Plasmodium berghei Purified Hemozoin Associated with DNA Strongly Inhibits P. berghei Liver-Stage Development in BALB/c Mice after Intravenous Inoculation

ABSTRACT

In the acidic lysosome-like digestive vacuole, Plasmodium parasites crystallize heme from hemoglobin into hemozoin, or malaria pigment. Upon release of progeny merozoites, the residual hemozoin is phagocytized by macrophages principally in the liver and spleen where the heme crystals can persist for months to years, as heme oxygenase does not readily degrade the crystal. Previous studies demonstrated hemozoin modulation of monocytes and macrophages. Hemozoin modulates immune function activity of monocytes/macrophages. Here, we used purified/washed hemozoin (W-Hz) isolated from murine Plasmodium berghei infections and intravenously (i.v.) injected it back into naive mice. We characterized the modulating effect of W-Hz on liver-stage replication. Purified washed hemozoin decreases P. berghei liver levels both at 1 week and 1 month after i.v. injection in a dose and time dependent fashion. The injected hemozoin fully protected in nine out of 10 mice given a 50 sporozoite inoculum, and in 10 out of 10 mice against 2,000 sporozoites when they were infected an hour or a day after hemozoin inoculation. DNase treatment at the hemozoin reversed the observed liver load reduction. The liver load reduction was similar in mature B- and T-cell-deficient RAG-1 knockout (KO) mice suggesting an innate immune protection mechanism. This work indicates a role for residual hemozoin in down modulation of Plasmodium liver stages.

INTRODUCTION

In 2020, 241 million cases of malaria were reported worldwide (1). Although malaria elimination efforts have mainly been focused on drug and vaccine development, certain immunological mechanisms behind Plasmodium infections remain to be fully understood. After asymptomatic liver-stage infection, Plasmodium parasites next invade erythrocytes in the symptomatic asexual blood stage. In this stage, the parasite relies on the host’s hemoglobin as its main source of amino acids. During hemoglobin digestion, a toxic form of heme is released. Plasmodium avoids parasitic oxidative stress by crystallizing free heme into a compound known as hemozoin, or malaria pigment (2, 3). The hemozoin crystallization pathway is the main target for quinoline drugs, which interfere with hemozoin formation, leading ultimately to the parasite’s death (2, 4).

Hemozoin is released into circulation after erythrocyte lysis and merozoite replication. After release into the bloodstream, hemozoin makes its way to the monocyte-macrophage system (5). In vitro and in vivo assays have shown that macrophages phagocytize hemozoin during Plasmodium infection (5, 6). Hemozoin has also been shown to modulate the function of monocytes in vitro (7). Over the last couple of decades, there have been contradictory results regarding hemozoin’s modulation of the immune response. Hemozoin has been described as an immunologically inert crystal (8, 9) or immunological activator (10–12) in mice. In the context of humans, hemozoin has been shown to accumulate in the tissues such as the liver, spleen, or brain (13–15), although its direct relationship with the immune system has yet to be described. Overall, the role of hemozoin in the host immune response to Plasmodium infection remains to be fully characterized. Here, we look at the immune-modulatory effects on liver stages of SDS-washed, Proteinase K-treated purified hemozoin in the absence of a previous Plasmodium infection.

RESULTS

SDS washed, proteinase K treated hemozoin (W-Hz) is present in the liver one week after i.v. inoculation.

Mice were inoculated with 750 nmol of W-Hz intravenously (i.v.) To confirm the presence of W-Hz in the liver, we harvested livers 1 week after W-Hz inoculation. Hemozoin is visible under low or high magnifications in the absence of fluorophores or dyes. We saw that there was no hemozoin present in naive mice (Fig. 1A), and that W-Hz was present in the liver 1 week after inoculation (Fig. 1B). On average, we recovered 90% of the 750 nmol that we inoculated 1 week prior, with 87% coming from the liver alone (data not shown). Our data show that W-Hz is still present in the liver 1 week postinoculation.

FIG 1.

W-Hz is present in the liver 1 week after i.v. inoculation. BALB/c WT mice were inoculated with 750 nmol of W-Hz i.v. 1 week before harvesting their livers. (A) Liver section at ×100 magnification of a naive BALB/c WT mouse stained with H&E. (B) Liver section at a ×100 magnification of a mouse inoculated with 750 nmol of W-Hz stained with H&E.

Inoculation with W-Hz decreases parasite liver load.

To determine if a W-Hz inoculation could influence the development of liver stages after a sporozoite challenge, we first injected mice with 750 nmol of W-Hz i.v.; 1 week or 1 month later, we infected them with 2,000 Pb-ANKA-Luc sporozoites i.v. The parasite load in W-Hz-laden livers was 24% of the parasite load in the naive group after 1 week (Fig. 2A), and 46% of the naive group after 1 month (Fig. 2B). To address whether W-Hz decreases parasite liver load depending on the dosage, we inoculated mice with 3 different doses of W-Hz (250, 500, or 750 nmol), and infected them with 2,000 Pb-ANKA-Luc sporozoites 1 week later. We saw that the parasite load in the group inoculated with 750 nmol of W-Hz was 13% of the naive group, and 33% in the group inoculated with 500 nmol (Fig. 2C). Additionally, the parasite liver load in the group that received 250 nmol of W-Hz was 56% of the naive control, but our nonparametric Mann-Whitney test showed that the difference was not statistically significant. Altogether, our data suggest that dose-dependent W-Hz inoculation decreases parasitic liver load 1 week or 1 month postinoculation.

FIG 2.

W-Hz inoculation decreases parasite liver load. (A and B) BALB/c WT mice were inoculated with 750 nmol of W-Hz i.v. and rested for 1 week or 1 month before receiving a 2,000-Pb-ANKA-Luc sporozoite i.v. injection. Parasite liver load was measured 42 h postinfection in photons per second (total flux). The nonparametric, unpaired Mann-Whitney test showed that there was a statistically significant difference between the naive control group and the group inoculated with W-Hz 1 week or 1 month before sporozoite infection. **, P ≤ 0.01;  n = 20. (C) Mice were inoculated with 250, 500, or 750 nmol of W-Hz i.v. and infected with 2,000 Pb-ANKA-Luc sporozoites 1 week later. Parasite liver load was measured 42 h postinfection in photons per second. The nonparametric, unpaired Mann-Whitney test gave a statistically significant difference between the naive control group and the groups inoculated with 500 or 750 nmol of W-Hz 1 week before sporozoite infection. The difference was higher in mice inoculated with 750 nmol of W-Hz. **, P ≤ 0.01; ns, P > 0.05; n  = 20. Error bars correspond to the standard error of the mean (SEM).

Inoculation with W-Hz fully protects against a sporozoite infection.

To address whether an infection with a dose of sporozoites similar to a mosquito feed (16, 17) could fully protect mice inoculated with W-Hz, we first injected mice 750 nmol of W-Hz i.v.; 1 week later, we infected them with 50 Pb-ANKA-Luc sporozoites i.v. We allowed the infection to progress to the erythrocytic cycle of the parasite and followed mice parasitemia for 13 days. No parasites were detected over the course of 13 days postmerogony in the group that was inoculated with W-Hz (Fig. 3A). We repeated the experiment and 4 out of 5 mice were fully protected in the group that was inoculated with 750 nmol of W-Hz (Fig. 3B). There is, however, a 2-day delay in patency between the single infected mouse with W-Hz and the naive group of mice. Full protection was confirmed by qPCR on day 13 postmerogony in the 4 fully protected mice (data not shown). We further verified full protection by performing a passive blood transfer on day 13 from the fully protected mice to naive mice. The naive mice that received the passive blood transfer remained uninfected for 15 days post-blood transfer (data not shown). Interestingly, mice inoculated with W-Hz and then challenged with 2,000 Pb-ANKA-Luc sporozoites 1 week later are not fully protected, but they display a 1-day delay in patency compared to the naive group (Fig. 3C), indicating approximately 10-fold lower liver-stage levels. These data suggest that inoculation with W-Hz can fully protect mice from when challenged with lower doses of sporozoites and can decrease Plasmodium liver levels after high sporozoite infections.

FIG 3.

Inoculation with W-Hz confers full protection against a sporozoite infection. (A and B) BALB/c WT mice were inoculated i.v. with 750 nmol of W-Hz before a 50 Pb-ANKA-Luc sporozoite infection 1 week later. Parasitemias were followed over the course of 13 days postmerogony. (A) 5/5 mice inoculated with W-Hz were fully protected. Each data point corresponds to the group’s mean. Error bars correspond to the SEM. n = 10. (B) 4/5 mice were fully protected. Each data point corresponds to an individual mouse. There was a 2-day delay in patency between the infected mouse inoculated with W-Hz and the naive controls. n = 10. (C) Mice inoculated i.v. with 750 nmol of W-Hz 1 before a 2,000-sporozoite infection showed a 2-day delay in patency compared to the naive group. Each data point corresponds to the group’s mean. Error bars correspond to the SEM. n = 8. (D) Groups of BALB/c WT mice were inoculated i.v. with 750 nmol of W-Hz before a 2,000-Pb-ANKA-Luc sporozoite infection 1 week, 1 day, or 1 h later. Parasite liver load was measured 42 h postinfection in photons per second (total flux). Error bars correspond to the SEM. The nonparametric, unpaired Mann-Whitney test showed that there was a statistically significant difference between the naive control group and the groups inoculated with W-Hz 1 week, 1 day, or 1 h before sporozoite infection. ***, P ≤ 0.001; n = 20. (E) Parasitemias were followed over the course of 13 days postmerogony in the naive group, and the groups infected 1 day or 1 h after W-Hz inoculation. 5/5 mice in the 1-day W-Hz group were fully protected, and 5/5 mice in the 1-h W-Hz group were fully protected. Each data point corresponds to the group’s mean. Error bars correspond to the SEM. n = 15.

To determine if we could achieve full protection against a 2,000-sporozoite infection, we shortened the time of sporozoite infection after W-Hz inoculation. First, we injected mice with 750 nmol of W-Hz i.v. Groups of mice were infected i.v. with 2,000 sporozoites 1 h, 1 day, or 1 week post W-Hz inoculation. The parasite load in W-Hz-laden livers was 25% of the liver load in the naive group after 1 week, 1.3% of the naive group after 1 day, and 1.2% of the naive group after 1 h (Fig. 3D). The numbers in photons/sec in the groups with 1 day and 1 h W-Hz incubation correspond to those of noninfected mice. Parasitemia was followed to verify full protection in those groups. No parasites were detected after 13 days postmerogony in the groups with 1 day or 1 h W-Hz incubation (Fig. 3E). We confirmed full protection by qPCR and passive transfer of blood (data not shown). Overall, these data suggest that inoculation with W-Hz can fully protect mice against a 2,000-sporozoite infection in a time-sensitive fashion, and also when mice are challenged with lower doses of sporozoites.

W-Hz decrease in parasite liver load is mediated by DNA and the innate immune response.

To determine if DNA was associated with W-Hz, we extracted DNA out of 1,000 nmol of W-Hz and amplified the Pb18S rRNA gene (Fig. 4A). To evaluate if the DNA associated with W-Hz played a role in decreasing liver load, we treated W-Hz with DNase and inoculated it i.v. in mice. Mice were inoculated with 750 nmol of W-Hz or DNase-Hz and rested for a week before challenge with 2,000 Pb-ANKA-Luc sporozoites. Importantly, there was a significant difference in parasite liver load between the W-Hz and the DNase-Hz groups (P ≤ 0.01), with values in W-Hz-laden livers corresponding to 25%, while the DNase-Hz group was 83% compared to the naive group (Fig. 4B). The difference in liver burden between the naive and the DNase-Hz group was not statistically significant. Additionally, we injected 50 mcgm of P. berghei DNA per mouse without hemozoin and saw no decrease in liver-stage P. berghei challenged after 1 week (Fig. 3). These results indicate that P. berghei DNA is associated with W-Hz, and the decrease in parasite liver load caused by W-Hz is largely mediated by DNA.

FIG 4.

W-Hz decrease in parasite liver load is mediated by DNA and the innate immune response. (A) DNA was extracted out of 1,000 nmol of W-Hz. P. berghei DNA was amplified from the remnants of the extracted DNA using qPCR. The individual samples were processed in 3 technical replicates. Each data point corresponds to the mean of the three replicates. Error bars correspond to the SEM. 1,000 nmol of hemin and ultrapure H₂O were used as a negative and nontemplate control, respectively. 1 × 10⁷ P. berghei genomic DNA copies/μL were used as positive control. (B) BALB/c WT mice were inoculated with 750 nmol of W-Hz or W-Hz treated with DNase I (DNase-Hz) i.v., and infected with 2,000 Pb-ANKA-Luc sporozoites 1 week later. Parasite liver burden was measured 42 h postinfection in photons per second. Error bars correspond to the SEM. The nonparametric, unpaired Mann-Whitney test showed that there was a statistically significant difference between the naive controls and the mice inoculated with W-Hz, and between the mice inoculated with DNase-Hz and the ones inoculated with W-Hz. The difference between the naive group and the DNase-Hz group was not significant. **, P ≤ 0.01; ns, P > 0.05; n = 15. (C) RAG-1 KO and BALB/c WT mice were inoculated with 750 nmol of W-Hz IV and infected with 2,000 Pb-ANKA-Luc sporozoites 1 week later. Parasite liver burden was measured 42 h postinfection. Error bars correspond to the SEM. The nonparametric, unpaired Mann-Whitney test showed that there was a statistically significant difference between both RAG-1 KO naive and WT naive, and the groups inoculated with W-Hz. **, P ≤ 0.01; n = 20. (D) Groups of BALB/c WT mice were inoculated with 750 nmol of W-Hz or W-PfHz, and infected with 2000 Pb-ANKA-Luc sporozoites 1 week later. Parasite liver burden was measured 42 h postinfection in photons per second. Error bars correspond to the SEM. The nonparametric, unpaired Mann-Whitney test showed that there was a statistically significant difference between the naive controls and the mice inoculated with W-Hz or W-PfHz. The difference between the W-Hz and the W-PfHz groups was not significant. **, P ≤ 0.01; *, P ≤ 0.05; ns, P > 0.05; n = 15.

Additionally, we wanted to verify if the decrease in parasite liver load was mediated by innate immune cells. We tested our W-Hz inoculation model in RAG-1 KO mice. One week after inoculation with 750 nmol of W-Hz, mice were infected with 2,000 Pb-ANKA-Luc sporozoites. We saw that the parasite liver load in RAG-1 KO inoculated with W-Hz was 21% of the liver load in the naive RAG-1 KO mice. In wild-type (WT) mice, we again saw that the parasitic load in W-Hz-laden livers was 25% of the naive mice’s livers. (Fig. 4C).

We also investigated the response with Plasmodium falciparum-derived Hz (W-PfHz). Mice were inoculated i.v. with 750 nmol of W-PfHz or W-Hz. They were then challenged with 2,000 Pb-ANKA-Luc sporozoites 1 week later. After measuring luminescence, we saw that the parasite liver load in the group inoculated with W-PfHz was 28% of the liver load in the naive group. We also saw that the liver load in the W-Hz group was 16% of the liver load in the naive group. The difference between the W-PfHz and the W-Hz groups was not statistically significant (Fig. 4D). Overall, these data suggest that the decrease in parasite liver load does not rely on B cells or T cells, and that it is likely mediated by the innate immune system. Although the overall liver burden in RAG-1 KO mice was slightly lower than the one in WT mice, the decrease in liver burden was consistent.

Inoculation with synthetic hemozoin (β-hematin) decreases parasite liver load in a DNA-dependent manner.

To evaluate if β-hematin could have a similar effect compared to W-Hz, we inoculated mice with 750 nmol of β-hematin synthesized from bovine hemin by the Egan method in 12 M sodium acetate pH 4.8, DNase-treated β-hematin, W-Hz, or DNase-Hz. Mice were challenged with 2,000 Pb-ANKA-Luc sporozoites 1 week later. The parasite liver load in mice inoculated with β-hematin was 36% of the liver load in the naive control group. The decrease was higher in the W-Hz group, with liver burden numbers corresponding to 29% of the naive control group (Fig. 5). The liver load in W-Hz-inoculated mice was 49% of the DNase-Hz group. DNase treatment in β-hematin had a similar effect, as liver load levels in the β-hematin group were 56% of those in the DNase-β-hematin group. The differences in liver load between the DNase-Hz and the naive groups, and the DNase-β-hematin and the naive groups were not statistically significant. Just as we saw with W-Hz, we conclude that β-hematin decreases parasite liver load, and that this effect is dependent on DNA.

FIG 5.

Inoculation with synthetic hemozoin decreases parasite liver load in a DNA-dependent manner. Mice were inoculated i.v. with 750 nmol of β-hematin, DNase-treated β-hematin, W-Hz, or DNase-Hz, and rested for 1 week before receiving a 2,000-Pb-ANKA-Luc sporozoite i.v. injection. Parasite liver load was measured 42 h postinfection in photons per second (total flux). Error bars correspond to the SEM. The nonparametric, unpaired Mann-Whitney test showed that there was a statistically significant difference between the naive control group and the groups inoculated with β-hematin or W-Hz. There was also a statistically significant difference between the DNase-Hz and the W-Hz groups, and the DNase-treated β-hematin and the β-hematin groups. The differences in parasite liver load between the naive group and the DNase-treated β-hematin group, or the DNase-Hz group were not significant. ns, P > 0.05; *, P ≤ 0.05; **, P ≤ 0.01; n = 24.

Differences in hemozoin and β-hematin.

We saw that 3D7 P. falciparum-derived hemozoin (100 × 100 × 300–500 nm) is larger in size compared to P. berghei-derived W-Hz (50–75 × 50–75 × 250 nm), and that synthetic β-hematin varies more in shape and size (200 to 1,000 nm in length with 50 to 3,000 nm in height and width) (Fig. 6).

FIG 6.

Different types of hemozoin viewed by Field Emission InLens Scanning Electron Microscopy (FEISEM). (A) P. falciparum-derived hemozoin. (B) P. berghei-derived W-Hz. (C) β-hematin (synthetic hemozoin). Bar is 100 nm.

DISCUSSION

Over the years, there has been no consensus about hemozoin’s role as an immune activator. Some studies suggested a link between hemozoin and NLRP3 or TLR9, while others suggested that it had an inactive role in the host’s immune response (8–12, 18). Most of these studies have focused on in vitro approaches. Our results suggest that P. berghei hemozoin is associated with DNA and its inoculation triggers the innate immune response to decrease or ablate parasite liver load in BALB/c mice.

We observed how, in the absence of a Plasmodium infection, a purified (washed) version of hemozoin had an immune-modulatory effect that decreased P. berghei sporozoite infection and proliferation in mice livers. This effect, however, is time and dose dependent. Parasite liver burden was lower when mice were infected 1 week after W-Hz inoculation, compared to 1 month. Hemozoin can remain within mice tissues over the course of 6 months (19). Therefore, we hypothesize that the difference in liver load levels between the 1-week and 1-month W-Hz groups is due to the absence of an adaptive immune response in our model, and the potential degradation of W-Hz-associated DNA over time. We also saw a small difference between naive groups in the 1-week and 1-month experiments. We believe that this difference was caused by mice ages, as it has been shown to be a relevant factor in the context of parasite liver load and infection. Older mice have a higher decrease in infectivity compared to younger ones (20–22).

Regarding W-Hz dosing, we have previously harvested and calculated a total of approximately 1,100 nmol of hemozoin in mice’s livers and spleens after 2 weeks of infection (data not shown). We also saw that parasite liver burden had a higher decrease in mice inoculated with 750 nmol, compared to mice inoculated with 500 or 250 nmol. After gathering these data, we ultimately decided to inoculate a maximum amount of 750 nmol of W-Hz, since we wanted to make certain that mice would tolerate it in a single dose.

Our results indicate that full protection can be achieved when W-Hz-inoculated mice are challenged after 1 h or 1 day with 2,000 sporozoites. In contrast, mice inoculated with W-Hz that rest for 1 week before challenge are still infected, although exhibiting the usual decrease in liver load. Additionally, mice inoculated with W-Hz that receive a lower dose of sporozoites 1 week later can also be fully protected. The number of parasites in this last experiment emulates an actual sporozoite dose given by a mosquito, although the numbers in rodent malaria species can range from 0 to over 200, with an average of 100 sporozoites (16, 17).

We have just begun to unravel the mechanism by which W-Hz confers partial or total protection against liver-stage Plasmodium, but our results suggest that it is mediated by parasitic DNA associated with the crystal, and likely by the innate immune response. We saw that P. berghei DNA was associated with W-Hz, and that treating W-Hz with DNase I prevented it from decreasing parasitic liver burden in mice challenged with P. berghei sporozoites. Our preparation of hemozoin is entirely devoid of proteins (see Fig. S1 in the supplemental material), and it should also be free from lipids, considering it is extensively washed with SDS, sodium bicarbonate, and treated with Proteinase K. Although the relevance of DNA in our model is clear, we do not think that the immune effect of W-Hz is dependent solely on i.v.-injected Plasmodium DNA, as extracellular DNA delivered i.v. without a vehicle is easily degraded and unable to induce a significant effect over the innate or adaptive immune arms of the host’s response (23–26). Furthermore, we believe that this mechanism is not specific to P. berghei DNA. Intravenous inoculation of P. falciparum-derived Hz has been shown to trigger the expression of proinflammatory cytokines in C57BL/6 mice (27). We verified our model and saw a decrease in parasite liver load in BALB/c mice inoculated with W-PfHz. More importantly, previous studies have reported innate immune activation in human PBMCs after stimulation with Pf-derived Hz (28). The responses to W-PfHz and Pf-derived Hz could be triggered by the presence of P. falciparum DNA associated with this type of hemozoin.

In accordance with Parroche and Golenbock (8), we hypothesize that hemozoin acts as a carrier to deliver parasitic DNA inside the cell. Macrophages phagocytize hemozoin (5, 6). Once in the endosomal compartment or the cytosol, the DNA can trigger innate immune receptors to the release of proinflammatory cytokines that will lead to the activation or recruitment of new cells that may play a role in the protective effect against sporozoites. Our results reinforce the idea that hemozoin’s protective role is mediated by the innate immune response. Importantly, the significant liver load decrease in Rag 1^−/− mice inoculated with W-Hz suggests that the protective effect is not mediated by T cells or B cells, major constituents of the adaptive immune response. In addition, we saw that the effect of W-Hz over parasite liver load waned when W-Hz incubated for a longer period inside mice before sporozoite challenge (1 week versus 1 month incubation). The innate immune response unfolds within the first few hours after cells engage with danger signals or pathogen-associated molecular patterns (PAMPs). This would also explain why mice that were challenged with 2,000 sporozoites 1 day or 1 h after W-Hz inoculation expressed a full protective phenotype. We postulate that, after reaching the liver, the inoculated W-Hz is first phagocytized by Kupffer cells or nonresident macrophages. The DNA associated with the W-Hz can trigger innate immune receptors like TLR9 or cGAS, which in turn would produce innate immune cytokines such as TNF, IL-6, IL-12, or IL-18. The production of most of these cytokines have been shown to be activated by hemozoin in vitro (8, 11, 12). Some of these cytokines can activate NK cells to produce IFN-γ, which has been previously associated with liver-stage protection (29). We do not rule out a much simpler mechanism of protection. For example, after being phagocytized by Kupffer cells, W-Hz could trigger the production of hepcidin, which has also been associated with protection against sporozoites (30). Hepcidin is synthesized in the liver (31) and thought to be upregulated in the presence of hemozoin (32).

Synthetic hemozoin has been associated with impaired T-cell and adaptive-immune function (18). However, the innate response triggered by natural (parasite-derived) hemozoin may still have a beneficial effect against a Plasmodium liver-stage infection. Furthermore, natural, and synthetic hemozoin are not identical, and their difference in shape and size may also modulate immune function in a different manner (33). Synthetic hemozoin, or β-hematin, is produced from bovine or porcine hemin, and our data suggest that foreign, animal DNA is attached to it. Importantly, we saw a significant decrease in parasitic liver load in mice inoculated with β-hematin, but there was no such decrease when β-hematin was treated with DNase. It is possible that some discrepancies in the field may be explained by foreign DNA bound even with synthetic hemozoin, although an inflammatory response may also be triggered by the crystal itself, in a similar fashion to alum (34). A previous study describes the survival of 1 out of 5 mice immunized with synthetic hemozoin that was challenged with a lethal dose of influenza virus (35). The lack of specificity in response to different types of bound-with-DNA hemozoin would explain these types of results.

Plasmodium parasites have remained elusive against drug or vaccine strategies. Finding new mechanisms by which the parasite can be impaired is key for the development of novel drug or vaccine candidates, or to enhance the efficacy of the ones already available. In addition, targeting bottlenecks in the parasitic life cycle, such as the number of sporozoites inoculated in the host’s skin, can prevent malaria infection and transmission (16). Our data may also help explain part of the behavior of reinfections in mice and humans. Mouse models have shown the importance of innate immunity in host resistance against Plasmodium reinfection (36). This study contributes to the idea that the partial protection seen in some reinfected subjects in malaria-endemic areas may be boosted by parasitic danger signals that trigger the innate immune response and may not be fully mediated by antibodies. A single bolus intravenous injection of W-Hz in mice is not comparable to what is occurring over the course of a blood-stage human infection where hemozoin is being continually produced with accumulating quantities linked to parasite density. During blood-stage infection, any early protective responses elicited by low-level hemozoin accumulation in the liver could be lost as hemozoin concentrations increase and potentially result in liver damage. Full protection is not observed in humans in malaria-endemic areas, but hemozoin might contribute to mechanisms decreasing suprainfections. In humans, remnant Plasmodium residual bodies containing hemozoin might either be shorter-lived or not in high enough dose in the liver.

MATERIALS AND METHODS

Animals and parasites.

Inbred, female 6 to 7-week-old wild-type BALB/c and RAG-1 KO BALB/c mice were purchased from the Jackson Laboratory (Bar Harbor, ME), and kept in the animal facility at the Johns Hopkins Bloomberg School of Public Health. All animal experiments were approved by the Johns Hopkins Animal Care and Use Committee (ACUC), under protocol MO21H419. P. berghei MRA-868 ANKA (Pb-ANKA-Luc) transgenic strain was used for all sporozoite infections. The Pb-ANKA-Luc parasite strain expresses a GFP-luciferase fusion gene. NF54 P. falciparum parasites at 4% hematocrit were cultured as described (37).

Hemozoin isolation and quantitation.

WT BALB/c mice infected with Pb-ANKA-Luc were euthanized 13 days postinfection. Livers and spleens were harvested and stored at −80°C until hemozoin isolation. Livers and spleens were individually resuspended in Milli-Q H₂O and sonicated until completely dissolved. Samples were subsequently washed 1× with 2% SDS/80 mM NaHCO3 and treated with 1 mg/mL Proteinase K at 56°C overnight. Samples were then washed 2× with 2% SDS and 7× with Milli-Q H₂O. For the DNase experiments, W-Hz was treated with DNase I at pH 7.6 (New England BioLabs Inc. catalog no. M0303S), heat-inactivated at 75°C, and washed with Milli-Q H₂O. W-PfHz was isolated from P. falciparum NF54 strain cultures at 4% hematocrit, following the same method as the one described for W-Hz. W-Hz and W-PfHz were quantitated by measuring absorbance at 405 nm after decrystallization, as previously described (38).

Synthesis of β-hematin.

Β-hematin was synthesized by the Egan method (39) from bovine hemin (Sigma catalog no. H9039) with a 36-h incubation in 12 M sodium acetate, pH 4.8 at 60°C. We washed the crystal product extensively with SDS/bicarbonate pH 9.1 and 2% SDS, followed by Proteinase K overnight, fresh 2% SDS washes, and Milli-Q H₂O.

Murine model of hemozoin inoculation and sporozoite infection.

Female Anopheles stephensi mosquitoes were fed with Pb-ANKA-Luc-infected blood for sporozoite development, as previously described (20). Mice were inoculated i.v. with W-Hz, W-PfHz, or DNase-Hz resuspended in 200 μL/mouse of 1× sterile-filtered PBS, before resting for 1 week or 1 month. A. stephensi salivary glands were dissected for sporozoite isolation in a 1× HBSS 2% FBS solution. After Hz incubation, mice were infected with 2,000 or 50 sporozoites resuspended in 200 μL of 1× HBB 2% FBS. Sporozoites were counted and diluted as previously described (20).

Parasite liver load imaging and blood-stage infection quantification.

Mice infected with Pb-ANKA-Luc sporozoites had their abdominal hair removed with depilatory lotion (Nair) 24 h postinfection. At 42 h after sporozoite infection, mice were anesthetized with isoflurane, and injected intraperitoneally with 200 μL of 15 mg/mL d-luciferin. Mice were subsequently imaged in the IVIS Spectrum in vivo imaging system. Luminescence was measured using the following settings: 15 s exposure, 1 min delay between sequences, and ten 15-s sequences. The data were captured in photons/second (total flux). Blood-stage infections were followed using a luciferase assay, and converting luminescence to parasites/μL, as previously described (40).

P. berghei Hz DNA extraction and amplification.

A total of 1,000 nmol of W-Hz were first resuspended in 200 μL of 1× PBS, and decrystallized by adding 20 μL of 1 M NaOH for 30 min. The pH was neutralized with 20 μL of 1 M HCl. DNA was extracted using a QIAmp DNA blood minikit and eluted in 60 μL of ultrapure H₂O. The remnant P. berghei DNA was amplified in triplicates by quantitative PCR (qPCR) with primers specific for the Pb18S gene.

Liver sectioning and H&E staining.

Mouse livers were harvested 1 week after inoculation with 750 nmol of W-Hz IV. Livers were fixed in a 10% formalin solution, neutral buffered (Millipore Sigma catalog no. HT5014-120ML) and sectioned with a 4 μm thickness. Liver sections were embedded in paraffin and stained with hematoxylin and eosin with the assistance of Johns Hopkins University Oncology Tissue and Imaging Services.

Statistical analysis.

All liver-burden experiments were analyzed with a nonparametric, unpaired Mann-Whitney test.