Inhibiting the Mammalian Target of Rapamycin Blocks the Development of Experimental Cerebral Malaria

To determine the effect of rapamycin on the development of ECM, C57BL/6 mice were infected with PbA and injected intraperitoneally with a vehicle control solution or with rapamycin (1.0 mg/kg of body weight) daily starting 1, 4, or 5 days p.i. All vehicle-treated mice developed severe neurological symptoms by day 6 p.i. and either died or were euthanized by day 9 p.i. (Fig. 1A). In contrast, mice that began daily rapamycin treatment on either day 1 or 4 p.i. did not develop symptoms of ECM and survived longer than untreated mice. Beginning rapamycin treatment on day 5 p.i. was considerably less effective, blocking ECM in only 20% of mice (Fig. 1A). Untreated mice died by day 9 with blood hemoglobin levels well above 6 g/dl. In contrast, mice that began rapamycin treatment on either day 1 or 4 p.i. showed dramatic drops in hemoglobin levels by day 9 p.i., indicating severe anemia (Fig. 1B). Using the criteria for ECM of clinical scores above 6 in severity and hemoglobin levels above 6 g/dl, it was apparent that all untreated mice and 80% of mice treated with rapamycin beginning on day 5 p.i. died of ECM (Fig. 1C), whereas all mice treated with rapamycin on day 1 p.i. and 90% of mice treated on day 4 p.i. were protected from ECM (Fig. 1C). Even though mice treated on day 1 or 4 p.i. were nearly equally protected from ECM, the mice treated with rapamycin beginning on day 1 p.i. died somewhat earlier than mice treated beginning on day 4 p.i. (Fig. 1A). Measurements of peripheral parasitemia showed that parasitemias in mice treated with rapamycin beginning on day 1 p.i. were significantly higher than mice treated with rapamycin beginning on day 4 or 5 p.i. and untreated mice (Fig. 1D). Thus, rapamycin treatment beginning on day 1 blocked the development of ECM despite enhancing peripheral parasite growth.

The presentation of neurological symptoms in mice with ECM is accompanied by focal hemorrhaging and disruption of the blood-brain barrier (BBB) (13). We assessed the effect of rapamycin treatment on the integrity of the BBB by measuring the exclusion from brain tissue of Evans blue (EB), a dye that binds to serum albumin that leaks into the brain parenchyma with BBB dysfunction (28, 29). PbA-infected mice were treated with rapamycin daily beginning on day 1 or 4 p.i. or with a vehicle control solution. Brains from mice infected with P. berghei NK65 (PbNK65), a parasite highly related to PbA but one that does not cause ECM and results in severe anemia (30, 31), served as a negative control. The PbNK65-infected control mice showed no leakage of EB into the brain by visual inspection of the brains (Fig. 2A). In contrast, PbA-infected, untreated mice showed severe leakage of EB into the brain both visually and as quantified by extracting EB from the brain. PbA-infected mice treated with rapamycin daily beginning on day 1 or 4 p.i. showed with minimal EB leakage into the brain (Fig. 2A). In separate experiments, rapamycin treatment starting on day 5 p.i. did not prevent BBB dysfunction, and the brains from most of these mice contained similar amounts of EB compared to the brains of PbA-infected mice treated with vehicle control (data not shown).

Uninfected mice and infected mice that were treated with rapamycin beginning on day 4 p.i. or left untreated were sacrificed on day 6 p.i., and their brains were removed for histological examination. Brain sections were stained with either hematoxylin and eosin (H&E) to detect petechial hemorrhages and sequestered iRBC hemozoin associated with ECM or with IgG-specific antibody to detect intraparenchymal IgG as a second measure of the loss of BBB integrity (32). The brains of infected, untreated mice showed petechial hemorrhages in H&E-stained brain sections (Fig. 2B). Quantification of hemorrhages showed greater hemorrhaging in the brains of PbA-infected, untreated mice (an average of 7 per brain section) compared to PbA-infected, rapamycin-treated mice (an average of 2 per section) (Fig. 2C). The brains of untreated, PbA-infected mice showed strong intraparenchymal staining for IgG, whereas the brains of mice treated with rapamycin beginning on day 4 p.i. resembled the brains of uninfected mice with no detectable intraparenchymal staining for IgG (Fig. 2B), which is consistent with results using EB exclusion as a marker for BBB integrity. iRBCs, identified by the presence of the parasite-produced hemozoin, were frequently observed in PbA-infected, untreated mice in H&E-stained brain sections along the venular endothelium (Fig. 2B) but only rarely observed in the brain vasculature of PbA-infected mice that were treated with rapamycin daily beginning on day 4 p.i. The parasite loads in the brains of saline-perfused mice on day 6 p.i. were quantified by quantitative PCR (qPCR) of 18S rRNA (Fig. 2D). We observed a 4-fold decrease in parasite 18S rRNA in the brains of PbA-infected mice treated daily with rapamycin beginning on day 4 p.i. compared to PbA-infected, untreated mice despite equivalent levels of peripheral parasitemia in rapamycin-treated compared to untreated PbA-infected mice (Fig. 1D). As expected, no qPCR amplification was detected in the brains of uninfected control mice with the same 18S ribosomal primers (data not shown). Taken together, these results show that administration of rapamycin as late as day 4 p.i. prevents the brain pathology associated with ECM, including breakdown of the BBB, brain hemorrhaging, and accumulation of iRBCs in the brain.

Because ECM is accompanied by the production of proinflammatory cytokines and chemokines, we determined the effects of rapamycin treatment on the serum levels of several inflammatory cytokines and chemokines, including interleukin 6 (IL-6), IL-10, IL-12p70, macrophage inflammatory protein 1α (MIP-1α), macrophage chemoattractant 1 (MCP-1), RANTES (regulated on activation, normal T cell expressed and secreted), IFN-γ, TNF-α, IL-1β, chemokine (C-X-C motif) ligand 1 (CXCL1), thymus and activation-regulated chemokine (TARC), and T cell alpha chemokine (TCA) using a cytokine array. Mice were uninfected or infected with PbA and left untreated or treated with rapamycin daily beginning on day 1 or 4 p.i. All mice were sacrificed on day 6 p.i., and sera were collected. The sera of PbA-infected, untreated mice had elevated levels of the proinflammatory molecules TCA, MIP-1α, CXCL1, MCP-1, RANTES, and IL-6 compared to uninfected mice, either treated with rapamycin or not treated with rapamycin (Fig. 3). The levels of IL-10, an anti-inflammatory cytokine often produced in response to inflammation (33), was also increased in PbA-infected, untreated mice as compared to uninfected controls. Remarkably, daily rapamycin treatment of PbA-infected mice beginning on day 4 p.i. resulted in significant increases in the serum levels of TCA, IFN-γ, MIP-1α, CXCL1, MCP-1, RANTES, and IL-6 (Fig. 3) compared to those of PbA-infected, untreated controls. The level of the anti-inflammatory cytokine IL-10, which was higher in PbA-infected mice than in uninfected mice, remained elevated in infected mice treated with rapamycin. In mice in which daily rapamycin treatment was started on day 1 p.i., we observed significantly higher levels of CXCL1 and significantly reduced levels of IL-10 (see Fig. S1 in the supplemental material) relative to PbA-infected, untreated controls. These findings provide evidence that rapamycin prevents ECM even though treatment augmented the PbA-induced systemic proinflammatory response.

ECM is accompanied by leukocyte sequestration in the brain, including CD4⁺ and CD8⁺ T cells, monocytes, neutrophils, and NK cells. We determined the number of total CD4⁺ and CD8⁺ T cells and the number of CD44⁺ effector/memory T cells, neutrophils, and NK cells in the brains and spleens of uninfected mice and PbA-infected mice that were either untreated or treated with rapamycin daily beginning on day 1, 4, or 5 p.i. All mice were terminally anesthetized on day 6 p.i. and transcardially perfused with phosphate-buffered saline (PBS). The brains and spleens were removed, processed, and analyzed by flow cytometry. The gating strategy used to identify total CD4⁺ and CD8⁺ T cells and effector/memory T cells, neutrophils, and NK cells is shown (see Fig. S2 in the supplemental material). We observed a large accumulation of CD8⁺ T cells in the brains of PbA-infected mice (1.0 × 10⁵ cells per brain) (Fig. 4A) and a somewhat smaller accumulation of CD4⁺ T cells (3.0 × 10⁴ cells per brain) (Fig. 4B) compared to uninfected mice. More than 80% of the CD8⁺ T cells that accumulated in the brain were CD44⁺ effector/memory T cells (Fig. 4C), and similarly, the majority of CD4⁺ T cells that accumulated in the brain, more than 60%, were CD44⁺ effector/memory T cells (Fig. 4D). Treatment of PbA-infected mice with rapamycin beginning on day 1 p.i. resulted in a dramatic reduction in the accumulation of both CD4⁺ T cells (an 88% reduction) and CD8⁺ T cells (a 99% reduction) in the brain (Fig. 4A to D). Treatment with rapamycin beginning on day 4 p.i. also resulted in reduced numbers of CD4⁺ and CD8⁺ T cells in the brain, and although these reductions were not as great as those observed in mice treated with rapamycin daily beginning on day 1 p.i., mice treated on day 1 or day 4 p.i. were similarly protected from ECM (Fig. 1C). There were no significant differences in the numbers of CD4⁺ or CD8 ⁺ T cells in the brains of mice treated with rapamycin on day 5 p.i. compared to untreated PbA-infected mice (Fig. 4A to D), which is consistent with the weak protection from ECM (20%) of mice treated with rapamycin on day 5 p.i. The effects of rapamycin on the expansion of CD8⁺ and CD4⁺ T cells in spleens were more modest than the effect on the accumulation of these cells in brains. The number of CD8⁺ effector/memory T cells increased approximately 3-fold in the spleens of PbA-infected mice as compared to uninfected mice. Rapamycin treatment inhibited this increase, with the most effective inhibition (67%) observed for mice treated with rapamycin beginning on day 1 p.i. (Fig. 4E). The number of CD4⁺ effector/memory T cells increased approximately 2-fold in the spleens of PbA-infected mice compared to uninfected mice, and rapamycin blocked this increase but only when treatment began on day 1 p.i. (Fig. 4F). Together these results suggest that rapamycin treatment functions to protect mice against ECM at least in part by blocking the expansion of effector/memory T cells in peripheral lymphoid tissues and their migration to the brain, resulting in greatly diminished numbers of T cells accumulating in the brains of infected mice. Compared to uninfected mice, the number of neutrophils dramatically increased (more than 100-fold) in the brains of PbA-infected mice, and rapamycin treatment beginning on day 1 or day 4 p.i. blocked this increase significantly. The number of neutrophils in the spleen did not increase with PbA infection, and rapamycin treatment resulted in increased numbers of neutrophils in the spleens of infected mice consistent with the ability of rapamycin to block neutrophil chemotaxis (34). The number of NK cells in the brains of PbA-infected mice increased 2.5-fold with PbA infection, but rapamycin treatment had no effect on this increase (data not shown).

To obtain a global view of how rapamycin treatment prevented cerebral disease in PbA infections, we carried out DNA microarray analyses of both brains and spleens at day 6 p.i. in mice that were treated with rapamycin beginning on day 1 p.i. or left untreated. Thus, four experimental groups were analyzed: uninfected mice, uninfected mice treated with rapamycin daily beginning on day 1 p.i., PbA-infected mice, and PbA-infected mice treated with rapamycin daily beginning on day 1 p.i. For the brain, the olfactory bulb, right cerebrum, and cerebellum were analyzed separately. The number of differentially expressed genes (DEGs) (cutoff criteria, absolute fold change of >1.5 and a false discovery rate [FDR] of <5%) for all comparisons are shown in Table S1 in the supplemental material (four mice in each group). Compared to uninfected mice, PbA infection alone induced marked transcriptional changes in the olfactory bulb (1,107 DEGs), right cerebrum (509 DEGs), and cerebellum (589 DEGs). However, rapamycin treatment of PbA-infected mice also induced substantial changes compared to untreated PbA-infected mice, with 350 DEGs in the olfactory bulb, 215 DEGs in the right cerebrum, and 306 DEGs in the cerebellum. Notably, the vast majority of genes differentially expressed in PbA-infected mice treated with rapamycin compared to untreated mice overlapped with genes differentially expressed in PbA-infected versus uninfected mice (81% in the olfactory bulb, 78% in the right cerebrum, and 75% in the cerebellum) (Fig. 5A). Thus, rapamycin appears to abrogate cerebral disease in large part by altering the expression of the very same genes affected by PbA infection. An even greater number of DEGs was observed in PbA-infected spleens relative to uninfected spleens (2,560 DEGs) with a similarly high overlap between these DEGs and the 1,037 DEGs in rapamycin-treated, PbA-infected mice compared to untreated PbA-infected mice (see Fig. S3A in the supplemental material). It is important to note that rapamycin treatment of uninfected mice had little effect on the gene expression profiles relative to saline treatment (146 DEGs) and no effect on any part of the brains of uninfected mice (0 DEGs) (Table S1).

Differentially expressed gene sets from the PbA-infected, rapamycin-treated mice versus PbA-infected, untreated mice were further analyzed to determine which pathways were most highly affected by rapamycin treatment in the brain (Fig. 5B) and in the spleen (see Fig. S3B in the supplemental material). Canonical pathway enrichment analysis identified several affected pathways that were mutually overrepresented among the differentially expressed gene sets from all three brain tissues. These all involved immune cell pathways and included the following: granzyme A signaling; role of pattern recognition receptors; granulocyte/agranulocyte adhesion, diapedesis, and activation of interferon regulatory factors by cytosolic pattern recognition receptors (Fig. 5B). The finding that the pathways most highly affected by rapamycin treatment were associated with immune cell functions likely reflects the marked decrease observed in the numbers of immune cells, both T cells and neutrophils, in the brains of rapamycin-treated, PbA-infected mice relative to untreated, PbA-infected mice (Fig. 4A to D). In contrast, a number of immune pathways that are involved during PbA infection are not affected by rapamycin treatment, suggesting that certain PbA-affected pathways are distinct from the mechanism by which rapamycin blocks ECM. For example, expression of genes in the complement pathway are affected by PbA infection compared to uninfected mice in all three parts of the brain, but rapamycin treatment had no effect on this pathway (Fig. 5B). The expression of subunits comprising the C1q complement complex was increased upon infection (for C1qa, 1.7-fold change and an FDR of 0.98% in the cerebrum, 1.6-fold change and an FDR of 3.0% in the olfactory bulb, and 1.7-fold change and an FDR of 1.3% in the cerebellum; for C1qb, 1.9-fold change and an FDR of 0.0024% in the cerebrum, 2.2-fold change and an FDR of <0.0001% in the olfactory bulb, and 2.7-fold change and an FDR of <0.0001% in the cerebellum) and unchanged by rapamycin treatment (not significant for C1qa and C1qb in all tissues). Although increased expression of C1q in the brain (35) and periphery (35, 36) has been suggested to play a role in pathogenesis in ECM, our results suggest that the development of ECM can be blocked despite such increases. Intriguingly, the antigen presentation pathway, which was significantly affected by PbA infection in all three brain parts, was not altered in any case by rapamycin treatment (Fig. 5B). Consistent with this finding, expression of Tap1, which encodes an ATP-binding cassette peptide transporter that is critical in antigen processing for presentation on major histocompatibility class I molecules, was significantly upregulated in the brains of PbA-infected mice versus uninfected mice (3.8-fold change and an FDR of <0.0001% in the cerebrum, 4.0-fold change and an FDR of <0.0001% in the olfactory bulb, and 3.6-fold change and an FDR of <0.0001% in the cerebellum), strongly implying increased antigen presentation in PbA-infected mice. However, rapamycin treatment had no effect on these changes in Tap1 expression. This finding suggests that although PbA infection profoundly increases antigen presentation in the brain during infection, rapamycin-mediated protection against cerebral disease may not require the restoration of antigen presentation to uninfected levels.

Significant overlap among the DEGs in the PbA-infected mice versus uninfected mice and the DEGs in the PbA-infected, untreated mice versus PbA-infected, rapamycin-treated mice suggests that rapamycin treatment may dampen specific subsets of genes that are induced by infection or vice versa. To identify changes in PbA-induced gene expression that were abrogated or reversed by rapamycin treatment, DEGs for each part of the brain and spleen were filtered on those genes that showed a >1.5-fold change in expression in the comparison of PbA-infected versus uninfected mice and a <−1.5-fold change in expression in the comparison of PbA-infected, rapamycin-treated versus PbA-infected, untreated mice (or vice versa) using an FDR of <5% for each comparison as the threshold for significance. Figure 6A shows the expression intensities of the discordantly regulated genes in the right cerebrum for uninfected mice, PbA-infected mice, and PbA-infected, rapamycin-treated mice. Pathway enrichment analysis was applied to this filtered set of discordantly regulated genes to predict the regulator effect networks that are reversed or abrogated by rapamycin treatment during PbA infection (Fig. 6B). This network analysis predicted inhibition of the following immunological functions with rapamycin treatment during PbA infection: cellular invasion, cellular chemotaxis, lymphocyte proliferation, and chronic inflammation (Fig. 6B). Thus, rapamycin treatment is predicted to reverse the effect of PbA infection by altering critical immune cell pathways, including trafficking to the brain, proliferation in the brain, and cerebral inflammation. A similar analysis of discordantly regulated genes predicted chronic inflammation to be inhibited in the olfactory bulb with rapamycin treatment. However, no other immunological functions were found among the top regulator effect networks in the cerebellum using the same analysis of discordantly regulated genes.

There are several genes of interest in the set of discordantly regulated genes (Fig. 6A) that provide insights into the mechanisms underlying rapamycin's effects. Expression of heme oxygenase 1 (HO-1; encoded by Hmox1) has been shown to be highly upregulated in PbA-infected, ECM-resistant mouse strains and to be required for protection against ECM (37). Unexpectedly, Hmox1 was among the discordantly regulated genes, demonstrating increased expression in PbA-infected mice relative to uninfected mice (2.8-fold change and an FDR of 0.008% in the cerebrum; 6.9-fold change and an FDR of <0.0001% in the olfactory bulb) but decreased expression (−2.0-fold change and an FDR of 2.7% in the cerebrum; −2.3-fold change and an FDR of 0.27% in the olfactory bulb) in rapamycin-treated, PbA-infected mice (Fig. 6A). This finding suggests that Hmox1 may not be absolutely required for protection from ECM. CD8⁺ T cells expressing granzyme B (encoded by Gzmb) have been demonstrated to be required for the development of ECM (15). Consistent with this, Gzmb expression significantly increased in the brains of PbA-infected mice exhibiting signs of ECM (in the right cerebrum, 5.7-fold change and an FDR of <0.0001%; in the cerebellum, 6.2-fold change and an FDR of <0.0001%; in the olfactory bulb, 4.9-fold change and an FDR of <0.0001%), but it significantly decreased in ECM-protected, PbA-infected mice treated with rapamycin at day 1 p.i. (in the right cerebrum, −2.8-fold change and an FDR of <0.0001%; in the cerebellum, −2.5-fold change and an FDR of 0.0001%; in the olfactory bulb, −1.8-fold change and an FDR of 0.077%), a result that could be explained in part by decreased CD8⁺ T cell recruitment in the brain (Fig. 4A and C).

An upstream regulator analysis of differentially expressed genes was carried out for all comparisons (PbA-infected mice versus uninfected mice and PbA-infected, rapamycin-treated mice versus PbA-infected mice). The results predicted genes involved in both type I and type II interferon responses (Ifng, Irf7, Ifna2, Irf3, Stat1, and Infb1) and proinflammatory genes (Tnf, Il1b, and Il6) as transcriptional activators in PbA-infected mice relative to uninfected controls both in the brain (Fig. 7A to C) and in the spleen (see Fig. S3C in the supplemental material). Consistent with the results of the serum cytokine analyses (Fig. 3 and Fig. S1), treatment of PbA-infected mice with rapamycin further increased activation of both interferon and proinflammatory pathways in the spleen (Fig. S3C). Notably, treatment of PbA-infected mice with rapamycin increased activation of interferon response genes in the brain (Irf7, Irf3, Ifna2, Ifnb1, and Ifng), but not proinflammatory genes downstream of the interferons (Tnf, Il1b, and Il6) (defined as a Z score of >2 and P < 0.01) (Fig. 7A to C). This suggests that, in the brain, rapamycin-mediated augmentation of inflammatory responses during PbA infection may be limited to the innate, interferon-related response and truncated prior to the initiation of the downstream mediators of inflammation. In contrast, rapamycin-mediated augmentation of the inflammatory response in the spleen during PbA infection involves both innate interferon-related responses as well as downstream mediators of inflammation.

The results presented thus far provide strong evidence that rapamycin treatment daily beginning on day 1 p.i. or as late as day 4 p.i. protects mice from the pathology of ECM. However, rapamycin treatment beginning at day 1 p.i. also resulted in an increase in peripheral parasitemia compared to untreated PbA-infected mice (Fig. 1D), resulting in a drop in hemoglobin levels (Fig. 1B). The effect of rapamycin treatment on parasitemia appeared to be specific for the ECM-causing PbA, as rapamycin treatment of mice infected with the closely related parasite PbNK65 did not result in a rise in peripheral parasitemia (Fig. 8A). Rapamycin treatment had some effects on parasitemia in PbNK65-infected mice, but the effects were small. Rapamycin treatment appeared to delay the rise in PbNK65 parasitemia compared to untreated mice (Fig. 8A), perhaps consistent with the reported inhibitory effect of rapamycin on parasite growth in vitro (26). Rapamycin treatment resulted in hemoglobin levels that dropped somewhat more slowly during the course of infection in PbNK65-infected, rapamycin-treated mice compared to untreated infected mice (Fig. 8B), and the rapamycin-treated mice survived longer (Fig. 8C). To determine whether rapamycin's effect on parasitemia was intrinsic to PbA or was dependent on the host's immune response, recombination-activating gene product 1 (RAG-1)-deficient mice (RAG-1 [KO {knockout}]) that lack B and T cells were infected with PbA and treated with rapamycin or a saline control. PbA infection does not cause ECM in RAG-1 [KO] mice, because the functions of CD8⁺ and CD4⁺ T cells are required for ECM pathology. Thus, PbA-infected RAG-1 [KO] mice survive longer than PbA-infected C57BL/6 mice and ultimately die of severe anemia caused by hyperparasitemia. Rapamycin treatment did not cause an increase in parasitemia in RAG-1 [KO] mice compared to untreated, PbA-infected RAG-1 [KO] mice (Fig. 8D), resulting in similar decreases in hemoglobin levels and survival in treated and untreated mice (Fig. 8E and F). Thus, rapamycin's effect on PbA parasitemia does not appear to be due to a direct effect on PbA growth in vivo but rather to an indirect effect on adaptive-immune-response-dependent mechanisms that control parasitemia.

For rapamycin treatment to be useful as an adjunctive therapy, it will be necessary to control the rapamycin-induced increase in parasitemia. To determine the feasibility of increasing survival following rapamycin treatment with antimalarials, PbA-infected mice treated with rapamycin beginning on day 1 or day 4 p.i. were treated with three daily doses of artesunate beginning on day 7 p.i. when parasitemias rise rapidly in rapamycin-treated, PbA-infected mice. Artesunate treatment reduced parasitemias, increased hemoglobin levels, and increased survival of PbA-infected, rapamycin-treated mice compared to mice not given artesunate (Fig. 9A to C). Treatment of PbA-infected mice with artesunate alone on day 7 p.i. was not feasible, as nearly all mice die by day 7 p.i. in the absence of rapamycin treatment. Thus, rapamycin treatment in conjunction with artesunate treatment results in a significant increase in survival of PbA-infected mice.

All experiments were approved by the National Institute of Allergy and Infectious Diseases Animal Care and Use Committee.

C57BL/6 and C57BL/6-[KO]RAG-1 female mice (7 to 10 weeks old) were obtained from the Jackson Laboratory. Mice were infected with either PbA or PbNK65 (New York line) by injecting 1 × 10⁶ PbA- or PbNK65-iRBCs obtained from infected C57BL/6 mice intraperitoneally (i.p.). Hemoglobin levels in blood samples taken from the tail vein (<10 µl/day) were determined using a HemoCue Hb201+ (HemoCue AB, Angelholm, Sweden). Peripheral parasitemia was determined in blood either by Wright-Giemsa-stained whole-blood smears or by flow cytometry as described below. Infected mice were monitored for the progression of experimental cerebral malaria (ECM) using a 10-point clinical scoring system that rates mice as symptomless (a score of 0), to moribund (a score of 10), as previously described (42). According to our animal protocol, mice with a clinical score of 6 or greater and severely anemic mice with a hemoglobin level below 2.5 g/dl were euthanized.

For rapamycin treatment, a stock solution of rapamycin (catalog no. R0395; Sigma Aldrich) was prepared by dissolving rapamycin in pure ethanol (25 mg/ml). For treatment of mice, which weighed approximately 20 g in these studies, the stock rapamycin solution was diluted in a solution of 5% polyethylene glycol 400 (Sigma), 4% ethanol, and 5% Tween 80 for a final concentration of rapamycin of 1 mg/ml. Mice were injected intraperitoneally with 1 mg/kg rapamycin every day starting on day 1, 4, or 5 p.i., unless otherwise noted. For artesunate treatment, 60 mg of artesunate (catalog no. A3731; Sigma) was dissolved in 1 ml of 5% sodium bicarbonate, to which 4 ml of 5% dextrose was added for a final concentration of 12 mg/ml. One hundred milligrams per kilogram was administered i.p. to mice on the specified days.

Mice were anesthetized with ketamine/xylazine 6 days p.i. and transcardially perfused with ice-cold PBS, and the brains and spleens were removed. The brains were dissected, minced, and digested with 1 mg/ml collagenase D for 30 min at 37°C. After the tissue was passed through 70-µm nylon mesh, homogenates were placed on a 90%−60%−40% discontinuous Percoll gradient and centrifuged for 18 min at 1,000 × g, and the cells at the 40%−60% interface containing mostly leukocytes were collected for analysis. The spleens were minced and forced through a 70-µm nylon mesh, and the cell suspension was incubated for 10 min in a solution of ammonium-chloride-potassium (ACK; Lonza) to lyse RBCs. The cells were washed and resuspended in RPMI 1640 with 10% heat-inactivated fetal bovine serum (FBS). The cells from both brain and spleen were stained in fluorescence-activated cell sorting (FACS) buffer (PBS plus 1% FBS). The following fluorescent-dye-conjugated antibodies specific for the following cell surface markers were used for staining: brilliant violet 421-conjugated NK1.1 (BV421–NK1.1) (BioLegend), BV605–CD4 (BioLegend), BV785–CD8 (BioLegend), phycoerythrin-conjugated Ly6G (PE–Ly6G) (BD Pharmingen), phycoerythrin-and-Cy7-conjugated CD3 (PE-Cy7–CD3) (eBioscience), allophycocyanin-conjugated Ly6C (APC–Ly6C) (BD Pharmingen), Alexa Fluor 700-conjugated CD44 (AF700–CD44) (eBioscience), APC-Cy7–CD45.2 (BD Pharmingen), and LIVE/DEAD (Aqua; Invitrogen). Gating of subsets is depicted in Fig. S1 in the supplemental material. Cell acquisition data were obtained on a BD LSRII flow cytometer. Data were analyzed with FlowJo software (Tree Star Technologies).

Evans blue (20 mg/kg) was injected intravenously on day 6 p.i., and 3 h later, the mice were anesthetized and perfused, and the brains were removed and immediately frozen at −80°C for later processing. EB was quantified by a modified version of the previously described protocol (43). Briefly, EB was extracted from brains with one perfused brain per 2-ml skirted screw cap tube (Greiner). Seven hundred microliters of N,N-dimethylformamide (DMF) (catalog no. D4551; Sigma) was added with three silica beads (2.3 mm) (catalog no. 11079125z; Biospec) per tube and homogenized for 1 min at room temperature (Minibeadbeater-16 model 607; BioSpec Products). This homogenized solution was centrifuged at 16,000 × g for 20 min at 4°C. The supernatant was transferred to a separate tube and spun again at the same speed and temperature for 10 min. Two hundred microliters of this supernatant was then quantified in duplicate using a Varioskan Flash fluorometer (620-nm excitation; 695-nm emission; Thermo Scientific). For quantification, a standard curve was generated by using a uninfected perfused brain prepared in the same way and adding EB at a twofold dilution starting at 1 mg/ml to 1 µg/ml.

Parasitemia was determined by flow cytometry using a modified version of a previously described method (44). Briefly, blood (approximately 0.6 µl) was obtained from mouse tail veins, fixed with 0.025% aqueous glutaraldehyde solution, washed with 2 ml PBS, resuspended, and stained with the following: the DNA dye Hoechst 33342 (Sigma) (8 µM), the DNA and RNA dye dihydroethidium (diHEt) (10 µg/ml), an APC-conjugated antibody (Ab) specific for CD45.2 (BioLegend), a pan-C57BL/6 lymphocyte marker, and APC-Cy7-conjugated Ab specific for Ter119 (BD Pharmingen), an RBC marker. The cells were analyzed on a BD LSRII flow cytometer equipped with UV (325-nm), violet (407-nm), blue (488-nm), and red (633-nm) lasers. Data were analyzed using FlowJo software (Tree Star Technologies). iRBCs were CD45.2⁻, Ter119⁺, Hoechst positive, and diHEt⁺. Overall parasitemia was calculated as the number of iRBCs/total number of RBCs.

After anesthetization and intracardiac perfusion of mice as described above, the brains were removed, immediately frozen in liquid nitrogen, and pulverized, and RNA was extracted using a Qiagen RNeasy minikit according to the manufacturer's instructions. Genomic DNA was digested on a column using a RNase-free DNase set (Qiagen), and the elimination of genomic DNA was further confirmed using no reverse transcriptase (no-RT) controls. cDNA was generated using a iScript cDNA synthesis kit (Bio-Rad Laboratories). SYBR green PCR master mix (Bio-Rad) was used to determine the relative expression of parasite 18S rRNA and of three host housekeeping genes, hprt, gapdh, and ppia. The Pb-18S primers and primer sequences were 5′-AAGCATTAAATAAAGCGAATACATCCTTAC-3′ and 5′-GGAGATTGGTTTTGACGTTTATGTG-3′. The mouse hprt, gapdh, and ppia primer sequences were 5′-TGCTCGAGATGTGATGAAGG-3′ and 5′-TCCCCTGTTGACTGGTCATT-3′, 5′-GTGGAGTCATACTGGAACATGTAG-3′ and 5′-AATGGTGAAGGTCGGTGTG-3′, and 5′-TTCACCTTCCCAAAGACCAC-3′ and 5′-CAAACACAAACGGTTCCCAG-3′, respectively.

The geometric means of the threshold cycle (C_T) values of housekeeping genes were used as the baseline for comparing the ΔC_T value of 18S gene amplification. The changes in gene expression were calculated by comparing the ΔC_T values of experimental and control groups using the comparative C_T method (2^−ΔΔCT).

Brain samples were fixed in 10% buffered formalin, embedded in paraffin, and sectioned. Sections were stained with hematoxylin and eosin (H&E) for ultrastructural examination and detection of hemorrhages and iRBC hemozoin. For detection of IgG, the slides were sequentially incubated in citrate for 20 min, 2% normal horse serum for 20 min, and biotinylated horse Ab specific for mouse IgG (Vector Labs, Burlingame, CA) for 30 min. Biotinylated horse Ab was detected using a biotin avidin peroxidase complex kit (Vector Labs). The slides were then examined by light microscopy with magnifications between ×5 and ×100, and the microscopic images were evaluated by two independent investigator in a blind manner (blind to the study design) as described previously (45). For quantitation of hemorrhages, 10 microscopic 40× power fields were examined, and the numbers of hemorrhages were counted and averaged.

Blood samples were collected on day 6 p.i., and sera were stored at −80°C until analyzed for IL-6, IL-10, IL-12p70, MIP-1α, MCP-1, RANTES, IFN-γ, tumor necrosis factor alpha (TNF-α), IL-1β, CXCL1, TARC, and TCA using the Q-Plex array mouse cytokine kit (Quansys Biosciences) according to the manufacturer's instructions.

C57BL/6 mice were infected with PbA or mock infected with saline vehicle and treated with either saline or rapamycin beginning on day 1 p.i. (4 conditions; 4 mice for each condition). On day 6 p.i., the mice were anesthetized and perfused with saline, and samples from 4 tissues (spleen, right cerebrum, cerebellum, and olfactory bulb; 64 samples total) were immediately frozen in liquid nitrogen. RNA was isolated. For each sample, labeled target was combined with 2× hybridization buffer, 3 nM B2 control oligonucleotide (catalog no. 900457; Affymetrix), 20× hybridization control stock (Affymetrix), and dimethyl sulfoxide (DMSO) making a final volume of 150 µl for the individual hybridizations to the Affymetrix GeneChip mouse gene 2.0 ST array containing the C57BL/6 mouse genome. The hybridization cocktail, including the components listed above, was denatured for 5 min at 99°C and then transferred to a 45°C heat block for an additional 5 min before transferring 130 µl of the cocktail onto the chip. The hybridization was carried out at a constant temperature of 45°C for approximately 40 h using an Affymetrix 640 hybridization oven. Upon completion of the hybridization step, each sample was removed from the chip and archived. Each chip was filled with approximately 160 µl of wash buffer A and then processed on the fluidics station 450. The reagents for the stain mixture consisted of 2× morpholineethanesulfonic acid (MES) stain buffer, 50 mg/ml of bovine serum albumin (BSA), 1 mg/ml of streptavidin phycoerythrin and water to make up a total volume of 600 µl for each stain. A holding buffer was added to make up a total volume of 800 µl for storage and scanning. Upon completion of the fluidics process, each sample was scanned using the Affymetrix GeneChip 3000 7Gplus scanner, and an expression console (Affymetrix version 1.3) was used to convert the data files to intensity (cel) files. The quality analysis was performed according to the “Quality Assessment of Exon and Gene Arrays” (Affymetrix revision 1.1). cel files representing individual samples were normalized using robust multiarray average (RMA) normalization followed by median normalization. Filtering was performed to remove any probe with mean log₂ expression of all samples below 5.0 or log₂ standard deviation of all samples below 1. Sample quality control was performed using principal component analysis (PCA) and sample-wise density plots in R. No outliers were identified in any of the aforementioned quality control methods used.

An empirical Bayes modified three-way repeated measure analysis of variance (ANOVA) was computed between the different treatment conditions using the limma package library in R to obtain false discovery rate (FDR)-adjusted P values and fold changes. The values for the probes were considered statistically significant if their FDR-adjusted P values were <0.05 and their absolute fold change was >1.5 except where otherwise noted. Gene symbols, log fold change ratios, P values, and false discovery rates from the empirical Bayes ANOVA were imported into Ingenuity Pathway Analysis (IPA) (Qiagen) to determine pathway enrichment scores and perform upstream regulator and regulator effect analyses. Network diagrams were exported from IPA, and heatmaps were generated with the pheatmap package and gplots libraries in R.

Statistical analyses of nonmicroarray data were computed using the latest versions of GraphPad Prism 6. Most comparisons are unpaired Mann-Whitney tests with Bonferroni's adjustments for multiple comparisons applied when appropriate. Student's t test was computed for the qPCR data to allow for easier fold change calculation. One-way ANOVA with Tukey posthoc adjustments for multiple comparisons was used for the log₁₀-transformed cell count data in Fig. 4. All survival curves are Kaplan-Meier curves with any log rank tests for any comparisons among curves.