We evaluated the immunogenic properties of four recombinant proteins representing MSP-3α and MSP-3β of P. vivax
; one representing the C-terminal region of PvMSP-3α and three representing different regions of PvMSP-3β. Initially, these recombinant proteins were compared for their ability to bind IgG antibodies in the serum of individuals exposed to P. vivax
malaria. We demonstrated that the frequencies of individuals with IgG antibodies to PvMSP-3α and at least one of the three recombinant proteins representing PvMSP-3β were relatively high (68.2% and 79.1%, respectively). For PvMSP-3α, these findings confirm two recent studies performed in distinct endemic areas of the Brazilian Amazon, where 78% 
and 58.4% 
of individuals presented IgG antibodies to this protein.
Previous studies evaluated the polymorphism in the C-terminal region of PvMSP-3α and found the region is highly conserved among natural isolates 
. This fact may account for the recognition by IgG antibodies from a relatively high percentage of individuals. In contrast to the C-terminal region of PvMSP-3α, a variable degree of polymorphism has been reported for the gene encoding PvMSP-3β 
. Despite the reported sequence diversity of PvMSP-3β, we found a significant percentage of individuals recognized PvMSP-3β recombinant proteins FP-2 and FP-3 (>60.0%). Nevertheless, we detected a lower frequency (26.3%) of responders to PvMSP-3β recombinant protein FP-1, suggesting most of the antibody responses were directed to the second moiety of the protein. The PvMSP-3α and PvMSP-3β polymorphic frequency in the studied areas is unknown. A recent study used PCR-RFLP to characterize the diversity of MSP-3α in 60 P. vivax
isolates from four geographic regions of the Brazilian Amazon. The results revealed a high diversity where three different fragment sizes were found 
The cause for differential recognition by human antibodies of the N- and C-terminal regions of PvMSP-3β (FP-2 and FP-3) does not seem to be related to the absence of proper folding. In a previous study involving circular dichroism experiments, we demonstrated FP-1 was better structured than FP-2, which was highly recognized by IgG antibodies 
The comparison of human antibody reactivities to different antigens revealed major correlations. Significant correlations were observed between FP-2 and FP-3 of PvMSP-3β, possibly because they share a number of common B and T cell epitopes. However, the lack of correlation in most cases reflects differential genetic control by human HLA molecules. This hypothesis is being tested.
To investigate the immunogenic properties of PvMSP-3α and PvMSP-3β as vaccine candidates, we tested their immunogenicity in the presence or absence of different adjuvant systems in pre-clinical vaccinations of mice. In the absence of adjuvant, some of PvMSP-3β recombinant proteins elicited a specific TLR4-independent antibody response. This observation may explain how these molecules are immunogenic during natural human infection. In contrast, PvMSP-3α did not induce antibody immune responses, indicating the presence of other molecules in the parasite providing the adjuvant signal.
Previous studies have demonstrated that a major challenge in the development of subunit vaccines for malaria is the identification of a safe and potent adjuvant capable of inducing immune responses high antibody titers 
. Antibody titers were very high in animals vaccinated with the C-terminus of PvMSP-3α or different regions of PvMSP-3β emulsified in IFA; thus, these recombinant proteins can be highly immunogenic. Our results indicate PvMSP-3α or a protein representing the majority of the PvMSP-3β sequence (FP-3) were immunogenic when administered in adjuvants other than IFA. The immunogenicity of PvMSP-3α was greater when administered in Quil A, a saponin derived from the bark of a Chilean tree, Quillaja saponaria
, than in Alum, FliC, or CpG ODN 1826 and was similar to IFA. In addition, the antigen in TiterMax generated antibody titers similar to that obtained in IFA. PvMSP-3β yielded high antibody titers in all tested adjuvants, although Alum and FliC failed to perform at the level of IFA; however, TLR-9 agonist CPG ODN 1826 improved their adjuvant activity. It is of interest to note that adjuvants such as Alum, FliC, TiterMax, and IFA tend to induce Th2 with high IgG1/IgG2a ratios, whereas CPG ODN 1826 and Quil A show a clear modulation of the IgG subclass response pattern to a more balanced Th1/Th2 response.
The cellular response to CpG DNA is mediated by TLR9, followed by induction of pro-inflammatory cytokines (e.g.
IL-12, TNF-α, and IFN-γ), and producing a strong Th1 response 
. We observed a response pattern favoring Th1 in all formulations containing CPG ODN (CPG ODN 1826 alone, Alum+CPG ODN 1826, and FliC+CPG ODN 1826). Clinical trials evaluating the adjuvant activity of CpG ODN with vaccines designed to prevent malaria have been reported 
. Co-administration of CpG with AMA-1 
of P. falciparum
increased the geometric media of antibodies by 5.5 or 8-fold, respectively, when compared to each protein alone.
The relevance of antibodies against PvMSP-3α or β in host protection remains untested. Evidence in favor of a protective role for anti-P. vivax
MSP-3 was obtained by clinical trials performed with P. falciparum
. The functional role(s) for the parasite and in the context of host immune responses remain to be determined for other members of the MSP3 family in each of these species 
. Such investigations as a whole should help guide decisions for the development of malaria vaccines based on these or alternative proteins, which could prove to be valuable in areas of the world afflicted with both P. falciparum
and P. vivax