ORIGINAL ARTICLE

S-nitrosoglutathione Prevents ExperimentalCerebral Malaria

Graziela M. Zanini & Yuri C. Martins & Pedro Cabrales &

John A. Frangos & Leonardo J. M. Carvalho

Received: 12 October 2011 /Accepted: 6 February 2012# Springer Science+Business Media, LLC 2012

Abstract Administration of the exogenous nitric oxide(NO) donor dipropylenetriamine-NONOate (DPTA-NO) tomice during Plasmodium berghei ANKA (PbA) infectionlargely prevents development of experimental cerebral ma-laria (ECM). However, a high dose (1 mg/mouse twice aday) is necessary and causes potent side effects such asmarked hypotension. In the present study we evaluatedwhether an alternative, physiologically relevant NO donor,S-nitrosoglutathione (GSNO), was able to prevent ECM atlower doses with minimal side effects. Prophylactic treat-ment with high (3.5 mg), intermediate (0.35 mg) or low(0.035 mg) doses of GSNO decreased incidence of ECM in

PbA-infected mice, decreasing also edema, leukocyte accu-mulation and hemorrhage incidence in the brain. The highdose inhibited parasite growth and also induced transienthypotension. Low and intermediate doses had no or onlymild effects on parasitemia, blood pressure, and heart ratecompared to saline-treated mice. PbA infection decreasedbrain total and reduced (GSH) glutathione levels. Brainlevels of oxidized (GSSG) glutathione and the GSH/GSSGratio were positively correlated with temperature and motorbehavior. Low and intermediate doses of GSNO failed torestore the depleted brain total glutathione and GSH levels,suggesting that ECM prevention by GSNO was probablyrelated to other effects such as inhibition of inflammationand vascular protection. These results indicate that ECM isassociated with depletion of the brain glutathione pool andthat GSNO is able to prevent ECM development in a widerange of doses, decreasing brain inflammation and inducingmilder cardiovascular side effects.

Keywords Cerebral malaria . Nitric oxide .

S-nitrosoglutathione . Plasmodium berghei

Introduction

Malaria still imposes a heavy burden to a large fraction ofthe World population. Estimates by the World Health Orga-nization points to the occurrence of 225 million cases ofmalaria worldwide in 2010, with 709,000 deaths, 85% ofwhich occurring in children under 5 years of age (WHO2010). Cerebral malaria (CM) is one of the most lethalcomplications of malaria by Plasmodium falciparum anddespite effort in this area there are no currently availableadjunctive therapies (John et al. 2010). Experimental (mu-rine) cerebral malaria (ECM) caused by Plasmodium

Graziela M. Zanini and Yuri C. Martins equally contributed to thisstudy.

G. M. Zanini :Y. C. Martins : P. Cabrales : J. A. Frangos :L. J. M. Carvalho (*)Center for Malaria Research, La Jolla Bioengineering Institute,3535 General Atomics Court suite 210,San Diego, CA 92121, USAe-mail: lcarvalho@ljbi.org

G. M. ZaniniParasitology Service, Evandro Chagas Clinical Research Institute,Fiocruz,Rio de Janeiro, Brazil

Y. C. MartinsLaboratory of Inflammation and Immunity,Institute of Microbiology, Federal University of Rio de Janeiro,Rio de Janeiro, Brazil

P. CabralesDepartment of Bioengineering, University of California,San Diego, CA, USA

L. J. M. CarvalhoLaboratory of Malaria Research, Oswaldo Cruz Institute, Fiocruz,Rio de Janeiro, Brazil

J Neuroimmune PharmacolDOI 10.1007/s11481-012-9343-6

berghei ANKA (PbA) is characterized by brain microcircu-latory complications leading to vasoconstriction, markeddecreases in cerebral blood flow and eventually vascularcollapse (Cabrales et al. 2010). These complications areassociated with a state of low nitric oxide (NO) bioavail-ability likely caused by the NO-scavenging action of gramamounts of cell-free hemoglobin in the plasma consequentto red blood cell (RBC) lysis during the parasite cycle,which can be reverted by the administration of high dosesof the potent NO-donor dipropylenetriamine-NONOate(DPTA-NO) (Gramaglia et al. 2006). Treatment withDPTA-NO during PbA infection reduces ECM incidence,improves cerebral blood flow and venous return, preventsblood brain barrier leakage, and decreases hemorrhage inci-dence, leukocyte and platelet adherence, and expression ofendothelial cell adhesion molecules including ICAM-1 andP-selectin in the brain (Cabrales et al. 2011; Zanini et al.2011). Protection is also associated with decreased levels ofinflammatory markers in the plasma and restoration of cyclicguanosine monophosphate levels in the brain (Gramaglia et al.2006).

Although DPTA-NO has been valuable in providingproof of concept for the role of NO in ECM, its furtheruse to explore potential therapeutic applications of thisconcept is hindered by a number of limitations. The highdose of DPTA-NO effective in preventing ECM (1 mg/mouse), combined with a fast and short-lived release ofNO by DPTA-NO in aqueous solutions, results in the gen-eration of a huge amount of NO in a short period of time,which leads to the occurrence of acute side effects such asmarked hypotension (Gramaglia et al. 2006). We have beenresearching alternative strategies to increase NO bioavail-ability during PbA infection, such as arginine and/or tetra-hydrobiopterin supplementation, with or without arginaseinhibition with nor-NOHA, but such compounds were noteffective in preventing ECM, at least in the doses andschemes used. Inhibition of phosphodiesterase 5 (PDE-5)by sildenafil was not effective by itself, however sildenafilgreatly lowered the amount of DPTA-NO needed to preventECM, and resulted in milder side effects, probably due toprolongation of the downstream effects of NO by mainte-nance of cyclic guanosine monophosphate (cGMP) levels(Martins et al. in press). These data show that strategies todecrease the amount of exogenous NO needed to preventECM and its consequent side effects are feasible. Based onthese results, we anticipated that slower-releasing NO-donors might also prove to be effective in preventing ECMby restoring the endogenous NO pools and improving endo-thelial and vascular function while generating mild sideeffects.

In the present study, we evaluated the efficacy of thenitrosothiol S-nitrosoglutathione (GSNO) in preventing

ECM in PbA-infected mice. GSNO has been shown to bea potent neuroprotective agent against stroke in rats, whereasfast-release NO-donors such as S-nitroso-N-acetyl-penicilla-mine, sodium nitroprusside, methylamine-hexamethylene-methylamine-NONOate, propylamine-propylamine-NONOate, and 3-morpholinosydnonimine (SIN-1) hadlimited or no protective effect (Khan et al. 2006). Inaddition to being a widely used NO-donor in a numberof experimental studies with therapeutic purposes inbrain inflammatory conditions (Khan et al. 2005, 2006;Haq et al. 2007; Nath et al. 2010), GSNO has the advantage ofbeing a naturally-occurring, physiologically-relevant mole-cule. GSNO is an important endogenous reservoir and trans-porter of NO, contributing to the regulation of NO physiology(Zeng et al. 2001; Zhang and Hogg 2002). In fact, GSNO aswell as L-glutathione reduced (GSH) and oxidized (GSSG)are potent endogenous anti-oxidant agents which were con-sidered candidates for the prevention of ECM, a diseasecharacterized by a state of oxidative stress (Wiese et al.2006). We tested different concentrations of GSNO to evalu-ate its effect on survival and on critical parameters such asbrain edema and inflammation, blood pressure and heart rate.

Methods

Animals, parasite and infection

Animal handling and care followed the NIH Guide for Careand Use of Laboratory Animals. All protocols were ap-proved by the La Jolla Bioengineering Institutional AnimalCare and Use Committee. Six to eight-week old C57Bl/6(Jackson Laboratories, ME) were inoculated intraperito-neally (IP) with 1×106 Plasmodium berghei ANKA (PbA)parasites expressing the green fluorescent protein (a dona-tion from the Malaria Research and Reference ReagentResource Center – MR4, Manassas, VA; deposited by CJJanse and AP Waters; MR4 number: MRA-865). Motorbehavior and rectal temperature were checked daily fromday 4 as previously described (Clemmer et al. 2011). Briefly, aset of six behavioral tests (transfer arousal, locomotor activity,tail elevation, wire maneuver, contact righting reflex andrighting in arena) adapted from the SHIRPA protocol(Lackner et al. 2006; Martins et al. 2010) was used toprovide a better estimate of the overall clinical status ofthe mice during infection. The performance in each testwas assessed and a composite score was built rangingfrom 0 to 23, where 23 indicates maximum performance and 0indicates complete impairment – usually coma. Body temper-ature was monitored using an Accorn Series Thermocouplethermometer with a mouse rectal probe (Oakton Instruments,Vernon Hills, IL). ECMwas defined as the presentation of one

or more of the following clinical signs of neurological in-volvement: ataxia, limb paralysis,, seizures, roll-over, coma.Parasitemia was determined by flow cytometry by detectingthe number of fluorescent GFP-expressing pRBCs in relationto 10,000 RBCs.

Drugs and treatments

GSNO and DPTA-NO were purchased from Cayman Chem-ical (Ann Arbor, MI). PbA-infected mice were treated twicea day starting on day 0 of infection until day 8 with freshlyprepared GSNO or saline. Groups of mice received doses ofGSNO of 3.5, 0.35 or 0.035 mg per mouse, in saline, via IP,100 μL per mouse. These doses are equivalent in NOgeneration to the dose of DPTA-NO (1 mg) shown to beeffective in preventing ECM, as well as 10 times and 100times lower, respectively. Treatment in all groups was dis-continued after day 8 and survivor mice were followed up today 12 of infection.

Exhaled NO, plasma nitrite levels and hematocrit

For determination of exhaled NO levels, one hour after themorning treatment on day 5 post infection mice were heldfor 1 min in a plastic chamber connected to a Sievers 280iNO-analyzer (Sievers, Boulder, CO). Air accumulated forone minute in the chamber was released by opening theports and the NO content of an exhaled air sample wasmeasured in parts per billion by the NO-analyzer. On day6 of infection, 30 μL of blood was collected from the tailvein in a heparinized micro-hematocrit capillary tube(Chase, Rockwood, TN) and centrifuged. Hematocrit wasmeasured and the plasma was separated and frozen at −20°C.Later, plasma was thawed, mixed with an equal amount ofmethanol to precipitate proteins and centrifuged at 10,000 gfor 10 min. Supernatant was collected and frozen (−80°C)until reading. Nitrite plasma content was measured using anENO-20 NOx Analyzer (Eicom, San Diego, CA) according tothe manufacturer’s instructions.

Measurements of brain glutathione (GSH-GSSG) levels

Total glutathione, GSH, GSSG concentrations in brain tis-sue were measured using enzyme-linked assay kits (CaymanChemical, Ann Arbor, MI) in accordance with the manufac-turer’s instructions. Mice were anesthetized (ketamine150 mg⁄kg; xylazine 10 mg⁄kg), the thoracic cavity wasopened, and perfusion was made with 5 ml of saline usinga gravity perfusion setup. The brain was carefully removed,weighed and one of the hemispheres was processed for themeasurement of total glutathione (GSH + GSSG) andGSSG. The brain sample was homogenized in 5 ml of

suggested cold buffer per gram of tissue using a T25 basicelectric homogenizer (IKA Works, Wilmington, NC). Theprecipitate was removed by centrifugation and supernatantwas collected and frozen (−80°C) until assaying.

Histology

Infected animals treated with saline or GSNO were euthan-ized at day 6 post-infection and their brains collected andprocessed for histology. The brain was carefully collected inthe same way described for glutathione measurement,weighed, and stored in 10% paraformaldehyde during 48 hfor fixation. Brains were cut in four coronal slices of 2–3 mm using a mouse brain blocker (David Kopf Instru-ments, Tujunga, CA) and each slice was embedded in par-affin. Sections of 5 μm were obtained at approximate400 μm intervals (four sections per slice), mounted in glassslides, and stained with hematoxylin-eosin (HE). The num-ber of leukocytes in each meningeal vessel, the number oftamponated vessels in parenchyma, and the number and areaof hemorrhages of each section was quantified using anocular grid calibrated with a 100× magnification (fielddimensions: 200×180 μm) in a Nikon microscope (EclipseE200). The whole area of each section was similarly quan-tified with the grid calibrated at 40× magnification. Quanti-fication was performed by an experienced investigator in ablinded fashion. Pictures were taken with a SPOT camera(Diagnostic Instruments Inc. USA; 1014×721 pixels).

Cardiovascular parameters

To evaluate the effects of GSNO on cardiovascular physiol-ogy, remote measurement of heart rate and blood pressures(systolic, diastolic and mean arterial pressures) was madewith Data System International telemetry devices (DSI, St.Paul, MN) on individually housed mice at room temperatureas described before (Astrand et al. 2004). Briefly, followinganesthesia an incision was made in the neck exposing theleft carotid artery and the catheter tip of a TA11PA-C20 unitwas inserted into the vessel and secured with 5-0 suture. Thetransducer device was then placed subcutaneously in theanimal’s back and the incision was stapled. The animalsreceived analgesics and were allowed to recover for a min-imum of 5 days. Before drug administration, animals wereallowed to acclimate on the telemetry platform for 30 min.Each animal then received an injection of saline or GSNO atdifferent concentrations and were immediately returned tothe telemetry platform. Hemodynamic parameters were con-tinuously recorded starting 30 min before to 60 min afterinjection. During the baseline and experimental periodsmice were left alone in the procedure room providing anoise free environment.

Statistical analyses

Results were expressed as mean and standard error of themean unless otherwise stated. The log-rank test was used tocompare different survival curves followed by post hoc log-rank tests for comparisons between different treatmentgroups. Two-way ANOVA with Bonferroni posttests wasused to analyze parasitemia. Curve analyses were made onlyat days 4–6 post-infection because mice started to die fromcerebral malaria after this period, introducing bias in thesample. Temperature, motor behavior, hematocrit, exhaledNO, plasma nitrite, body and brain weight, and brain histo-pathology parameters were initially analyzed using one-wayANOVA or Kruskal-Wallis test if the variances betweengroups differed significantly (Bartlett’s test for equal varian-ces). Body and brain weight, temperature, behavior, exhaledNO, and plasma nitrite data were further analyzed byplanned post hoc t-tests with Welch’s correction for compar-isons between different treatment groups. Due to the highnumber of comparisons, post-tests p-values were correctedusing Šidák-Holm’s p-value adjustment, to reduce the num-ber of false positive tests. Tukey’s multiple comparison test

was used to check for differences in hematocrit and brainhistopathology parameters between experimental groups.When multiple concentrations of GSNO were tested, post-tests to check for linear trend following log-rank, one-wayANOVA, and Kruskal-Wallis tests were also performed.Spearman correlation coefficients and two-tailed p-valueswere calculated to estimate the degree and significance ofassociation between two parameters. A p-value <0.05 wasconsidered significant. Heart rate, systolic, diastolic, andpulse pressures data were averaged as 10 s bins for eachanimal and the average for each bin from the same timepoint in each group was determined. Means from binsrepresenting the first 30 min before injection were thenaveraged to calculate a baseline for each group. Data fromeach bin was converted and plotted as percentage of base-line. Lowess curves were then calculated to show the trendof the data in each group. Normal ranges for each parameterwere calculated based in values obtained from four salinetreated animals. The range of values falling within the meanplus and minus two standard deviations (SD) was consid-ered normal and treatments that decrease or increase theparameter values outside this range were considered to

Fig. 1 GSNO prevents ECMdevelopment. Cumulativesurvival (a), course ofparasitemia (b), rectaltemperature (c), and motorbehavior score (d) of PbA-infected mice treated with sa-line (n055) or GSNO at 0.035(n018), 0.35 (n028), and 3.5(n020) mg/mouse. Bodyweight (e, n04–6 per group)and hematocrit (F, n0cumula-tive survival) were also evalu-ated. Rectal temperature, motorbehavior score, body weightand hematocrit were measuredon day 6 of infection. *p<0.05,**p<0.01, ***p<0.001, arrowsindicate the presence of a lineartrend

affect the parameter. All statistics were calculated usingGraphPad Prism 4.01 (GraphPad Software, San Diego,CA) and Stata 9.0 (StataCorp LP, College Station, TX).

Results

GSNO protects against ECM

Treatment with GSNO resulted in significant protectionagainst the development of ECM with the three doses used(3.5 mg, 0.35 mg and 0.035 mg) (Fig. 1a). The higher dose(3.5 mg/mouse) was chosen to be equivalent in NO molarterms to the dose of DPTA-NO (1 mg/mouse) shown toafford protection against ECM. GSNO at doses 10 timesand 100 times lower was also able to afford significantprotection against ECM (Fig. 1a). Survival curves of micesupplemented with 0.035 or 0.35 mg were not significantlydifferent, indicating that both doses were equally efficient toprevent ECM (Fig. 1a). Treatment with 3.5 mg per mousecaused inhibition of parasite growth whereas the low andintermediate doses of GSNO had no effect on the course ofparasitemia compared to saline-treated mice (Fig. 1b). Thedecrease in parasite growth caused by the high dose ofGSNO also decreased disease morbidity as shown by themaintenance of temperature and behavior score at similarlevels to uninfected mice and the mild decrease in bodyweight by day 6 of infection (Fig. 1c–e). Supplementationwith 0.035 and 0.35 mg per mouse showed a trend toimprove temperature and body weight in a dose responsemanner (Fig. 1c and e). However, a similar pattern was notobserved for the behavior score with mice treated with theintermediate dose not showing improvement in this param-eter at day 6 of infection (Fig. 1d). Treatments with thedifferent doses of GSNO had opposite effects on hematocrit.The group of mice receiving the higher dose presented hemat-ocrit levels similar to uninfected control animals (Fig. 1f).This effect on preventing RBC destruction can be ascribedto the inhibitory effect on parasite growth induced by thisdose. On the other hand, the group receiving the intermediatedose of GSNO had hematocrit levels significantly lower com-pared to the group receiving saline (Fig. 1f), despite similarlevels of parasitemia. The lower dose of GSNO had no effecton hematocrit compared to saline.

Effects of GSNO on exhaled NO and plasma nitrite levels

Treatment with GSNO resulted in a trend for increasedlevels of exhaled NO and plasma nitrite (Fig. 2a, b). How-ever, the increase in exhaled NO and plasma nitrite inrelation to the saline group was statistically significant onlyfor the intermediate and higher dose GSNO-treated mice(Fig. 2a, b). As expected mice receiving saline presented

low levels of exhaled NO when compared to uninfectedmice (Fig. 2a). These results indicate that GSNO supple-mentation of PbA-infected mice resulted in increased NObioavailability in a dose response manner.

GSNO supplementation prevents brain edema and decreasesbrain inflammation and hemorrhages

PbA-infected mice showed brain edema on day 6 of infection,which was prevented by any of the three doses of GSNO, witha trend for less edema with increasing GSNO doses (Fig. 3a).GSNO treatment also prevented or decreased the accumula-tion of leukocytes in brain vessels (Fig. 3b, c) and the occur-rence of brain hemorrhages (Fig. 3d, e), again with a trend formore robust effects with increasing GSNO doses. Represen-tative brain histology images from uninfected mice andinfected mice given saline or GSNO (0.35 and 3.5 mg/mouse)are shown in Fig. 4a–h. Interestingly, the few adherent leuko-cytes in mice receiving 3.5 mg of GSNO were observedmostly in meningeal vessels, being the vast majority of brainparenchymal vessels in this group totally clean (Fig. 4g, h).

Fig. 2 GSNO increases NO bioavailability in PbA infected mice.Exhaled NO (a) of uninfected and PbA-infected mice treated withsaline or GSNO at 0.035, 0.35, and 3.5 mg/mouse was measured 1 hafter the morning treatment on day 5 of infection (n≥5 per group).Plasma nitrite (b) of uninfected and PbA-infected mice treated withsaline or GSNO at 0.035, 0.35, and 3.5 mg/mouse was measured onsamples collected prior to the morning dosing on day 6 of infection (n≥7 per group). *p<0.05, **p<0.01, ***p<0.001, arrows indicate thepresence of a linear trend

Effect of PbA infection on brain total glutathione, GSH,and GSSG levels

PbA-infected mice receiving saline showed amarked decreasein brain total glutathione and GSH levels on day 6 ofinfection (Fig. 5a, b). Treatment with the lower and inter-mediate doses of GSNO did not affect total glutathioneand GSH levels in the brain of infected animals, but micereceiving the higher dose showed an improvement in brainGSH levels (Fig. 5a, b). Mean GSSG levels and GSH/GSSG ratio did not differ among the groups on day 6 ofinfection (Fig. 5c, d). However, the variance in theseparameters in mice receiving saline was significantly greaterthan in uninfected and GSNO-supplemented mice (p<0.05)indicating that the former population was not homogeneous.A correlation matrix was built to test if brain GSH and GSSGlevels in saline-treated mice on day 6 of infection could berelated to body and brain weight, behavior score, temperature

or parasitemia. Brain GSSG levels were positively cor-related with behavior score and temperature (Fig. 5e, f),GSH/GSSG ratio was negatively correlated with behav-ior score (Fig. 5g), and as expected, temperature waspositively correlated with behavior score (Fig. 5h). In-terestingly GSH levels were not correlated with behav-ior score or temperature. These data suggest that thebrain of mice with ECM is seriously affected by a lossof its anti-oxidant defense mechanisms, with a markeddecrease in total glutathione and GSH levels, and low GSSGlevels appear to be particularly an indicator of diseaseseverity.

Low and intermediate doses of GSNO do not inducehypotension

Injection of the higher (3.5 mg) dose of GSNO induced asharp but transient decrease in mean arterial pressure (MAP)

Fig. 3 GSNO decreases brainedema and inflammation inPbA infected mice. Micereceiving GSNO at 0.35 and3.5 mg/mouse did not show anincrease in total brain weight(a) upon infection. Treated micealso presented a smaller numberof adherent leukocytes permeningeal vessel (b), a lessernumber of tamponated vessel inbrain parenchyma (c), and asmall hemorrhagic burden asreflected by a decreased numberof hemorrhages/mm2 (d) andhemorrhagic area (e, defined asthe area occupied byhemorrhages in relation to thetotal area analyzed) of brainparenchyma in a dose responsemanner on day 6 of infection.n04–5 per group, *p<0.05,**p<0.01, ***p<0.001, arrowsindicate the presence of a lineartrend

(Fig. 6a). By 60 min MAP returned to normal. A similarpattern was observed for the systolic and diastolic pressures(Fig. 6b, c), whereas heart rate presented a less abrupt butlonger lasting decrease (Fig. 6d). The intermediate dose in-duced no significant changes in MAP or systolic and diastolicpressures (Fig. 6a–c), however changes in heart rate and pulsepressure were similar to those induced by the higher dose.Saline injection as well as the lower dose of GSNO caused nochanges in any of the parameters (Fig. 6a–d).

Discussion

Both human and murine CM have been shown to be asso-ciated with endothelial dysfunction and brain microcircula-tory complications including vascular occlusion, reducedblood flow, vascular collapses or ghost vessels, vasocon-striction and inflammation (Beare et al. 2009; Cabrales et al.2010). In ECM, exogenous NO has been shown to bebeneficial in preventing the syndrome (Gramaglia et al.

Fig. 4 Inhibition ofinflammation in the brain ofPbA-infected mice treated withGSNO. (a and b) Pial (a) andparenchymal (b) vessels of un-infected control mice, showingno accumulation or adhesion ofleukocytes; (c and d) pial (c)and parenchymal (d) vessels ofmice receiving saline withECM, plugged with leukocytes;an extensive subarachnoidhemorrhage is seen in panel c;(e and f) pial (c) and parenchy-mal (f) vessels of mice receiv-ing GSNO at 0.35 mg/mouseshowing a less intense leuko-cyte adhesion; (g and h) pial (g)and parenchymal (h) vessels ofmice receiving GSNO at3.5 mg/mouse, a small numberof adherent leukocytes is shownin panel g, and a clean paren-chymal vessel is shown in panelH. Magnitude of leukocyte ac-cumulation and adherence washeterogeneous among vessels inmice receiving saline andGSNO at 0.35 and 3.5 mg/mouse. All sections werestained with H&E(magnification, X400)

2006; Cabrales et al. 2011). Nevertheless, the NO donorused in the murine studies, DPTA-NO, presents a numberof limitations including rapid clearance and the induction ofpotent side effects including a major drop in arterial pres-sure. While it remains a useful tool to increase NO bioavail-ability in mechanistic studies, the adverse effects limit itsclinical potential. We considered whether GSNO could be abetter alternative NO-donor based on the fact that GSNO isan important endogenous reservoir and carrier of NO, con-tributing to regulate its physiological roles (Chiueh andRauhala 1999). Physiologically, GSNO is stable, releasesNO only under certain conditions, and its auto-oxidation isslow (Singh et al. 1996; Zhang and Hogg 2002). Much ofthe NO in GSNO can actually be used to transnitrosylate

protein thiols which can act either as an endogenous reser-voir for NO extending its bioavailability or have physiolog-ical roles on their own (Radomski et al. 1987; Foster et al.2009).

Here we showed that GSNO was indeed superior toDPTA-NO, as it was able to decrease ECM incidence andincrease survival of PbA-infected mice when administeredat doses 10–100 times lower than the comparable effectivedose of DPTA-NO. While the higher dose of GSNO (3.5 mgper mouse) induced transient hypotension as does DPTA-NO, the lower and intermediate doses had no or only mildeffects on blood pressure or heart rate. These results makeGSNO a more promising candidate to be used with thepurpose of restoring NO bioavailability. It should be

Fig. 5 PbA infection decreasesbrain total glutathione and GSHlevels, and low GSSG levels arecorrelated with ECMdevelopment. Brain levels oftotal glutathione (a), GSH (b),and GSSG (c) and GSH/GSSGratio of uninfected and PbAinfected mice receiving salineor GSNO at 0.035, 0.35, and3.5 mg/mouse. Brain GSSGlevels were positivelycorrelated with motor behaviorscore (e) and temperature (f),and negatively correlated withthe GSH/GSSG ratio (g) in PbAinfected mice. Temperature wasalso positively correlated withmotor behavior score (h). Allmeasurements were made onday 6 of infection. *p<0.05,**p<0.01, r0spearmancorrelation coefficient

stressed, however, that the present study was not designed toverify its potential efficacy as an adjunctive therapy inECM, and additional studies are necessary to address thisissue.

GSNO has several properties that may be relevant for itsprotective effect on ECM, which may be linked to its NOmoiety but may also be modulated by, or independent of, itsglutathione backbone. ECM is associated with endothelialdysfunction, vascular obstruction, decreased cerebral bloodflow and a vasospasm-like phenomenon develops in the pialvessels (Cabrales et al. 2010). Diffuse microhemorrhagesand eventually larger sub-arachnoid hemorrhages are ob-served (Carvalho et al. 2000; Cabrales et al. 2011). Thesefeatures can be ameliorated by exogenous NO (DPTA-NO)and therefore GSNO may act similarly. These ECM featuresalso resemble stroke, and GSNO has been shown to bebeneficial in experimental stroke by reducing oxidativestress and inhibiting the activation of both endothelial cellsand macrophages in the brain (Khan et al. 2005). In addi-tion, GSNO treatment increased cerebral blood flow, de-creased mRNA expression of E-selectin, and inhibited theexpression of inducible NOS, TNF-α, IL-1β, ICAM-1, andLFA-1 in a rat model of stroke (Khan et al. 2006).

ECM is also characterized by brain inflammation, withincreased expression of cell adhesionmolecules and leukocyteand platelet migration/adherence to brain vessels (Carvalho etal. 2000; Bauer et al. 2002; Cabrales et al. 2010, 2011). Weshowed that GSNO supplementation decreased brain leukocyte

accumulation and edema during PbA infection, indicating thatsome of these anti-inflammatory mechanisms could be claimedto explain the phenomenon. GSNO has indeed been shown toattenuate central nervous system inflammatory syndromes suchas experimental autoimmune encephalomyelitis (Nath et al.2010) as well as experimental autoimmune uveitis (Haq et al.2007). Furthermore, GSNO is also an inhibitor of plateletaggregation (Radomski et al. 1992).

Oxidative stress, particularly affecting brain endothelialcells, plays an important role in ECM (Wiese et al. 2006).Here we show that PbA-infected mice presented markeddecreases in the brain levels of GSH, indicating that indeedinfection depleted one of the major antioxidant and redoxbuffer systems in the brain (Aoyama et al. 2008), whichlikely contributes significantly to ECM pathogenesis.GSNO is a potent antioxidant against peroxynitrite andprotects neurons against oxidative stress (Rauhala et al.1998). However, our data suggest that the protective effectof GSNO was not derived directly from the restoration ofGSH pools in the brain since animals treated with the lowerand intermediate doses still presented glutathione depletion.Indeed, it has been shown that GSNO is not able to cross theblood brain barrier when administered i.p. in healthy mice(Peyrot et al. 2005). Accordingly, only 0.5% of radiolabeledGSH administered by intra-carotid injection was detectablein brain extracts (Cornford et al. 1978). Taken together thesedata suggest that the resistance to ECM development con-ferred by GSNO supplementation is likely related to effects

Fig. 6 GSNO effects on cardiovascular parameters. Changes in meanarterial pressure (MAP, a), systolic pressure (b), diastolic pressure (c),and heart rate (d) following one IP injection of saline (black dots andlines), and GSNO at 0.035 (red dots and lines), 0.35 (blue dots andlines), and 3.5 (green dots and lines). Vertical doted lines represent thetime when the IP injection was given and separate the baseline period

from the experimental period. Results are expressed as the percentagechange in relation to the mean of the baseline period for each group.Horizontal doted lines represent the range of values falling within themean plus and minus two standard deviations (SD) of the baselinevalue calculated for saline treated group. N04 per group

for instance on the endothelium, an interpretation supportedby the decreased level of inflammation observed in GSNO-treated mice. In addition, 8-oxoguanine staining of the brainrevealed that oxidative stress during ECM affects mainly theendothelial cells (Wiese et al. 2006). The effect of the higherGSNO dose in preventing glutathione depletion was likelysecondary to the anti-malarial activity of this dose.

Alternatively, GSNO promotes the formation of S-nitroso-(bCys93) hemoglobin (SNO-Hb) which right shiftsthe oxygen dissociation curve of hemoglobin favoring theoxygen offload and providing an alternative, vasorelaxation-independent mechanism for improving oxygen delivery(Patel et al. 1999). For example, SNO-Hb can mediateoxygen tension (pO2)-dependent vasodilatory activity(Singel and Stamler 2005). The fact that GSNO ornitrosylated proteins store NO making it available upondemand may help to explain why only the intermediateand high doses of GSNO were able to generate detect-able increased levels of exhaled NO and plasma nitrite. Inaddition, although GSNO is considered an NO-donor, in vitroNO generation accounts for only a part of the decompositionproducts of GSNO by endothelial cells (Zeng et al. 2001) andnitrosothiols may exert its biological effects through NO-independent mechanisms.

The higher dose of GSNO also showed a strong anti-parasite effect, keeping parasitemia low compared to saline-treated animals. It has been previously shown that GSNOas well as other NO donors can have anti-plasmodialeffects in vitro (Rockett et al. 1991; Totino et al. 2008).Interestingly, treatment with DPTA-NO, which generatessimilar amounts of NO, does not inhibit PbA growth invivo (Gramaglia et al. 2006; Cabrales et al. 2011). Thereasons for the differential anti-parasite effect of thesetwo donors are unclear. Malaria parasites grow readilyin saturated NO solutions but are highly sensitive to S-nitrosothiols (Rockett et al. 1991). In any case, the factthat GSNO presents anti-plasmodial activity also adds toits potential for anti-malarial therapies.

In summary, the present study shows that GSNO in awide range of doses can prevent the development ofCM in PbA-infected mice. At a higher dose it can alsoinhibit parasite growth but induces transient hypoten-sion. Lower doses are also effective in preventing CMwithout inhibiting parasite growth and without changingblood pressure. These features reveal that GSNO plays arelevant role in ECM pathogenesis. While it is clear thatthe data shown here with GSNO used as a prophylacticintervention in ECM allow no inferences about its po-tential in a clinically relevant setting as an adjunctivetherapy intervention, the characteristics above make GSNOa promising compound for further studies to determine itsefficacy in CM-rescuing treatments with anti-malarialdrugs.

Acknowledgements This study was supported by NIH grants R01-HL087290 and R01-AI082610. GMZ was recipient of a CNPq (Brazil)post-doctoral fellowship. We thank Dr John Nolan (LJBI) for grantingaccess to the flow cytometry facilities, Wisam Barkho for performingsurgeries to implant the TA11PA-C20 units, and Diana Meays foranimal care.

Competing interests The authors declare that they have no compet-ing interests.

References

Aoyama K, Watabe M, Nakaki T (2008) Regulation of neuronalglutathione synthesis. J Pharmacol Sci 108(3):227–238

Astrand A, Bohlooly YM, Larsdotter S, Mahlapuu M, Andersen H,Tornell J, Ohlsson C, Snaith M, Morgan DG (2004) Mice lackingmelanin-concentrating hormone receptor 1 demonstrate increasedheart rate associated with altered autonomic activity. Am J PhysiolRegul Integr Comp Physiol 287(4):R749–R758. doi:10.1152/ajpregu.00134.2004

Bauer PR, Van Der Heyde HC, Sun G, Specian RD, Granger DN(2002) Regulation of endothelial cell adhesion molecule expres-sion in an experimental model of cerebral malaria. Microcircula-tion 9(6):463–470. doi:10.1038/sj.mn.7800159

Beare NA, Harding SP, Taylor TE, Lewallen S, Molyneux ME (2009)Perfusion abnormalities in children with cerebral malaria and ma-larial retinopathy. J Infect Dis 199(2):263–271. doi:10.1086/595735

Cabrales P, Zanini GM, Meays D, Frangos JA, Carvalho LJ (2010)Murine cerebral malaria is associated with a vasospasm-like mi-crocirculatory dysfunction, and survival upon rescue treatment ismarkedly increased by nimodipine. Am J Pathol 176(3):1306–1315. doi:10.2353/ajpath.2010.090691

Cabrales P, Zanini GM, Meays D, Frangos JA, Carvalho LJ (2011)Nitric oxide protection against murine cerebral malaria is associ-ated with improved cerebral microcirculatory physiology. J InfectDis 203(10):1454–1463. doi:10.1093/infdis/jir058

Carvalho LJ, Lenzi HL, Pelajo-MachadoM,Oliveira DN, Daniel-RibeiroCT, Ferreira-da-Cruz MF (2000) Plasmodium berghei: cerebralmalaria in CBA mice is not clearly related to plasma TNF levelsor intensity of histopathological changes. Exp Parasitol 95(1):1–7.doi:10.1006/expr.2000.4508

Chiueh CC, Rauhala P (1999) The redox pathway of S-nitrosoglutathione, glutathione and nitric oxide in cell to neuroncommunications. Free Radic Res 31(6):641–650

Clemmer L, Martins YC, Zanini GM, Frangos JA, Carvalho LJ (2011)Artemether and artesunate show the highest efficacies in rescuingmice with late-stage cerebral malaria and rapidly decrease leuko-cyte accumulation in the brain. Antimicrob Agents Chemother 55(4):1383–1390. doi:10.1128/AAC.01277-10

Cornford EM, Braun LD, Crane PD, Oldendorf WH (1978) Blood-brain barrier restriction of peptides and the low uptake of enke-phalins. Endocrinology 103(4):1297–1303

Foster MW, Hess DT, Stamler JS (2009) Protein S-nitrosylation inhealth and disease: a current perspective. Trends Mol Med 15(9):391–404. doi:10.1016/j.molmed.2009.06.007

Gramaglia I, Sobolewski P, Meays D, Contreras R, Nolan JP, FrangosJA, Intaglietta M, van der Heyde HC (2006) Low nitric oxidebioavailability contributes to the genesis of experimental cerebralmalaria. Nat Med 12(12):1417–1422. doi:10.1038/nm1499

Haq E, Rohrer B, Nath N, Crosson CE, Singh I (2007) S-nitrosoglutathione prevents interphotoreceptor retinoid-bindingprotein (IRBP(161-180))-induced experimental autoimmune uve-itis. J Ocul Pharmacol Ther 23(3):221–231. doi:10.1089/jop.2007.0023

John CC, Kutamba E, Mugarura K, Opoka RO (2010) Adjunctivetherapy for cerebral malaria and other severe forms of Plasmodi-um falciparum malaria. Expert Rev Anti Infect Ther 8(9):997–1008. doi:10.1586/eri.10.90

Khan M, Sekhon B, Giri S, Jatana M, Gilg AG, Ayasolla K, Elango C,Singh AK, Singh I (2005) S-Nitrosoglutathione reduces inflam-mation and protects brain against focal cerebral ischemia in a ratmodel of experimental stroke. J Cereb Blood Flow Metab 25(2):177–192. doi:10.1038/sj.jcbfm.9600012

Khan M, Jatana M, Elango C, Paintlia AS, Singh AK, Singh I (2006)Cerebrovascular protection by various nitric oxide donors in ratsafter experimental stroke. Nitric Oxide 15(2):114–124.doi:10.1016/j.niox.2006.01.008

Lackner P, Beer R, Heussler V, Goebel G, Rudzki D, Helbok R, TannichE, Schmutzhard E (2006) Behavioural and histopathological alter-ations in mice with cerebral malaria. Neuropathol Appl Neurobiol32(2):177–188. doi:10.1111/j.1365-2990.2006.00706.x

Martins YC, Werneck GL, Carvalho LJ, Silva BP, Andrade BG, SouzaTM, Souza DO, Daniel-Ribeiro CT (2010) Algorithms to predictcerebral malaria in murine models using the SHIRPA protocol.Malar J 9:85. doi:10.1186/1475-2875-9-85

Martins YC, Zanini GM, Frangos JA, Carvalho LJ. Efficacy of differ-ent nitric oxide-based strategies in preventing experimental cere-bral malaria by Plasmodium berghei ANKA. PLoS One (in press)

Nath N, Morinaga O, Singh I (2010) S-nitrosoglutathione a physiolog-ic nitric oxide carrier attenuates experimental autoimmune en-cephalomyelitis. J Neuroimmune Pharmacol 5(2):240–251.doi:10.1007/s11481-009-9187-x

Patel RP, Hogg N, Spencer NY, Kalyanaraman B, Matalon S,Darley-Usmar VM (1999) Biochemical characterization ofhuman S-nitrosohemoglobin. Effects on oxygen binding andtransnitrosation. J Biol Chem 274(22):15487–15492

Peyrot F, Grillon C, Vergely C, Rochette L, Ducrocq C (2005) Phar-macokinetics of 1-nitrosomelatonin and detection by EPR usingiron dithiocarbamate complex in mice. Biochem J 387(Pt 2):473–478. doi:10.1042/BJ20040828

Radomski MW, Palmer RM, Moncada S (1987) Comparative phar-macology of endothelium-derived relaxing factor, nitric oxide

and prostacyclin in platelets. Br J Pharmacol 92(1):181–187

Radomski MW, Rees DD, Dutra A, Moncada S (1992) S-nitroso-glutathione inhibits platelet activation in vitro and in vivo. Br JPharmacol 107(3):745–749

Rauhala P, Lin AM, Chiueh CC (1998) Neuroprotection by S-nitrosoglutathione of brain dopamine neurons from oxidativestress. FASEB J 12(2):165–173

Rockett KA, Awburn MM, Cowden WB, Clark IA (1991) Killing ofPlasmodium falciparum in vitro by nitric oxide derivatives. InfectImmun 59(9):3280–3283

Singel DJ, Stamler JS (2005) Chemical physiology of blood flowregulation by red blood cells: the role of nitric oxide and S-nitrosohemoglobin. Annu Rev Physiol 67:99–145. doi:10.1146/annurev.physiol.67.060603.090918

Singh SP, Wishnok JS, Keshive M, Deen WM, Tannenbaum SR (1996)The chemistry of the S-nitrosoglutathione/glutathione system.Proc Natl Acad Sci U S A 93(25):14428–14433

Totino PR, Daniel-Ribeiro CT, Corte-Real S, de Fatima Ferreira-da-CruzM (2008) Plasmodium falciparum: erythrocytic stages die byautophagic-like cell death under drug pressure. Exp Parasitol 118(4):478–486. doi:10.1016/j.exppara.2007.10.017

WHO (2010) World malaria report. WHO, GenevaWiese L, Kurtzhals JA, Penkowa M (2006) Neuronal apoptosis, metal-

lothionein expression and proinflammatory responses during ce-rebral malaria in mice. Exp Neurol 200(1):216–226. doi:10.1016/j.expneurol.2006.02.011

Zanini GM, Cabrales P, Barkho W, Frangos JA, Carvalho LJ (2011)Exogenous nitric oxide decreases brain vascular inflamma-tion, leakage and venular resistance during Plasmodium ber-ghei ANKA infection in mice. J Neuroinflammation 8:66.doi:10.1186/1742-2094-8-66

Zeng H, Spencer NY, Hogg N (2001)Metabolism of S-nitrosoglutathioneby endothelial cells. Am J Physiol Heart Circ Physiol 281(1):H432–H439

Zhang Y, Hogg N (2002) Mixing artifacts from the bolus addition ofnitric oxide to oxymyoglobin: implications for S-nitrosothiol for-mation. Free Radic Biol Med 32(11):1212–1219