Intravenous administration of DOX prevented ECM of P. berghei ANKA infected mice

We addressed the question whether the antibiotic DOX might influence the outcome of PbA infection and ECM in C57BL7/6 mice. For this, we infected C57BL/6 mice with P. berghei ANKA parasitized red blood cells and confirmed successful infection of the mice (Figure A in S1 Fig). To study the effect of DOX on the course of ECM, we administered in accordance to pharmacokinetic studies from Prall et al. 80 mg/kg DOX per mouse intravenously (i.v.) on a daily basis from day 4 until day 6 post PbA infection (dpi 4–6) [22]. Ninety percent of untreated PbA-infected control mice developed ECM (dpi 6–9), whereas all DOX-treated PbA-infected mice were completely protected from ECM and survived this period (Fig 1A and 1B). As expected, parasitemia was reduced in blood and brain tissue of PbA-infected mice that were treated with DOX (Figures B and C in S1 Fig). Stopping the treatment after 6 dpi led to hyperparasitemia (Figure D in S1 Fig) or anaemia in DOX-treated animals around dpi 20 and therefore mice were sacrificed according to predefined ethical criteria.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. DOX administration reduced ECM development and stabilized the blood-brain barrier after PbA infection.

(A, B) C57BL/6 mice received 5*10⁴ PbA-infected erythrocytes (PbA-iRBC) and indicated mice were injected i.v. with 80 mg/kg DOX dpi 4–6. All animals were then monitored for (A) survival and (B) ECM score (ECM phase dpi 6–9 marked in grey). (C, D) C57BL/6 mice received 5*10⁴ iRBC and indicated groups of mice were treated orally with either 80 mg/kg/day or 250 mg/kg/day from dpi 4–6. Mice treated i.v. with 80 mg/kg/day DOX (dpi 4–7) served as reference treatment group. All mice were monitored for (C) survival and (D) ECM score (ECM phase dpi 6–9 marked in grey). Data sets are representative for 2–3 individual experiments with n = 10 mice/group. Survival data were analyzed with log-rank (Mantel-Cox) test. ECM scores are displayed as median. (E, F) Mice were infected with 5*10⁴ PbA-iRBC ± 80 mg DOX/kg/day. The integrity of the BBB was analyzed with an Evans Blue assay on dpi 6. All mice were injected i.v. with 2% Evans Blue dye and one hour later, extravasation of the dye into the brain was determined. (E) Photo documentation of the discoloration and (F) quantification of dye extravasation into the brain by measuring the absorbance of brain tissue at 620nm. Data show representative results of 1 from 2 independent experiments. *p<0.05, t-test on data after analysis of normal distribution.

https://doi.org/10.1371/journal.pone.0192717.g001

In contrast to i.v. injection of DOX, oral administration of either the same dose (80 mg/kg) or an elevated dose of 250 mg/kg given daily from dpi 4–6 did not influence ECM progression, nor did it improve the survival of PbA-infected mice (Fig 1C and 1D). Subsequent experiments were performed with i.v. injection of 80mg DOX/kg from dpi 4–6. Thus, we could show that i.v. administration of DOX prevented efficiently ECM in already infected mice.

BBB integrity was intact after DOX treatment

We next investigated whether brains of DOX-treated mice that were protected from ECM showed differences in the stability of the blood-brain barrier (BBB), since a BBB disruption, increased vascular permeability and the accumulation of activated T cells in the brain have been considered to be important prerequisites for the development of ECM. Therefore, vascular leakage into brain tissue was visualized and quantified with Evans Blue dye. Brains of ECM-positive PbA-infected mice exhibited on dpi 6 strong coloration and enhanced extravasation of Evans Blue (Fig 1E and 1F). In contrast, brains of DOX-treated mice were marginally stained and the concentration of Evans Blue dye in the brain tissue was low (Fig 1E and 1F). Thus, DOX-treated mice successfully maintained the integrity of the BBB, which corresponded with the absence of ECM-induced pathology.

DOX altered inflammatory responses and the activity of tissue degrading enzymes in brain tissue of ECM-negative PbA-infected mice

We then studied whether immune responses in the CNS and in the periphery were altered in DOX-treated mice that were protected from ECM. Interleukin 10 (IL-10) is a key anti-inflammatory factor associated with the capacity to control pathology also in ECM [23]. To study the role of IL-10 in DOX-induced protection, we investigated IL-10 deficient mice. Importantly, PbA-infected IL-10-deficient animals were still protected from ECM after DOX treatment, which clearly excluded any role of this strong anti-inflammatory cytokine in this setting (Figure E in S1 Fig). For further investigation of protective or alternatively activated immune reactions induced by DOX resulting in prevention of ECM-related pathology, we extracted RNA from the brains of naïve mice, PbA infected WT mice and from DOX-treated PbA-infected WT mice to perform a PCR-array. Brains from control-infected B6 mice without DOX treatment revealed an enhanced gene expression of well-described effector molecules in ECM such as chemokines (CCL2, CCL5, CXCL1, CXCL2), adhesion molecules (ICAM1) and cytokines (IL-6 and TNF) in the brain tissue of dpi 6 infected mice in comparison to brains of naïve mice (S1 Table).

Importantly, a comparison of PbA-infected mice versus DOX-treated PbA-infected animals revealed reduced gene expression of several inflammatory mediators in the brain tissue of DOX-treated mice. The most prominent down-regulation was detected in IL-6 and in a group of chemokines, in particular CCL2, CXCL2 and CCL5 (Fig 2A).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. DOX regulates inflammatory responses and tissue degrading enzyme activity in brain cells of PbA infected animals.

C57BL/6 mice were left naïve or infected with PbA as described above and indicated groups received DOX (80 mg/kg/day) from dpi 4–6. Six days post PbA infection, brain tissue of naïve mice and PbA infected mice ± DOX was examined for inflammatory responses and MMP activity. (A) RNA was prepared from brain tissue and analyzed via PCR array. Volcano plot analysis of the comparison PbA versus PbA DOX: Shown is the logarithmic fold-change (x-axis) against the negative logarithmic p-value (y-axis). The dashed lines indicate where p ≤ 0.05 and FC ≥2. (B) Additionally, IFN-γ was analyzed by RT-qPCR in the brain tissue of all groups (C) Quantification of MMP-2-specific zymography. Brain tissue extracted on dpi 6 was subjected to zymography and quantified by scanning densitometry of the gelatinolytic bands of MMP-2. Representative experiments are shown with 6–7 mice/ group. Relative scanning units of MMP-2 are shown as fold change against expression of naïve animals displayed as median (D) Characterization of brain leukocytes on dpi 6 from naïve, PbA and PbA+DOX-treated animals. Flow cytometry gating scheme to identify CD8⁺T cells is shown for representative animals from naïve and infected ± 80 mg/kg/day DOX mice. (E) Absolute numbers of brain infiltrating CD45^hiCD8a⁺ cells as determined by FACS. (F) Brain tissue supernatant from dpi 6 was analyzed with an ELISA for granzyme B. (B, C, E, F) *p<0.05, **p<0.01, ***p<0.001, ANOVA with Tukey’s post test on data after analysis of normal distribution.

https://doi.org/10.1371/journal.pone.0192717.g002

Furthermore, the volcano plot revealed in brain tissue of DOX-treated mice a significantly reduced expression of adhesion molecules L-selectin and P-selectin (Fig 2A), which have been associated with pathogenesis of severe malaria [24]. Although we had observed reduced gene expression of CCL5, which is together with IFN-γ associated with cerebral malaria in humans and mice, protein levels of this mediator were reduced in supernatants from brain tissue cultures from DOX-treated mice and control-infected mice, but the difference was not statistically significant (Figure A in S2 Fig). In contrast, mRNA levels of IFN-γ measured by RT-qPCR in brain tissue of the respective groups was significantly elevated in PbA-infected mice but not in those treated with DOX after infection (Fig 2B). Supernatants of brains from both PbA-infected groups contained enhanced levels of TNF family proteins compared to those from naïve animals, but DOX-treatment reduced the extent of this increase significantly (Figure B in S2 Fig). TNF is relevant for the induction of metalloproteinases, which are important molecules involved in tissue degradation and destruction [25]. To analyse their activity in the brains of PbA-infected mice in comparison to brains of naïve or control-infected mice, we isolated gelatinolytic matrix-metalloproteinases 2 and 9 (MMPs) from brain tissue of all groups of mice to perform a zymography assay. We detected enzymatic degradation of gelatine in SDS gels in all three groups. While in brains of DOX-treated mice MMP-2 levels were significantly reduced in comparison to brains of untreated control-infected mice (Fig 2C), MMP-9 activity was unaffected by DOX treatment (Figures B and C in S2 Fig).

DOX prevented cellular influx into the brains of PbA-infected animals

As we had observed a stabilized BBB and less production of inflammatory molecules in DOX-treated mice, we hypothesized that DOX-treatment interfered with the recruitment of detrimental effector cells to the brain. Therefore, we analysed the cellular CD45^hi infiltrates and production of effector molecules in the brains of the DOX-treated versus untreated mice. Granzyme B represents an important molecule involved in tissue damage and cell destruction, which is essential for the development of ECM [26]. Among the CD45^hi cells, DOX-treated mice contained fewer CD8⁺ T cells (Fig 2D and 2E). The increased frequency and gMFI of intracellular granzyme B from these CD8⁺ T cells that was observed after PbA-infection did not change after DOX treatment (Figures E and F in S2 Fig). However, we found significantly lower levels of secreted granzyme B in by collagenase-digestion isolated brain cells from DOX-treated PbA-infected mice compared to cells isolated from PbA control-infected mice (Fig 2F).

It has been shown that cross-presentation of parasite-specific peptides by brain microvessels to parasite-specific cytotoxic T cells is a key determinant of ECM development [27]. Therefore, we analysed whether DOX alters local cross-presentation in the brain, which subsequently could influence the cytotoxic activity of brain-infiltrated CD8⁺ T cells. After DOX treatment, the capacity of brain microvessels to cross-present PbA-derived antigen to TCR-transduced NFAT-lacZ reporter cells was slightly decreased compared to that from untreated PbA infected animals, but did not reach statistical significance (Figure G in S2 Fig).

Our data indicate that DOX dampened infection-induced inflammatory processes in the brain of PbA-infected mice, whereas PbA infection in untreated mice triggers brain inflammation, characterized by recruitment of effector cells, activity of tissue-degrading enzymes and secretion of soluble inflammatory mediators. We conclude that DOX treatment reduced such infection-induced responses on the mRNA level as well as on protein level and enzymatic activity, thereby ameliorating inflammation in the brain.

Increased numbers of CD8⁺ T cells and decreased lytic activity in the spleens of PbA infected mice after DOX treatment

Priming and generation of immune effector cells in blood-borne diseases primarily takes place in the spleen [28]. Since the brains of DOX-treated PbA-infected mice contained less infiltrated T cells in comparison to untreated PbA-infected mice (Fig 2D and 2E), the DOX treatment might have an impact on effector cells, their local activity in the spleen and /or the cell recruitment into the brain. We therefore analysed the effector cell populations in the spleens of PbA-infected mice with and without DOX-treatment. Importantly, DOX-treated PbA-infected mice presented a splenomegaly with elevated total splenocyte counts as well as increased numbers of splenic CD8 T cells compared to spleens of naïve and PbA infected mice (Fig 3A and 3B and Figure A in S3 Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Splenic CD8⁺ T cell responses are altered after DOX administration in PbA infected mice.

Analysis of splenocytes on dpi 6 in naïve, PbA ± 80 mg/kg/day DOX-treated C57BL/6 mice. (A) Total splenocyte number on dpi 6. (B) Absolute cell numbers of CD8⁺ splenocytes. (C) Geometric mean fluorescence intensity (gMFI) of CD69 on CD8⁺T cells. (D) gMFI of CCR5 on CD8⁺T cells. (E) Frequency of CD8⁺T cells expressing granzyme B intracellularly. (F) Parasite-specific lysis 18h after injection of PbA-infected ± DOX mice with SIINFEKL-pulsed syngeneic target cells. (G) Ex vivo IFN-γ production of splenocytes and (H) IFN-γ levels in serum from the same animals was measured by ELISA. One data set shows representative results from 2–3 independent experiments with 5–6 animals/group. Graphs display the mean and data were statistically tested for normal distribution and then with 1-way ANOVA and Tukey’s Post hoc test (A, B, C, E, G) or with Kruskal-Wallis followed by the Dunn’s post-hoc test, if data were not normally distributed (D, F, H). *p<0.05, **p<0.01,***p<0.001.

https://doi.org/10.1371/journal.pone.0192717.g003

Although surface expression of C-C chemokine receptor type 5 (CCR5) on splenic T cells (Fig 3D) and the frequency of T cells expressing intracellular granzyme B (Fig 3E) were not altered between DOX-treated mice and untreated PbA-infected mice, surface expression of CD69, a marker for general T cell activation, was significantly reduced on CD8⁺ T cells after DOX treatment (Fig 3C), suggesting a reduced T cell activation. Other markers that are associated with T cell activity were either found to be unchanged between PbA-infected groups with or without DOX treatment, such as PD-1 (Figure B in S3 Fig) or were slightly reduced after DOX treatment without reaching statistical significance, such as ICAM-1 and CD40L (Figures B and C in S3 Fig).

Next, we analysed the impact of the DOX-treatment on parasite-antigen-specific activities of splenic CD8⁺ T cells. Although the in vivo antigen-specific CTL lysis and the capacity of splenic CD8⁺ T cells from DOX-treated mice to produce IFN-γ was not significantly impaired (Fig 3F and 3H), in contrast, IFN-γ serum levels were decreased in DOX-treated mice compared to control-infected mice (Fig 3H), indicating a reduced level of Th1 cytokines.

Thus, although DOX-treated mice contained peripheral CD8+ T cells displaying a similar activation phenotype compared to T cells from untreated control-infected mice suffering from ECM, these T cells accumulated in their spleens and were hardly detectable in the brains—in contrast to untreated control mice. We conclude that this apparently impaired recruitment of effector cells upon DOX treatment—together with a reduced antigen load in the brain—contributed to an attenuated inflammation and ECM protection.

Elevated parasite dose could not reverse DOX-mediated ECM protection

As we demonstrated above, DOX treatment of PbA-infected mice led to complete protection from ECM and also to an arrest of parasite growth on day 6 p.i. (Fig 4A), that is in line with its well-described anti-parasitic effects. We therefore addressed the question whether the impact of DOX-induced anti-inflammatory effects could be dissected from the anti-parasitic effects in our model. To rule out that the DOX-mediated decrease of parasitemia was solely responsible for the ECM protection, we determined the parasitemia of untreated PbA-infected mice on day 6, which is the time point of ECM onset. These control-infected mice displayed on dpi 6 an average parasitemia of 8% (Fig 4A).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Elevated initial infective parasite dose could not reverse DOX-mediated ECM protection.

C57BL/6 mice were infected with either 5*10⁴ or 1*10⁶ PbA-iRBCs and received from dpi 4–6 80 mg DOX/kg/day i.v. or not. (A) Blood parasite burden was assessed from dpi 5 on. (B) Survival and (C) ECM score (ECM phase dpi 6–9 marked in grey) were analyzed over the course of infection with low or high dose of parasites and DOX treatment. Representative data from 2–3 individual experiments with n = 10 mice/group are shown. Parasitemia is displayed as the mean/SD and ECM score as median. Survival data was analysed with log-rank (Mantel-Cox) test.

https://doi.org/10.1371/journal.pone.0192717.g004

Titration experiments regarding the infectious dose revealed that an infection with 1*10⁶ iRBC, which is a 20-fold increase of the parasite dose compared to the previous experiments, was sufficient to establish a parasitemia of approximately 8% on day 6 even after DOX treatment, thereby equivalent to the parasitemia of untreated mice infected with 5x10⁴ iRBC, which all died from CM (Fig 4A, see PbA 5e4 versus PbA 1e6 +DOX on 6 dpi). Survival and ECM development of mice that had received an elevated dose of parasites did not differ from those mice that had been infected with the lower dose of parasites that has been used in the other experiments (Fig 4B and 4C). Importantly, DOX administration promoted survival (Fig 4B) and inhibited ECM development in mice infected with either dose of PbA (Fig 4C). Furthermore, we demonstrated maintained BBB stability in DOX-treated mice infected with high dose PbA (S4 Fig). Thus, DOX treatment was able to protect PbA-infected mice from ECM even in the presence of an elevated parasite load, which strongly supports anti-inflammatory effects of DOX in addition to anti-parasitic effects.

Taken together, these results showed that DOX-induced protection from ECM was stable, indicating that DOX exerts strong immune-regulatory functions–in addition to its anti-parasitic effects—that inhibited disease progression by modulation of the inflammatory responses against PbA.

Ethical statement

The study includes work with experimental animals and was designed under consideration of the 3R rules and the ARRIVE checklist [65]. Mouse studies were approved by the local regulatory agency Landesamt fuer Natur, Umwelt und Verbraucherschutz (LANUV (84–02.04.2012A295). Research staff was trained according to FELASA B guidelines. Female 6–8 weeks old C57BL/6 N mice (average weight of 20g) were purchased from Janvier (Le Genest Saint Isle, France). IL-10 ko mice [66] were bred in the House of Experimental Therapy (HET) of the University of Bonn according to institutional animal guidelines. All mice were housed in autoclaved Tecniplast^® cages under specific pathogen free conditions with common animal bedding. A 12 hour light/dark cycle was maintained and water and food were provided ad libitum. Mice were allocated randomly to cages with n = 4–6 mice per group accordingly to the individual experimental groups. Survival experiments were performed with 10 mice per group and ex vivo experiments were performed with n = 4–6 animals per group and 2–3 times repeated, accordingly to sample size determination performed before by statistical power calculation. Infection, treatment and assessment of the health status were performed sequentially.

Longterm anaesthesia for ex vivo analysed experimental mice was given before perfusion by intramuscular injection of 10μl Rompun^® (2%, Bayer, Germany) + 40μl Ketamin (50mg/ml, Ratiopharm GmbH Germany) per 25g mouse. In order to meet humane endpoints, critically sick mice were euthanized by cervical dislocation under isofluran inhalation anaesthesia.

Parasites, infection and disease assessment

In all experiments, P. berghei infected red blood cells (iRBCs) were used to inoculate mice. Mice had an average weight of 20 g and were infected intravenously (i.v.) with 5x10⁴ iRBCs or 1x10⁶ iRBCs (high dose), obtained from mice that had been infected intra-peritoneally (i.p.) with stock solution. Stock solution contained 1x10⁷ iRBCs in glycerine and was stored in liquid nitrogen. Stock mice were of the same background as experimental animals and parasites were isolated on d5 p.i. Experiments were conducted with either Plasmodium berghei ANKA (PbA) [67] or P, berghei expressing OVA (PbTG) [10], the latter being kindly provided by Rachel Lundie. All experiments were conducted with P. berghei ANKA except analysis of in vivo cytotoxicity, IFN-γ release and detection of antigen-specific T cells (PbTG). Experimental setup and used Plasmodium strains are denoted in the figure legends. From day 4 post infection (p.i.), mice were punctured at the tail vein for daily blood-smears. Before day 4, parasite levels were almost undetectable (d1 and d2 p.i.) or very low (d3 p.i.). For evaluation of disease onset, mice were monitored twice daily for ECM symptoms. ECM development was scored according to the following symptoms: 0 = without symptoms, 1 = ruffled fur, 2 = hunching, 3 = wobbly gait, 4 = limb paralysis, 5 = convulsions, 6 = coma adapted from Amante et al. [68]. Mice did not succumb to infection or any other cause with a score of ≤4; mice that reached a score of >4 were sacrificed immediately accordingly to predefined ethical criteria. Graphs display a timeframe in which ECM development occurs, all infected animals were sacrificed by dpi 20 or immediately upon development of hyperparasitemia or anaemia.

Treatment and drug administration

If not otherwise stated mice were treated i.v. with 80 mg/kg DOX dissolved in 200 μl PBS daily starting from d4 till d6 p.i. Mice had an average weight of 20 g. Oral (p.o.) treatment was performed with a special gauge, mice received 80 mg/kg or 250 mg/kg DOX dissolved in 200 μl H₂O.

Perfusion and lymphocyte isolation

In order to remove non-adhering and non-infiltrating erythrocytes and blood-lymphocytes long-term anaesthetised mice were intracardially perfused for 5 minutes with 1x PBS. Spleens and brains were removed, placed in MACS buffer (1x PBS containing 1% FCS and 2 mM EDTA) or collagenase A buffer (Sigma-Aldrich, St. Louis, USA), respectively, and placed on ice. Spleens were gently pressed through a sieve, washed with MACS buffer and centrifuged (10 min., 1500 rpm), cell numbers were determined and adjusted to 2x10⁷ per ml medium, and 50 μl were plated into 96 well plates. OVA-peptide S8L was added in a final concentration of 2 μM to spleen cell culture. Brain tissue was cut into small pieces and incubated in 2 ml collagenase A buffer for 30 min. at 37°C in a water bath. After homogenization in MACS buffer single cell suspensions were then centrifuged at 1500 rpm for 8 min. at 4°C. 500 μl of primary brain cells were adjusted to 2.5x10⁷ cells per ml RPMI and 100/μl plated into 96 well plates for cell culture. The cell pellet was suspended in 5 ml 30% percoll that was then layered onto a two-step percoll gradient made of 3 ml 37% percoll and 3 ml 70% percoll. Samples were centrifuged at room temperature for 20 minutes without brakes at 2000 rpm. After percoll separation, two white interphases were isolated and transferred into a fresh tube and washed with MACS buffer. After a second centrifugation step, cells were resuspended in 300 μl FACS buffer (1x PBS, 1% FCS) and counted.

In vivo cytotoxicity, FACS analysis and ELISA

Cytotoxic T cell activity was determined as described before [69]. Briefly, five days post infection splenocytes from naive syngeneic donor mice were prepared, split into equal populations, loaded with or without 1μM OVA peptide S8L (SIINFEKL) and with CFSE (1 μM or 0.1 μM respectively). Both populations were mixed at equal numbers, every mouse received 1x10⁷ of total cells. 18h later, mice were sacrificed and spleens were isolated. CFSE positive cells were acquired by flow cytometry and both cell populations were determined. The specific lysis was determined as follows: 100-(ratio sample/ratio average of naive) x 100. Splenocytes and brain leukocytes were stained and analyzed for expression of CD4, CD8, CD11b, CD11c, CD45, CD69, CCR5 and granzyme B in a FACS Canto II flow cytometer (BD, San Jose, USA).

Data were analysed with FlowJo Software (Treestar Inc., Ashland, USA). ELISA IFNγ (e-Bioscience, San Diego, USA), TNF, CCL5 and Granzyme B (R&D, Minneapolis, USA) ELISA-kits were used according to the manufacturers protocol. The TNF antibody has not been tested by the company for crossreactivity with lymphotoxin (LT).

RNA-isolation, RNA-array and RT-PCR

Brain tissue samples were collected from all groups six days post infection, where some of the control mice already showed signs of cerebral pathology. Samples were snap frozen in liquid nitrogen and stored at -80 degrees until RNA extraction. Samples were transferred to tubes with 1.4mm Precellys-Ceramic beads and homogenized in TRI-Reagent^® (Sigma-Aldrich, St. Louis, USA) with the Precellys^® 24 (PEQLAB, Erlangen, Germany) homogenizator. After bromchlorpropane/isopropanol isolation samples were treated with DNase and RNA quantity and quality was examined by the OD260 value.

One microgram of total RNA was reverse transcribed to cDNA.

IFN-γ forward: TCAAGTGGCATAGATGTGGAAGAA,

IFN-γ reverse: TGGCTCTGCAGGATTTTCATG.

PbA 18srRNA forward: CTAACATGGCTTTGACGGGTA,

PbA18srRNA reverse: TGTCACTACCCTCTTATTT.

PCRs were performed with the Rotor- Gene 3000 or 6000 (Corbett Research, Hilden, Germany) with conditions set to: 15 min at 95 °C followed by 45 cycles of 15 s at 94°C, 20 s at 58 °C and 20 s at 72 °C. Rate of temperature change was 0.8 °C/s. Copy numbers were determined by a reference plasmid. CCL3 and CCL4 RNA levels were measured with the Qiagen Quantitect Primer-Assay and the Qiagen RotorGene SybrGreenkit (Qiagen, Hilden, Germany). For SABioscience (SABioscience, Hilden, Germany) RT- RNA-Array cDNA samples were isolated with the above described protocol but prior to DNase digestion purified with the Qiagen RNeasy Mini Kit. Samples were eluted in 40 μl RNase free Water and the quality of the samples were spectrometrically tested and ribosomal RNA band integrity measured by the Experion (Bio-Rad, Hercules, USA).1 μg RNA was transcribed into cDNA with the RT-First Strand Kit from SABioscience according to the manufacturer’s protocol. Samples were stored at -20°C until usage. We performed the Mouse Endothelial Cell Biology RT-Profiler PCR Array from SABioscience in a 96-Well-Plate as described in the corresponding protocol. The sample was mixed with RT-SYBR Green / ROX qPCR (SABioscience) and added to the plate; the PCR was performed on a LightCycler 480 from Roche (Basel, Switzerland). Analysis of the array was done with the PCR Array Data Analysis Web Portal from SABioscience (http://www.SABiosciences.com/pcrarraydataanalysis.php). Analysis was performed in R version 3.1.0. Pba and Dox samples were compared using Mann-Whitney statistic, a non-parametric test without assumptions on the distribution of Ct values. Adjustment for multiple testing was performed by Benjamini-Hochberg correction.

Zymography

Half of a mouse brain was homogenized with 1.4 mm Precellys-Ceramic beads with the Precellys^® 24 (PEQLAB, Erlangen, Germany homogenized material was centrifuged in a fresh tube for 5 minutes at 12 000 g at 4°C. 0.5 ml of the supernatant was incubated with 50μl of gelatine-Sepharose 4B (GE-Healthcare, Munich, Germany) for 1 hour at 4°C. After a brief centrifugation at 500 g for 2 minutes supernatant was washed 3 times with working buffer (50 mmol/L Tris-HCl, pH 7.6; 150 mmol/L NaCl; 5 mmol/L CaCl2; 0.05% BRIJ-35; 0.02% NaN3; 1% Triton X-100). Incubation of the supernatant with 150 μl elution buffer (working buffer + 10% DMSO) for 30 minutes at 4°C, followed by 2 minutes centrifugation at 500 g. 15 μl sample together with 2x SDS loading buffer and a prestained standard (Novex^® Sharp Pre-stained Protein Standard; Invitrogen, Paisley UK) (8μl + 5μl loading buffer) were applied to a SDS-page gel containing 0.5% gelatine (Sigma-Aldrich, St. Louis, USA). The gel was run at 125 V for 1.5 h. 1 h incubation in renaturing buffer (2.5% TritonX 100), followed by 30 minutes incubation at RT in developing buffer (10 mmol/L Tris-Base; 50 mmol/L Tris-HCl; 150 mmol/L NaCl; 5 mmol/L CaCl2; 0.05% BRIJ-35; adjusted to pH 7.6). Replace buffer with fresh developing buffer and incubate for 48h at 37°C. Stain overnight with 0.5% coomassie stain. Destain for 1 hour with 30% Methanol. For quantification of relative gelatinolytic activity the optical density of bands was analyzed with FIJI graphic software. Scanned images were converted into gray scale pictures and the colours were reversed before a black and white picture was generated. The size of the bands was analyzed and normalized to the naive group (fold change against naive).

Cross-presentation assay of brain microvessels

Brain microvessels were isolated as described before [27]. Briefly, the brain was minced, homogenized and fractionated by 15% dextran gradient centrifugation. Microvessels in the pellet were retained on a 40 μm cells strainer. After collagenase and DNaseI digestion for 90 min, the microvessels were plated on to a filter plate. LR-BSL8.4 cells are transduced with a T cell receptor (TCR) recognizing the SQLLNAKYL epitope from P. berghei and contain a NFAT-lacZ reporter cassette. After overnight coculture of LR-BSL8.4 cells and microvessels the wells were stained with X-gal and after 6 hours blue spots were imaged and counted in an ImmunoSpot Analyser.

Statistical analysis

Survival was analysed with the Mantel-Cox log-rank test. ECM scores are displayed as median. Parasitemia is displayed as mean/SD.

PCR-array analysis was performed in R version 3.1.0. PbA and PbA/DOX samples were compared using Mann-Whitney statistic, a non-parametric test without assumptions on the distribution of Ct values. Adjustment for multiple testing was performed by Benjamini-Hochberg correction.

If the data met criteria for normal distribution, they were tested with one-way ANOVA followed by the Tukey post-hoc test to check differences between groups. Non-parametric data were tested with Mann-Whitney U test or with the Kruskal-Wallis test followed by the Dunns post-hoc test to correct for multiple groups. For data organization and statistical analysis, Graphpad^® Prism 5.0 software was used.