TEXT

Top Abstract Text References

Despite the significant progress that has been made in the identification of vaccine candidates, an effective vaccine against human malaria has not yet been developed. Antigen selection and characterization have been hampered by the lack of a readily available challenge system for Plasmodium falciparum. Consequently, rodent models for malaria have attracted much attention and proved to be useful in assessing the antigenicity and immunogenicity of vaccine candidates. To date, the most effective vaccine components are proteins in exposed locations on the parasite, such as the merozoite surface, the rhoptries, and the surface of the infected red blood cell (3). Studies in the P. yoelii system showed that immunization with MSP1 was capable of triggering protective responses (14). Antibodies were suggested to play a major role in protection (7, 22), and they were directed mainly against the C-terminal portion of the MSP1 molecule, which contains two epidermal growth factor (EGF)-like domains that seem to be essential for protection (5, 13). Nevertheless, protection is limited to homologous challenge (17, 18).

Recently, two novel antigens, each containing a single EGF-like domain, were identified in P. falciparum: MSP4 (15, 24) and MSP5 (16). The syntenic region of the genome in rodent malaria species contains only a single gene with one EGF-like domain. This antigen, MSP4/5, has structural features in common with PfMSP4 and PfMSP5 but little sequence similarity outside the EGF-like domain (4, 11). We wanted to determine whether this protein may have efficacy in inducing protective immunity against lethal challenge using the P. yoelii model. In this study, we show that immunized mice are partially protected against malaria infection and that immunization induces high levels of antibody. Our findings also suggest that reduction and alkylation have only a small effect on the protective efficacy of recombinant PyMSP4/5.

The full-length PyMSP4/5 sequence lacking the predicted signal peptide and glucosylphosphatidylinositol (GPI) anchor was expressed as a His₆-tagged recombinant protein (PyMSP4/5-His) and purified on Talon Metal Affinity Resin (Clontech, Palo Alto, Calif.) as described previously (11). Groups of female BALB/c mice were immunized with 25 µg of either nonreduced (NR) or reduced and alkylated (RA) PyMSP4/5-His emulsified in complete Freund adjuvant (Difco Laboratories, Detroit, Mich.) administered intraperitoneally (i.p.). Two subsequent boosters of 25 µg of antigen emulsified in incomplete Freund adjuvant were delivered i.p. at monthly intervals. Control mice were injected with phosphate-buffered saline (PBS) emulsified in adjuvant. Sera were collected prior to the initial injection and 2 days before challenge. At 12 to 14 days after the second boost, mice were challenged i.p. with 10⁵ P. yoelii YM parasitized red blood cells (PRBC). Parasitemia was monitored microscopically by Giemsa-stained thin blood smears fixed with methanol. Blood for smears was collected each day from day 2 to day 30 postinfection. Indirect enzyme-linked immunosorbent assays (ELISAs) were performed for antibody determination as previously described (24). The optical density (OD) was read at 405 nm, and the background OD values from PBS-coated plates were subtracted from values obtained from antigen-coated plates.

Four separate vaccination trials were performed using the protocol described above. The results are summarized in Table 1. All mice in the control groups developed fulminating infections, and all except one mouse died on days 5 to 10, with a mean parasitemia of >80% (Table 1). In general, animals in the control groups had detectable levels of parasitemia on day 2 postchallenge, which then increased rapidly. The sole surviving mouse in the control group developed a peak parasitemia of 40%, which then cleared after 6 days (data not shown). The prechallenge antibody responses in the control groups showed no reactivity to PyMSP4/5-His when tested by ELISA (data not shown). The immunized mice as a whole showed clear evidence of induced protection, although the level of protection differed between individual animals. The peak parasitemia in immunized groups ranged from 0.2 to 75% in protected mice (i.e., mice able to survive challenge and clear the infection). Out of a total of 33 immunized mice, 5 mice died. Three of these mice developed fulminant infections similar to those observed in control animals, whereas the other two mice had parasitemia levels comparable to those of other immunized animals. However, after several days they succumbed to the infection with parasitemias of 28 and 50%, respectively. In these four experiments, none of the animals immunized with PyMSP4/5-His developed sterile immunity. The prepatent period in immunized groups varied from 2 days (similar to controls) to up to 15 days, and the clearance time ranged from 11 to 24 days postchallenge. The duration of infection varied from transient (3 days) to prolonged (23 days). Surviving animals from trials 1 and 4 were rechallenged 2 weeks after recovery with a higher dose of parasites. In the case of trial 1, all surviving mice were rechallenged with 10⁶ PRBC. Blood smears were examined for 13 days after rechallenge, but no patent parasitemia was observed (data not shown). Surviving mice from trial 4 were also rechallenged with 10⁶ PRBC, and blood samples were examined on day 7 postinjection; however, no parasites were detected. Whole blood from these mice (0.3 ml) was transferred to naive mice. These animals did not develop patent parasitemia either, indicating full recovery of PyMSP4/5-immunized mice after the initial infection and subsequent development of sterile immunity to malaria. The significance of the differences in the number of surviving mice in immunized and control groups was determined using Fisher's exact probability test. The P value obtained for all mice (33 immunized animals and 34 controls) was <0.0001. The Mann-Whitney test was used to determine significance in differences in peak parasitemias between the two groups, and the P value was <0.0005.

In two vaccination trials (trials 3 and 4; Table 1), groups of mice were immunized with RA PyMSP4/5-His recombinant protein to determine the effects of the disruption of disulphide bonds in the EGF-like domain of the molecule on protective efficacy. Purified PyMSP4/5-His was reduced and alkylated as previously described (2) and then extensively dialyzed against PBS. Western blot analysis was used to assess the efficiency of reduction and alkylation of the recombinant protein (Fig. 1). The anti-PyMSP4/5 antibodies detected high-molecular-mass complexes in the NR sample, which we believe are aggregates of PyMSP4/5-His molecules. In contrast, such multimers were not detected in the RA sample. Furthermore, there was a clear difference in mobility on sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gels between the monomer form of PyMSP4/5-His (which migrates at 36 kDa) in the NR sample compared to the RA protein. Differences in mobility on SDS-PAGE gels have been shown previously for reduced and alkylated MSP1 protein (14). This result suggests that PyMSP4/5-His was reduced successfully and did not refold. Reduction and alkylation did not abolish protection induced by PyMSP4/5-His, although there was some suggestion of reduced efficacy, as indicated by increased numbers of deaths (Table 1). Additionally, the peak parasitemia was elevated in mice given RA antigen (28.4 versus 19.5% in trial 3 and 56.7 versus 41% in trial 4). Nevertheless, the course of parasitemia in these mice was similar to that observed in mice immunized with intact PyMSP4/5-His, except that a longer period of time was needed to clear parasites from the blood. The statistical analysis showed no significant difference either in peak parasitemia or in survival numbers between groups of mice immunized with RA or NR protein material. However, there was a statistically significant difference between the control group and the group immunized with RA recombinant protein, when determined either by Fisher's exact test (P < 0.0003) or Mann-Whitney test (P = 0.0003). Furthermore, ELISA analysis of prechallenge antibody responses in the sera from immunized mice from both groups showed no differences when checked on either NR or RA target antigen. The OD₄₀₅ values were analyzed using the Mann-Whitney test, but the P value obtained was not statistically significant.

View larger version (51K): [in this window] [in a new window]   FIG. 1.   Immunoblot of recombinant PyMSP4/5 with anti-PyMSP4/5 sera (1:5,000 dilution). Lane 1, RA PyMSP4/5; lane 2, NR PyMSP4/5. The monomeric form of PyMSP4/5-His migrates with a molecular mass of 36 kDa (marked by an arrow). Molecular mass standards (in kilodaltons) are shown on the left side of the panel.

Antibody responses induced after immunization with PyMSP4/5-His were determined using serum samples collected prior to challenge and assayed for specific anti-PyMSP4/5 activity by ELISA. In general, mice with the highest antibody response (specific to PyMSP4/5) showed the best protection, whereas mice with the lowest antibody levels showed poor protection or succumbed to the malaria infection. The association of prechallenge antibody levels (OD₄₀₅ values) was assessed using the Spearman rank correlation test, with a correlation coefficient of r = 0.59 (P = 0.0001), indicating significant correlation (Fig. 2). We attempted to estimate the average titer of anti-PyMSP4/5 antibodies induced by preparing a series of twofold dilutions of the individual sera from mice immunized with NR PyMSP4/5-His. All immunized animals showed increased antibody levels compared to control animals, and the average titer was 1:1,600,000.

View larger version (11K): [in this window] [in a new window]   FIG. 2.   Scatter diagram showing the relationship between prechallenge antibody levels and the peak parasitemia in 33 mice immunized with 25 µg of PyMSP4/5-His. The Spearman rank correlation coefficient r is 0.59 (P = 0.0001). OD405 values (1:5,000 dilution) were used rather than titers to calculate the correlation coefficient to avoid the effect of a large number of ties.

We have demonstrated that immunization with PyMSP4/5 offers protection against challenge with a lethal dose of P. yoelii YM. Survival rates varied from 75 to 100%, but the majority of immunized animals, although protected, did develop patent parasitemia. The immunization was highly protective when death was considered as the readout (P < 0.0001; Fisher's exact probability test), and it was also capable of significantly reducing peak parasitemia (P < 0.0001; Wilcoxon rank test). The levels of parasitemia and the survival rates observed in this study are comparable to those previously obtained in immunization trials with MSP1 (8, 13, 23) and AMA1 (2, 6). All animals immunized with PyMSP4/5-His showed high levels of anti-PyMSP4/5 antibodies prior to challenge, and mice with high levels of antibodies directed against PyMSP4/5 were better protected than those with low levels. In contrast to studies with MSP1 and AMA1, we found that RA PyMSP4/5-His protein was still capable of inducing substantial levels of protection in immunized mice. The level of protection as measured by survival or peak parasitemia was less than that induced by NR antigen, but the differences did not reach statistical significance. For MSP1 and AMA1, such treatments essentially ablated the capacity of the immunogen to induce protection (6, 14). Our previous data (11) showed that the addition of reducing agent to parasite samples greatly reduced the levels of antibody recognition of native MSP4/5 by antisera raised to nonreduced recombinant proteins. However, in all three cases (P. chabaudi, P. berghei, and P. yoelii) recognition was not abolished. It may be that the very high levels of antibody produced here were still capable of recognizing the target protein, although at a lower efficiency. Alternatively, it may be that some protective epitopes are not reduction sensitive. The use of E. coli-prepared material raises the possibility that bacterial endotoxin may contribute to the protective effect noted here. Although the antigen preparations were washed extensively while bound to the affinity matrix, it is possible that some endotoxin might be present. We would suggest, however, that endotoxin effects are likely to be insignificant in the presence of a strong adjuvant such as complete Freund adjuvant. Such questions would, of course, have to be addressed if primate or clinical trials were commenced using the P. falciparum homologues of PyMSP4/5.

It is generally agreed that immunity to P. yoelii is predominantly antibody mediated (9, 12). We have not performed experiments here that directly address the mechanism underlying the observed protection. It may be that other mechanisms, including natural-killer-cell immunity (19), T-cell immunity (20), or cell-mediated immunity (10), are also operating. The general trend that mice with lower ELISA values within an experiment had poorer outcomes does support an involvement of antibodies. The isotype distribution of anti-PyMSP4/5 antibodies in prechallenge sera indicated high levels of immunoglobulin G2a (IgG2a) and IgG2b subtypes (data not shown). The high level of IgG2a may be significant since it has been shown that antibodies of IgG2a isotype are responsible for antimalarial protection in passive-transfer studies (26) and are believed to play a crucial role in antimalarial immunity (1, 21). Furthermore, other investigators have shown that passive transfer of this particular subtype protects animals against infection with Trypanosoma musculi (25).

The results obtained here demonstrate that immunization with recombinant PyMSP4/5 can induce immune responses in mice that are strongly protective against challenge with a lethal dose of P. yoelii YM. Questions that need to be addressed include whether such immunity extends to heterologous challenge and whether such immunity can be induced using adjuvants suitable for human use. The first question will need to be addressed in the P. chabaudi model since we have been unable to identify a P. yoelii strain that expresses variant forms of PyMSP4/5. In P. falciparum the MSP4/5 homologues are two separate proteins, PfMSP4 and PfMSP5. Assuming that these proteins are capable of inducing protection against P. falciparum infection, it is unknown whether one or both proteins would be required. At present, we would suggest that the proteins need to be administered together, perhaps in combination with other asexual stage antigens. We suggest that these results provide strong evidence for the inclusion of these proteins in vaccine trials.