Naturally acquired humoral and cellular immune responses to Plasmodium vivax merozoite surface protein 8 in patients with P. vivax infection

Gene sequence of Pvmsp8 was obtained from the PlasmoDB website (http://​plasmodb.​org; accession no. PVX_097625). Protein domains were further predicted using the Simple Modular Architecture Research Tool (SMART) (http://​smart.​embl-heidelberg.​de/​) [15, 16]. Recombinant PvMSP8 (rPvMSP8) with a truncated signal peptide (SP) and GPI anchor was expressed and purified using a wheat-germ cell-free (WGCF) expression system [17]. Briefly, the specific primers: PvMSP8F, 5′-GGGCGGATATCTCGAGGGAAACGTTAGCCCACCC-3′; and PvMSP8R, 5′-GCGGTACCCGGGATCCTTAGCAGTATATTCCGTCTCCCTCA-3′ were used for DNA amplification. Then, the PvMSP8 DNA was cloned into the pEU-E01-His-TEV-MCS vector (CellFree Sciences, Matsuyama, Japan). The rPvMSP8 protein was expressed using a WGCF expression system and purified using a Ni–NTA agarose column (Qiagen, Hilden, Germany), as described elsewhere [17]. The production of rPvMSP8 protein was separated using 12% SDS–PAGE and detected via Western Blot using an anti-Penta-His antibody (Qiagen).

Female BALB/c mice, at 6–8 weeks of age (DaehanBiolink Co., Eumsung, Korea) were immunized with 20 μg of rPvMSP8 and phosphate buffered saline (PBS, pH 7.4) with complete Freund’s adjuvant (Sigma-Aldrich, St. Louis, MO, USA) using intraperitoneal route of administration. Three and 6 weeks after immunization, the equal volume of antigen with incomplete Freund’s adjuvant (Sigma-Aldrich) was boosted. Mouse blood samples were taken after the final booster injection, 2 weeks later. The antisera against PvMSP1 was also produced following the same procedure as PvMSP8 [17]. Animal experimental protocols were approved by the Kangwon National University, and the experiments were performed according to the Ethical Guidelines for Animal Experiments of Ehime University and Kangwon National University.

The schizont-stage-rich parasites of P. vivax were collected from malaria patients in Thailand, as described previously [18]. Briefly, the slides were blocked with PBS containing 5% nonfat milk, incubated with rabbit anti-MSP1-19 (1:200 dilution) [18] and mouse anti-MSP8 (1:100 dilution) as primary antibodies, followed by incubation with Alexa Flour 546-conjugated goat anti-rabbit IgG or Alexa Flour 488-conjugated goat anti-mouse IgG antibody (Invitrogen, Carlsbad, CA, USA) and nuclear staining with DAPI (Invitrogen). Then, the slides with ProLong Gold antifade reagent (Invitrogen) were mounted. The parasites were observed under oil immersion using a confocal laser scanning FV200 microscope (Olympus, Tokyo, Japan).

One hundred and twelve serum samples from vivax-infected patients (mean age 25.5 years; range 18–42 years) from endemic countries and 80 samples from healthy individuals, assessed as being negative using microscopy, were collected in the Republic of Korea. The Myanmar samples (n = 56) were collected in 2012 from patients in Shwe Kyin area of Myanmar, and the Thai samples (n = 56) were collected from symptomatic, smear-positive patients from the Mae Sod district of western Thailand. For the 1-month follow-up study, 25 patients from the Shwe Kyin area were assessed at 7, 14, 21, and 28 days after treatment. The serum samples from healthy individuals with a vivax malaria history were collected from Chinese residents who had an episode of vivax malaria infection in malaria endemic areas of Anhui Province, China, in 2012, and who did not have a reinfection episode of vivax malaria in the preceding 5 (n = 30), 12 (n = 30), or 30 (n = 30) years (Table 1).

For the cellular immunity study, 10 mL of heparinized blood from P. vivax subjects (n = 15) who had recovered from P. vivax infection after 8–10 weeks were collected at malaria clinics in Tha Sae, Chumphon Province, which is located in the southern peninsular region of Thailand. Moreover, heparinized blood from healthy individuals (n = 15) who had no history of exposure to malaria were collected. The samples were used for peripheral blood mononuclear cell (PBMC) preparation.

Amine (NH₂-)-coated slides were prepared as previously described [19]. Serum samples from 112 cases of vivax malaria and 80 healthy individuals were used for humoral immune response analysis via well-type amine arrays. The chips were probed, scanned, and analysed as described previously [19]. The prevalence of an IgG isotype specific to PvMSP8 in the sera of 50 vivax patients and 10 healthy individuals was evaluated. Briefly, rPvMSP8 (50 μg/mL) was used for coating, followed by blocking and addition of human sera. Then these coated proteins were incubated with each IgG isotype for isotyping assay. The reaction was detected and analysed as described above.

An array of 22 peptides, 18-mer each (Additional file 1: Table S1) overlapping by nine amino acids with >90% purity and spanning the conserved C terminus of PvMSP8 Sal-1 sequences was custom synthesized, purified, and used for epitope mapping (Peptron Co., Ltd., Daejeon, Korea). In the process of coating step, 1 μL of peptides (10–40 μg/mL) in 200 mM 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC, Thermo Fisher Scientific Inc., Rockford, IL, USA) and 50 mM N-hydroxysuccinimide (NHS, Thermo Fisher Scientific Inc.) with coupling buffer were coated on amine-coated slide, respectively. After blocking, rPvMSP8-immunized mouse sera, pre-immunized mouse sera at 1:200 dilution, pooled sera from 10 high IgG titers of vivax-infected patients at 1:50 dilution, or 10 vivax-unexposed human sera at 1:50 dilution was added, respectively. Alexa Fluor 546 goat anti-mouse IgG (50 μg/mL, Invitrogen) or Alexa Fluor 546 goat anti-human IgG (10 μg/mL, Invitrogen) antibodies were used for detection of binding activity.

Lymphocyte stimulation assay was carried out for measuring of cellular immunity in 15 recovered P. vivax subjects. Briefly, 2.5 × 10⁵ PBMCs/well were stimulated with 10 μg/mL of rPvMSP8, 2% v/v of PHA, or complete RPMI 1640 medium only. After 96 h of stimulation, cytokine levels in the lymphocyte culture supernatant were measured using a BD OptEIA kit (BD Biosciences, San Jose, CA, USA). The phenotypes of T cells that responded to rPvMSP8 were analysed by intracellular cytokine staining after in vitro stimulation with 10 μg/mL of rPvMSP8, 20 ng/mL of PMA/ionomycin, or medium alone. Cells were stained with monoclonal antibodies (mAbs) against surface determinants of CD3, CD4, and CD8 and intracellular cytokine markers of anti-IFN-γ and anti-IL-10 antibodies, according to the manufacturer’s instructions.

A splenocyte proliferation assay was performed using splenocytes removed from mice that had been immunized with rPvMSP8. The cells were stimulated with rPvMSP8 (2.5 μg/mL), Con A (5 μg/mL), LPS (10 μg/mL), or medium alone. Culture supernatants were collected after 72 h of incubation and assayed using a BD CBA Flex Set kit (BD Biosciences), according to the manufacturer’s instructions.

The correlation between the antibody reactivity of different concentrations of the recombinant proteins and duplicate spots of protein arrays was observed using GraphPad Prism software, version 5.0 (GraphPad, San Diego, CA, USA) and PASW Statistics 18.0 (SPSS Inc., Chicago, IL, USA). Sensitivity and specificity were measured by the percentage of patients who had a positive test result and the percentage of healthy individuals who had a negative test result, respectively. The significance of differences in mean fluorescence intensity (MFI) values between every two groups were performed with the Mann–Whitney U test.

The predicted molecular weight of PvMSP8 (PVX_097625) is 54.7 kDa, with 487 amino acids containing a signal peptide, asparagine-rich region in N-terminal region and two EGF domains and GPI-motif in C-terminal region (Fig. 1a). The recombinant protein encoding the truncated PvMSP8 (ΔSP/GPI) was successfully expressed and purified as soluble form using a WGCF expression system. The purity of the purified recombinant PvMSP8 proteins were assessed, and the protein migrated as a single band of 55 kDa from SDS-PAGE analysis (Fig. 1b). The corresponding immunoblots probed with an anti-His tag antibody, anti-rPvMSP8 immune mouse serum, or a mixture of vivax patient sera revealed a similar and specific pattern of migration for PvMSP8 (Fig. 1c), whereas preimmune mouse sera and healthy individuals serum samples in non-endemic areas were used as negative controls; recombinant Pv41 did not react with either reagent.

To determine PvMSP8 localization, IFA was carried out using anti-PvMSP8 and anti-PvMSP1-19 immune sera. In blood-stage parasites of P. vivax, PvMSP8 was localized to the outer circle of pigments associated with the FV (Fig. 2, arrow heads). A comparison with PvMSP1-19 in different intraerythrocytic-stage parasites indicated that PvMSP8 localized on pigments densely from the ring to late trophozoite stage (Fig. 2a, b) and on scattered pigments from mature to late-schizont-stage parasites (Fig. 2c, d).

Antibody responses against rPvMSP8 in serum samples from 112 P. vivax-infected patients and 80 healthy individuals were determined. Seropositivity of anti-rPvMSP8 antibodies was 73.2% in sensitivity and 96.2% in specificity (Additional file 1: Table S2). These sera from P. vivax-exposed individuals exhibited a significantly higher MFI than did those from malaria-naïve subjects (Fig. 3a, P < 0.0001). Furthermore, PvMSP8 produced similar IgG titers among isolates of vivax malaria patients from Korea, Myanmar, and Thailand, as well as from malaria-naïve individuals (Fig. 3b). The sera of P. vivax-exposed individuals had a significantly higher MFI than sera from malaria-naïve subjects.

Protective antibodies against blood-stage parasites have been shown to belong to cytophilic classes; hence, the IgG subclass distribution was analysed. The prevalence of specific IgG subclasses was extremely high: >80% of the individuals had IgG antibodies that reacted with rPvMSP8 (Fig. 3c). Antibody responses against rPvMSP8 in vivax malaria patients were predominantly non-cytophilic IgG2 responses, which indicated that the efficacy of their protection may be poor against blood-stage parasites of P. vivax.

The isotypic distribution of anti-rPvMSP8 antibodies from immunized mice was analysed, assuming that the cytophilic IgG2a and IgG2b mouse isotypes would correspond to a Th1 response, whereas the non-cytophilic IgG1 and IgG3 would correspond to a Th2 response (Fig. 3d). Cytophilic antibodies (IgG2a and IgG2b) against PvMSP8 were the major components of the antibody response, especially IgG2b, observed in immunized mice. Non-cytophilic IgG antibodies were predominant in patients however cytophilic IgG antibodies were predominant in immune mice.

The immune reactivity of the antibody response from serum samples during the 1-month follow-up after treatment in 25 patients with acute vivax malaria was assessed. Antibody response to PvMSP8, when detected, generally increased between the day of presentation, peaking on day 7, and was followed by a slight decrease from days 7 to 28. However, no significant difference was observed regarding IgG titers within 1 month (Fig. 4a; Additional file 1: Table S3).

The longevity of anti-PvMSP8 antibody responses was also analysed using long-term after exposure to vivax parasites (5-year recovery, 12-year recovery, and 30-year recovery samples) from China. A significantly higher IgG reactivity was observed in 5- and 12-year recovery sera, although the IgG titer was lower than that observed for acute infection. Moreover, in most cases, IgG antibody levels in individuals who infected malaria 30 years previously had decreased to baseline levels (Fig. 4b; Additional file 1: Table S3). These results indicate that IgG antibody responses against PvMSP8 are stably sustained in this population.

A peptide was considered part of the potential epitope that there was significantly greater antibody reactivity for the antibodies in the vivax patient group compared with the vivax-unexposed human group, or greater reactivity for antibodies in rPvMSP8-immunized mice compared with preimmunized mice. Based on the reactivity of the pooled sera from vivax-infected patients, four peptides (nos. 2–5) located in the N terminus were highly reactive (Fig. 5a); based on those of immune mouse sera, two peptides (nos. 3 and 5) located in the central region and one peptide (no. 13) located in the C-terminal part were also highly reacted (Fig. 5b). Intriguingly, the two groups of sera were highly reacted with two peptides (nos. 3 and 5) located in the central region of PvMSP8. As a negative control, antibodies from preimmunized mice or unexposed human individuals that reacted with individual peptides were defined (all intensity values <1500).

The determination of IFN-γ or IL-10 cytokine levels after in vitro rPvMSP8 stimulation showed that IFN-γ was significantly upregulated (fourfold) in comparison to that observed in the absence of antigen stimulation (rPvMSP8 = 188.10 pg/mL, no stimulation = 27.68 pg/mL, P < 0.05; Fig. 6a). Moreover, the levels of IL-10 in rPvMSP8 cultures were significantly higher (twofold) in comparison to that observed in the absence of stimulation (rPvMSP8 = 119.17 pg/mL, no stimulation = 65.44 pg/mL, P < 0.05; Fig. 6a). The PBMCs from P. vivax subjects produced higher levels of IFN-γ and IL-10 in response to PHA and rPvMSP8 than did those of the healthy controls. The phenotyping of IFN-γ- or IL-10-producing cells that responded to the PvMSP8 antigen showed that CD8⁺ T cells were IFN-γ-producing cells; they produced IFN-γ at levels three times higher than those observed in un-stimulation (rPvMSP8 = 0.059%, no stimulation = 0.022%; Fig. 6b). In contrast, IFN-γ levels were low in CD4⁺ T cells (rPvMSP8 = 0.010%, no stimulation = 0.012%; Fig. 6b). Neither CD4⁺ nor CD8⁺ T cells produced IL-10 cytokine in response to the rPvMSP8 antigen (Fig. 6b). The cytokine responses from rPvMSP8-immunized mouse splenocytes stimulated by various concentrations of the rPvMSP8 protein, ConA (positive control), LPS (positive control), and culture medium alone (negative control) (Additional file 1: Figure S1A), respectively, showed predominant IFN-γ and TNF secretion compared with IL-2, IL-10, and IL-4 levels (Additional file 1: Figure S1B).