Endothelin-1 Mediates Brain Microvascular Dysfunction Leading to Long-Term Cognitive Impairment in a Model of Experimental Cerebral Malaria

Plasmodium falciparum infection causes a wide spectrum of diseases, including cerebral malaria, a potentially life-threatening encephalopathy. Vasculopathy is thought to contribute to cerebral malaria pathogenesis. The vasoactive compound endothelin-1, a key participant in many inflammatory processes, likely mediates vascular and cognitive dysfunctions in cerebral malaria. We previously demonstrated that C57BL6 mice infected with P. berghei ANKA, our fatal experimental cerebral malaria model, sustained memory loss. Herein, we demonstrate that an endothelin type A receptor (ETA) antagonist prevented experimental cerebral malaria-induced neurocognitive impairments and improved survival. ETA antagonism prevented blood-brain barrier disruption and cerebral vasoconstriction during experimental cerebral malaria, and reduced brain endothelial activation, diminishing brain microvascular congestion. Furthermore, exogenous endothelin-1 administration to P. berghei NK65-infected mice, a model generally regarded as a non-cerebral malaria negative control for P. berghei ANKA infection, led to experimental cerebral malaria-like memory deficits. Our data indicate that endothelin-1 is critical in the development of cerebrovascular and cognitive impairments with experimental cerebral malaria. This vasoactive peptide may thus serve as a potential target for adjunctive therapy in the management of cerebral malaria.


Introduction
Malaria, caused by infection with the intraerythrocytic parasite Plasmodium falciparum, is a potentially life threatening disease, with significant morbidity and devastating economic consequences in developing countries [1]. Acute infection causes a spectrum of disease including cerebral malaria (CM), the most severe manifestation of infection. Despite extensive research, there remains a critical gap in our understanding of the mechanisms that promote the CNS pathology, and the ensuing neurocognitive deficits that result from infection. Consequently, the case fatality from CM remains high at 20%, and one in every four survivors develops longterm neurological sequelae even after successful parasite eradication [2][3][4].
The actions of ET-1 are mediated through two G-protein-coupled receptors ET A and ET B , which modulate inflammation, BBB integrity, and vascular tone [13][14][15][16]. Our published data demonstrate elevated mRNA expression of ET-1 and of the ET receptors in the brains of mice with ECM, in association with reduced CBF [6]. These reports mirror observation in patients with severe P. falciparum malaria [17,18].
We recently reported that ET A receptor antagonism reduced the incidence of brain hemorrhage in ECM [12]. Using a multidisciplinary approach, we have now extended these studies to demonstrate, for the first time, a mechanistic role for ET-1 in the development of endothelial dysfunction and the generation of ECM. Herein we report that ET-1 is essential in mediating the vascular dysfunction and the subsequent associated cognitive impairment during ECM. In this regard, we demonstrate that ET A receptor antagonism prevents visual memory impairments as a consequence of attenuating cerebral vasoconstriction, BBB disruption, and vascular congestion.

Results
ET A receptor antagonism improves survival and disease severity during the course of acute P. berghei ANKA (PbA)-infection We examined the effect of the selective ETA receptor antagonist, BQ123, on uninfected and PbA-infected mice. Despite having no antimalarial properties (Fig 1A), BQ123 treatment significantly prolonged survival following PbA infection (Fig 1B). BQ123 also improved disease severity and the development of neurological signs associated with ECM, in PbA-infected mice, as measured by the rapid murine coma and behavior scale (RMCBS) (Fig 1C). Interestingly, BQ123 had no effect on weight or on body temperature in healthy uninfected mice, nor on the body weight or temperature fluctuations in PbA-infected mice (S1 Fig). ET A receptor antagonism prevents memory dysfunction during acute PbA-infection Our laboratory previously proposed that elevated levels of ET-1 in the brain may contribute to ECM pathogenesis [6,12]. Corroborating our previous studies, we herein demonstrate that C57BL/6 mice infected with PbA developed visual memory loss during the acute phase of infection (Fig 1) [9]. PbA-infected mice performed significantly worse during the Object Recognition (OR) test than uninfected age-matched controls (Fig 2A; 71.6 ± 3% in uninfected animals vs. 54.4 ± 3% in PbA-infected animals; P < 0.01). Furthermore, a significantly lower proportion of infected mice successfully identified the novel object in the OR test when compared to uninfected controls ( Fig 2B).
To determine whether blocking the ET A receptor would prevent the ECM-induced neurocognitive impairment, infected mice were treated with the selective ET A receptor blocker (ETARB), BQ123. Infected mice treated with BQ123 performed significantly better in the OR test than infected mice treated with saline (Fig 2A; 77.2 ± 5% vs. 54.4 ± 3%; P < 0.001). Unlike untreated infected mice, all PbA-infected mice treated with BQ123 successfully completed the OR test, similar to healthy uninfected controls ( Fig 2B).
Memory impairment in ECM mice was not associated with any abnormalities in motor performance or exploratory behavior as all mice had similar motor performance and total exploration time (S2A and S2B Fig).
ET A receptor antagonism prevents persistent visual memory loss in convalescent mice after successful antimalarial treatment Both human and experimental CM studies demonstrate that even after successful treatment of malarial infection, cognitive deficits persist [3,10,19]. Adjunctive BQ123 treatment in PbAinfected mice treated with the antimalarial agent, artemether, significantly improved survival, aiding in the recovery from ECM (Fig 3). Consistent with our previous publications, artemether-treated PbA-infected mice displayed significant visual memory loss in the OR test of visual memory 10 days after the cessation of anti-malarial treatment, despite parasite eradication (Fig 4A; 46.4 ± 9.9% in PbA infected mice vs. 82.3 ± 3% in uninfected mice; P < 0.01). A significantly higher percentage of the artemether-treated PbA-infected mice demonstrated memory impairment than uninfected control mice ( Fig 4B). Since treatment with BQ123 prevented cognitive dysfunction during acute illness, we tested whether this protection would be sustained in mice with persistent deficits after antimalarial treatment. Adjunctive therapy with BQ123 prevented visual memory deficits in ECM mice 10 days after antimalarial treatment (Fig 4A and 4B; 71.9 ± 6% in PbA+BQ123+Artemether vs. 46.4 ± 9.9% in PbA+Artemether; P < 0.05). BQ123 adjunctive therapy prevented memory impairment in a significantly higher percentage of infected mice treated with artemether ( Fig 4B). Mice treated with adjunctive BQ123 exhibited similar preference scores to controls (Fig 4A; 71.9 ± 6% in PbA+BQ123+-Artemether vs. 82.3 ± 3% in uninfected mice; P = NS).
An estimated 20-25% of patients who survive an episode of CM exhibit long-term cognitive deficits, underscoring an important need for adjunctive therapies in the management of CM [20][21][22][23]. As we previously observed, the cognitive deficits detected early after the cessation of anti-malarial therapy were persistent in ECM mice one month later. Forty days after the cessation of artemether therapy, visual memory deficits were evident following a 60-minute retention interval (Fig 4C; 49 ± 3% in infected vs. 68.5 ± 4% in uninfected; P < 0.01), and a significantly higher percentage of infected mice exhibited random exploration in the object recognition test (Fig 4D). Preference scores of ECM mice that received adjunctive therapy with BQ123 were comparable to those of uninfected controls (Fig 4C; 68.5 ± 4% in uninfected vs. 59.9 ± 4% in BQ123 treated; P = NS). However, BQ123 adjunctive treatment conferred only a partial long-term benefit in convalescent ECM mice. Although a significantly higher percentage of ECM mice treated with BQ123 preferentially explored the novel object than ECM mice treated with artemether alone (Fig 4D; 58% in BQ123-treated vs. 37% in saline treated; P < 0.05), BQ123 treatment conferred only a partial protection of visual working memory in artemether treated ECM mice at 40 dpt (Fig 4C; 59.9 ± 4 in BQ123 treated vs. 49 ± 3 in saline treated; P = 0.058).
Memory impairment in artemether-treated ECM mice was not associated with abnormalities in motor performance or exploratory behavior as all mice had similar motor performance and total exploration time at 10 and 40 dpt (S3A- S3C Fig). BQ123 effects on parasitemia, survival, and disease severity. (A) BQ123 treatment had no effect on parasitemia in PbA-infected mice; Two-way ANOVA. (B) BQ123 treatment significantly prolonged survival in PbA-infected mice; Log-rank (Mantel-Cox) test. (C) PbA-infected mice demonstrated gradual decreases in the rapid murine coma and behavior scale (RMCBS) after infection with significant changes occurring at 6 and 7dpi. Although BQ123-treated infected mice also scored significantly lower in the RMCBS than uninfected controls at 7dpi, BQ123 significantly dampened the decline in RMCBS scores in infected mice; Two-way ANOVA; ** = p < 0.01, *** = p < 0.001 and **** p < 0.0001. For graph C: * = Con vs. PbA; # = Con vs. PbA+BQ123; † = PbA vs. PbA+BQ123. n = 10/group.  Endothelin-1 and Experimental Cerebral Malaria ET A receptor antagonism has a positive impact on cerebral microvascular constriction Vascular dysfunction including large vessel infarcts, cerebral edema, and impaired tissue perfusion in the retinal microvasculature are important components of CM pathogenesis [24][25][26]. The neurological impairments present in ECM are often associated with vascular complications including vasoconstriction, vascular collapse, and vascular congestion ultimately leading to disruption of the BBB [7,27]. In order to determine the association between ET-1 induced microvascular disease and adverse cognitive outcomes during ECM, we assessed cerebral microvascular patency using intravital microscopy through a closed cranial window. As demonstrated in Fig 5A, mean baseline vessel diameters did not vary significantly between the different experimental groups (Fig 5; 46.5 ± 4 in control v. 51.9 ± 5.5 in PbA v. 46.7 ± 5.9 in PbA +HJP272; P = NS). In corroboration with previous studies [5,28], mice with ECM had a significant reduction in pial arterial diameter at 6 dpi, while healthy uninfected animals maintained stable vessel diameters over the course of the experiment (

ET A receptor antagonism ameliorates endothelial activation and vascular congestion during PbA-infection
During CM, there is an increase in brain microvascular endothelial cell activation, critical for parasitized red blood cell (pRBC), platelet and leukocyte adhesion, resulting in sequestration, capillary obstruction, localized hypoxia, tissue injury and subsequent neurocognitive impairment or death [29][30][31]. Endothelial activation and microvascular obstruction are hallmark features of CM, and are associated with coma in patients with CM [24]. Using a semiquantitative scoring system we assessed the degree of cellular retention within the cerebral vasculature, measured as percent obstruction of vessel lumen. Mice with ECM displayed a significant degree of vessel congestion, measured by the degree of vessel lumen obstruction (Fig 6C and 6D), that was not evident in uninfected mice in several brain regions, including the brainstem (  Pro-inflammatory cytokines and cell adhesion molecules have been shown to contribute to the CM disease process by inducing the sequestration of pRBCs and the recruitment and adhesion of leukocytes to vascular beds [32][33][34]. These inflammatory mediators are increased in response to ET-1 [14,35,36], and likely participate in the vascular obstruction observed in our mice. Quantitative real-time PCR at 7 dpi demonstrated that PbA infection resulted in significantly higher expression of tumor necrosis factor-α (TNFα, Fig 6I; Log 10 mRNA: 0.18 ± 0.04 vs. 0.03 ± 0.02; P < 0.01) and the monocyte chemoattractant, CCL2 (Fig 6J; Log 10 mRNA: 0.32 ± 0.04 vs. -0.05 ± 0.07; P < 0.0001) when compared to uninfected mice. In addition, immunoblot and analysis of mice with ECM demonstrate significantly higher vascular cell ET A receptor antagonism stabilizes the blood brain barrier during ECM Endothelial activation and vascular obstruction contribute to the loss of vascular integrity during CM pathogenesis. During ECM, ET-1 expression is increased in the brain. ET-1 has been shown to induce disruption of the blood brain barrier (BBB), increasing paracellular permeability [37,38]. Using a small animal in vivo imaging system, we measured BBB leakage by quantifying near-infrared fluorescence intensity of Tracer-653 (Fig 7). PbA infection resulted in impaired vascular integrity, significantly increasing BBB leakage as early as 4 dpi (Fig 7B; 2.52 ± 0.5 in PbA vs. 1 ± 0.2 in control; P = < 0.01), and sustained at 7 dpi, during terminal disease (Fig 7D; 1.42 ± 0.05 in PbA vs. 1 ± 0.03 in control; P < 0.0001). BQ123 treatment in PbAinfected mice prevented the BBB disruption observed in ECM mice treated with saline (Fig 7), Endothelin-1 and Experimental Cerebral Malaria and protected BBB integrity at both earlier (Fig 7B; 1.39 ± 0.2 in PbA+BQ123 vs. 2.52 ± 0.5 in PbA; P < 0.05) and later (Fig 7D; 1.14 ± 0.05 vs. 1.42 ± 0.05; P < 0.01) stages of disease.

ET-1 induces cognitive dysfunction in a non-CM model of malarial infection
Infection of C57BL/6 mice with P. berghei NK65 (PbN), a model often used as a non-CM control for PbA induced ECM, results in a syndrome of severe malaria, but does not cause ECM [8,39]. While PbA-infected mice display significantly higher expression of all the components of the ET-1 pathway when compared to uninfected controls [6], PbN-infected mice had significantly lower ET-1 expression than ECM mice (Fig 8; 0.94 ± 0.19 PbN vs. 1.65 ± 0.2 PbA; P < 0.05), with levels comparable to those in uninfected controls (Fig 8; 0.94 ± 0.19 vs. 0.86 ± 0.07; P = NS).
Unlike ECM mice, PbN-infected mice did not exhibit visual memory deficits at 6dpi (Fig  9A; 53.53 ± 2.9% in ECM vs. 69.07 ± 5.8% in non-CM, PbN-infected mice; P < 0.05). In addition, a greater proportion of non-CM mice were able to successfully completed the object recognition test of visual memory compared to ECM mice (Fig 9B; 86% non-CM vs. 39% ECM; P < 0.001).PbN-infected mice performed as well as uninfected controls in the object recognition test (Fig 9A; 69.07 ± 5.8% in PbN mice vs. 71.59 ± 2.7% in uninfected controls; P = NS). To establish whether the deleterious effects of ECM were in fact induced by increased production of ET-1, we treated PbN-infected mice with exogenous ET-1 and tested their visual memory. Similar to PbA-infected mice, ET-1 treated non-CM mice displayed impaired visual learning (Fig 9A; 45.22 ± 7.7% PbN+ET-1 vs. 71.59 ± 2.7% controls; P < 0.01). A significantly lower proportion of non-CM mice treated with ET-1 were able to successfully complete this memory task when compared to PbN-infected mice treated with saline ( Fig 9B; 86% in PbN Endothelin-1 and Experimental Cerebral Malaria infected mice vs. 25% in PbN+ET-1 mice; P < 0.001), supporting our theory that cognitive impairment in ECM is induced by ET-1.

Discussion
CM has long been thought to cause a vascular inflammatory process in the brain [11,27]. Yet a comprehensive delineation of the precise mechanisms by which malarial infection promotes cerebral vasculopathy, BBB disruption, CNS pathology, and the ensuing neurocognitive deficits remains elusive; impeding successful efforts to develop adjunctive therapies. Previous research has shown that ECM results in imbalances in vascular tone and vasoactive mediators [27,40,41]. Furthermore, pRBCs, platelets and leukocytes have been shown to sequester to cerebrovascular beds secreting cytokines and chemokines further obstructing cerebral blood flow triggering a vasoactive, pro-inflammatory environment [24,31,40,[42][43][44]. This environment leads to neurological and cognitive complications, which are known to persist long after the resolution of infection [3,4,21,22].
Increased concentrations of ET-1 have been reported in patients with P. falciparum infection [17,45]. Interestingly, TNF-α reduces cerebral blood volume, and disrupts the BBB and tissue homeostasis in Wistar rats in an endothelin dependent manner [13]. Our group previously demonstrated that ET-1 was elevated in the brains of mice with ECM [6]. In the current study, we demonstrated that cerebrovascular constriction, the activation of certain cytokines and chemokines, vascular congestion, and BBB disruption, as well as the ensuing increase in mortality and acute and chronic visual memory impairments during ECM were mitigated by treatment with an ET A receptor blocker. More importantly, our observations were substantiated by the induction of visual memory deficits in non-CM (PbN-infected) mice upon administration of exogenous ET-1. While the direct effects of P. falciparum infection on the expression and function of ET receptors have not previously been investigated; altogether, these data strongly support a role for ET-1 in the pathogenesis of ECM. Treatment with ET A receptor antagonists have been effective in preventing cognitive impairment in other animal models [46,47]. Although we acknowledge that in the absence of genetic knockout experiments to confirm selectivity, the possibility of off-target effects of ET-1 receptor antagonism are inherent limitations in the study, the protective effects of ET A receptor antagonism with Endothelin-1 and Experimental Cerebral Malaria ECM observed herein can be ascribed to a reduction in malaria-induced cerebrovascular constriction and congestion as well as prevention of ET-mediated endothelial activation and BBB leakage.
A major physiological role of ET-1 is the regulation of vascular tone. ET-1 binding to the ET A receptor induces vasoconstriction and reduction of CBF by inhibiting nitric oxide (NO) production and increasing intracellular concentrations of Ca 2+ in vascular smooth muscle cells [28,[48][49][50]. Previously, our laboratory group demonstrated an association between increased ET-1 expression and a decrease in CBF in mice with ECM [6]. Previous intravital microscopic studies demonstrate that ECM resulted in the constriction of pial vessels in the brain, contributing to diminished CBF and vascular collapse reminiscent of subarachnoid hemorrhage [7,41]. Using intravital microscopy through a closed cranial window we demonstrated that ET A receptor antagonism prevented vasoconstriction during ECM. ET-1 has been shown to increase the concentration of intracellular Ca 2+ from internal and external stores, mediating cell death [51]. In several studies, blockage of the ET A receptor prevents intracellular mobilization of Ca 2+ , and reduces intracellular accumulation of Ca 2+ , ultimately preventing vessel constriction [52,53]. Vasoconstriction in CM is likely a result of ET-1 mediated increases in Ca 2+ mobilization in the CNS [7,27,41].
In the current studies, ET A antagonism also resulted in suppression of the infection-induced inflammatory response. Furthermore, ET A antagonism blunted the associated vascular obstruction resulting from leukocyte, pRBC and platelet accumulation. VCAM-1 expression was only partially reduced in the brains of PbA infected mice treated with the ETARB BQ123 toward uninfected levels. This may indicate that additional factors beyond ET-1 are involved in the induction of VCAM-1 during ECM, or alternatively, that increases in VCAM-1 could have limited reversibility during PbA infection. Activation of the cerebrovascular endothelium contributes to the development of both human and murine models of CM [30]. In this regard, cerebral vascular congestion has been shown to correlate with coma in the setting of CM [24]. The upregulation of cytokines, such as TNFα, which results in the increased expression of leukocyte binding adhesion molecules and chemokines [54], is a hallmark of CNS infection [34]. In this regard, Wassmer et. al. demonstrated that cultured vascular endothelial cells isolated from patients with CM significantly upregulated their production of CCL-2 in response to TNFα stimulation [55]. This increased production may be dependent on ET-1. The ET-1-mediated increase in CCL-2 production observed in our model likely increases the recruitment and diapedesis of immune cells into the brain parenchyma, further exacerbating neuroinflammation and ultimately disrupting BBB integrity. Administration of ET A receptor antagonists in our mice prevented ECM-induced accumulation of cells in the brain microvasculature, ultimately maintaining cerebral vascular patency.
Our observations support the concept that ET-1 plays a central pathogenic role in the development of CM. The current study provides evidence that endothelial activation and inflammation, as well as vascular constriction, occlusion, and permeability contribute to the cognitive impairment and mortality following PbA-infection. We propose that ET A antagonism may be beneficial in CM by inhibiting the vasoconstriction, endothelial activation, and BBB disruption that occur with infection. The efficacy of this therapy in our study is likely due to the observed protective effects on the microvascular endothelium, as the induction of cellular adhesion molecules has been associated with ETA signaling on endothelial cells [14,56], and on the inflammatory response in the brain [13,35,36,55]. By reversing endothelial activation and neuroinflammation, ETARB may restore vascular patency and CBF, ultimately improving oxygen delivery to the brain. We previously observed that with resolution of acute P. berghei infection, there remained significant brain microvascular damage during the early convalescent period [12], which was resolved by 40 dpt [10]. As BQ123 does not cross the BBB [57], the partial protection of neurological function at 40 dpt, conferred by BQ123 adjunctive therapy in the current study, may be due to possible residual CNS damage downstream of the microvasculature. Additional studies, directly targeting CNS ET-1 and ET receptors, should help determine the full extent of ET-1 induced CNS damage during ECM.
Since the discovery of ET-1 in the late 1980s it has been implicated in the pathogenesis of many disease states. These include congestive heart failure, subarachnoid hemorrhage, preeclampsia, cancer, hypertensive crisis and primary pulmonary hypertension [58]. In the realm of infectious diseases bedsides malaria ET -1 has also been linked to bacterial sepsis, meningitis and pneumonia [28]. In clinical trials, ET receptor antagonism has been promising in the management of primary hypertension [58], skin microcirculation and limb ischemia in diabetic microangiopathy [59,60], and ET receptor blockers have been the mainstay of therapy in the management of primary pulmonary hypertension [58,61].
While we recognize that further studies are required to determine other mechanisms by which ET-1 participates in persistent cognitive deficits or in mortality during acute illness, we conclude that there is therapeutic potential in using the ET system as a target for adjunctive therapy. In conjunction with anti-malarials, pharmacologic antagonism of the ET system may be of value in the management of CM.

Ethics statement
All experimental protocols were carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health, reviewed and approved by the Institutional Animal Care and Use Committee of the Albert Einstein College of Medicine (protocol number: 20130602). All efforts were made to minimize suffering.
Disease severity was scored according to a grading system described by Carroll et al [63]. A quantitative rapid murine coma and behavior scale (RMCBS) was used to assess the development of ECM. The RMCBS comprised of 10 parameters in which hygiene-related behavior, gait, body position, exploratory behavior, and balance [63] were monitored daily throughout the study for health status, as were weight and temperature. Parasitemia (percentage of pRBCs), assessed by examination of tail blood smears stained with Giemsa, was also monitored daily.
Intraperitoneal osmotic minipumps (model #1002, Alzet) were placed into PbA-infected mice, which were then randomly assigned to receive continuous infusion of either BQ123 (120/ug/kg/d, selected based on dose titration experiments; Sigma) or vehicle (sterile normal saline) via the pumps for a total of 10 days. Mice undergoing cranial window surgery were treated once daily with either 50 mg/kg of HJP-272 via IP injection or saline, initiated at 3dpi for a total of 10 days. In some experiments, PbA-infected mice were allowed to manifest the signs of ECM, as previously described [10], then treated with Artemether (at a dose of 50mg/ kg; Kind gift from Dafra Pharma GmbH, Switzerland) dissolved in coconut oil and administered by i.p. injection, for a total of 5 days [12].
In a separate set of experiments PbN-infected mice were treated daily with either exogenous ET-1, dissolved in normal saline (7.5ug/mouse; Millipore), or with normal saline, administered by i.p. injection beginning at day 3 after infection.
Mice were sacrificed and intracardially perfused with ice-cold PBS. Brains were excised and stored at -80°C for analysis.

Cognitive tests
Behavioral tests were performed to determine early and persistent cognitive deficits present during ECM [9,10]. Animals underwent cognitive testing at 3 different time points: 6 days post infection (dpi), 10 days and 40 days after the cessation of Artemether treatment (S4 Fig). Object recognition tests were performed to assess visual memory.
During the object recognition test animals underwent a training and test trial, as previously described [9,10,64]. This test is based off the inherent tendency of mice to preferentially explore novel objects [9,10,64]. In the training trial mice were gently placed in an open field with two identical objects. Mice were allowed to habituate and object exploration of each object was recorded. Mice were then returned to their home cage for a retention interval of 60min. After a 60min retention interval mice were returned to the open field with the one familiar object and one novel object. Again exploration of each object was recorded.
Healthy mice preferentially explore novel objects. Data are represented as the amount of novel object exploration divided by the total time exploration of both objects. A preference score of 50% indicates chance performance, whereas a score greater than 55% indicates novel object preference. Test objects were counterbalanced to prevent confounds of preferential exploration of an object. Examiners were blind to infection and treatment conditions of the mice.

Cranial window surgery and intravital microscopy
A closed cranial window was placed as previously described [7]. Briefly, mice were anesthetized, the scalp was then removed and a craniotomy of 3-4 mm in diameter was created, using a surgical drill. The exposed tissue was covered with a 5mm glass coverslip. Two weeks after surgery, a panoramic photograph of vessels was taken and a map of selected vessels was created. Approximately 7-10 vessels were imaged in each animal at baseline and 6 dpi. Vessel diameters at 6 dpi were compared to baseline measurements in the same mice for comparison. Data are represented as percent change in vessel diameter to baseline measurements.

Histology and immunohistochemistry
Mice were anesthetized by isoflurane inhalation and perfused through the heart with saline. Brains were excised, fixed and embedded in paraffin. Brains were cut into 4 μm sagittal sections and stained with hematoxylin-eosin. A congestion score adapted from Fauconnier et. al. was performed to quantify the degree of obstruction in the cerebral vasculature after perfusion [66]. Briefly, approximately 500 vessels were analyzed per group. Congestion scores were based on a scale of 0-4; 0 = No cells present within the vessel lumen; 1 = vessels with a minimal amount of cells, including RBCs, platelets, and adherent leukocytes, obstructing less than 10% of the vessel lumen; 2 = vessels with a moderate degree of cellular obstruction, with 10% to 50% occlusion of the vessel lumen; 3 = vessels with an extensive amount of cellular obstruction, with 50% to 90% occlusion of the vessel lumen; 4 = completely occluded vessel lumen. Immunohistochemistry was performed on sagittal sections as previously described [6]. VCAM-1 (Santa Cruz, Dallas, TX) expression was assessed in experimental animals. Isotype control staining was performed to account for non-specific binding to tissue.

Statistical analysis
Statistical analyses were performed using one-way ANOVA, with post-hoc group comparisons tests. In cases where the variances between samples were significantly different, and/ or the sample distributions were significantly skewed, i.e., violated normal Gaussian curves, a logarithmic transformation was performed on the samples to yield variance homogeneity and/ or normal distribution, and ANOVA were calculated on the transformed data. Success rates in cognitive tests were analyzed by Chi Square tests. Results are shown as mean ± S.E.M. Analyses were conducted with the GraphPad Prism Software (GraphPad Software Inc., La Jolla, CA) or with the IBM SPSS Statistics v23 software (IBM Corporation, Armonk, NY). A P < 0.05 was considered significant.
Additional methods can be found in S1 Text. At day 0, mice were injected with either parasitized RBCs (pRBCs) or uninfected RBCs. Treatment with BQ123 was started at 3 dpi for a total of 10 days. For acute illness assessment, cognitive testing was performed 6 dpi (acute cognitive testing). Treatment with artemether was initiated after day 7 and continued for 5 days. Cognitive testing was again performed 10 days after the cessation of treatment (cognitive testing 10 dpt), and again 40 days after the cessation of treatment (cognitive testing 40 dpt) to examine long-term cognitive function.