Suppression of Lethal Plasmodium yoelii Malaria following Protective Immunization Requires Antibody-, IL-4-, and IFN-γ-Dependent Responses Induced by Vaccination and/or Challenge Infection¹

There are 300–500 million new cases of malaria per year that result in 1–3 million deaths, predominantly in children under the age of five (1). In humans, malaria is caused by one of four protozoan parasites with Plasmodium falciparum causing the most severe disease. Clinical disease occurs when parasites invade and replicate within host RBCs and can range from flu-like symptoms to life-threatening anemia, cerebral malaria, and/or organ failure (2). Prior studies indicate that repeated infection of individuals living in endemic areas induces acquired immune responses that limit malaria morbidity and mortality (3, 4). However, there has been limited success in the development of a blood-stage vaccine that effectively reduces parasite burden and/or limits disease severity.

Merozoite surface protein-1 (MSP-1)³ and apical membrane Ag-1 (AMA-1) are the two leading blood-stage vaccine candidate Ags (5, 6). Both MSP-1 and AMA-1 are essential for parasite growth, are expressed on the surface of merozoites, the invasive form of blood-stage parasites, and are necessary for parasite invasion of host erythrocytes (7, 8, 9). Through in vitro studies of P. falciparum and in vivo studies using animal models of malaria, it has been established that Abs against conformational epitopes of MSP-1 (10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20) and AMA-1 (19, 20, 21, 22, 23, 24) can neutralize merozoites and provide significant protection. Even so, it has been difficult to define and measure vaccine-induced Ab responses that correlate with protection against blood-stage parasites in vivo. The magnitude of the AMA-1 and MSP-1 vaccine-induced Ab response is clearly important, but higher titers of Ag-specific Abs are not always predictive of increased protection against homologous parasite challenge. To correlate vaccine-induced responses with protective efficacy, the use of a functional assay that measures Ab-mediated inhibition of P. falciparum growth in vitro may be useful (15).

In attempting to define immune correlates of protection in these vaccine trials, it is also important to realize that blood-stage vaccines are not likely to prevent infection of individuals living in malaria endemic areas. In extensive MSP-1 and AMA-1 preclinical studies in Plasmodium yoelii (17, 18) and Plasmodium chabaudi (19, 20, 21, 22) rodent models and of P. falciparum in Aotus monkeys (11, 12, 14, 15, 23), immunized and protected animals typically get infected but suppress blood parasitemia more quickly following challenge infection. Data from both active and passive immunization studies involving MSP-1 indicate that an active immune response following challenge infection is still necessary for the complete control of parasite growth (17, 18). In addition to Ab responses to other blood-stage Ags, these infection-induced responses may also include Ab-independent, CD4⁺ T cell-mediated protective responses (25, 26, 27).

Studies of the immune responses of naive animals to blood-stage malaria parasites indicate that the host response varies depending on the strain of parasite and genetic background of the host (28, 29, 30, 31, 32, 33). In particular, it has been shown that the magnitude, kinetics of production, and overall balance of proinflammatory and anti-inflammatory cytokines are critically important to infection outcome (34, 35, 36, 37). Several studies demonstrate the involvement of the Th1 associated cytokine, IFN-γ, in both protection and pathology during blood-stage malaria (37). Limited data are available on the impact of these infection-induced immune responses on suppression of parasitemia in an animal that has been immunized to induce the production of high titers of Abs, a typically Th2-associated response. Most malaria vaccine studies in animal models and in human clinical trials focus only on the evaluation of immunization-induced immune responses present before challenge and do not adequately consider the host’s response to infection. Is it possible that two individuals with similar profiles of Ab responses elicited by MSP-1 or AMA-1 immunization are not equally protected against malaria because they respond differently to infection? If the answer is yes, we can expect that problems in defining predictive correlates of vaccine-induced immunity will continue.

MSP-8 is another blood-stage Ag that is structurally similar to MSP-1, as both possess two related C-terminal epidermal growth factor-like domains (38, 39). In previous studies using the P. yoelii model, we showed that immunization with recombinant Plasmodium yoelii MSP-8 (rPyMSP-8) protects mice against lethal malaria (38, 40). As with MSP-1, vaccine-induced protection primarily depends on the production of Abs that recognize conformation-dependent epitopes of PyMSP-8. In contrast to naive control mice, rPyMSP-8-immunized and -challenged animals show an increase in prepatent period but do typically develop parasitemia that is eventually suppressed late in infection. In the present study, we have used this model of rPyMSP-8 immunization followed by P. yoelii challenge to evaluate the importance of IL-4 and IFN-γ in 1) the production of protective PyMSP-8-specific Abs elicited by immunization and 2) the host response to challenge infection that is necessary to ultimately control P. yoelii 17XL parasitemia. Combined, the data obtained from studies of C57BL/6 and BALB/c strains of mice highlight the importance of IFN-γ in both vaccine-induced and infection-induced responses in mice immunized and protected against lethal P. yoelii malaria. Importantly, the data indicate that the host response to infection has a significant impact on the measurement of vaccine efficacy and needs to be considered when defining immune correlates of protection.

Five-week-old male C57BL/6 and BALB/c mice, as well as IL-4^−/− and IFN-γ^−/− mice on a C57BL/6 or BALB/c genetic background, were purchased from The Jackson Laboratory. All animals were housed in the Animal Care Facility of Drexel University College of Medicine under specific pathogen-free conditions. The lethal strain of Plasmodium yoelii 17XL was originally obtained from William P. Weidanz (University of Wisconsin, Madison, WI) and maintained as cryopreserved stabilates. All animal studies were reviewed and approved by the Institutional Animal Care and Use Committee of Drexel University College of Medicine.

Production and purification of rPyMSP-8 have been previously described in detail (38, 40). Mice (4–5 per group) were immunized subcutaneously with 5 μg of purified rPyMSP-8 with 25 μg of Quil A as adjuvant (Accurate Chemical & Scientific Corporation). Control animals were immunized with Quil A alone. Mice were boosted twice at 3-wk intervals with the same doses of Ag and adjuvant. Approximately 10–12 days following the third immunization, a small volume of serum was collected from each animal. When larger volumes of serum were needed, two sets of animals were immunized: one for the collection of serum and the second set for efficacy studies. Two weeks following the final immunization, mice were infected with P. yoelii 17XL parasitized RBC (pRBCs). To maintain a comparable disease course in the two strains of mice, infections were initiated by i.p. injection of 1 × 10⁵ or 1 × 10⁶ P. yoelii 17XL pRBCs for BALB/c and C57BL/6 mice, respectively. Resulting parasitemia was monitored by determining the percentage of pRBCs in thin tail-blood smears stained with Giemsa. In compliance with Institutional Animal Care and Use recommendations, mice were removed from the study and euthanized when parasitemia exceeded 50% or when animals became moribund. Such infections were scored as lethal.

C57BL/6 and BALB/c mice were immunized with rPyMSP-8 and Quil A as adjuvant or with Quil A alone as described above. Two weeks after the final immunization, mice were exsanguinated and sera from each group were collected and pooled. For passive protection assays, naive C57BL/6 and BALB/c mice received i.p. injections of 0.45 ml of immune or control sera on days −1, 0, and +1 relative to challenge infection. Blood-stage infection was initiated by i.v. injection of 1 × 10⁴ P. yoelii 17XL parasitized erythrocytes. Parasitemia and mice were monitored as described above.

PyMSP-8-specific Abs present in prechallenge sera were quantified by ELISA. To determine the level of Abs against conformational and linear epitopes, the reactivity of each serum with refolded rPyMSP-8 and reduced and alkylated (R/A) rPyMSP-8 was compared as described previously (40). Equivalent binding of refolded and R/A rPyMSP-8 to ELISA plates (0.25 μg/well) was monitored using a mAb that recognized the N-terminal His₆ tag (Novagen). Ag-coated wells were washed and blocked for 1 h with TBS (25 mM Tris-HCl (pH 8) and 150 mM NaCl) containing 5% nonfat dry milk. The reactivity of each serum, serially diluted (1/32,000 to 1/2,048,000) in TBS containing 0.1% Tween 20 and 1% BSA, was determined. Bound Abs were detected using HRP-conjugated rabbit Ab specific for mouse IgG (H+L chain) (Zymed Laboratories) and ABTS as substrate. Serum samples were run in duplicate and absorbance read at 405 nm. Values obtained with adjuvant control sera were comparable to background values obtained with normal mouse sera and have been subtracted. A pool of high titer sera obtained from rPyMSP-8-immunized mice was included on each assay as an internal reference to normalize the data between assays.

The isotypic profile of Ag-specific Abs present in prechallenge sera was determined using wells coated with rPyMSP-8 (0.25 μg/well) (20). Serum from each animal was assayed over a range of dilutions (1/1,000 to 1/1,024,000) as appropriate for each isotype. Ag-specific Abs were detected using HRP-conjugated rabbit Ab specific for mouse IgG1, IgG2a, or IgG2b (Zymed Laboratories) or HRP-conjugated goat anti-mouse IgG2c (IgG2a b allotype) or IgG3 (Southern Biotechnology Associates) and ABTS as substrate. Serum samples were run in duplicate and absorbance read at 405 nm. In each assay, wells coated with purified IgG1, IgG2a, or IgG2b myeloma proteins (Zymed Laboratories) (16 ng/ml to 1 μg/ml) were used to generate a standard curve. The IgG2c and IgG3 standard curves were generated using a purified monoclonal IgG2c (BD Biosciences) or IgG3 Ab. The isotype standard curves were used to quantify Ag-specific Ab present in each serum sample, using the highest dilution of serum that yielded an OD at 405 nm between 0.1 and 1.0. The concentration of IgG isotype was expressed in U/ml where 1 U/ml is equal to 1 μg/ml myeloma standard. The mean absorbance of sera from adjuvant control mice was subtracted as background.

Serum IFN-γ and TGF-β were measured by sandwich ELISA with paired capture and detection Abs from the mouse IFN-γ or human TGF-β1 BD OptEIA sets (BD Biosciences). Serum from each animal was assayed in duplicate at dilutions of 1/10, 1/50, and 1/100 according to the manufacturer’s instructions. A standard curve using recombinant mouse IFN-γ or human TGF-β1 was used to quantify the amount of cytokine present in serum with a lower limit of detection of ∼10 pg/ml or ∼200 pg/ml, respectively. Values obtained from the assay of normal mouse sera were subtracted as background.

Groups of BALB/c mice were immunized as described above with rPyMSP-8 formulated with Quil A as adjuvant or with Quil A alone. Following challenge with 1 × 10⁵ P. yoelii 17XL parasitized erythrocytes, immunized and control mice (n = 4) were sacrificed every other day of infection, and total splenic RNA was isolated. Spleens were harvested and homogenized into single cell suspensions using 70 μm nylon filters. RBCs were removed by hypotonic lysis, and splenocytes were washed, pelleted, and resuspended in the TRIzol reagent (Invitrogen Life Technologies). Total RNA was extracted, precipitated, and purified using an RNeasy RNA isolation kit (Qiagen) according to the manufacturer’s protocol. For baseline comparisons, total RNA was isolated from the spleens of a set of uninfected BALB/c mice.

Real-time PCR was used to quantify the levels of IL-4 and IFN-γ mRNA present in the spleens of rPyMSP-8-immunized BALB/c mice following P. yoelii 17XL challenge. Customized primer sets were generated for each gene of interest using Primer3 software (primer3_www.cgi v 0.2; Whitehead Institute for Biomedical Research). Primer sequences were as follows: IL-4 (forward) 5′-TCACAGCAACGAAGAACACC-3′, IL-4 (reverse) 5′-TGCAGCTCCATGAGAACACT-3′; IFN-γ (forward) 5′-GCTTTAACAGCAGGCCAGAC-3′, IFN-γ (reverse) 5′-GGAAGCACCAGGTGTCAAGT-3′; and β-actin (forward) 5′-GCGCAAGTACTCTGTGTGGA-3′, β-actin (reverse) 5′-CATCGTACTCCTGCTTGCTG-3′. Reverse transcription (RT) of 2 μg of total RNA was conducted with an oligo(dT)₁₅ primer (Promega) and the Omniscript Reverse Transcription kit (Qiagen), and cDNA was amplified using the SYBR Green PCR Master mix (Applied Biosystems). All reactions were run in duplicate using an Applied Biosystems 7300 Real-Time PCR system. Data were analyzed with the Applied Biosystems Sequence Detection software (v.1.2.3). Serial dilutions of input cDNA (0.5–10 ng/well) were used to generate standard curves. β-actin was used as an endogenous reference. The relative expression level of each target gene normalized to β-actin was determined based on the threshold cycle of product detection. The fold-change in the expression of each gene from rPyMSP-8-immunized mice and adjuvant controls infected with P. yoelii 17XL was calculated relative to their expression in the spleens of uninfected mice.

The statistical significance of differences in Ab responses and in mean peak parasitemia between groups was calculated by ANOVA. The significance of differences in the number of surviving animals between groups was determined by the Mantel-Haenszel log rank test (GraphPad Prism 4.0).

In an effort to define immune correlates of vaccine-induced immunity and the pathways that lead to the production of protective Abs, C57BL/6 (Th1 responder) and BALB/c (Th2 responder) mice were immunized three times with rPyMSP-8. Following the third immunization and before challenge infection, the quantity and isotype profile of Abs induced by rPyMSP-8 immunization were measured by ELISA. As shown in Fig. 1,A, both C57BL/6 and BALB/c mice responded well to immunization with rPyMSP-8 with the production of high levels of Ag-specific IgG. There was no significant difference in the quantity of Ag-specific IgG in the serum of these two strains of mice (p > 0.05). However, analysis of the IgG subclasses of PyMSP-8-specific Abs present in prechallenge sera revealed differences between the two strains (Fig. 1 A). C57BL/6 mice produced a significantly higher level of PyMSP-8-specific Abs of the IgG2b isotype relative to IgG1 (p < 0.01) or IgG2a/c (p < 0.05). In contrast, BALB/c mice produced high levels of PyMSP-8-specific Abs of the IgG1 isotype, significantly greater than IgG2b (p < 0.01) or IgG2a/c (p < 0.01). Extremely small quantities of Ag-specific IgG3 were observed in both strains of mice. Between strains, Ag-specific IgG1 levels were significantly higher in BALB/c relative to C57BL/6 mice (p < 0.01). Compensatory higher levels of IgG2b were present in C57BL/6 mice compared with BALB/c mice (p < 0.01).

PyMSP-8-induced protection is B cell dependent and requires the production of Abs that recognize conformation-dependent epitopes (40). As such, differences in the level of Abs in the serum of immunized C57BL/6 and BALB/c mice that recognize protective epitopes of rPyMSP-8 were also measured. The reactivity of prechallenge, immune sera with refolded rPyMSP-8 or R/A rPyMSP-8 was compared by ELISA. As above, C57BL/6 and BALB/c mice produced high and comparable levels of IgG that were reactive with refolded rPyMSP-8. However, a greater proportion of IgG produced in C57BL/6 mice also reacted with R/A rPyMSP-8 (Fig. 1 B). The ratio of reactivity to R/A rPyMSP-8 over refolded rPyMSP-8 was 0.56 ± 0.12 for C57BL/6 sera, significantly higher than that of 0.13 ± 0.04 for sera from BALB/c mice (p < 0.01). In contrast to C57BL/6 mice, immunization of BALB/c mice with rPyMSP-8 elicited the production of Abs that primarily reacted with nonlinear, disulfide-dependent epitopes of rPyMSP-8. Combined, these data indicate that C57BL/6 and BALB/c mice produced similar levels of PyMSP-8-specific IgG in response to immunization, but these Abs differed significantly with respect to isotype profile and epitope specificity.

To determine whether the observed differences in vaccine-induced Ab response had an impact on protective efficacy, rPyMSP-8-immunized C57BL/6 and BALB/c mice were challenged with P. yoelii 17XL parasites. As shown in Fig. 2, all C57BL/6 and BALB/c mice immunized with adjuvant alone were unable to control parasite burden and succumbed to infection by day 13 or 10, respectively. With the exception of one animal, rPyMSP-8-immunized C57BL/6 mice suppressed an otherwise lethal infection with clearance of blood-stage parasitemia by day 20 postchallenge. Peak parasitemia in rPyMSP-8-immunized C57BL/6 mice occurred between days 10 and 14 of infection with a mean peak of 17.2 ± 10.2%. The course of P. yoelii malaria in rPyMSP-8-immunized BALB/c mice was not significantly different. All rPyMSP-8-immunized BALB/c mice were protected against lethal P. yoelii malaria. Blood parasitemia in rPyMSP-8-immunized BALB/c mice peaked between days 12 and 16 of infection with a mean of 16.7 ± 10.0%. These data indicate that, despite qualitative differences in the Ab response to rPyMSP-8 immunization, C57BL/6 and BALB/c mice are similarly protected against lethal P. yoelii 17XL malaria.

To focus primarily on the contribution of Abs induced by immunization with rPyMSP-8 to subsequent protection against P. yoelii 17XL challenge, a set of passive protection assays were completed. Sera from rPyMSP-8-immunized C57BL/6 or BALB/c mice or the corresponding adjuvant controls were passively transferred to naive C57BL/6 and BALB/c recipients. Passively immunized mice were then challenged with P. yoelii 17XL pRBCs and parasite replication in each group was compared. As shown in Table I, immune sera from C57BL/6 and BALB/c donor mice significantly delayed the onset of P. yoelii 17XL parasitemia when the recipients were naive C57BL/6 mice as indicated by >20-fold reduction in parasitemia on day 8 of infection relative to recipients of control sera. C57BL/6 mice passively immunized with rPyMSP-8-specific sera from C57BL/6 and BALB/c donors all cleared their infection with mean peak parasitemia of 10.9 ± 6.9% and 3.2 ± 2.5%, respectively. C57BL/6 mice passively immunized with sera from Quil A-immunized donors developed fulminant P. yoelii malaria resulting in 60–100% mortality (Fig. 3, A and B). In this set of assays, no difference in the protective capacity of Abs obtained from rPyMSP-8-immunized C57BL/6 and BALB/c mice was apparent.

When rPyMSP-8-immune sera from C57BL/6 and BALB/c donor mice were passively transferred to naive BALB/c recipients, the outcome following P. yoelii 17XL challenge differed. As shown in Table I, passive transfer of immune sera from BALB/c donors to naive BALB/c recipients delayed the onset of P. yoelii 17XL parasitemia, with a 4-fold reduction in parasitemia observed on day 8 of infection relative to controls. In contrast, there was no alteration in P. yoelii 17XL growth in naive BALB/c mice passively immunized with rPyMSP-8 sera obtained from C57BL/6 donors. These data suggest that the differences in isotype and/or epitope specificity in rPyMSP-8 Abs present in sera from C57BL/6 and BALB/c mice may affect protective capacity, but this depends on the genetic background of the recipient strain. In addition and unlike C57BL/6 recipients, all BALB/c mice passively immunized with rPyMSP-8-immune sera eventually succumbed to P. yoelii infection (Fig. 3, C and D). These data suggest that in actively immunized BALB/c mice, exposure to P. yoelii 17XL blood-stage parasites is necessary to maintain production of PyMSP-8-specific Abs and/or to elicit additional protective immune responses that ultimately suppress parasite growth.

To focus primarily on the contribution of key cytokines to protection in mice actively immunized with rPyMSP-8-, IL-4-, or IFN-γ-deficient C57BL/6 mice were immunized and challenged. In immunized IL-4^−/− C57BL/6 mice, the PyMSP-8-specific Ab response was affected resulting in ∼17-fold (p < 0.05) decrease in PyMSP-8-specific IgG1 as compared with immunized, intact mice (Fig. 4,A). Likewise, the IgG2a/c response in IFN-γ^−/− C57BL/6 mice was severely impaired with at least a 100-fold reduction (p < 0.05) in the production of specific IgG2a/c following rPyMSP-8 immunization. In both strains of cytokine-deficient mice, compensatory increases in the production of Ag-specific Abs of other IgG isotypes above levels observed in similarly immunized, intact C57BL/6 mice did not occur (Fig. 4,A). Upon challenge with P. yoelii 17XL, intact C57BL/6 mice immunized with rPyMSP-8 showed an early control of parasite replication, with a mean day 8 parasitemia of 12.2 ± 10.2% relative to 25.8 ± 10.2% in adjuvant controls (Table II). Unlike adjuvant control mice which succumbed, 4 of 5 C57BL/6 mice immunized with rPyMSP-8 ultimately controlled their P. yoelii 17XL infection (Fig. 4,B). In contrast, no delay in ascending P. yoelii 17XL parasitemia was observed in IL-4^−/− C57BL/6 mice (day 8, 22.7 ± 11.3%) or IFN-γ^−/− C57BL/6 mice (day 8, 30.8 ± 0.8%) that had been immunized with rPyMSP-8 (Table II). Likewise, overall survival of rPyMSP-8-immunized IL-4^−/− or IFN-γ^−/− C57BL/6 mice following P. yoelii 17XL challenge was not significantly different from respective adjuvant control mice (Fig. 4, C and D). These data suggest that in the absence of IL-4 or IFN-γ, the immune response elicited by rPyMSP-8 immunization in C57BL/6 mice is impaired and insufficient to significantly suppress P. yoelii 17XL growth and reduce mortality.

Actively immunized BALB/c mice primarily produce PyMSP-8-specific IgG1 Abs, characteristic of an IL-4, Th2 associated activation pathway. Accordingly, immunization of IL-4^−/− BALB/c mice with rPyMSP-8 resulted in a dramatic, ∼20-fold decrease in the production of Ag-specific IgG1 Abs relative to similarly immunized, intact BALB/c mice (p < 0.01) (Fig. 5,A). Due to a compensatory increase in the production of PyMSP-8-specific IgG2a/c and IgG2b, however, total, Ag-specific IgG was not significantly reduced in rPyMSP-8-immunized IL-4^−/− BALB/c mice. Upon challenge with P. yoelii 17XL parasites, all rPyMSP-8-immunized wild type and IL-4^−/− BALB/c mice controlled their infection with parasite clearance occurring by day 20 postchallenge (Table II, Fig. 5, B and C). Mean peak parasitemia in rPyMSP-8 immunized IL-4^−/− mice was 11.4 ± 6.9%, comparable to that of 19.0 ± 8.5% in immunized, intact BALB/c mice. P. yoelii 17XL parasitemia was fulminant and uniformly lethal in mice immunized with adjuvant alone. These data clearly indicate that in BALB/c mice, IL-4 is not necessary for the protection induced by rPyMSP-8 immunization. Furthermore, the responses that are elicited in immunized BALB/c mice upon P. yoelii 17XL challenge and required for final parasite clearance are IL-4 and IgG1 independent.

To assess the contribution of Th1 associated responses to vaccine-induced protection, a parallel set of immunization and challenge studies was completed in IFN-γ^−/− BALB/c mice. As shown in Fig. 5,A, the lack of IFN-γ during rPyMSP-8 immunization resulted in a marked decrease in the production of Ag-specific IgG2a/c (p < 0.01) and a boost in the production of IgG1 relative to similarly immunized, intact BALB/c mice. Again, total Ag-specific IgG levels were not significantly different between rPyMSP-8 immunized intact and IFN-γ^−/− BALB/c mice. Upon challenge infection, P. yoelii 17XL growth was initially suppressed as indicated by a mean parasitemia of 9.1 ± 5.4% on day 8 in rPyMSP-8-immunized, IFN-γ^−/− mice compared with 37.0 ± 11.8% in IFN-γ^−/− adjuvant control mice (Table II). All adjuvant controls succumbed to P. yoelii 17XL malaria by day 15. Following the initial period of control, P. yoelii parasitemia in rPyMSP-8-immunized IFN-γ^−/− mice gradually increased with 7 of 10 animals eventually succumbing during the 3^rd week of infection (Fig. 5 D). Combined, these data suggest that unlike C57BL/6 mice, IFN-γ is not critical for the early protection induced by rPyMSP-8 immunization in BALB/c mice. However, without the ability to produce IFN-γ upon subsequent P. yoelii 17XL challenge, rPyMSP-8-immunized BALB/c cannot mount and/or sustain a protective immune response that suppresses parasite growth.

Given the protective role of immune responses elicited during the course of P. yoelii 17XL malaria in rPyMSP-8-immunized BALB/c mice, the magnitude and kinetics of the production of several cytokines of interest, including IFN-γ, were directly measured. Total RNA was isolated from the spleens of both rPyMSP-8-immunized and adjuvant control mice on alternate days following P. yoelii 17XL challenge. Changes in the level of splenic cytokine mRNA were measured by quantitative RT-PCR. Increased expression of both IL-4 and IFN-γ was detected early during P. yoelii malaria in rPyMSP-8-immunized BALB/c mice. As shown in Fig. 6, B and C, 6-fold and 10-fold higher levels of IL-4 and IFN-γ mRNA, respectively, were present on day 6 of infection relative to uninfected controls. As shown in Fig. 6,D, this increase in splenic mRNA was also reflected in significantly higher levels of serum IFN-γ at a comparable time point following challenge (p < 0.01). These early increases in IL-4 and IFN-γ production occurred in both rPyMSP-8-immunized and adjuvant control mice infected with P. yoelii 17XL. As adjuvant control mice succumbed to infection by day 8, the immune responses associated with these increases in cytokine production were not adequate to control infection in the absence of rPyMSP-8 vaccine-induced Abs. In rPyMSP-8-immunized and -protected mice, a second, modest increase in both IL-4 and IFN-γ production was noted between days 14 and 18 of P. yoelii infection (Fig. 6, B and C). This second wave of cytokine production coincided with the period of descending P. yoelii 17XL parasitemia in rPyMSP-8-immunized, intact BALB/c mice (Fig. 6,A) but sustained parasite growth in immunized, IFN-γ^−/− BALB/c mice (Fig. 5 D). These data demonstrate that IFN-γ, which is an essential component of the immune response in rPyMSP-8-immunized mice, is produced upon exposure to blood-stage parasites at two distinct times during P. yoelii 17XL malaria.

Given the pattern in the regulation of IFN-γ production during the course of infection in rPyMSP-8-immunized BALB/c mice, the production of the anti-inflammatory cytokines TGF-β and IL-10 was also measured. IL-10 was not detected in the serum of either rPyMSP-8-immunized or adjuvant control mice during the course of P. yoelii 17XL malaria (data not shown). However, the decline in the early peak of IFN-γ production coincided with an increase in serum TGF-β levels. As shown in Fig. 6,E, marked increases in serum TGF-β were noted on days 6–8 of infection in rPyMSP-8-immunized BALB/c mice. Elevated serum TGF-β levels were maintained for the remaining 10 days of P. yoelii 17XL malaria. Interestingly, the initial increase in serum TGF-β correlated with both the decrease in IFN-γ production and the rise in P. yoelii 17XL blood parasitemia in rPyMSP-8-immunized mice (Fig. 6).

One of the many challenges for the malaria vaccine development effort has been to define measurable parameters of immune responses that correlate with protection against infection and/or disease. The lack of such correlates has made it difficult to critically evaluate responses induced in humans during vaccine trials and to use these data to guide the effort to optimize vaccine formulations. For blood-stage vaccines based on merozoite surface Ags, including MSP-1 and AMA-1, it has been established that high titers of vaccine-induced Abs are important for protection (10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24). With current Ag/adjuvant formulations, however, it has been difficult to induce robust Ab responses in human subjects (41, 42, 43, 44, 45). Importantly, ELISA based assays of Ab concentrations and functional assays measuring growth inhibition of P. falciparum in vitro have only partially correlated with protective efficacy (15). Often missing from vaccine efficacy studies is an evaluation of any positive or negative effects of responses induced by infection in an immunized and challenged subject.

We used rPyMSP-8 as a model MSP vaccine to dissect the immunization- and infection-induced responses that collectively contribute to protection against lethal P. yoelii malaria. We know that like AMA-1 and MSP-1, rPyMSP-8 vaccine-induced immunity is B cell dependent and requires the production of Abs against conformational determinants (40). In an attempt to further define immune correlates of protection, we completed an initial set of immunization and challenge studies using C57BL/6 and BALB/c mice, two strains that tend to develop somewhat polarized Th1- and Th2-type responses to protein immunization and/or pathogen exposure (46, 47, 48). Although we detected differences in the isotypic profile and epitope specificity of Ag-specific Abs, both strains of mice were similarly protected against P. yoelii 17XL challenge. As such, we could not correlate any measurable difference in vaccine-induced responses with protection. Our initial assumption was that the protective responses elicited in C57BL/6 and BALB/c mice were similar and not significantly influenced by host genetic background.

Studies of passively immunized C57BL/6 mice were consistent with our initial assumption. When rPyMSP-8-immune sera from either C57BL/6 or BALB/c mice were passively transferred to C57BL/6 mice, recipients were similarly protected against P. yoelii 17XL challenge and controlled their infections. Protection was not influenced by differences in the isotypes of rPyMSP-8-specific Abs present in C57BL/6 sera vs BALB/c sera. In actively immunized IL-4^−/− C57BL/6 mice, there was a marked reduction in IgG1 responses and a loss in protection against P. yoelii 17XL malaria. Somewhat to our surprise, a similar result was obtained when IFN-γ^−/− C57BL/6 mice were immunized and challenged. In the absence of IFN-γ, the IgG2a/c response elicited by rPyMSP-8 immunization was impaired and protection against P. yoelii 17XL challenge abrogated. Combined, these data from active and passive immunization studies in C57BL/6 mice indicated 1) that the magnitude of the vaccine-induced IgG response, but not a particular Ab isotype, was important for protection and 2) that adequate protection could only be induced by immunization when IL-4 and IFN-γ were both present.

Our inability to correlate PyMSP-8-induced protection with the isotype of Ag-specific Abs is consistent with several active and passive immunization studies in the P. yoelii and P. chabaudi models involving MSP-1 and AMA-1 (19, 20, 49, 50). The conclusion from these studies was that Ab-mediated neutralization of extracellular merozoites was IgG isotype and Fc receptor independent. In contrast, cytophilic IgG1 and IgG3 Abs have been repeatedly correlated with protection against P. falciparum malaria in human subjects (51, 52, 53). Recently, passive protection mediated by a human IgG1 Ab specific for P. falciparum MSP-1₁₉ was shown to require the expression of FcγRI, presumably on macrophages, in a model system that combined Fc receptor transgenic mice and P. falciparum-Plasmodium berghei chimeric parasites (54). The corresponding cytophilic isotype in mice, IgG2a/c, has also been associated with protection in studies of vaccine-induced and infection-induced responses in the P. yoelii (55, 56, 57, 58) and P. chabaudi (58, 59) models. Differences in methodology and/or Ab-mediated effector mechanisms active in human subjects vs mice may in part explain the apparent discrepancy in these studies. We must also consider that similar Fc receptor-mediated effector mechanisms can be linked to two different IgG isotypes (i.e., IgG1 and IgG2a/c in mice) but are active at dramatically different IgG concentrations due to differences in the affinity of each isotype for the set of activating and inhibitory Fc receptors engaged (60). It is also possible that multiple effector mechanisms associated with a polyclonal Ab response can contribute to protection, some of which require Fc receptors for IgG and/or the activation of complement. For the immunization of human subjects with MSP-based vaccines, it will be important to determine which mechanisms are most effective at a concentration of Ab that can be readily induced by subunit vaccines and whether such mechanisms can be specifically targeted by immunization.

In parallel with our work in C57BL/6 mice, we completed an additional series of studies in BALB/c mice that further highlighted the challenge in defining protective immune mechanisms associated with vaccine efficacy when host genetic background differs. First, IL-4^−/− BALB/c mice immunized with rPyMSP-8 were protected against P. yoelii 17XL to the same degree as immunologically intact BALB/c mice. A dramatic reduction in the high PyMSP-8-specific IgG1 titers in the absence of IL-4 was balanced by an increase in production of IgG2a/c and IgG2b Abs that was adequate to maintain vaccine efficacy. Second, immunization of IFN-γ^−/− BALB/c mice resulted in a reduction in IgG2a/c and IgG2b levels that was balanced by a boost in the IgG1 response. rPyMSP-8 immunized IFN-γ^−/− BALB/c mice showed an initial delay in the onset of parasitemia upon P. yoelii 17XL challenge comparable to immunized, intact BALB/c mice. In agreement with our findings in C57BL/6 mice, these data further support the conclusion that the magnitude, not the isotype of the vaccine-induced Ab response, is a key determinant of protection. In contrast to C57BL/6 mice, however, an adequate response can be induced in BALB/c mice in the absence of either IL-4 or IFN-γ. Finally, the initial suppression of P. yoelii 17XL parasitemia in rPyMSP-8-immunized IFN-γ^−/− BALB/c mice was followed by sustained parasite growth and decreased survival. This reflected what we also observed in BALB/c mice passively immunized with rPyMSP-8-immune sera from BALB/c donors when parasite growth was initially controlled but animals eventually succumbed to infection. Even in the presence of protective levels of rPyMSP-8-specific Abs at the time of challenge, suppression of P. yoelii 17XL parasitemia required an immune response elicited upon infection that was dependent on IFN-γ.

In naive mice, IFN-γ has been shown to be essential for immunity to blood-stage P. yoelii (31, 32, 36, 61) and P. chabaudi (61, 62, 63) parasites. In P. yoelii, a lack of IFN-γ early during P. yoelii malaria has been correlated with disease progression and lethality (32, 36). In previous active and passive immunization studies, the need for an ongoing immune response upon infection in P. yoelii MSP-1 immunized mice was demonstrated (17, 18). A specific contribution for IFN-γ to protection in P. yoelii MSP-1-immunized mice has also been suggested. In a study by Matsumoto et al. (64), mice were immunized with a recombinant strain of Mycobacterium bovis bacillus Calmette-Guérin (rBCG) engineered to express and secrete the C-terminal 15 kDa fragment of P. yoelii MSP-1. Immunization of C3H/He mice with rBCG-PyMSP-1 resulted in protection against lethal P. yoelii malaria. Neutralization of IFN-γ by Ab treatment following P. yoelii 17XL challenge abrogated the ability of rBCG-PyMSP-1-immunized mice to control their infection. These previous studies with MSP-1 and our current data with MSP-8 emphasize the need to consider infection-induced production of IFN-γ in immunized subjects when attempting to define immune correlates of protection associated with MSP-based vaccines.

In intact, rPyMSP-8-immunized, BALB/c mice infected with P. yoelii 17XL, increases in the production of IFN-γ were readily detected by quantitative analysis of splenocyte gene expression and by the analysis of serum cytokine levels. The first wave of IFN-γ production occurred early during infection (days 6–8) in both immunized and adjuvant control mice suggesting that rPyMSP-8-specific T cells primed by immunization were not the IFN-γ source. The presence of IFN-γ alone at this time point was insufficient to control infection in adjuvant control mice. It is possible that such production of IFN-γ helped to sustain protective Ab levels and/or was necessary for responses specifically required for MSP-specific Abs to be effective. Alternatively, the critical role for this early IFN-γ production may be to enhance Ag processing and presentation by APCs and promote the initiation of protective immune responses to additional parasite Ags (33, 65). A second wave of IFN-γ production in rPyMSP-8 immunized mice was noted late during infection as parasitemia began to resolve. Interestingly, this second wave of IFN-γ occurred near the time of death of rPyMSP-8-immunized IFN-γ^−/− BALB/c mice. It will be necessary to determine whether this late IFN-γ production is required for Ab-mediated and/or cell-mediated effector functions needed to completely clear P. yoelii 17XL blood-stage parasites.

The positive influence of infection-induced IFN-γ on the measurement of rPyMSP-8 vaccine efficacy prompted us to consider other cytokines. Accumulating data in rodent models (36, 66) and in human subjects (34, 35, 67, 68) indicate that the interplay of proinflammatory and anti-inflammatory cytokines in a nonimmune host during blood-stage malaria contributes significantly to disease severity and infection outcome. In rPyMSP-8-immunized BALB/c mice challenged with P. yoelii 17XL, we did not detect significant production of TNF-α or IL-10 (data not shown). In contrast, there was a sharp increase in the production of TGF-β by day 6 of infection, and serum levels persisted through day 18. It was striking to see that this increase in TGF-β production coincided with the decline in the early wave of IFN-γ production and with the period of increased replication of P. yoelii 17XL blood-stage parasites. As such, it appears that the response of an immunized host exposed to blood-stage malaria parasites and the resulting balance in cytokines produced has a direct bearing on the measurement of MSP-based vaccine efficacy.

The primary goal of preclinical vaccine studies in rodent models is to acquire data that will provide guidance as clinical studies in human subjects are considered. We believe the data from our present study raise several concerns for MSP-based malaria vaccines developed to immunize a diverse population of human subjects. First, vaccine adjuvants must concurrently promote the induction of IgG responses by both IL-4- and IFN-γ-dependent pathways to achieve protective levels of Abs in the largest proportion of vaccinees. Second, most agree that innate and acquired immune responses to blood-stage malaria parasites differ from individual to individual according to age, genetic background, immune status, concurrent infections, transmission dynamics, etc. Such differences in response to infection will influence vaccine efficacy even if adequate MSP-specific Ab responses are induced by immunization. This must be considered in the effort to correlate immunization-induced responses with protection. The production of IFN-γ and TGF-β, two key cytokines, should be evaluated. In the end, it may be necessary to incorporate a significantly large number of parasite proteins into blood-stage vaccine formulations such that the response to infection in an immunized subject is dominated by a diverse set of protective responses primed by immunization and not by those induced by blood-stage parasites to favor their growth.

We thank Dr. William P. Weidanz, University of Wisconsin-Madison, for helpful discussions and critical review of this manuscript.

The authors have no financial conflict of interest.