Female C57BL/6 and BALB/c mice (6 to 8 weeks old) were purchased from Beijing Animal Institute (Beijing, China). All experimental animals were implemented according to the Animal Care Institute of China Medical University and complied with the Institution Committee-approved protocols. P. berghei (ANKA strain 2.34) was maintained by serial mechanical passages and prepared for challenge infection [13].

The PbTIP genomic sequences were searched from the PlasmoDB (http://www.plasmodb.org). Putative signal peptides and transmembrane domains were analyzed using SMART online server (http://smart.emblheidelberg.de). Multiple sequences were aligned using the Clustal W. The TMHMM online server (http://www.cbs.dtu.dk/services/TMHMM/) was used to predict the expressed segment of PbTIP protein.

For the expression of PbTIP, a segment of it encoding 204–335 amino acids was amplified from P. berghei genomic DNA with forward primer (3′-sequence-5′) and reverse primer (3′-sequence-5′). Using restriction enzymes BamHI and XhoI, rPbTIP was cloned into the expression vector pET32a (+) (Novagen, Darmstadt, Germany). The recombinant plasmid was transformed into Escherichia coli BL-21 (Novagen, Darmstadt, Germany). His-tagged rPbTIP was expressed at 20°C for 12 hr with 1 mmol isopropyl-β-D-thiogalactopyranoside (Sigma-Aldrich, St. Louis, Missouri, USA). Ultrafiltration was performed to remove the endotoxin from the bacterially produced protein (endotoxin<1 Eu/μg). The soluble rPbTIP was purified by Ni-NTA His-Bind Superflow (Invitrogen, Carlsbad, California, USA) in accordance to the manufacturer’s instructions. Purified rPbTIP was dialyzed with phosphate buffered saline (PBS) for 4 hr at 4°C. Then the purified recombinant PbTIP was analyzed on 10% SDS-PAGE gel.

To obtain polyclonal anti-rPbTIP antisera, 10 female BALB/c mice were immunized with rPbTIP (50 μg/mouse) using complete Freund’s adjuvant (Sigma-Aldrich). Mice were given booster immunizations of rPbTIP (25 μg/mouse) twice at 3-week intervals with incomplete Freund’s adjuvant. In the control group, mice were immunized with PBS in adjuvant formulations. After immunization, blood was collected via cardiac puncture from mice under anesthesia. Immune responses to antisera of rPbTIP was detected by ELISA as previously described [14].

After 4 days post-infection (p.i.), blood-stage parasites from the infected mice were cultured using blood-stage culture medium with the following: RPMI 1640, 25% (v/v) fetal calf serum (FCS), 50 mg/L penicillin and streptomycin at 37°C overnight [15]. On the following day, the culture was fractionated on 55% (v/v) Nycodenz-RPMI (Axis-Shield Poc, Olso, Norway). The gray layer of the interface consisted of concentrated schizonts, which was then collected and washed with PBS. On day 4 of p.i., mice were treated with sulfadiazine (Sigma, 20 mg/L, drinking water) for 2 days to collect the gametocytes. Blood was then purified on 48% (v/v) Nycodenz-RPMI, and the interface layer was collected and washed [16]. Parasitaemia was allowed to reach 1–3% on day 3 of p.i. for ookinetes purification. Collected blood was diluted to 1:10 with ookinetes medium (RPMI 1640, 25% [v/v] FCS, 50 mg/L penicillin and streptomycin at pH 8.3) at 19°C for 24 hr. Then the ookinetes were separated on 62% (v/v) Nycodenz-RPMI. In order to obtain the lysate, the purified parasites were diluted in 5 ml of 0.17 mol NH₄Cl for 10 min on ice to lyse the iRBCs. Then the cytoplasmic, nuclear and membrane proteins of PbTIP from separated schizonts, gametocytes and ookinetes were extracted using Minute^™ Plasma Membrane Protein Isolation and Cell Fractionation Kit (Invent Biotechologies, Plymouth, Minnesota, USA) according to the manufacturer’s instructions.

The isolated Plasmodium protein (cytoplasmic, nuclear and membrane protein), with 30 μg of 3 representative stages (schizonts, gametocytes and ookinetes) were electrophoresed on 10% SDS-PAGE gel and transferred on to 0.22 μm PVDF membranes (Bio-Rad, Hercules, California, USA). Membranes were then incubated with 5% non-fat milk with TBST for 2 hr and incubated with anti-rPbTIP antisera (1:500). After washing 3 times with TBST, the membranes were incubated with HRP-conjugated goat-anti-mouse antibodies. Proteins on the blot were visualized with a Pierce ECL Western blotting kit (Thermo Scientific, Waltham, Massachusetts, USA).

Purified parasites were fixed with 4% paraformaldehyde (Sigma-Aldrich) in PBS for 30 min at room temperature [17]. Slides were incubated with anti-rPbTIP antisera (1:100) for 1 hr at 37°C. The slides were then washed twice with PBS and incubated with FITC-conjugated goat anti-mouse antibody (1:500) (Invitrogen) for 30 min at 37°C. Nuclei were stained with 4, 6-diamidino-2-phenylindole (DAPI; Invitrogen) 1 μg/ml. Stained parasites were mounted with ProLong® Gold antifade reagent (Invitrogen), and captured on fluorescence microscopy (Olympus, Tokyo, Japan).

Forty female C57BL/6 mice were infected with intraperitoneal (i.p.) injection of 5×10⁶ WT P. berghei ANKA pRBCs [18]. Then the experimental mice were divided into 2 groups of 20 each: PBS i.v. injection group on days 1, 0, 1, 3, 5, and 7; rPbTIP i.v. injection group 2 mg/kg on days 1, 0, 1, 3, 5, and 7 [11]. The percentage parasitaemia of Giemsa-stained thin smears from tail blood were monitored by light microscopy and determined by the formula (the number of pRBCs/total number of RBCs)×100. ECM symptoms evaluated were as follows: ruffled fur, respiratory distress, limb paralysis, wobbly gait, hunching, paroxysmal shivering and coma [19]. When the ECM symptoms were performed, BBB was evaluated by i.v. injection using 200 μl of 2% (wt/vol) Evans blue solution (Sigma-Aldrich) at 6 days p.i. of 3 mice in each group. After 1 hr, the mice were euthanized and perfused with 20 ml of 0.9% NaCl. Brains were taken out and immersed in 2 ml of formamide for 48 hr at 37°C to extract Evans Blue, which was determined by measuring the absorbance at 620 nm [20,21]. When the mice were evaluated for neurological symptoms on day 6, brains of 5 mice from each group were removed, fixed in 4% paraformaldehyde for 24 hr and immersed into paraffin. Serial 4 μm thick brain sections were made and stained with hematoxylin and eosin (HE) to examine for micro-vascular obstruction and leakage. Along the endothelial lining, immunohistochemical staining was detected with specific polyclonal antibodies against ICAM-1 (Abcam, Cambridge, UK), VCAM-1 (Santa Cruz Biotechnology, Santa Cruz, California, USA) and CD36 (Santa Cruz Biotechnology) as previously described [22,23]. Followed by incubation with biotin-conjugated secondary antibodies, and streptavidin-conjugated peroxidase and the color development was shown using 3-amino-9-ethylcarbazole as the substrate. Nuclei were counterstained with hematoxylin. Positive vessels were visualized by microscopy at ×400 and counted in 20 fields per mouse.

Cytokine levels of IFN-γ, TNF-α, IL-1, IL-12, IL-10, and TGF-β in splenocyte culture or serum were measured by ELISA (R&D Systems, Minneapolis, Minnesota, USA). For quantification of splenocyte cytokines, 10⁶ splenocytes were obtained from 5 mice of each group were isolated on days 0, 2, 4, 6, and 8 p.i. and then cultured for 48 hr ex vivo. Then the culture supernatants or serum were collected and measured by ELISA according to the manufacturer’s protocol. All the experiments were repeated 3 times.

Five mice from each group were euthanized on days 0, 2, 4, 6, and 8 p.i. Total RNA was extracted from the brain and spleen samples by using TRIzol (Invitrogen). Contaminated DNA was digested by DNAase. Total RNA was transcribed with oligo (dT) primers for reverse transcription (RT) (Takara, Kusatsu, Japan). The resulting cDNA of IFN-γ, TNF-α, IL-1, and IL-12 from the spleen samples and VCAM-1, ICAM-1, CD36, CXCL9, CXCL10, and CXCR3 from the brain samples were used for qRT-PCR assays with SYBR Green quantitative PCR kit (Takara) by 40 cycles in ABI Quant Studio 6 Flex (Applied Biosystems, Foster City, California, USA). Primer sequences were listed in Table 1. β-actin was used as an internal control for measuring the relative ratio of each target gene. Normal group values were taken as 100% and treated sample values were calculated based on the control. The specificity of qRT-PCR was confirmed by melting-curve analysis. qRT-PCR analysis was determined with 3 replicates.

Spleens from 5 mice of each experimental group were harvested on days 0, 2, 4, 6, and 8 p.i. The spleen of day 0 was labeled as control. To measure the activation of regulatory T cells (Tregs), 10⁷ fresh splenocytes were stained with FITC-anti-CD4, and PE-anti-CD25 for surface staining in cell staining buffer (Biolegend, SanDiego, California, USA). After the cells were fixed and permeabilized, intracytoplasmic staining by labeling with fluorochrome conjugated antibody of APC-anti-Foxp3 was performed. After that, the resuspended, fixed and intracellularly labeled cells were analyzed on FACSCalibur flow cytometers (BD Bioscience, San Jose, California, USA). Data were analyzed with FlowJo software (Treestar). The experiments were replicated 3 times.

For comparisons between 2 groups, statistically significant differences were analyzed by Student’s t-test or 2-way analysis of variance (ANOVA) (SPSS 17.0 software, SPSS Inc., Chicago, Illinois, USA). For comparisons among 3 groups, statistical significance was determined using Kruskal-Wallis test. Significant differences in survival analysis between the groups were calculated by Kaplan-Meier long rank test. Data are presented as means±SEM. P-values<0.05 were considered to be significant.

Through bioinformatic searches from Plasmodium database in PlasmoDB (http://www.plasmodb.org), The predicted protein contains a putative signal peptide and transmembrane domain (Fig. 1A). The PbTIP gene is located on chromosome 12 and encodes a protein of 703 AA with a calculated molecular weight of 80.4 kDa. We identified the P. berghei ANKA homolog of TIP (PBANKA_124360), which has a similar genome and structure with TIP of Plasmodium falciparum (Fig. 1B). Multiple sequence alignments using Clustal W showed that the gene was conserved among Plasmodium species (Fig.1B). Based on the above discussion, we annotated this conserved P. berghei protein as PbTIP.

Anti-rPbTIP antisera determined native PbTIP to be a membrane protein in P. berghei. We screened PbTIP protein in TMHMM online server (http://www.cbs.dtu.dk/services/TMHMM/) (Supplementary Fig. S1A). According to the following criteria, a segment of PbTIP protein corresponding to 204–277 AA were selected for VCBS using Pfam database (http://pfam.xfam.org/) (Supplementary Fig. S1B). PbTIP belongs to protein families PF13517 (Family: VCBS). The VCBS of PbTIP protein sequences were also mined in NCBI database (Supplementary Fig. S1C). Besides that, we also showed SCOP fields among the comparison species. Among such genes and bio-informatics comparisons, we finally selected 204–335 AA for the specific segment for PbTIP. Subsequently, a segment of rPbTIP that corresponds to 204–335 AA was selected to express in E. coli BL-21. The rPbTIP had a molecular weight of −35 kDa through SDS-PAGE analysis, which was correlated with the predicted molecular size (Fig. 2A). After ultrafiltration and purification, rPbTIP was used to immunize BALB/c mice to obtain polyclonal antibody. ELISA showed that the mice induced strong antibody responses compared to PBS control (Fig. 2B) (t-test, P<0.01). rPbTIP-specific IgG titers reached 1:204,800 after third immunization (Fig. 2C) (t-test, P<0.01).

To determine whether the polyclonal antisera can react with native PbTIP in P. berghei, Western blot was performed with isolated cytoplasmic, nuclear and membrane protein of PbTIP from separated schizonts, gametocytes and ookinetes (Supplementary Fig. S2). The results demonstrated that the anti-rPbTIP antisera recognized native PbTIP protein bands at approximately 80 kDa as a membrane protein on the outer surface of P. berghei in 3 representative stages, which was as expected from the PbTIP expression profile (Fig. 3A). The protein bands confirmed the increased antisera against native PbTIP in P. berghei, which was consistent with the predicted molecular weight of PbTIP present on the surface of P. berghei.

Next, we confirmed the expression of PbTIP, and determined the localization of PbTIP by IFA using immune sera of rPbTIP. The consequences revealed strong fluorescence on schizonts, female gametes, male gametes, zygotes, retorts and ookinetes (Fig. 3B). Pbs21mAb served as a positive control. The localization by IFA indicated that PbTIP was membrane protein on the outer surfaces in asexual and sexual stages of P. berghei.

Comparisons of parasitaemia on day 8, 10, and 12 p.i. in the 2 groups demonstrated that parasitaemia in rPbTIP i.v. injection group showed significant difference than parasitaemia in PBS i.v. injection group (2-way ANOVA; P<0.05) (Fig. 4A). Nevertheless, comparison of parasitaemia on day 6 p.i. in the 2 groups demonstrated that parasitaemia (6%) in PBS i.v. injection group showed no significant difference than parasitaemia (5.2%) in rPbTIP i.v. injection group (2-way ANOVA; P>0.05) (Fig. 4A). These results did not provide any information about the different mortality rates associated with parasitaemia in the 2 groups. Two groups of 20 WT P. berghei ANKA infected mice each were injected with either PBS or rPbTIP 2 mg/kg on days −1, 0, 1, 3, 5, and 7. We observed rPbTIP i.v. injection group could alleviate the severity of ECM. Two groups demonstrated significant differences in the survival curves by daily monitoring (log-rank test; P<0.05) (Fig. 4B). Approximately on day 6, 80% of mice in PBS i.v. injection group demonstrated apparent neurological symptoms and death occurred between days 6 and 7 for ECM. Nevertheless, 30% of mice in rPbTIP i.v. injection group died due to ECM during days 6 to 7 (log-rank test; P<0.05) (Fig. 4B).

On day 6 of p.i., a substantial loss of Evans blue dye as indicated by BBB leakage in rPbTIP i.v. injection group was lower than that in PBS i.v. injection group (Kruskal-Wallis test; P<0.05) (Fig. 5A). Immunostaining of VCAM-1, ICAM-1 and CD36 on the endothelium of microvessels showed less intense in rPbTIP i.v. injection group (Kruskal-Wallis test; P<0.05) (Fig. 5A). Immunocytochemistry revealed that the amount of leukocytes per microscopic field in rPbTIP i.v. injection group showed significant reduction than PBS i.v. injection group (Kruskal-Wallis test; P<0.05) (Fig. 5A). Then the corresponding adhesive molecules and chemokines were identified by qRT-PCR analysis in the brains on day 6 p.i. qRT-PCR studies have revealed significantly lower expression levels of VCAM-1, ICAM-1 and CD36 in rPbTIP i.v. injection group, which were correlated with the above immunostaining results (2-way ANOVA; P<0.01). mRNA expression levels of chemokines CXCL9, CXCL10 and receptor CXCR3 were much lower in rPbTIP i.v. injection group (2-way ANOVA; P<0.01) (Fig. 5B). These results suggested that rPbTIP i.v. injection group showed significant reduction of inflammation and less severe cerebral pathology than PBS i.v. injection group.

Detection of cytokines in splenocyte culture supernatants by ELISA showed lower splenic cytokine levels in rPbTIP i.v. injection group than PBS i.v. injection group (2-way ANOVA; P<0.05) (Fig. 6A–D). Correspondingly, qRT-PCR analysis revealed that the pro-inflammatory cytokines in spleens demonstrated similar results as in ELISA. The results showed smaller number of cytokines in rPbTIP i.v. injection group than that in PBS i.v. injection group (2-way ANOVA; P<0.05) (Fig. 6E–H).

Compared with PBS i.v. injection group, CD4⁺CD25⁺Foxp3⁺ regulatory T cells (Tregs) was promoted in rPbTIP i.v. injection group as absolute numbers and percentages (2-way ANOVA; P<0.01) (Fig. 7). A few potential mediators of Tregs includes Foxp3, TCR4, CD25, cytotoxic T lymphocyte Ag-4 (CTLA-4), anti-inflammatory cytokines IL-10, and transforming growth factor β (TGF-β) [24–31]. In recent studies, IL-10 and TGF-β have shown essential impact in the regulation of Th1 response. So, we measured the expression of IL-10 and TGF-β in splenocyte culture and serum by ELISA. The results showed that the expression levels of IL-10 and TGF-β from splenocyte culture were elevated in rPbTIP i.v. injection group (2-way ANOVA; P<0.05) (Fig. 8A and B). Furthermore, the concentration of IL-10 and TGF-β in sera of rPbTIP i.v. injection group was higher than PBS i.v. injection group (2-way ANOVA; P<0.05) (Fig. 8C and D). Collectively, these results suggested that the host immune response from pro-inflammatory milieu to anti-inflammatory milieu was increased by the amount of anti-inflammatory cytokines of Tregs, IL-10 and TGF-β in ECM, reducing the severity of ECM.