Antibodies to Plasmodium falciparum and Plasmodium vivax Merozoite Surface Protein 5 in Indonesia: Species-Specific and Cross-Reactive Responses

Study subjects and samples

Patients were recruited in Timika, a lowland region of Papua, Indonesia, that has endemic unstable transmission of both P. falciparum and P. vivax, a similar ratio of P. falciparum to P. vivax (57:43) year-round, and an annual malaria incidence of 938 cases per 1000 person-years [9]. Subjects were enrolled in trials of chloroquine and sulfadoxine-pyrimethamine therapy or artemisinin combination therapy (ACT) after providing informed consent [9]. Papuans with fever or a history of fever within 48 h of enrollment, with no alternative cause of fever identified, and with microscope-identified P. falciparum (82 subjects), P. vivax (86 subjects), or mixed P. falciparum and P. vivax infection (85 subjects) were included in this study. Eighty-seven Plasmodium-exposed healthy adults who had resided in Timika for at least 2 years and who had not had fever or symptoms of malaria during the preceding 2 weeks were also evaluated as exposed asymptomatic control subjects. Venous blood was collected at the time of presentation and, for the groups with malaria, at ~7 and 28 days after treatment. Plasma was separated and cryopreserved in Timika and transported to the Menzies School of Health Research for analysis. Plasma samples from 36 anonymous healthy Australian blood donors (provided by the Australian Red Cross Blood Service) were tested as negative controls. The study was approved by the ethics committees of the National Institute of Health Research and Development, Ministry of Health, Jakarta, Indonesia; the Menzies School of Health Research; and the Australian Red Cross Blood Service.

MSP5 antigen

Recombinant PfMSP5 was expressed and purified as described elsewhere [10, 11]. Recombinant PvMSP5 protein of the Salvador I isolate was expressed from Escherichia coli BL21 (DE3) strain (Novagen) containing a pTrcHis-A recombinant plasmid [7] by induction with 2 mmol/L isopropyl-β-D-thiogalactopyranoside (Progen Industries). Lysate was first passed through a metal affinity resin column (Amersham) for binding of the hexahistidine-tagged protein, in accordance with the manufacturer’s instructions, and the eluate was then subjected to ion-exchange chromatography on a Sepharose column (GE Healthcare), in which proteins were eluted on a linear gradient of sodium chloride. Analysis of purified protein by SDS-PAGE and Western blotting using anti-His antibody (Roche) and rabbit anti-PvMSP5 polyclonal serum identified a 62-kDa protein band corresponding to PvMSP5 (data not shown). E. coli host cell protein content was determined using an immunoenzymometric assay (Cygnus Technologies) and was found to be 0.004% of total protein content. In the total IgG ELISA, PfMSP5 was used at 1 μg/mL and PvMSP5 at 0.25 μg/mL, and both were used at 0.75 μg/mL in the competition ELISA.

MSP5 ELISA

Nunc plates coated with MSP5 protein were blocked with 5% skim milk in PBS containing 0.05% Tween (Sigma) (PBS-T) and washed with PBS-T. Then, 50 μL of plasma, diluted 1:800 in PBS-T, was added to the plate, and the assay was incubated for 1 h. Anti–human total IgG horseradish peroxidase (HRP; 1:2000 dilution; Zymed), anti–human total IgM (1:2000 dilution; Chemicon), or a 1:500 dilution of anti–human IgG₁, IgG₂, IgG₃, and IgG₄ (Zymed; clones HP6069, HP6014, HP6047, and HP6025) was added, and the assay was incubated for 1 h. The assay was developed using tetramethyl benzidine solution (Zymed), the color reaction was stopped with 1 mol/L hydrochloric acid, and the plate was read at 450 nm. The binding of antibodies in plasma from 36 unexposed Australian blood donors was used to define a cutoff (mean optical density plus 3 SDs) for positive and negative responses to each antigen. ChromPure human IgG and IgM (myeloma) whole molecules (Jackson ImmunoResearch Laboratories) were used as standards for quantitation of antibody responses, after dilution in PBS to 0–300 ng/mL for IgG or 0–125 ng/mL for IgM. Our cutoffs for positive responses were ≥23 μg/mL PfMSP5 and 24.8 μg/mL PvMSP5 for IgG and 7.9 μg/mL PfMSP5 and 7.2 μg/mL PvMSP5 for IgM.

Competition ELISA

First, 120 μL of plasma (diluted 1:400 in PBS-T) was incubated with serial dilutions of PfMSP5 and PvMSP5, respectively, at 0.02–10 μg/mL in separate 96-well round-bottom tissue culture plates (Becton Dickinson) and incubated at 37°C for 2 h. Then, 50 μL from each well was transferred into Nunc Maxisorb plates coated with 0.75 μg/mL PfMSP5 and PvMSP5, respectively, after blocking with 5% skim milk in PBS-T and incubated for 1 h at room temperature. The ELISA was performed as described above, using anti–human total IgG HRP (1:2000 dilution; Zymed). IgG responses were considered to be cross-reactive if preincubation with PfMSP5 resulted in reduced antibody reactivity in the PvMSP5 ELISA, and vice versa.

Statistical analysis

Statistics were calculated using GraphPad Prism software (version 5; GraphPad) and SPSS for Windows (version 15; SPSS). Fisher’s exact test was used to compare proportions of responders between 2 groups. Antibody titers were compared between groups by the nonparametric Mann-Whitney U test. Spearman’s rank correlation was used to compare antibody titers and parasitemia or age. The significance level used was α = .05 for all statistical tests. The therapeutic response was defined on the basis of the cumulative incidence on day 42, calculated by the Kaplan-Meier method. The risk of treatment failure was compared by the Mantel-Haenszel log-rank test and the hazard ratio presented. In multivariate analysis, any variable found to be significantly associated with the dependent variable in univariate analysis was entered into a Cox regression model, and the model was constructed using all factors.

Frequent recognition of MSP5

Plasma was available from 253 Papuans treated for malaria (82 with P. falciparum infection, 86 with P. vivax infection, and 85 with mixed P. falciparum and P. vivax infection) and from 87 Papuan control subjects without malaria symptoms (table 1). Both PfMSP5 and PvMSP5 were frequently recognized by both IgG and IgM antibodies from Papuans. The overall prevalence of IgG responses was similar across groups with asymptomatic or symptomatic infection, with 39%–50% of subjects seropositive for PfMSP5 and 42%–52% seropositive for PvMSP5 (table 2). High antibody levels, with mean IgG responses of 32–35 μg/mL for PfMSP5 and 39–47 μg/mL for PvMSP5, were found in each group, similar to the levels of antibody induced in the clinical trial of another malaria vaccine candidate, MSP1₄₂ (mean IgG level, 48.3 μg/mL) [12]. There was no significant difference in the magnitude of IgG responses to PfMSP5 or PvMSP5 between patients with acute symptomatic P. falciparum, P. vivax, or mixed P. vivax and P. falciparum infection and the exposed asymptomatic subjects (figure 1) and no significant association between antibody responses and parasitemia (data not shown).

Mean IgM levels were lower than mean IgG levels for both PfMSP5 (16–23 μg/mL) and PvMSP5 (12–16 μg/mL) across all groups. There was no significant difference in the magnitude of IgM responses between groups with acute symptomatic malaria and exposed asymptomatic individuals (figure 1). Surprisingly, we found that IgM responses to PfMSP5 but not to PvMSP5 were significantly and substantially more frequent in subjects with acute P. vivax or mixed P. falciparum and P. vivax infection (71% and 72%, respectively) than in the exposed asymptomatic control subjects (44%) (P < .05) (table 2), even though the magnitude of IgM responses did not differ between groups.

MSP5 IgG cross-reactivity

In our cohort of 340 individuals, 69%–71% of subjects with symptomatic infection or asymptomatic exposure who had IgG to MSP5 responded to both PfMSP5 and PvMSP5. Plasma samples from 107 subjects with dual IgG responses were tested for cross-species reactivity by competition ELISA. This involved preincubation of plasma with MSP5 from 1 species at serial protein dilutions followed by ELISA using MSP5 from the other species. Cross-reactive immunoglobulin would bind to MSP5 during the preincubation step and, with increasing concentrations of competing MSP5 protein, would reduce signal in the ELISA. Homologous proteins were used as controls for the assay. Figure 2 shows representative results for samples reactive to either PfMSP5 or PvMSP5 only, reactive to both but with no cross-reactivity, or with cross-recognition of both proteins. We found that 7 (7%) of the 107 subjects tested had cross-reactive IgG responses, indicating cross-species recognition, but the majority of IgG responses recognized PfMSP5 or PvMSP5 exclusively (data not shown). These data show that natural exposure to Plasmodium species induces MSP5 IgG responses that cross-react with P. falciparum and P. vivax, albeit infrequently. The 7 subjects with cross-reactive antibodies were from 3 different cohorts (3 exposed asymptomatic subjects, 3 with mixed infection, and 1 with P. vivax infection).

The IgG subclass of MSP5 humoral responses

Because malaria-specific IgG₁ and IgG₃ responses have been associated with lower parasitemia and a reduced risk of clinical malaria [13–16], we evaluated all PfMSP5 and PvMSP5 IgG responders to determine the subclass of response by ELISA. Across all groups tested, the majority of anti-PfMSP5 antibodies (53%–69%) were of the IgG₁ and IgG₃ subclass, with 20%–31% of subjects producing both IgG₁ and IgG₃ (table 3). PfMSP5-specific IgG₂ was detected in 6%–8% of subjects with acute malaria, compared with 23% for exposed subjects without clinical malaria (P = .08; χ² test). No IgG₄ responses were detected in any of the groups.

A different distribution pattern of IgG subclass was observed for antibodies to PvMSP5, because IgG₃ responses were on average 2.5 times more prevalent than IgG₁ responses across all groups. Similar to findings for PfMSP5, 12%–21% of subjects had both IgG₁ and IgG₃ antibodies (table 3). PvMSP5 IgG₂ responses were rare in all groups (0%–7%), including in exposed individuals with no clinical malaria, and no IgG₄ was detected. Overall, we found no relationship between the frequency of IgG subclass response and clinical presentation (i.e., asymptomatic or acute). The differing ratios of IgG₁ and IgG₃ isotype responses to PfMSP5 and PvMSP5 highlight a species-specific difference in the MSP5 antibody response, which occurred irrespective of past exposure or current infection.

MSP5 humoral responses and recurrence of parasitemia

The similar magnitudes of antibody responses to PfMSP5 and PvMSP5 across the different clinical cohorts led us to question whether the MSP5 antibody responses detected during acute infection contribute to parasite clearance and/or impair reinfection. The potential clinical relevance of MSP5 IgM and IgG antibody responses was evaluated in 2 groups: (1) a subset of 26 patients with malaria who received chloroquine and sulfadoxine-pyrimethamine treatment [17] and (2) 191 patients who received ACT [9], all of whom were monitored for up to 45 days after treatment for disease recurrence. Because significant drug resistance to both chloroquine and sulfadoxine-pyrimethamine has emerged [17], whereas ACT clears parasites rapidly [9], the correlation with therapeutic response was assessed separately in both groups. In the 26 patients who received chloroquine and sulfadoxine-pyrimethamine treatment, the magnitude of PfMSP5 IgM and IgG antibody responses were compared between the 15 subjects who successfully cleared P. falciparum (monitored for 20–45 days) and the 11 patients in whom reinfection was detected between days 7 and 31. IgG responses did not differ significantly between the 2 groups (data not shown). IgM responses, however, were significantly stronger in the group that successfully cleared parasites (P = .005; data not shown), but this difference did not remain after controlling for age.

In the second group, a clinical response after antimalarial therapy was documented in 96% of patients (243/253) with symptomatic malaria treated with ACT. Of the 78% of patients (191/243) who received ACT, 47 were infected with P. falciparum, 71 were infected with P. vivax, and 73 were infected with both species. There was no significant relationship between the magnitude of total IgM or IgG antibody levels and the age of the patient or the therapeutic response (data not shown).

Longitudinal analysis of MSP5 IgM and IgG humoral responses

To determine whether IgM and IgG antibodies to PfMSP5 and PvMSP5 were induced and/or boosted during acute malaria and/or after recovery, we compared the magnitude of antibody responses detected at treatment with those detected 7 and 28 days later. The pattern of total antibody responsiveness was monitored over time in 20 patients with acute uncomplicated P. falciparum, P. vivax, or mixed malaria. Antibody responses to MSP5, when detected, generally increased between the day of presentation and peaked on day 7, in accordance with antibody responses to other P. falciparum merozoite antigens [18]. In general, 2 separate patterns of responsiveness were observed: (1) IgM but no IgG antibody responses, associated with immune priming (9/20 subjects [45%]); and (2) IgM and IgG antibody responses, representing immune boosting of a secondary or later infection (11/20 subjects [55%]) (figure 3). Surprisingly, in 5 of 9 subjects singly infected with P. falciparum, increased PfMSP5 IgG responses (corresponding to the current infection) and boosted PvMSP5 IgG responses were detected. Induction or boosting of PvMSP5 IgM was observed in 7 of 9 individuals with a diagnosis of P. falciparum infection. The data show that a current Plasmodium infection can prime and/or boost MSP5 antibody responses. The data also suggest an elevation in cross-species MSP5 immune responses. The subclass of MSP5 IgG responses, once induced, appeared to be durable until day 28 after treatment in the 20 subjects with a range of ages (5–60 years) tested, providing no evidence of IgG class switching in the time frame analyzed (data not shown). The MSP5-specific antibodies, whether IgM or IgG, appeared to be short-lived, with their levels declining by day 28.