CD19(+) B Cells Confer Protection against Experimental Cerebral Malaria in Semi-Immune Rodent Model

In African endemic area, adults are less vulnerable to cerebral malaria than children probably because of acquired partial immunity or semi-immune status. Here, we developed an experimental cerebral malaria (ECM) model for semi-immune mice. C57BL/6 (B6) mice underwent one, two and three cycles of infection and radical treatment (1-cure, 2-cure and 3-cure, respectively) before being finally challenged with 104 Plasmodium berghei ANKA without treatment. Our results showed that 100% of naïve (0-cure), 67% of 1-cure, 37% of 2-cure and none of 3-cure mice succumbed to ECM within 10 days post challenge infection. In the protected 3-cure mice, significantly higher levels of plasma IL-10 and lower levels of IFN-γ than the others on day 7 post challenge infection were observed. Major increased lymphocyte subset of IL-10 positive cells in 3-cure mice was CD5(−)CD19(+) B cells. Passive transfer of splenic CD19(+) cells from 3-cure mice protected naïve mice from ECM. Additionally, aged 3-cure mice were also protected from ECM 12 and 20 months after the last challenge infection. In conclusion, mice became completely resistant to ECM after three exposures to malaria. CD19(+) B cells are determinants in protective mechanism of semi-immune mice against ECM possibly via modulatory IL-10 for pathogenic IFN-γ production.


Background
Malaria caused an estimated 655,000 deaths in 2010, mostly among African children [1]. In sub-Saharan Africa, malaria might account for 40% of pediatric admissions to some hospitals, 10% of which may be due to cerebral malaria [2]. The prevalence of cerebral malaria in endemic areas in Africa (Zambia, Kenya, Tanzania and Malawi) was 1.12 cases per/1000 children per year [3]. Mortality from cerebral malaria remained between 10% and 14% in sub-Saharan and Southeast Asia [4][5][6].
In high transmission areas, the adults, as semi-immune individuals, were less vulnerable to cerebral malaria than children because of acquired partial immunity [7,8]. The incidence rate of cerebral malaria in patients aged from 16 years and above was less than 10% as compared to 34% of patients aged under 5 years [7]. Naturally acquired immunity to malaria minimizes malaria morbidity and mortality in older children and adults living in intensive Plasmodium sp. transmission regions [9,10]. Noticeably, cerebral malaria rarely occurs in very young children at their first malaria infection, but often attacks young children undergoing second or subsequent malaria infections [11,12]. However, the mechanisms leading to protection versus death from cerebral malaria and why cerebral malaria was more common in children than in adults [13,14] are not yet fully understood. It has been proposed that T cells but not antibody response induced the immunity against malaria in human experiment of repeated malaria infection [15]. Immunomodulatory mechanism was involved in preventing ECM in mouse model [16]. Recently, IL-10 producing regulatory CD19(+) B cells has been intensively studied in abrogating immune-pathologies [17][18][19] and modulatory IL-10 has been beneficial for prevention of auto-immune diseases [20][21][22][23].
In mouse model, it is known that C57BL/6 mice are susceptible to ECM induced by virulent Plasmodium berghei ANKA (PbA) [14,24] and usually die from typical neurological symptoms between days 6-10 after infection [25][26][27][28]. Approximately, 95% to 100% of those mice exhibited the symptoms of ECM such as coma, convulsion, paralysis and hemiplegia, and died within 14 days of infection [29][30][31]. The C57BL/6 mouse model was wellestablished to identify the host determinants in protective immunity and malaria parasite mediated immunopathology [32,33]. Previously, we have established a semi-immune malaria model using PbA and C57BL/6 mice delineating the mechanism of destruction of erythrocytes and severe malaria anemia [34,35], but ECM was not mentioned. Here, we asked the question whether the semi-immune could prevent the development of ECM in the PbA-infected C57BL/6 model, and what are the factors in this regulatory and protective mechanism. We demonstrated that three exposures to malaria infection provided the resistance to ECM. Our data indicated the involvement of CD19(+) cells in the protection of ECM in semi-immune model.

Results
Three cycles of malaria infection and treatment are necessary for mice to become completely resistant to cerebral malaria Semi-immune mice were generated by performing 0, 1, 2, and 3 cycles of infection and radical anti-malaria treatment, namely 0cure, 1-cure, 2-cure, and 3-cure mice. We found that within 7-9 days post challenge infection (PI), 100% of 0-cure mice, 67% of 1cure mice and 37% of 2-cure mice (Fig.1A) succumbed to cerebral malaria with neurological signs including hemi-paresis, ataxia, paralysis, seizures and coma. In addition, the post-mortem brain histopathology images showed vascular wall disruption, hemorrhage, and sequestration of infected erythrocytes (Fig.S1). Conversely, 100% of infected 3-cure mice survived beyond day 14 and did not develop ECM. There was no reduction in mortality caused by ECM between 0-and 1-cure mice (p = 0.3094, log-rank test), whereas the survival rate noticeably increased among 2-cure mice compared to 0-cure mice (p = 0.0053, log-rank test). However, two cycles of PbA infection and anti-malaria treatment were not sufficient to produce an optimal immune status to protect mice completely from ECM. The survival rate was significantly higher in 3-cure mice compared with 0-cure controls (p = 0.0046, logrank test). The result was reproduced in another independent experiment. Thus, three infection-cure rounds were needed to establish the ECM-resistant mice. Additionally, the frequency of resistance in ECM-susceptible B6 mice increased dramatically after each cycle of infection and treatment (Fig.1A), suggesting that protective mechanism was further remarkably enhanced after each one.
Next, we examined the parasitemia of all mice in the different cure groups. Parasitemia increased steadily in the 0-and 1-cure mice as well as in the 2-and 3-cure mice (Fig.1B). However, highly ECM-resistant mice, including 2-and 3-cure mice, had significantly lower parasitemia during the first six days of infection, as indicated on day 4, 5 and 6 (p,0.05; Fig.1B), showing that the mice exposed to more cycle of infection were able to control parasitemia more efficiently than 0-and 1-cure.
Inhibition of ECM development in 3-cure mice was not due to the reduction in parasitemia level during the early stage of infection Since parasitemia levels were inhibited during early period of infection, we examine the possibility that 3-cure protection was derived from lowered peripheral parasite. We generated high parasitemia in 3-cure mice by challenging them with high dose (10 7 ) of PbA-pRBCs while 0-cure mice were inoculated with 10 4 PbA-pRBCs (as the control). Although all of 3-cure mice infected with higher PbA-pRBCs dose showed comparable parasitemia level to the 0-cure controls, the high-dose PbA-pRBCs infected mice did not develop ECM ( Fig.2A, B). Significant difference in survival between 0-and 3-cure mice was recorded over the course of infection (p = 0.0002, using log-rank test for survival (p.0.05, Fig.2A). Additionally, similarity in parasitemia was observed between two mice-groups throughout the first week of infection (p.0.05, Fig.2B), clearly demonstrating that ECM developed in 0cure mice at higher parasitemia levels at which ECM did not occur in 3-cure counterparts. Alternatively, parasitemia was not correlated to ECM-resistance in 3-cure mice. Taken these data together, we concluded that there was an alternative protective mechanism other than lowered-parasitemia in 3-cure mice.

PbA-specific IgG antibody levels day 7 PI in ECM-resistant and susceptible mice
To investigate the humoral immune response in semi-immune mice, we initially assessed the role of malaria-specific antibody. As shown, from day 0 to day 5 PI, the levels of all IgG, IgG1 and IgG2a in 2 and 3-cure mice were significantly higher than in 0and 1-cure mice (p,0.05, using Mann-Whitney U test) (Fig.3A). Although on day 7, all types of malaria-specific antibody drastically increased in 1-cure mice whereas 2-and 3-cure mice did not show any change (Fig.3A), most of the ECM-susceptible mice (1-cure) became moribund around this day. The significantly higher antibody amounts in 1-cure on day 7 PI indicate the normal booster response whereas 2-and 3-cure mice antibody response was marginal or even suppressed. There was no significant difference in antibody responses (Fig.3A) between 2cure and 3-cure mice. Although not statistically significant, survival rate was slightly lower in 2-cure mice (63%) than 3-cure mice (100%) (p = 0.1328) (Fig.1A), suggesting that malaria specific antibodies could not provide 100% of protection. Further study with larger number of mice will confirm this issue. Vigorous humoral immune response in 1-cure mice at late stage of ECM could not rescue them from ECM.
Plasma level of IFN-c was suppressed on day 5 PI and that of IL-10 vigorously increased on day 7 PI in ECMresistant mice ECM has been suggested as the result of excessive immune responses leading to overproduction of pro-inflammatory cytokines in which IFN-c is essential in ECM pathogenesis. To look for the possible suppression to immune responses in ECM-resistant mice, we further assess Th1, Th2 activation and immunomodulators in mice with different degrees of immunity. We examined profile of IFN-c, a pathogenic Th1 cytokine of ECM [36][37][38], IL-4, a Th2 cytokine, TGF-b and IL-10, regulatory cytokines [39] during the first week of infection. The amount of IFN-c was similar in all groups on day 0 PI, increased and became higher in 0-and 1-cure, but suppressed in 2-and 3-cure mice on day 7 PI (p = 0.021, except for 1-cure vs. 2-cure with p = 0.149) (Fig.3B). IL-4 levels were similarly low in all groups during the course of observation, slightly increased on day 5 PI and also inhibited on day 7 PI in 2 and 3-cure mice (Fig.3B). Together, Th1 and Th2 cytokines were suppressed in ECM-resistant mice on day 7 PI. Active TGF-b, a regulatory cytokine, was present at low levels and appeared to be negatively correlated with IL-10 on day 7 PI in all groups of mice (Fig.3B). By contrast, IL-10 levels started increasing on day 5 PI in all mice-groups, but dramatically elevated in 3-cure mice on day 7 compared to the others (p = 0.021) (Fig.3B). There was a gradual increase of IL-10 levels from 0-to 2-cure mice but a big jump in 3-cure mice around day 7, implying the enhancement of IL-10 production after each cycle of malaria infection and treatment. Remarkably, higher level of plasma IL-10 in ECM resistant 3-cure mice versus its much lower level in 0-, 1-, and 2-cure mice suggested that the inhibition of exacerbated inflammatory responses was possibly via IL-10 in 3cure mice.
CD19(+) B cells are a major source of IL-10, in which IL-10(+)CD5(2)CD19(+) cells appear as the main subset We further sought the major source of IL-10 in the spleen of ECM-susceptible, 0-cure and ECM-resistant, 3-cure mice. The median of total splenocytes' numbers of day 7 PI 3-cure mice (9.17610 7 ) was 2.2 fold higher than those of 0-cure mice (3.81610 7 ), indicating that repeated exposure of malaria enhanced the proliferation of splenocytes (Fig.4A). Because the increased IL-10 level in the plasma was suggested to be related to the protection of ECM in 3-cure mice, we investigated IL-10 producing spleen cells of 3-cure mice. Splenocytes were collected from 0-cure and 3cure mice on day 7 PI and were cultured with parasitized erythrocytes, lipopolysaccharides (LPS), phorbol 12-myristate 13acetate (PMA), ionomycin, and monensin (Golgi plug) for 5 hours. Intracellular cytokine staining revealed that the median percentage and absolute number of IL-10(+) cells were 1.94-fold and 3.62-fold higher in 3-cure (0.97% and 726,930) than those in 0-cure mice (0.5% and 200,421) on day 7 PI, respectively ( Fig.4B-D, p,0.05).

Naïve B6 mice receiving CD19(+) B cells transfer from 3cure mice were resistant against ECM
To test the hypothesis that CD19(+) B cells derived from 3-cure mice splenocytes may confer the abrogation of immune-mediated pathology, CD19(+) B cells in spleen were isolated and passively transferred to naive mice. After first separation of CD11c(2) splenocytes of 3-cure and 0-cure mice at day 7 PI, we then sorted CD11c(-) cells based on antibody to CD19 conjugated magnetic beads to isolate CD19(+) cells. Before transfer experiment, purity of CD19(+) cells was assessed. The average cell numbers of CD19(+) cells were 3610 7 and 1.5610 7 in one spleen of 3-cure and 0-cure mice, respectively. The obtained CD19(+) cells were then intravenously injected at high dose (3610 7 cells per mouse) and low dose (1.5610 7 cells per mouse) into naïve mice followed by intra-peritoneal infection of 10 4 PbA-pRBCs. As shown in Fig.6, 100% and 80% of mice receiving high and low dose of 3cure CD19(+) cells, respectively survived compared with 0% of survival in PBS treated controls, indicating that CD19(+) cells from 3-cure mice controlled the pathology. On the contrary, in mice treated with low dose of 0-cure CD19(+) cells, survival rate was much reduced, only 20% (Fig.6), revealing that CD19(+) cells of 0cure mice were insufficient to inhibit ECM. However, transfer of 0-cure CD19(+) cells at high dose improved survival rate of recipient mice to 75%, suggesting dose-dependent protection of 0cure CD19(+) cells.

Protection against ECM in semi-immune mice existed throughout life
For understanding of the longevity of protection against ECM in the absence of persisting exposure, 3-cure mice were rechallenged with 10 4 PbA-pRBCs twelve and twenty months after the last challenge. It was recorded that 100% of the aged mice were still resistant to ECM compared with 100% ECM-induced death of age-matched 0-cure counterparts (Fig.7A, p,0.01). No significant difference in parasitemia was observed between aged 3cure mice and 0-cure controls during the first week of infection (Fig.7B), once again indicating a resistant mechanism independent to parasitemia.

Discussion
In this work, we established a semi-immune model in ECMsusceptible mouse strain with virulent PbA. This was achieved by repeated infection and radical anti-malaria treatment, mimicking the natural boost infection and recovery in humans in areas of intense P. falciparum transmission. We found an absolute protection against ECM in 3-cure mice. Simultaneously, we also demonstrated that CD19(+) B cells of the semi-immune mice promote their acquired immunity to ECM.
Splenic CD19(+) B cells, including regulatory and memory B cells [44,18], of infected 3-cure mice were proved to completely protect naïve B6 mice against ECM in the passive transfer experiments, confirming the beneficial effects of CD19(+) cells in pathology. High survival rate (75%) of mice transferred with CD19(+) cells from two spleens of infected 0-cure mice highlighted that protective CD19(+) cells were present in 0-cure mice but their number was not enough to prevent ECM.
Apart from the correlation with immunopathogenesis in autoimmune disease [47], understanding of CD19(+) is still unclarified in infectious disease, particularly in ECM. CD19(+) cells have been known to play a protective role in the control of pathogenic T cell activation in previous studies [18]. It was found that CD19 (2/2) mice had augmented experimental autoimmune encephalomyelitis responses and develop severe non-remitting form of the disease [48]. In addition, similar benefits of CD19(+) were evident in CD19(2/2) mice which had increased contact hypersensitivity responses after subsequent contact with the sensitizing antigen [49]. Our data were in consistent with these findings that CD19(+) cells enhanced the protective immunity against the immune-pathologic diseases including ECM.
IL-10 appears to oppose the immunopathological process [50][51][52][53] which is driven by IFN-c, a key pathogenic factor of ECM [37,36,38]. It was discovered that three exposures to malaria increased mouse's splenocytes number drastically compared to 0cure counterparts. The abrupt increase of IL-10(+) cells in splenocytes of 3-cure mice than that in 0-cure controls, could account for the lack of ECM in 3-cure mice via the augmentation of peripheral IL-10 concentration. The in-vitro experiments revealed that CD19(+) B cells in spleen of 3-cure mice were the major source of IL-10 upon 5 hour-stimulation. So far, no study has been reported on protective regulatory B cells in ECM. Besides, CD3(+) cells expressed a considerable amount of IL-10, but significantly less than CD19(+) B cells. IL-10(+)CD19(+) B cells and IL-10(+)CD3(+) T cells in infected spleen were significantly increased in number after being exposed to malaria three times compared to infected 0-cure controls.
Regarding the subset of IL-10(+)CD19(+) B cells, majority of these cells were negative in CD5 and Foxp3 expression. As shown, Foxp3(+)IL-10(+)CD19(+) cells accounted for a low percentage of IL-10(+) cells but significantly higher in 3-cure than 0-cure mice. Regulatory B cells, including Foxp3(+)CD5(+)CD19(+) and regulatory CD1d hi CD5(+)CD19(+) B cells (B1a) have been intensively studied in recent years [54][55][56][57][58][59][60][61]19,18]. TLR-mediated negative modulation of B cell IL-10 evolved as a mechanism that prevents an exacerbation of acquired antigen-specific responses (T cell responses) [17]. IL-10 secreted from these B cells could then down-regulate the local immune responses by a 'negative-feedback loop' resulting in the limitation of excessive self-tissue injuries [17]. Nonetheless, CD5(2)CD19(+) cells (B1b cells) were pivotally responsible for IL-10 production in 3-cure mice. Our results diverged from studies which suggested B1a cells as the major IL-10 secreting cells [18]. IL-10 production by Foxp3(+)CD19(+) have not been mentioned previously [62]. It appeared to be the first time Foxp3(+)IL-10(+)CD19(+) cells were identified as a new subset of IL-10 producing B cells in our mouse model study. This minor population will be further investigated for the role of immune-modulation. Collectively, CD19(+) B cells prevented ECM in naïve mice possibly via regulatory IL-10 derived mostly from B1b cells. The second source of IL-10 in 3-cure mice was CD3(+) cells which were considered as the predominant IL-10 producers in infected naïve B6 mice in previous malaria studies [63,64]. In line with one of our findings (data not shown), Th1 cells were reported as the source of IL-10 in Plasmodium chabaudi chabaudi AS infection to protect mice from pathology during acute phase [65]. Additionally, macrophages, mast cells, NK cells, neutrophils and eosinophils were also reported as origins of regulatory IL-10 [66,62,65,67,68]. The method used here for stimulating IL-10 secreting cells is cocktail of parasitized red blood cells (pRBC), lipopolysacharide (LPS), phorbol 12-myristate 13-acetate (PMA), ionomycin. This method was modified from previous studies [18,69]. Neither pRBC nor cocktail of LPS, PMA and ionomycin resulted in stimulation of IL-10. It may be attributed to the differences between in vitro and in vivo condition. Using IL-10 reporter mice could be better for identification of IL-10 production from all myeloid and lymphoid cell types [70]. In 3-cure mice on day 7 PI, profoundly high plasma levels of IL-10 and suppressed levels of IFN-c, known to play a key role in the development of ECM [71,72], implied that IL-10 was a much potential modulator in this protection. Of notes, activation of Th1 cytokines including IFN-c was modulated by IL-10 [73,52,53]. It was previously found that both IFN-c and IFN-c receptordeficient mice showed resistance to the development of ECM [74,75] while IL-10 has been strongly assumed as immunemodulator [76][77][78]68,79]. Previously, filaria-infected B6 mice was protected from full progression to ECM via suppressing the overwhelming inflammatory reaction by IL-10 [80]. 40% of Balb/ c ECM-resistant mice developed ECM when treated with IL-10 neutralizing antibody and 70% of recombinant IL-10 treated ECM-susceptible CBA mice were protected from ECM [81]. Accordingly, IL-10, mostly from CD19(+) B cells, could be a defense barrier for 3-cure mice against ECM.
Reducing ECM in naïve mice via low parasitemia has been reported in PbA infected C57BL/6 mouse model, in which the  parasitemia was correlated with reduced levels of pathogenic cytokines, chemokines, CD8+ T cells and parasites in the brain [28,82,83]. In contrast, other studies have reported that parasite burden does not appear to play a pivotal role in the development of ECM in the PbA model [84][85][86]16,87]. It was reported that malaria specific antibodies inhibited parasite proliferation [88]. In our study, 3-cure mice, having lower parasitemia and higher malaria specific antibodies than 0-cure counterparts, did not develop ECM during the course of infection. However, high dose of PbA infection in 3-cure mice induced parasitemia similar to 0cure controls but none of 3-cure mice manifested symptoms of ECM, demonstrating that parasitemia-independent factors contributed to ECM resistance in semi-immune mice. Instead, regulatory IL-10 cytokine should work on the pathogenic process mainly provoked by Th1 inflammatory cytokines.
We found that after two or three exposures to malaria, mice developed antibody response. However, the antibody amounts drastically increased in 1-cure mice on day 7 PI more than the others (Fig.3A) but they almost die due to ECM, implying that parasite-specific antibodies at late stage could not prevent ECM in this mouse model study. The balance between consumption and production of antibody was maintained in 3-cure mice, presumably due to the suppression of IL-10 to Th2 cell activation, inhibiting the antibody production of plasma cells [89].
In our study, the role of IL-10 derived from CD19(+) cells, malaria specific antibodies and the mechanism how the donor CD19(+) cells interact with other cells in the recipient that develops primary malaria infection have not been confirmed. CD19(+) cells could be considered as an endogenous source of IL-10 in the recipient. Additionally, effector cells transformed from CD19(+) B cells is also able to play a protective role. Antigen specific antibodies and IL-10 secreting CD19(+) cells in semiimmune mice should be studied for inhibition of ECM. Similar transfer experiment of IL-10 producing B cells generated by nonpathogenic stimulus will determine their applicability in treatment of cerebral malaria.
Three-cure mice were also absolutely protected from ECM during the twelve and twenty months after the last challenge infection of malaria, demonstrating that the resistance against ECM could persist throughout life. This may explain why cerebral malaria occurs less commonly in adult and elderly people than in children in African endemic areas. Long term antibody production or T cell memory could prevent ECM in aged 3-cure mice. Previously, exposed children who had documented P. falciparum infections several years ago, but minimal exposure since then had similar frequencies of memory B cells to those of persistently malaria exposed children [90]. The chronic presence of the parasite might drive expansion of memory B cells in P. falciparumexposed adults and children compared with naïve adults [91]. Therefore, memory B cells are one of candidates in suppressing ECM in aged 3-cure mice. Together, the expansion and long existence of malaria specific memory B cells in semi-immune mice is our expectation.

Conclusion
Three times of repeated malaria infection were required for susceptible B6 mice to develop a complete resistance against ECM. The repeated malaria infection promoted the immunity against ECM in mouse model by splenic CD19(+) cells possibly via regulatory IL-10. The major source of IL-10 was CD5(2)CD19(+) cells (B1b). Further study will be performed to elucidate functional heterogeneity of semi-immune CD19(+) B cells in ECM. The protection against ECM in semi-immune mice was long lived despite of no further exposure to malaria. Studies targeting on human IL-10 producing B cells in people living in malariaendemic regions should be conducted to clarify the whole spectrum of similar resistant mechanism in humans.

Materials and Methods
Mice, Parasites and establishment of experimental cerebral malaria C57BL/6 (B6) mice were supplied by SLC laboratories, Fukuoka, Japan and were housed under the barrier conditions at the Animal Center of the Institute of Tropical Medicine. Mice used in all experiments were female and used between 6-8 weeks of age. Plasmodium berghei ANKA (PbA) strain was originally obtained from T.Yanagi at National Bioresource Center at the Institute of Tropical Medicine Nagasaki University and used in all experiments after passaged once through naïve Balb/c mice. The PbA was selected for its capacity of ECM induction in B6, with neurological signs (ataxia, paralysis, deviation of the head, and convulsion) appearing 6 to 10 days after infection [92]. Blood stages of the parasite were stored in liquid nitrogen at a concentration of 50% parasitized red blood cells in phosphatebuffered saline (PBS) containing 10% glycerol (Wako Pure Chemical Industries, Ltd, Japan). Cerebral malaria was induced by intra-peritoneal injection of 10 4 PbA-infected red blood cells (pRBCs). Parasitemia was monitored on Giemsa-stained thin blood smear from the day 1 to day 8 post infection (PI) and after that, every two days. The parasitemia was expressed as a percentage of more than 1000 RBCs. The day of patency was determined following microscopic examination of 10,000 RBCs.

Generation of the semi-immune mice and challenge infection
The method based on a previous description [35] was pivotally modified. Briefly, C57BL/6 mice were infected with 10 4 PbA-pRBCs and then treated on day 5 after infection with intra- peritoneal injection of chloroquine (20 mg/kg/day) and pyrimethamine (20 mg/kg/day) daily for 7 days. Before subsequent rounds of infection, the parasite-clearance was confirmed and mice were rested for two weeks and then re-challenged with 10 4 PbA-pRBCs. Four groups of mice underwent zero (the control group without any infection and treatment), one, two, and three cycles of infection and treatment (designated as 0-cure, 1-cure, 2-cure and 3-cure) before finally challenged with 10 4 PbA-pRBCs without treatment. The course of infection was monitored twice daily for neurological symptoms including ruffled fur, abnormal postural responses, reduced reflexes and grip strength, coma, and convulsions. Mice that presented all these symptoms were considered as having ECM [25,93]. The mice being either judged to have ECM or moribund to infection were euthanized. Mice were inhaled with lethal dose of diethyl ether (Wako Pure Chemical Industries, Ltd.) and then were exsanguinated via cardiac puncture. Three-cure mice were also intra-peritoneally challenged with 10 7 of PbA-pRBCs to observe high peripheral parasite, while 0-cure mice were inoculated with 10 4 PbA-pRBCs as the controls. The whole set of experiments were performed twice.

Histopathological experiments
All animals were monitored closely for cerebral symptoms. Dead mice of 0-, 1-, 2-cure and sacrificed 3-cure mice were used for histopathological analysis. Brains were carefully removed after death and fixed by immersion in 10% paraformaldehyde in phosphate-buffered saline for at least 7 days. After fixation, preserved brain was transversally sectioned. These specimens were suspended in absolute alcohol, absolute xylene overnight, and then embedded on paraffin for one hour. Sections were cut to 5 mm thickness in four different positions: frontal lobe, temporal lobe, parietal lobe and occipital lobe, then stained with hematoxylineosin (HE) and analyzed by light microscopy. For each brain sample, all microscopic fields at 1006were examined for any histopathological changes including: blocked vessels, hemorrhages, edema perivascular cuffing and disruption and damage to cerebral vasculature endothelial linings [94][95][96][97][98].

Plasma harvesting
On day 0, 3, 5 and 7 after challenge infection, four mice in each group of 0-cure, 1-cure, 2-cure and 3-cure were euthanized to collect blood samples via sterile cardiac puncture into tubes containing 5 UI of heparin as anticoagulant. Plasma was harvested after centrifugation of the whole blood. All plasma samples were stored in 280uC until further use for antibody and cytokine assays. Three-cure mice infected on day 7 after challenge infection and naïve mice were also euthanized to collect plasma for passive transfer experiment.

Measurement of plasma PbA-specific IgG and its subtype levels
The plasma levels of parasite-specific IgG and its isotype antibodies on day 0, 3, 5 and 7 PI of 0-cure, 1-cure, 2-cure and 3cure B6 mice were estimated by ELISA as described previously [98]. Plasma was serially added in a dilution of 1/50-1/1,600 and a dilution was then chosen whereby all samples fell in the linear range of the curve; for all IgGs 1/100, dilution was applied. Briefly, the PbA crude antigen was used at a protein concentration of 0.03 mg in 100 ml (0.3 mg/ml) of coating buffer (pH 9.6) per well to coat polystyrene plates (Lot number 091611, Nunc, Copenhagen, Denmark) and kept at 4uC overnight. The plates were washed thrice with 0.05% Tween-20-PBS. Optimum blocking conditions for non-specific binding was achieved using 200 ml per well of 0.1% blocking buffer (0.1% Tween-20/PBS, pH 7.2), and incubated for 1 hour at 37uC. Plates were washed thrice with PBS containing 0.05% Tween-20. The antigen in coated plates was then reacted with the plasma samples obtained from 0-cure, 1cure, 2-cure and 3-cure mice at 1/100 dilutions, in duplicates. After two hours incubation at 37uC, plates were washed four times with 0.05% Tween-20/PBS. Later, 100 ml of diluted horse radish peroxidase (HRP)-conjugated goat anti-mouse IgG, IgG1 and IgG2a (Santa Cruz Biotechnology) was added to each well and incubated for 1 hour at 37uC, then washed thrice with 0.05% Tween-20/PBS. For color development, 3, 39, 5, 59-tetramethylbenzidine (TMB, stabilized chromate, Catalogue number 72673125A, Nitrogen, CA, USA) was used and prepared according to the manufacturer's instructions. The reaction was then interrupted after 15 minutes by the addition of 50 ml 1N H2SO4. Optical density (OD) was determined at 450 nm using EIA-reader (Bio-Rad, Hercules, CA).

Plasma cytokine quantification
Determination of cytokine levels in the plasma of all mice was performed on day 0, 3, 5 and 7 PI. Cytokine (IFN-c, IL-4, IL-10, and TGF-b) levels in plasma of all mice were measured using ProcartaH Cytokine Assay Mouse Plex Kits (Affymetrix, Inc. Santa Clara, CA, USA) in duplicate, according to the manufacturer's instructions. The plate was read on a Luminex instrument, LABScan 100 (Luminex Corporation. Austin, USA). The detection threshold was 0.37 pg/ml. Data are presented as arithmetic means of duplicate values.

Cell adoptive transfer experiments
Zero-cure and 3-cure mice were euthanized on day 7 PI for isolating splenic CD19(+) cells. Cells were sorted by using a magnetically activated cell sorting system (autoMACS, Miltenyi Biotec). Initially their spleens were homogenized in collagenase, re-suspended in PBS (pH 7.4). After incubation with 25 ml of anti-CD11c microbeads (Miltenyi Biotec), 225 ml of cell suspension from one spleen was first sorted for negative selection. CD11c(2)cell fraction was secondly incubated with 25 ml of anti-CD19 microbeads (Miltenyi Biotec) and then sorted into CD19 positive and negative population. The enriched CD19(+) cell fraction were isolated at .90% purity. On average, 3610 7 and 1.5610 7 of CD19(+) cells from one spleen of day 7-infected 3cure and 0-cure mice, respectively, were harvested. Finally, 1.5610 7 (low dose) or 3610 7 (high dose) of CD19(+) cells of 0cure or 3-cure mouse, re-suspended in total 250 ml of PBS, were injected intravenously (i.v.) into one naive mouse 24 hours before infection with 10 4 of PbA-pRBCs. PBS treated mice were used as controls.

Statistical analysis
Differences in survival of semi-immune groups were analyzed using the log-rank test. Differences in parasitemia, antibody and cytokine levels were analyzed for statistical significance using the nonparametric Mann-Whitney U test. Values of parasitemia, antibodies and cytokines indicated in the text represent the median and inter-quartile range of the data. For all statistical tests, p,0.05 was considered as significance. The results of data were analyzed by GraphPad prism software (version 5.0).