splenocytes from LPS-preconditioned mice (Fig. S2B) and cocultured these cells with pRBCs for 16 h. Unlike the liver MNCs, the CD11b^high F4/80^low spleen macrophages and other splenocytes did not augment the phagocytic clearance of pRBCs compared to that of the whole splenocytes of the control mice (Fig. S1B). These findings suggested that LPS preconditioning stimulated liver MNCs, particularly CD11b^high F4/80^low liver macrophages, encouraging them to augment the phagocytic clearance of pRBCs.

LPS preconditioning did not reduce the plasma TNF levels after P. yoelii 17XL infection but delayed the peak of plasma IFN-γ after infection.

LPS preconditioning markedly reduces the tumor necrosis factor (TNF) and gamma interferon (IFN-γ) secretion in mice after bacterial infection (11). We then examined these cytokine responses to P. yoelii 17XL infection in the LPS-preconditioned mice. Unlike bacterial infection, no significant difference in the plasma TNF levels was observed after P. yoelii 17XL infection between the control and LPS-preconditioned mice with 5 or 500 μg/kg LPS (Fig. 4A and B). Nevertheless, LPS-preconditioned mice tended to increase the plasma TNF levels beyond 7 days after P. yoelii 17XL infection, albeit not to a significant degree (Fig. 4A and B). The plasma IFN-γ levels peaked at 3 days after PyLGFP infection in control mice, whereas LPS preconditioning with 5 and 500 μg/kg LPS both delayed those IFN-γ peaks, which instead peaked at 5 days after infection (Fig. 4C and D).

FIG 4.

Changes in the plasma cytokine levels after PyLGFP infection in LPS-preconditioned mice and control mice. LPS-preconditioned mice with 5 μg/kg LPS (A, C) or 500 μg/kg LPS (B, D) and control mice were i.v. infected with 5 × 10⁴ PyLGFP-parasitized RBCs to examine the plasma IFN-γ and TNF. Data are shown as the mean ± SE from 10 mice in each group (2 separate experiments with 5 mice per group). Data were evaluated by nonparametric Mann-Whitney U test (A to D). **, P < 0.01 and *, P < 0.05 versus control (at the same time point).

Parasitized RBC-phagocytosing macrophages were obviously observed in the liver of LPS-preconditioned mice 5 days after P. yoelii 17XL infection.

In the pathological examination of hematoxylin-eosin (H.E.) staining, samples were obtained from the lateral left lobes in the liver. RBC phagocytosis by macrophage, which was indicated as a hemosiderin-laden macrophage, was markedly observed around the mid-zone of the liver in LPS-preconditioned mice (with 500 μg/kg LPS) compared with control mice (Fig. 5). To confirm the phagocytosis of PyLGFP-parasitized RBCs in the murine liver, we performed the immunohistochemical staining of GFP in the liver of PyLGFP-infected mice. Consistent with H.E. staining, positive staining of GFP in liver macrophages, which indicates phagocytosis of pRBCs, was also obviously observed in the LPS-preconditioned mice compared with the control mice (Fig. 6).

FIG 5.

Results of a pathological examination of the liver after PyLGFP infection. Liver sections of LPS (500 μg/kg)-preconditioned mice and control mice were stained with H.E. Data are representative of three mice in each group with similar results. Right columns show magnified images of the yellow squares in the left columns. Arrowheads indicate RBC-phagocytosed macrophages.

FIG 6.

Results of an immunohistochemical analysis of pRBC-phagocytosed macrophages in the liver. Liver sections of LPS (500 μg/kg)-preconditioned mice and control mice were stained with anti-GFP Ab, as described in Materials and Methods. Arrowheads indicate cells with GFP-positive staining, which are pRBCs. Data are representative of three mice in each group with similar results. Right columns show magnified images of the squares in the left columns. The lower right frame is a magnified image of pRBCs phagocytosed by macrophages.

Adoptive transfer of CD11b^high F4/80^low liver macrophages from LPS-preconditioned mice induced resistance to P. yoelii 17XL infection in control mice.

To confirm the effect of CD11b^high F4/80^low liver macrophages, which are induced by LPS preconditioning, on resistance to P. yoelii 17XL infection, we performed the adoptive transfer of CD11b^high F4/80^low liver macrophages from the LPS-preconditioned mice (with 500 μg/kg LPS) to control mice and then infected these mice with PyLGFP. Adoptive transfer of CD11b^high F4/80^low liver macrophages from the LPS-preconditioned mice to the control mice significantly prolonged the survival time after P. yoelii 17XL infection, although the adoptive transfer of other liver MNC subset cells did not improve survival (Fig. 7A). There were no significant differences in the plasma IFN-γ levels among the groups (Fig. 7B). However, transfers of CD11b^high F4/80^low liver macrophages from LPS-preconditioned mice delayed the peak of plasma IFN-γ at 5 days after P. yoelii 17XL infection (Fig. 7B). Transfers of other liver MNCs from LPS-preconditioned mice also delayed the plasma IFN-γ peak at 5 days after infection (Fig. 7B), although they did not affect mouse survival (Fig. 7A). Adoptive transfer of CD11b^high F4/80^low liver macrophages from LPS-preconditioned mice tended to increase the plasma TNF levels beyond 7 days after infection in comparison to those of the other groups, albeit not to a significant degree (Fig. 7C). These results support the notion that CD11b^high F4/80^low liver macrophages, which are a key fraction in the induction of LPS preconditioning, play a major role in resistance to Plasmodium infection.

FIG 7.

Adoptive transfer of CD11b^high F4/80^low liver macrophages from LPS-preconditioned mice to control mice. CD11b^high F4/80^low liver macrophages and CD45⁺ liver MNCs (except for the CD11b^high F4/80^low subset) were obtained from the LPS-preconditioned mice (with 500 μg/kg LPS) and transferred to control mice. Thereafter, mice were i.v. infected with 5 × 10⁴ PyLGFP-parasitized RBCs to monitor their survival (A), plasma IFN-γ (B), and TNF (C) levels. The number of mice in each group is indicated (three separate experiments with three or four mice per group). Data are shown as the mean ± SE. Significant differences in survival rate were analyzed using log rank (Kaplan-Meier) test (A). *, P < 0.01 versus control mice and P < 0.05 versus mice with transferred CD45⁺ liver MNCs except for CD11b^high F4/80^low cells. Statistical analyses were also performed using Mann-Whitney U test (B, C), whereas no significant differences were observed among three groups.

DISCUSSION

LPS preconditioning significantly augmented the phagocytic clearance of PyLGFP-parasitized RBCs by CD11b^high F4/80^low liver macrophages (Fig. 3B), which are monocyte-derived macrophages increased by LPS preconditioning (11), resulting in the improvement of the murine survival after PyLGFP infection (Fig. 1). We confirmed the beneficial effect of CD11b^high F4/80^low liver macrophages on the survival of PyLGFP-infected mice through the adoptive transfer of these macrophages (Fig. 7).

Although the T cell-mediated immune response to initial Plasmodium infection is well described, other immune cells, such as dendritic cells and monocytes/macrophages, have been shown to modulate immune activation and the severity of disease as well (12, 13). However, mice lacking tissue-resident macrophages experience increased malaria-related complications, such as disruptions in the blood-brain barrier, increased vascular permeability in the liver, and increased accumulation of hemozoin pigment in the lung (14). These studies imply a critical role for macrophages in the initial response to Plasmodium infection. The liver stage of malaria is the first phase of infection in the host, and the liver-resident macrophages, namely, Kupffer cells, play important roles in ameliorating the severity of Plasmodium infection and preventing parasite release into the blood circulation (15).

Spleen macrophages also play a predominant role in clearing pRBCs during Plasmodium infection. Nevertheless, the current LPS preconditioning induced by intraperitoneal LPS injections did not augment the phagocytosis of pRBCs by splenocytes (see Fig. S1A in the supplemental material). We previously reported that LPS preconditioning is more effective in the liver than in other organs, such as the spleen or lung (11), although the current LPS preconditioning regimen (500 μg/kg of LPS) increased the proportion of CD11b^high F4/80^low macrophages in the spleen (LPS preconditioning, 12% versus normal control, 5% on average) as well as liver (18% versus 10%). However, unlike the CD11b^high F4/80^low liver macrophages, the CD11b^high F4/80^low spleen macrophages did not augment the phagocytosis of pRBCs (Fig. S1B). Intraperitoneally administered LPS is presumably absorbed into the mesenterium and reaches the liver via portal vein. It may be one of the reasons why the LPS preconditioning activated pRBC phagocytosis of liver macrophages rather than spleen macrophages. However, further study should be performed to reveal the differences in the characteristics of liver and spleen macrophages, in particular CD11b^high F4/80^low macrophages induced by LPS preconditioning, while we should also take into careful consideration the manner of LPS administration. We also hope to develop an adjuvant therapy that potently enhances the phagocytic activity of spleen macrophages and intend to investigate its effect on Plasmodium infection in the future.

LPS preconditioning may stimulate the CD11b^high F4/80^low liver macrophages to augment phagocytic activity against bacteria, which are external pathogens (11). These monocyte-derived macrophages also showed strongly upregulated phagocytic activity for pRBCs by LPS preconditioning, thereby reducing the growth of parasites in mice during the early stage of infection (Fig. 2 and 3). A previous report showed that opsonin-independent phagocytosis contributes to the regulation of infected erythrocytes in the early stage of Plasmodium infection (16). The lack of parasite-specific antibody production during the early stages of infection and the results of the present in vitro phagocytosis assay suggest that opsonin-dependent phagocytosis is likely not involved in controlling parasite burden during early infection stages in the liver and then suggest that a role for opsonin-independent phagocytosis in controlling parasite burden. This indicates the importance of opsonin-independent efficient phagocytosis of infected erythrocytes, which was induced by the phagocytic activity of macrophages enhanced by LPS preconditioning.

Although LPS preconditioning markedly reduces the response of proinflammatory cytokines, such as TNF, to bacterial stimuli (10, 11), it did not decrease TNF secretion after P. yoelii 17XL infection in mice (Fig. 4). LPS preconditioning tended to conversely increase the plasma TNF levels beyond 7 days after infection, albeit not to a significant degree (Fig. 4). Such tendency of the elevation of plasma TNF beyond 7 days was also observed in the mice receiving adoptive transfer of CD11b^high F4/80^low liver macrophages from LPS-preconditioned mice (Fig. 7). Regarding bacterial infection, challenge with bacteria, such as Escherichia coli, evokes a steep elevation of plasma TNF in mice immediately after the challenge, which causes endotoxin shock, and LPS preconditioning drastically reduces this steep elevation of TNF (10, 11). Although macrophages evoke a certain proinflammatory response at the initial stage of blood parasitemia, Plasmodium infection does not induce such a steep elevation of TNF like that after bacterial challenge (17, 18), so the drastic reduction of TNF secretion by LPS preconditioning may not be induced in the current Plasmodium infection.

IFN-γ is one of the most powerful proinflammatory cytokines for host defense against intracellular parasitic pathogens (19). An IFN-γ-induced tissue inflammatory reaction is required for effective bacterial elimination by the host; however, an IFN-γ-induced exaggerated inflammatory response can also cause sepsis and multiorgan injuries (20). LPS preconditioning did not reduce the peak of plasma IFN-γ but did delay its peak from 3 days to 5 days after P. yoelii 17XL infection (Fig. 4C and D). Such delay in the peak of plasma IFN-γ after P. yoelii 17XL infection was also observed in the adoptive transferred mice with CD11b^high F4/80^low liver macrophages from LPS-preconditioned mice (Fig. 7). The result showing the delay in IFN-γ peak production with LPS preconditioning may be tied to the reduction in parasite burden seen in these mice. However, mice deficient in IFN-γ or its receptor were reported to have strongly suppressed protective immunity against infection with nonlethal malaria parasites (21, 22). In contrast, these mice have shown significant reductions in cerebral malaria and liver damage caused by infection with lethal malaria parasites (23–25). Thus, IFN-γ production during Plasmodium infection may have a significant impact on both protective immunity and disease severity in Plasmodium infection. However, the rationale of this delayed peak of IFN-γ induced by LPS preconditioning should be investigated further in a future study.

Finally, we should address the limitations associated with the current study. We do not believe that the current LPS preconditioning regimen can be directly applied to the clinical setting. Precise investigations into the timing and effective duration of LPS preconditioning are important for clinical application. However, we do not have precise information on how long LPS preconditioning could sustain the activation of innate immunity. In the clinical settings, a possible first step would be a short-term application to people living outside of areas where malaria is endemic, such as travelers. In addition, toxicity due to LPS (even low-dose inoculation) should be avoided. Additionally, although a correlation was observed between the priming dose of LPS and the overall effects on survival, no dose dependency of the effect of LPS priming on the parasitemia or cytokine response was observed. This suggests that other mechanisms, not addressed in this study, may be involved in the improvement of survival. We should address these important issues in a future study. Although we confirmed that LPS preconditioning was effective in E. coli infection in female mice (unpublished data), we have not examined the effect of LPS preconditioning on Plasmodium infection in female mice. It is widely known that females and males have different immune responses and susceptibilities to infections and diseases. Moreover, nonspecific effects of vaccines and trained immunity have been shown to be different in females than in males. Therefore, the effect of LPS preconditioning in female mice should be investigated.

In conclusion, LPS preconditioning rendered mice resistant to rodent Plasmodium infection. CD11b^high F4/80^low liver macrophages that are induced by LPS preconditioning enhance the phagocytic clearance of parasitized RBCs, which may contribute to improving the survival in Plasmodium-infected mice. Thus, LPS preconditioning potently augmented phagocytic activity of liver macrophages (CD11b^high F4/80^low) against not only invading bacteria but also Plasmodium-infected RBCs and improved the survival of infected hosts (animals). Training/enhancement of the host innate immunity by LPS preconditioning may become an effective prophylaxis strategy against not only bacterial but also plasmodium infection.

MATERIALS AND METHODS

The Ethics Committee of Animal Care and Experimentation in National Defense Medical College Japan approved all requests for animals and the intended procedures of the present study (permission number 18025).

Animals and reagents.

Male C57BL/6 mice (8 weeks old, body weight of 20 g) were purchased from Japan SLC (Hamamatsu, Japan) and used for this study. LPS (E. coli 0111:B4) was purchased from Sigma-Aldrich (St. Louis, MO, USA) to use for LPS preconditioning. The P. yoelii 17XL lethal strain has been kept in our laboratory as described elsewhere (26). We created a stable GFP-expressing P. yoelii 17XL (PyLGFP) by transfection of an uncloned population of P. yoelii 17XL parasites with the vector pL0016 as described previously (27, 28).

We obtained pRBCs of PyLGFP from donor mice after intravenous (i.v.) inoculation with a frozen stock of parasites. After i.v. injection of PyLGFP, we checked the parasitemia of the donor mice daily. Thereafter, the pRBCs of PyLGFP were obtained at the proliferation phase of parasitemia for use in the current experiment (parasite rates of PyLGFP were 10% to 15% of RBCs).

Infection of rodent malaria (PyLGFP) in mice.

After obtaining pRBCs of PyLGFP from the donor mice, 5 × 10⁴ pRBCs with 0.2 ml RPMI 1640 medium were i.v. injected into recipient mice 24 h after the last injection of LPS (LPS-preconditioned mice). The percentage of pRBCs in the recipient mice was monitored by thin tail blood smears stained with Giemsa stain. The murine survival was monitored every day.

Induction of in vivo LPS preconditioning.

LPS preconditioning was induced in mice by an intraperitoneal (i.p.) injection of 5, 50, or 500 μg/kg of LPS (dissolved in 0.5 ml saline) once daily for three days as we previously described (11). Control mice were similarly i.p. injected with saline (0.5 ml) three times.

Analyses of PyLGFP-parasitized RBCs using flow cytometry.

RBCs were obtained from the PyLGFP-infected mice at 5 days after infection. Percentages of GFP-positive RBCs were evaluated as fluorescein isothiocyanate (FITC) intensity using flow cytometer (ACEA Biosciences, San Diego, CA, USA). After gating RBCs, GFP-positive RBCs were detected as FITC positive.

Isolation of liver and spleen MNCs from mice.

Liver MNCs, including macrophages, were obtained from mice as described elsewhere (29). In brief, under deep isoflurane anesthesia, the liver was removed and minced with scissors. After shaking with 10 ml of Hanks’ balanced salt solution containing 0.05% collagenase (type IV; Sigma-Aldrich) for 20 min at 37°C, liver specimens were filtered through mesh, suspended in 33% Percoll solution (Sigma-Aldrich) containing 10 U/ml heparin, and centrifuged for 15 min at 500 × g at room temperature. After lysing RBCs, the remaining cells were washed twice to obtain the liver MNCs. The spleen was then removed and filtered through the mesh. After lysing RBCs again, the specimens were washed twice to obtain spleen MNCs.

Sorting CD11b^high F4/80^low liver and spleen macrophages.

LPS preconditioning increases the number of CD11b^high F4/80^low macrophages in the murine liver and potently augments their bactericidal activity (11). To sort this CD11b^high F4/80^low subset from the liver MNCs, obtained liver MNCs were stained with FITC-conjugated anti-F4/80 monoclonal antibody (MAb) (clone BM8; eBioscience, San Diego, CA, USA), PE-conjugated anti-CD11b MAb (clone M1/70, eBioscience), and APC-conjugated anti-CD45 MAb (clone 30-F11; eBioscience). Thereafter, CD11b^high F4/80^low CD45⁺ cells were sorted using a Sony SH800 cell sorter (Sony, Tokyo, Japan) (see Fig. S2 in the supplemental material). We also sorted CD45⁺ MNCs except for the CD11b^high F4/80^low cell subset from the liver MNCs using a cell sorter. CD11b^high F4/80^low CD45⁺ cells as well as other subsets in the spleen MNCs were similarly sorted using a cell sorter.

In vitro phagocytic clearance of PyLGFP-parasitized RBCs by liver/spleen MNCs or CD11b^high F4/80^low liver/spleen macrophages.

To examine the pRBC phagocytic clearance by liver or spleen MNCs, liver or spleen MNCs (5 × 10⁵ cells/200 μl) were obtained from LPS-preconditioned mice 24 h after the last LPS injection or control nontreated mice and then cocultured with 1 × 10⁷ pRBCs of PyLGFP (PyLGFP parasite rates were 10% to 15% of RBCs) in antibiotic-free RPMI 1640 medium for 16 h. The CD11b^high F4/80^low liver macrophages or other CD45⁺ liver MNCs were sorted from the liver MNCs in mice 24 h after the last LPS injection. Thereafter, these sorted cells (5 × 10⁵ cells/200 μl) were similarly cocultured with 1 × 10⁷ pRBCs for 16 h. After coculture for 16 h, the erythrocytes were obtained from the culture medium in order to count the total number of RBCs using an Erma PCE 170 hematology analyzer (Erma, Tokyo, Japan). Thereafter, these cells were stained with Giemsa stain and over 300 RBCs were counted using a light microscope to calculate the proportion of residual parasitized erythrocytes. We then obtained the number of residual PyLGFP-parasitized RBCs in the culture medium. We also similarly sorted CD11b^high F4/80^low spleen macrophages or other CD45⁺ spleen MNCs and cocultured them with 1 × 10⁷ pRBCs for 16 h to examine their phagocytic clearance of pRBCs.

Adoptive transfer of CD11b^high F4/80^low liver macrophages from the LPS-preconditioned mice to control mice.

To examine the effect of CD11b^high F4/80^low liver macrophages that are induced by LPS preconditioning on the P. yoelii 17XL infection, CD11b^high F4/80^low CD45⁺ cells were sorted from the liver MNCs of the LPS-preconditioned mice using the cell sorter. The CD45⁺ liver MNCs except the CD11b^high F4/80^low subset were also sorted from the liver MNCs of the LPS-preconditioned mice. Thereafter, these sorted cells (1 × 10⁶ cells/200 μl phosphate-buffered saline [PBS]) were adoptively transferred into the recipient normal mice, and 5 × 10⁴ pRBCs (with 0.2 ml RPMI 1640 medium) were subsequently i.v. injected into recipient mice.

Measurements of plasma cytokines in mice.

Blood samples were obtained from the mice via submandibular bleeding using 5-mm GoldenRod animal lancets (Medipoint Inc., Mineola, NY, USA). Plasma TNF and IFN-γ levels were measured using their respective enzyme-linked immunosorbent assay (ELISA) kits (BD OptEIA; BD Biosciences, San Diego, CA, USA).

Pathological analyses of the liver in the rodent malaria-infected mice.

Under deep isoflurane anesthesia, the livers were removed from the mice 5 days after PyLGFP infection. Samples were obtained from the lateral left lobe and fixed in 4% formaldehyde, embedded in paraffin, and sectioned at a thickness of 3 μm. The sections were subjected to H.E. staining or an immunohistochemical analysis. Regarding immunohistochemical staining of GFP, anti-GFP polyclonal antibody (1:500; MBL, Nagoya, Japan) was used as the primary antibody, horseradish peroxidase (HRP)-labeled anti-rabbit IgG antibody (Nichirei, Tokyo, Japan) was used as the secondary antibody, and diaminobenzidine was used for the colorimetric reaction. Thereafter, nuclei of hepatocytes and liver macrophages were counterstained with hematoxylin. These procedures were outsourced to LSI Medience (Tokyo, Japan).

Statistical analyses.

Statistical analyses were performed using the GraphPad Prism software program version 8 (GraphPad Software, San Diego, CA, USA). The data were shown as the means ± standard error (SE) in Fig. 2C and D and Fig. 3A and B or were shown as the medians and quartiles in Fig. 4A to D and Fig. 7B and C. The survival curves were made using the Kaplan-Meier methods, and differences in survival rates were compared by a log rank test. Mann-Whitney U test was performed for comparison between two groups, and a one-way analysis of variance (ANOVA) followed by Bonferroni’s post hoc comparison test was performed for comparisons among more than three groups. Appropriate statistical tests were performed as described in the figure legends. Statistical significance levels are indicated as follows: *, P < 0.05; **, P < 0.01; ***, P < 0.001; NS, not significant.