Caspase-1 Activation of Interleukin-1β (IL-1β) and IL-18 Is Dispensable for Induction of Experimental Cerebral Malaria ^{▿ †}

Animals.

Procedures with experimental animals were performed according to the German Animal Protection Law, United Kingdom Home Office, and European regulations. C57BL/6AnNCr mice (6 to 10 weeks old) were obtained from Charles River Laboratories, Sulzfeld, Germany. Mice deficient in MyD88 (obtained from S. Akira [1]), TLR2/4 (originally obtained as singly deficient mice from S. Akira [47]), caspase-1 (obtained from BASF [30]), IL-1β (obtained from D. Chaplin [43]), or IL-18 (obtained from S. Akira [46]) in the C57BL/6 background were bred in a pathogen-free animal facility at the Max-Planck-Institute for Infection Biology and have been backcrossed to the C57BL/6 background for at least 10 generations.

P. berghei ANKA infection.

The complete life cycle of P. berghei ANKA was maintained by passage through naïve mice and Anopheles stephensi mosquitoes. The P. berghei ANKA parasites utilized in this study were clonal parasites which are passaged at regular intervals through the entire life cycle and originate from clone 15cy1, which had been engineered to express the green fluorescent protein (17). Infections were initiated by intravenous (i.v.) inoculation of 10⁴ sporozoites, purified from mosquito salivary glands, or similar numbers of infected red blood cells (iRBCs), obtained from infected C57BL/6 mice. Infected mice were observed daily for the manifestation of ECM, which was defined as the sudden onset of ataxia, paralysis, convulsion, or coma (12). The presence of ECM signs served as the ethical endpoint criterion for culling of animals. Spleens and brains were removed for analysis by flow cytometry. Infected mice that did not develop ECM until day 18 after infection were considered resistant and were sacrificed in order to avoid suffering from high parasitemia and severe anemia. The presence of blood-stage parasites was determined by daily examination of Giemsa-stained blood smears by oil immersion microscopy. For all infection experiments, parasitemia was assessed every other day by counting up to 50 microscopy fields.

Quantification of parasite development in the liver.

Total RNA was extracted from mouse livers at 44 h after infection using the RNeasy kit (Qiagen) and treated with DNase I. cDNA was synthesized using the RETROscript kit (Ambion). Hepatic parasite development was assessed by quantitative reverse transcription-PCR (RT-PCR) using SYBR green I master mix (Applied Biosystems). Reverse transcripts of P. berghei ANKA 18S rRNA were amplified in parallel with murine glyceraldehyde-3-phosphate dehydrogenase (mGAPDH) in a final volume of 25 μl in 96-well plates (Applied Biosystems). Oligonucleotide primers (Eurofins MWG Operon) for mGAPDH were 5′-CGT CCC GTA GAC AAA ATG GT-3′ (forward) and 5′-TTG ATG GCA ACA ATC TCC AC-3′) (reverse), and those for P. berghei ANKA 18S rRNA were 5′-AAG CAT TAA ATA AAG CGA ATA CAT CCT TAC-3′ (forward) and 5′-GGA GAT TGG TTT TGA CGT TTA TGT G-3′ (reverse).

Cytokine measurement.

Cytokine measurement was performed using the mouse inflammation cytometric bead array kit from BD Biosciences according to the manufacturer's instructions on a LSR II cell analyzer (BD Biosciences). Raw data were analyzed with FlowJo software (Tree Star). Plasma samples for in vivo measurement of cytokines were prepared from heparinized blood collected from naïve and infected mice.

Preparation of brain-sequestered lymphocytes.

Brain-sequestered lymphocytes were isolated as described previously (6). Perfused brains were digested in collagenase D buffer and lymphocytes isolated on a 30% Percoll (GE Healthcare) gradient. The isolated brain-sequestered lymphocytes were surface stained according to standard protocols. Antibodies for cell surface staining were obtained from eBiosciences (anti-mouse CD4 [GK1.5] and CD8 [53.6-7]). Antibodies for intracellular staining of IFN-γ (XMG1.2) were obtained from eBioscience. Cells were analyzed using an LSR II cell analyzer (BD Biosciences) and FlowJo software (TreeStar).

Statistical analysis.

Statistical significance was assessed using the Mann-Whitney test, with a P value of <0.05 taken as indicating a significant difference. Survival curves were compared by using the log rank (Mantel-Cox) test. All statistical tests were computed with GraphPad Prism 5 (GraphPad Software).

ECM pathogenesis is dependent on MyD88 and TLR2/4 signaling.

We first evaluated whether the MyD88 and TLR2/4 signaling pathways are involved in the development of ECM. For this purpose, we compared the courses of Plasmodium berghei ANKA infection in WT, MyD88-deficient, and TLR2/4-deficient C57BL/6 mice and contrasted infections caused by inoculation of 10⁴ sporozoites or the same number of infected red blood cells (iRBCs) (Fig. 1).

All infections initiated with sporozoites in WT, MyD88^−/−, and TLR2/4^−/− mice became patent at 3 to 4 days after infection (Fig. 1A). The majority of the sporozoite-infected WT mice (n = 25) developed ECM on day 7 or 8 after sporozoite inoculation (Fig. 1B). These results indicate that a blood-stage infection of as short as 4 to 5 days is sufficient to induce ECM. In sharp contrast, none of the sporozoite-infected MyD88^−/− (n = 18) and TLR2/4^−/− (n = 13) mice developed ECM. All animals that did not develop ECM were carefully culled on day 18 postinfection to avoid severe anemia and high parasitemia. There was no difference in parasitemia levels during the first 7 days of infection among sporozoite-inoculated WT, MyD88^−/−, and TLR2/4^−/− mice (Fig. 1A). Thus, we conclude that the MyD88-dependent TLR2/4 pathway is critically involved in induction of ECM upon a sporozoite-induced malaria infection.

We next compared disease progression in infections initiated with iRBCs that bypasses the initial liver stage of infection (Fig. 1D to F). All of the WT mice developed ECM on day 7 or 8 after infection. In contrast, 38% (6/16) of the infected MyD88^−/− mice had a significant delay in the onset of neurological signs (day 9 to 11 after infection; P < 0.0001 by log rank [Mantel-Cox] test), and 44% (7/16) did not develop ECM. Similarly, 33% (4/12) of the infected TLR2/4^−/− mice had a significant delay in the onset of neurological signs (day 9 to 11 after infection; P = 0.0002), and 42% (5/12) did not develop ECM. The survival curves for infected MyD88^−/− and TLR2/4^−/− mice did not differ significantly (P = 0.8). We did not observe any difference in parasitemia levels during the first 7 days of infection between WT, MyD88^−/−, and TLR2/4^−/− mice (Fig. 1D). A follow-up of the remaining mutant mice, which stayed free of ECM signs, revealed an increasing variability of parasitemia levels after day 7. All animals developed severe anemia and parasitemia above 20% by day 18 postinfection (data not shown).

One possible explanation of why we obtained disparate ECM levels when infection was initiated by sporozoite versus blood stages is that liver-stage development is different in WT versus MyD88^−/− or TLR2/4^−/− mice. To examine this possibility, we infected WT, MyD88^−/−, or TLR2/4^−/− mice with 10⁴ sporozoites. The livers of the animals were excised at 42 h postinfection and processed for quantitative real-time PCR to measure parasite development in the liver. As shown in Fig. 1G and H, we did not observe any difference in the development of the parasite in the liver.

Taken together, our results concur with reports indicating a partial dependency of ECM on MyD88 and TLR2/4 signaling following an iRBC-induced infection (11, 20). Strikingly, an absolute dependency is observed when infection is initiated with sporozoites, leading to a complete protection against ECM in the MyD88^−/− and TLR2/4^−/− mice. This difference cannot be attributed to differential development of parasites in the liver when infection is initiated with sporozoites.

Correlation of disease outcome with plasma cytokine levels.

ECM is characterized by excessive production of proinflammatory cytokines (41). We measured the systemic levels of various cytokines in the plasma of sporozoite-infected animals (Fig. 2). Infected WT mice had significantly elevated levels of plasma IFN-γ and monocyte chemotactic/chemoattractant protein 1 (MCP-1), which peaked on day 5 after sporozoite infection. Expression of TNF-α was significantly elevated on day 5 and even further increased on day 7. The anti-inflammatory cytokine IL-10 only showed a significant increase on day 7.

In infected MyD88^−/− mice, however, significantly lower IFN-γ and MCP-1 levels (P = 0.036 and P < 0.0001, respectively, by the Mann-Whitney test), which corresponded to significantly higher IL-10 levels (P = 0.0072), were observed on day 5 postinfection. In contrast, infected TLR2/4^−/− mice presented significantly elevated plasma concentrations of IFN-γ, TNF-α, and IL-10 (P = 0.0046, P = 0.042, and P = 0.0097, respectively) that exceeded those of infected WT mice on day 5 after infection. While beyond the scope of the current study, it will be interesting to analyze the underlying molecular mechanisms that can explain these differential effects.

Inoculation of iRBCs induced plasma cytokine kinetics (see Fig. S1 in the supplemental material) that were comparable to those after infection initiated with sporozoites. However, while day 5 MCP-1 levels were significantly lower in MyD88^−/− mice than in WT mice, no difference in the IFN-γ and IL-10 levels in the two groups was observed. The disparate IFN-γ and IL-10 levels in infections initiated by different stages of the Plasmodium parasite in WT and MyD88^−/− mice may possibly explain the partial susceptibility of mutant mice to ECM when infected with blood stages. Overall, our data demonstrate that the innate host defense system contributes critically to ECM.

ECM pathogenesis is independent of caspase-1, IL-1β, and IL-18.

Other receptors that are known to signal through MyD88 are the IL-1 and IL-18 receptors (1). Synthesized as propeptides, mature IL-1β and IL-18 are products of activated caspase-1, which in turn is tightly regulated by the Nalp3 inflammasome, another innate immune pathway involved in inflammatory responses (32).

To determine a role for caspase-1 activation of IL-1β and IL-18 during ECM, we compared the courses of P. berghei ANKA infection in WT, caspase-1^−/−, IL-1β^−/−, and IL-18^−/− mice inoculated with sporozoites (Fig. 3 A to C). Infected enzyme/cytokine-deficient mice developed ECM as WT mice did (Fig. 3B). Furthermore, we did not observe any difference in parasitemia levels during the first 7 days of infection in the different groups (Fig. 3A). Similar results were obtained when infections were initiated with iRBCs (Fig. 3D to F). Collectively, these results indicate that caspase-1 activation of IL-1β and IL-18 is dispensable for the development of effector responses associated with ECM.

Consistent with the above results, the plasma cytokine profiles of IFN-γ, MCP-1, TNF-α, and IL-10 in sporozoite-inoculated caspase-1^−/−, IL-1β^−/−, and IL-18^−/− mice were indistinguishable from those in infected WT mice (Fig. 4). Similarly, using iRBCs for infection, we did not observe any significant differences in the cytokine profiles among the different groups (see Fig. S2 in the supplemental material).

To examine the induction of IFN-γ responses, we additionally performed short-term stimulation of day 7 splenocytes from WT and enzyme/cytokine-deficient mice in the presence of phorbol myristate acetate (PMA)/ionomycin. In good support of our findings from the plasma cytokine profiling, intracellular staining analysis of CD4⁺ and CD8⁺ cells from infected WT and enzyme/cytokine-deficient mice showed comparable frequencies of cells producing IFN-γ (see Fig. S3 in the supplemental material).

The IFN-γ-controlled trafficking of pathogenic CD8 T cells in the brain is a canonical feature of ECM (6, 7, 23). Therefore, we isolated brain-infiltrating lymphocytes from WT and enzyme/cytokine-deficient mice at the time the animals were committed to neurological signs (Fig. 5). As predicted from the clinical outcome, WT and enzyme/cytokine-deficient mice had similar proportions of brain-infiltrating CD8⁺ T cells. This infiltration was independent of the induction of infection by either sporozoites (Fig. 5A and B [left panel]) or iRBCs (Fig. 5B [right panel]). Therefore, we conclude that caspase-1 activation of IL-1β and IL-18 plays no pathological role in ECM in vivo.

Interference with IL-1β signaling does not offer an experimental therapeutic approach.

Our findings of a dispensable role for IL-1β signaling in ECM further substantiate previous work showing that IL-1R-deficient mice succumb to ECM indistinguishably from WT mice (38). To obtain further independent evidence and to examine a potential drug target to prevent severe forms of malaria, we tested whether treatment with the IL-1β antagonist anakinra (24) modulates the clinical outcome of a malaria infection (Fig. 6). High doses of anakinra (250 mg/kg) were administered to mice for 7 consecutive days following infection with P. berghei ANKA sporozoites. In agreement with our data from the knockout mice, treated and untreated mice showed similar courses of infection (Fig. 6A). Importantly, animals in both groups displayed signs of ECM at 7 days after infection to similar degrees (Fig. 6B). As a consequence, no protection against ECM was detected in anakinra-treated animals (Fig. 6C). These findings independently reject the hypothesis that the IL-1β signaling pathway plays an important role in sporozoite-induced ECM.