Background

Approximately 2.5 billion people, or one-third of the world’s estimated 2018 population live in regions of stable or unstable malaria transmission, and are at risk for infection [1, 2]. In sub-Saharan Africa, Plasmodium falciparum has been the major focus of treatment and intervention strategies because of the high mortality associated with infection. Three additional species of human malaria, two Plasmodium ovale sub-species and Plasmodium malariae, share much of the same geographic range in Africa yet are considered less important because prevalence estimates based on microscopic detection of parasites in blood films are generally low [3]. However, mounting clinical evidence suggests that malaria infection with species other than P. falciparum is not benign and that infection prevalence may be increasing in children, even in areas where anti-malarial drug therapy is regularly administered [4,5,6,7,8]. Similarly, the risk of Plasmodium vivax infection in sections of sub-Saharan and central Africa has been considered to be nil because large fractions of the populations in these regions lack the Duffy receptor used by the parasite for red blood cell invasion [9,10,11,12]. New evidence, however, suggests that low levels of P. vivax transmission in Africa may be occurring in susceptible Duffy-positive residents and that some level of infection is also occurring in Duffy-negative individuals by another reticulocyte invasion mechanism [13,14,15,16]. Because these non-P. falciparum infections are frequently sub-patent and their symptoms may be masked by the overwhelming levels of P. falciparum parasitaemia, accurate mapping and the estimation of prevalence levels in this population are difficult using traditional microscopic or PCR methods.

Serologic assays that detect IgG antibodies to specific P. falciparum and P. vivax antigens have been used in multiple studies in many parts of the world to estimate infection incidence and immunity levels (reviewed in [17,18,19]). Antibody data from cross-sectional surveys can be used to calculate the community-level seroconversion rates [20,21,22,23,24,25,26,27], and longitudinal and cross-sectional data provide similar estimates of community seroconversion rates [28]. Serologic assays using species-specific antigens could identify individuals who either are currently infected or have been previously infected with different malaria species, even if the infections were sub-patent [29]. Advances in multiplex assay technology make serologic antibody assays for multiple malaria antigens more attractive because antibody responses to a range of malaria antigens can be detected in a single well from a small volume of blood or serum and because malaria-specific assays can be integrated with assays for other antibody responses of public health interest [30,31,32,33].

One target antigen frequently used in malaria serologic antibody studies is the 19-kDa carboxy-terminal sub-unit of the merozoite surface protein 1 (MSP119) [34,35,36], a glycosylphosphatidylinositol-anchored fragment of the larger MSP1 protein that is found in abundance on the parasite surface (reviewed in [37]). Although the MSP119 proteins from P. falciparum, P. malariae, P. ovale, and P. vivax share 48–58% identity at the amino acid level [38], many of the conserved residues are cysteines and other hydrophobic amino acids that are unlikely to be exposed to the immune system [39]. Despite the sequence similarity, Cook et al. [24] were able to demonstrate unique seroprevalence curves for the P. falciparum and P. vivax MSP119 antibody responses in areas of reduced transmission in Vanuatu. Similarly, Bousema et al. [40] did not observe any correlation between the P. falciparum and P. vivax MSP119 antibody responses in ELISA studies of sera from a population living in a Somalian region of low endemicity for both parasites. In a study of malaria antibody responses in adult Cambodian women, Priest et al. [33] found that 79% of sera from women who were positive for antibodies to malaria reacted with the MSP119 antigen from only one species. However, all of these three studies were conducted in regions of relatively low transmission, and it is important to determine whether the MSP119 antigen-based assays will be species specific in regions of high transmission with multiple circulating species of malaria parasite.

During a bed net intervention study in a high malaria transmission region of northern Mozambique [41], numerous samples from individuals were assayed and found to have very high IgG antibody responses to multiple malaria species MSP119 antigens, including the P. vivax antigen. These samples and samples from two low transmission regions in a multiplex assay format to expand on the MSP119 competition ELISA studies of Amanfo et al. [42].

Methods

Human sample sets

Anonymous serum samples (N = 88), collected prior to 2000 from US citizens with no history of foreign travel, were presumed to be negative for antibodies to Plasmodium spp. and were used to define the cut-off values for the various assays. Anonymized, residual sera submitted to the Centers for Disease Control and Prevention between 1995 and 2011 for malaria diagnostic testing were used for the assessment of multiplex assay sensitivity. The panel included sera from patients having microscopically confirmed and/or PCR confirmed infections with P. falciparum (N = 33), P. malariae (N = 6), P. ovale (N = 7), or P. vivax (N = 35) [43]. The timing of sample collection relative to malaria infection or symptom development was not known. In addition to a pan-Plasmodium spp. immunofluorescence assay (IFA) positive serum pool (CDC Lot 8), mono-specific infection IFA serum controls were available for P. falciparum (CDC Lot 6) and P. malariae (CDC Lot 2).

Sera or dried blood spots previously identified by multiplex assay as having high levels of IgG antibodies to one or more MSP119 proteins were selected for the specificity studies. This set included: 3 anonymous, adult blood donor samples collected in 1998 from a region of Haiti with a low prevalence of P. falciparum infection [43]; 9 sera from an integrated serologic study of immune status to vaccine-preventable diseases and neglected tropical diseases conducted in 2012 among women 15–39 years of age in Cambodia [33, 44, 45]; and, 20 dried blood spots from participants (4–60 years of age) in a long-lasting insecticide-treated bed net impact study conducted in 2013–2014 in a high malaria transmission province of northern Mozambique [41]. The sample set from Mozambique was biased towards individuals with a positive antibody response to the P. vivax, P. ovale and P. malariae antigens.

Ethics statement

Residual malaria diagnostic sera were made anonymous under a protocol approved by the CDC Institutional Review Board. Written informed consent was obtained prior to enrolment and participation in the Cambodian sero-survey, and the study protocol was reviewed and approved by the National Ethics Committee in Cambodia [33, 44, 45]. Written informed consent was obtained prior to enrolment and participation in the Mozambique bed net study and sero-survey, and the study was approved by the National Bioethics Committee in Mozambique. For both of these studies, CDC researchers had no access to personal identifiers, and CDC staff were not considered to be engaged with human research subjects.

Banked chimpanzee sera

Banked sera from malaria studies conducted in experimentally infected chimpanzees prior to 2000 were included in this report. Chimpanzees Bit and Klimatis were infected with the Uganda I strain of P. malariae [46, 47], Alpert was infected with the Nigeria I strain of P. ovale [48], and Duff was infected with the Salvador I strain of P. vivax [49]. As previously described, all animals had been splenectomized before they were inoculated intravenously with heparinized, infected blood.

Antigens for multiplex assay

The cloning of the 3D7 strain P. falciparum MSP119 in pGEX 4T-2 plasmid (GE Healthcare, Piscataway, NJ, USA) as a fusion protein with Schistosoma japonicum glutathione-S-transferase (GST) and the purification of the MSP119-GST fusion protein have been described elsewhere [50].

Using the protocol of Priest et al. [33] and a new reverse PCR primer, a P. vivax MSP119 clone lacking the carboxy-terminal, hydrophobic anchor sequence was generated in pGEX 4T-2 plasmid (GE Healthcare), and the MSP119-GST fusion protein was purified. The target sequence was amplified from Belem strain DNA using a reverse deoxyoligonucleotide PCR primer with the following sequence: 5′-GCG GAA TTC TTA GCT GGA GGA GCT ACA GAA AAC TCC C-3′. The underlined sequence reverse primer identifies an EcoRI restriction endonuclease site used in directional cloning, and the italicized bases identify an introduced in-frame stop codon. All other cloning conditions remained as previously described [33]. The clone was sequenced using BigDye Terminator V3.1 chemistry (Applied Biosystems/Thermo Fisher Scientific, Foster City, CA, USA).

Cloning, expression and purification of a P. ovale MSP119-GST fusion protein from Nigeria I strain DNA was accomplished using the strategy described in Priest et al. [33] with the following deoxyoligonucleotide primers: forward, 5′-CGC GGA TCC TCT ATG GGA TCT AAA CAT AAA TGT-3′ and reverse, 5′-GCG GAA TTC TTA ACT TGA TGA GCC ACA GAA AAC ACC-3′. The underlined sequence in the forward primer identifies a BamHI restriction endonuclease site used in directional cloning. These primer sequences were based on the sequence of the Cameroon OMA1A P. ovale isolate sequence (GenBank accession number FJ824670) described by Birkenmeyer et al. [38].

Cloning of the P. malariae MSP119 coding sequence from China I strain DNA required two PCR reactions. The first reaction used long PCR primers (forward, 5′-AAT ATT AGC GCA AAA CAT GCA TGT ACC GAA ACA-3′; reverse, 5′-ACT TGA AGA ACC ACA GAA AAC ACC TTC AAA TAT AG-3′) and the amplification conditions previously described [33]. These primer sequences were based on the sequence of the Cameroon MM1A P. malariae isolate sequence (GenBank accession number FJ824669) described by Birkenmeyer et al. [38]. A total of 5% of the purified primary product (StrataPrep PCR purification kit, Stratagene, LaJolla, CA, USA) was used in a second amplification reaction with the following primers: forward, 5′-CGC GGA TTC AAT ATT AGC GCA AAA CAT GCA TGT-3′; reverse, 5′-GCG GAA TTC TTA ACT TGA AGA ACC ACA GAA AAC ACC-3′. This final PCR product was cloned in pGEX 4T-2 plasmid (GE Healthcare), and a GST fusion protein was expressed and purified using the protocol of Priest et al. [33].

Expression and purification of the control GST protein with no fusion partner has been described elsewhere [51]. A synthetic 20 amino acid peptide, (NANP)5-amide, corresponding to the carboxy-terminal repeat of the P. falciparum circumsporozoite protein (PfCSP peptide) [52, 53] was cross-linked to GST using the glutaraldehyde protocol of Benitez et al. [54]. Tetanus toxoid antigen from Massachusetts Biologic Laboratories (Boston, MA, USA) was exchanged into buffer containing 10 mM Na2HPO4 and 0.85% NaCl at pH 7.2 (PBS) [44].

Comparison of Plasmodium malariae MSP119 sequences from other geographic locations

Ten nanograms of DNA from P. malariae strains Greece I, Guyana, and Uganda I were PCR amplified using the forward and reverse long deoxyoligonucleotides described above and the Expand High Fidelity PCR system (Roche Applied Science, Indianapolis, IN, USA). Cycle conditions were as follows: 94 °C for 5 min, 35 cycles of 95 °C for 30 s, 55 °C for 30 s, and 68 °C for 1 min, and a final extension step of 68 °C for 5 min. Products were purified (StrataPrep PCR purification kit, Stratagene) and sequenced as described above.

Antigen coupling and multiplex bead assays

Antigens were coupled in 1.0 ml of buffer containing 25 mM 2-(N-morpholino)-ethanesulfonic acid (MES) at pH 5.0 with 0.85% NaCl using the following amounts of protein for 12.5 × 106 SeroMap microspheres (Luminex Corp, Austin, TX, USA): MSP119-GST fusion proteins, 30 μg; GST control protein, 15 μg; PfCSP peptide-GST, 30 μg; and tetanus toxoid, 12.5 μg. The coupling protocol and bead storage buffer have been described previously [55].

Blood spots were collected on filter paper disks (Cellabs, Sydney, Australia). A single tab containing 10 μl whole blood (approximately 5 μl of serum) was eluted overnight at 4 °C in 200 μl of buffer containing PBS with 0.05% Tween-20 and 0.05% NaN3 for a 1:40 serum protein dilution [56]. Samples were further diluted 1:10 (final 1:400 serum dilution) in PBS buffer (pH 7.2) containing 0.3% Tween-20, 0.02% NaN3, 0.5% casein, 0.5% polyvinyl alcohol (PVA), 0.8% polyvinylpyrrolidone (PVP), and 3 μg/ml Escherichia coli extract (Buffer A) [33, 57]. Test sera were diluted 1:400 in Buffer A. BSA was not included in the dilution buffer as it was found to be unnecessary for blocking when casein was present.

The multiplex bead assay protocol for total IgG has been described elsewhere [55, 58]. Assays were run in duplicate wells, and each plate included a buffer only blank. The reported “median fluorescent intensity minus background” value (MFI-bg) is the average of the 2 median fluorescent intensity values minus background blank values from two replicate wells. Negative MFI-bg values were set to 0.

In the multiplex IgG sub-class assays, serum antibodies were bound to beads using the previously described multiplex assay protocol [55, 58]. Washed beads were then incubated for 45 min at room temperature with 50 μl/well of a 1:500 dilution in Buffer B (0.5% BSA, 0.05% Tween-20, and 0.02% NaN3 in PBS at pH 7.2) of biotinylated monoclonal mouse anti-human IgG subclass secondary antibody to IgG1 (clone HP6025), IgG2 (clone HP6002), IgG3 (clone HP6047), or IgG4 (clone HP6025) (all from Zymed/Invitrogen, South San Francisco, CA, USA). Wells were developed with R-phycoerythrin-labelled streptavidin and read on a BioPlex 200 instrument (BioRad, Hercules, CA, USA) as described above.

Assessment of coupling efficiency

To determine whether the Plasmodium spp. GST-MSP119 fusion proteins were coupled to the SeroMap beads with similar efficiencies, multiplex assays were run using a serial dilution of a goat anti-GST polyclonal IgG antibody (GE Healthcare) as the primary antibody to detect the fusion protein on the bead. The initial dilution of anti-GST antibody was 1:1000 in modified Buffer A lacking the E. coli extract (50 μl/well), and the final dilution was 1:1.0 × 107. Bound anti-GST antibody was detected with 50 μl/well of a biotinylated rabbit anti-goat IgG secondary antibody (1:500 dilution in Buffer B; Invitrogen) and wells were developed with R-phycoerythrin-labelled streptavidin and read on a BioPlex 200 instrument (BioRad) as described above.

MSP119 competition assays

Serial dilutions of purified MSP119-GST competitor proteins were generated from a 0.5 mg/ml stock solution using PBS buffer at pH 7.2. A 96-well incubation plate (V-bottom, Fisher Scientific) was set up such that wells contained 3 μl of the competitor MSP119-GST fusion protein dilution and 147 μl of serum dilution in Buffer A for 1:50 dilution of competitor protein and a negligible further dilution of the serum. Final competitor protein concentrations in the serum dilution ranged from 10 μg/ml to as low as 0.025 μg/ml. The plate was incubated at room temperature for 1 h, and then each well of the incubation plate was used to load duplicate multiplex bead assay wells at 50 μl each. The standard total IgG assay protocol was then followed as described above. The standard MSP119 competition assay used a final competitor protein concentration in diluted serum of 10 μg/ml, and a ≥ 30% reduction in multiplex assay signal was considered to be evidence of antibody cross-reactivity.

MSP119-specific antibody binding and elution

Using the standard coupling protocol [55], individual MSP119-GST fusion proteins were coupled to magnetic beads (region 14, Luminex) in 100 μl of MES/NaCl buffer at pH 5.0 at 4.5 μg for 1.25 × 106 microspheres (a 50% increase in protein compared to SeroMap bead amounts). Coupled beads were re-suspended in 120 μl of storage buffer with protease inhibitors [55]. A 1:200 dilution of serum in Buffer A or a 1:5 dilution of blood spot eluate in Buffer A (approximately 1:200 serum dilution) was incubated for 1 h at room temperature with 20 μl of coupled beads (washed 1× with 0.05% Tween-20 in PBS prior to use). Beads were collected by magnetic capture, the used serum or blood spot dilution was removed, and the beads were washed 4× with 200 μl 0.05% Tween-20 in PBS. To elute the bound antibodies, beads were re-suspended for 10 min at room temperature in 100 μl of buffer containing 3 parts of 4 M MgCl2 in 100 mM N-hydroxyethylpiperazine-N′-2-ethanesulfonic acid (HEPES) at pH 8.0 and 1 part ethylene glycol [59]. The beads were collected by magnetic capture, and the supernatant was removed and diluted into 0.9 ml of buffer containing 50 mM tris(hydroxymethyl)-aminomethane (Tris) at pH 7.5 and 0.85% NaCl. The antibody elution process was repeated once. The 2 ml of eluted antibody in Tris/NaCl buffer was concentrated to 50 μl using a Centricon-50 centrifugal filter device as directed by the manufacturer (Millipore, Bedford, MA, USA). The concentration procedure was repeated following a 2 ml Tris/NaCl buffer dilution and again after a 1 ml PBS buffer dilution. The final 30–50 μl of concentrate was diluted with > 3 volumes of Buffer A, and duplicate multiplex assays were performed using half of the eluted antibody per well.

Antibody avidity determinations

To determine the avidity of IgG antibody binding, MSP119-GST fusion protein coated SeroMap beads incubated for 1 h with 1:400 serum dilutions were immediately washed with 100 μl of 6 M urea in PBS for 5 min at room temperature [60]. The urea wash was repeated once followed by three 100 μl washes with 0.05% Tween-20 in PBS. The normal total IgG development protocol was then followed. An avidity index was calculated by dividing the 6 M urea-treated MFI-bg value by the untreated MFI-bg value.

Data analysis

Protein sequences were aligned using COBALT [61]. The means plus 3 standard deviations of the MBA responses from 88 adult US citizens with no history of foreign travel were used to define potential cutoffs for the MSP119 protein and CSP peptide assays. Receiver-operating characteristic (ROC) curves were also used to generate potential cut-offs for the MSP119 assays. The ROC analysis [62] and Spearman rank order correlation analysis were performed using SigmaPlot 13.0 (Systat Software, Inc., San Jose, CA, USA). The J-index [63] was calculated from the sensitivity and specificity values.

Results

MSP119 sequences

The DNA sequence of the P. malariae China I strain MSP119 clone (deposited in GenBank as MH577182) differed from the Cameroon sequence of Birkenmeyer et al. [38] at 3 nucleotide base positions, leading to 2 amino acid substitutions in the deduced amino acid sequence: G41E and Q51K (numbering based on mature MSP119 protein sequence). As shown in Fig. 1, the deduced amino acid sequence of the China I strain was identical to that reported for the Brazil I11 strain [64] and was also identical to that of the Greece I strain (GenBank MH577183). Compared to the Cameroon strain, the Uganda I strain of P. malariae contained only a G41Q amino acid substitution (GenBank MH577184), while the Guyana strain contained only a G41E substitution (GenBank MH577185).

Fig. 1
figure 1

Alignment of predicted Plasmodium spp. MSP119 protein sequences using COBALT [61]. Residues in the P. malariae sequence that differ from the Cameroon sequence of Birkenmeyer et al. [38] are shaded. Predicted protein sequences resulting from the oligonucleotides used in PCR amplification are underlined. The positions of residues conserved among all the presented MSP119 protein sequences are indicated in the consensus with divergent residues indicated by a dot. GenBank accession numbers are MH577181, P. ovale Nigeria I strain; MH577182, P. malariae China I strain; MH577183, P. malariae Greece I strain; MH577184, P. malariae Uganda I strain; and MH577185, P. malariae Guyana strain

The DNA sequence of the P. ovale Nigeria I clone (GenBank MH577181) matched the GenBank sequence reported for the Cameroon OMA1A P. ovale isolate (FJ824670) by Birkenmeyer et al. [38]. The Nigeria I strain likely belongs to the newly identified Plasmodium ovale curtsi species as the MSP119 predicted protein sequence has a Ser at position 23 rather than a Pro [65]. The sequence of the P. vivax clone matched that found in GenBank (accession number AF435594.1) [66]. Alignment of the deduced MSP119 amino acid sequences of the P. malariae[38, 64], P. falciparum 3D7 strain [67], P. vivax Belem strain[66, 68], and P. ovale Nigeria I strain proteins in Fig. 1 showed conservation of 32 amino acids among the four species including 10 cysteines and 5 hydrophobic residues.

Assessment of coupling efficiency

The multiplex response titration curves using dilutions of the anti-GST antibody as the primary antibody in the multiplex reaction were similar for all 4 proteins (Fig. 2). In contrast, the multiplex response titration curves for the GST control bead (coupled at half the protein concentration of the MSP119-GST reactions) and for the cross-linked P. falciparum CSP peptide-GST bead were indicative of lower amounts of bound target protein compared to the MSP119 beads.

Fig. 2
figure 2

Assessment of coupling efficiency using a dilution series of goat anti-GST IgG antibody. Goat anti-GST IgG antibody was diluted 1:1 × 103, 1:1 × 104, 1:5 × 104, 1:1 × 105, 1:5 × 105, 1:1 × 106, 1:5 × 106, and 1:1 × 107 in modified Buffer A lacking the E. coli extract. Following a 1-h incubation (50 μl/well) at room temperature, bound anti-GST antibody was detected with a biotinylated rabbit anti-goat IgG secondary antibody (1:500 dilution in Buffer B, 50 μl/well, 1 h at room temperature). Wells were developed with R-phycoerythrin-labelled streptavidin and read on a BioPlex 200 instrument (BioRad) as described in “Methods

Cut-off determinations

One outlier from the group of 88 US citizens with no history of foreign travel with a MBA response of 8690 MFI-bg units was censored from the P. vivax cut-off calculation, and one outlier with a MBA response of 10,377 MFI-bg units was censored from the P. falciparum CSP peptide calculation. The cut-offs in MFI-bg units were: P. falciparum CSP peptide, 1351; P. falciparum MSP119, 313; P. malariae MSP119, 397; P. ovale MSP119, 65; and, P. vivax MSP119, 86. In an analysis of the residual diagnostic serum panel that included sera from patients having microscopically confirmed and/or PCR confirmed infections, 26 of 33 (79%) of P. falciparum, 5 of 6 (83%) of P. malariae, 5 of 7 (71%) P. ovale, and 33 of 35 (94%) of P. vivax were positive by multiplex assay. The sensitivity of the P. falciparum CSP peptide assay was not determined. Specificities measured from the presumed negative US citizen panel were ≥ 96% for each assay.

Cut-offs determined from all MSP119 values (no outliers censored) using ROC curve analysis [62] were lower for P. falciparum and P. malariae (111 and 237 MFI-bg units, respectively) but higher for P. ovale and P. vivax (175 and 203 MFI-bg units, respectively). J-index analysis [63] yielded identical cut-off values. The ROC cut-off values had no impact on the sensitivities of the assays for P. malariae, P. ovale or P. vivax and either had no impact (P. malariae) or resulted in increases of 3% (P. ovale) or 2% (P. vivax) in specificity. Sensitivity and specificity for the P. falciparum assay using the ROC cut-offs were 88 and 94%, respectively.

In order to maximize specificity and to estimate seropositivity conservatively, the higher of the cut-off values from the various methods of analysis for the MSP119 assays (in MFI-bg units) were chosen: P. falciparum MSP119, 313; P. malariae MSP119, 397; P. ovale MSP119, 175; and, P. vivax MSP119, 203.

MSP119 multiplex assay specificity

If closely related antigens coupled on different beads share common epitopes and compete for the same pool of antibodies, the values from a multiplexed assay would be expected to be lower than the values from assays using a single bead only. To test this hypothesis, each Plasmodium spp. MSP119 was assayed in isolation (individual monoplex), and the results were compared to values obtained when all of the beads were included in the routine multiplex format. As shown in Table 1, responses from 2 defined sera (Pan Plasmodium spp. Lot 8 and P. malariae Lot 2), 3 elutions from individual Mozambique blood spots, and one high-titre elution from a combination of Mozambique blood spots were essentially identical regardless of the bead complexity of the assay (Spearman rank order correlation coefficient = 0.999; P < 0.001). The results from this limited panel of samples suggest that a response dilution effect in the multiplex assay format is not a universal feature of the MSP119 protein family and that the multiplex assay may be useful in infection species determinations.

Table 1 Impact of bead complexity on multiplex bead assay response values using beads coated with MSP119 proteins from four malaria species

That some MSP119 antibody responses are species specific can also be demonstrated using sera from experimentally infected chimpanzees (Table 2). Sera from chimpanzees Klimatis (P. malariae infection) and Duff (P. vivax infection) had high antibody response values to the corresponding species-specific MSP119 protein and had no responses to MSP119 antigens from other species. In contrast, other animals such as chimpanzees Alpert (P. ovale infection) and Bit (P. malariae infection) reacted strongly with the MSP119 antigen corresponding to the species of the infecting parasite, but they also had weak responses to P. vivax MSP119 and strong responses to P. falciparum MSP119. The wild-caught chimpanzees used in the experimental infection studies were never exposed to P. falciparum sporozoites in the laboratory and were never experimentally infected with P. falciparum. Thus, the presence of a P. falciparum CSP peptide response suggests that the P. falciparum MSP119 response likely arose by natural infection with a closely related species of chimpanzee malaria, such as Plasmodium reichenowi [69], rather than with an experimentally-induced cross-reactivity. Similarly, the weak P. vivax antibody responses in these 2 chimpanzees may also reflect prior exposure to P. vivax-like parasites in the wild. These unexpected responses highlight the difficulty of differentiating historic infection from true antibody cross-reaction.

Table 2 Multiplex bead assays results using sera from chimpanzees experimentally infected with a single species of malaria parasite

Specificity analysis using antigen competition

An alternative approach to assess the specificity of the MSP119 antibody response relies on the ability of soluble antigen to saturate the antibody in a pre-incubation step so as to prevent antibody binding to MSP119 coated beads during the multiplex assay. To determine the concentration of competitor protein necessary to prevent antibody binding to MSP119 coated beads, sera that had high antibody responses were incubated with 0.025–10 μg/ml of competitor protein prior to multiplex assay as described in “Methods”. The GST control, PfCSP peptide-GST, and all 4 MSP119-GST protein-coated beads were included in the multiplex assay, but only the homologous MSP119 was used as competitor. Thus, the P. falciparum Lot 6 defined human serum was competed using soluble P. falciparum MSP119-GST fusion protein, while sera from chimpanzees Klimatis (P. malariae), Alpert (P. ovale), and Duff (P. vivax) were each competed using the corresponding species-specific MSP119-GST fusion proteins. Figure 3 shows that the multiplex responses to all 4 MSP119 proteins were reduced > 97% following pre-incubation with 2.5 μg/ml competitor protein, and all 4 MSP119 antibody responses were below their respective cut-off values when sera were pre-incubated with 5 μg/ml competitor protein.

Fig. 3
figure 3

Antibody competition titration assays using homologous MSP119 proteins. Dilutions (1:400) of P. falciparum Lot 6 defined human serum or of sera from chimpanzees experimentally infected with either P. malariae (Klimatis), P. ovale (Alpert) or P. vivax (Duff) were incubated with the indicated concentrations of the homologous MSP119 competitor protein for 1 h at room temperature. Multiplex bead assays were performed as described in “Methods”, and the multiplex responses in MFI-bg units are plotted versus the competitor concentration

Next, sera from chimpanzees Klimatis (P. malariae), Alpert (P. ovale), and Duff (P. vivax) (1:400 dilution of each serum) were combined, and the competitor titration assays were repeated. Figure 4 shows the multiplex responses in the presence of various concentrations of the 4 MSP119 competitor proteins and expressed as a percentage of the PBS control. In Panel A, addition of the P. falciparum MSP119 competitor protein had no effect on the P. malariae, P. ovale or P. vivax multiplex responses. Similarly, heterologous MSP119 competitor proteins had no effect on multiplex responses (Fig. 4b–d), while multiplex response curves for the homologous species of competitor MSP119 proteins in Fig. 4b–d resemble the individual curves previously shown in Fig. 3. The chimpanzee multiplex responses in the presence of homologous species competitor protein showed > 97% suppression at 2.5 μg/ml and were below the respective cut-off values at 5 μg/ml of competitor.

Fig. 4
figure 4

Antibody competition titration assays using MSP119 proteins from four Plasmodium species. A combined dilution (1:400 of each serum) containing sera from chimpanzees experimentally infected with either P. malariae (Klimatis), P. ovale (Alpert) or P. vivax (Duff) was incubated with the indicated concentrations of the MSP119 competitor protein for 1 h at room temperature. Competitor proteins used were: a P. falciparum MSP119; b P. malariae MSP119; c P. ovale MSP119; d P. vivax MSP119. Multiplex bead assays were performed as described in “Methods” and the multiplex response in MFI-bg units are plotted versus the competitor concentration. Multiplex responses are presented as a percentage of the assay results for the PBS control

Finally, combined human sera (pan-Plasmodium spp. positive serum pool and P. malariae mono-specific infection serum control, each at 1:400 dilution) competitor studies showed similar heterologous and homologous titration profiles except that the P. malariae response was reduced by approximately 27–29% at the 10 μg/ml of heterologous MSP119-GST competitor protein concentrations (Additional file 1). For the human multiplex responses in the presence of homologous species competitor protein, values were below the respective cut-off values at 2.5 μg/ml of competitor. Based on these studies, a competitor concentration of 10 μg/ml was selected to maximize suppression of antibody binding in the multiplex assay.

Assay specificity in low malaria incidence settings

Representative competition assay results for a serum sample set from two regions of low malaria incidence (Haiti and Cambodia) are presented in Table 3. Additional results from this sample set can be found in Additional file 2. Most of the samples chosen from these areas had positive antibody responses to only one or two MSP119 proteins, and only one person (Cambodia 5) reacted with MSP119 proteins from three malaria species. Antibody responses to the P. falciparum CSP peptide were mostly negative, and, when present, were < 4000 MFI-bg units (median = 64.5; range 14–3930). Addition of GST control protein to the competition assay at a concentration of 10 μg/ml had no effect on any of the antibody responses. One person had an antibody response to the GST coupled control bead, but the response was not inhibited by pre-incubation with soluble GST protein. This response was probably unrelated to the presence of the GST protein as the P. ovale MSP119-GST response was consistently < 50 MFI-bg units.

Table 3 Representative MSP119 competition assay results using sera from low incidence settings

In 9 of the 12 serum samples tested, all of the malaria MSP119 antibody responses appeared to be species specific as only homologous competitor MSP119 protein completely eliminated the antibody response (highlighted in italics in Table 3 and Additional file 2). For two of the tested sera (Table 3) multiplex assay response values to the P. vivax antigen demonstrated a weak heterologous competition effect with P. falciparum MSP119 competitor protein, but the effect did not meet the 30% threshold definition (approximately 25% reduction; Cambodia 4 and 9). Interestingly, the sample from Cambodia donor 4 had no antibodies to either P. falciparum antigen by MBA. Only one donor had a heterologous competition assay response reduction of > 30%: for Cambodia 5 (indicated in bolditalics in Table 3), addition of P. falciparum MSP119 competitor protein reduced a weak P. malariae response by 41% while completely eliminating a strong homologous P. falciparum MSP119 response (> 28,000 MFI-bg units). The MFI-bg value for the heterologous competition assay remained above the respective P. malariae cut-off.

Assay specificity in a high malaria incidence setting

Representative competition assay results for a sample set from a high malaria incidence region of Mozambique are presented in Table 4 with additional values shown in Additional file 3. These 20 samples were selected from the parent study [41] because they exhibited high IgG antibody responses to one or more MSP119 antigens by multiplex bead assay, and the selection was biased towards samples that had a strong positive responses to the P. vivax, P. ovale or P. malariae proteins. Historically, rates of vivax malaria have been expected to be low in East African populations lacking the Duffy antigen [12], and an antibody response to the P. vivax MSP119 might be indicative of antibody cross-reactivity. In contrast to the samples from low prevalence areas described above, eluted blood spot samples from Mozambique often had strong antibody responses to the P. falciparum CSP peptide (median = 23,407; range = 2833–27,033).

Table 4 Representative MSP119 competition assay results using sera from a high incidence setting

Species-specific anti-MSP119 antibody responses, as indicated by the presence of complete homologous competition and the absence of heterologous competitor effects, were observed in 10 of the 20 samples tested (Table 4 and Additional file 3). In one additional sample (Mozambique 20 in Table 4), species-specific responses were observed, but the suppression of the P. malariae MSP119-specific antibody responses was incomplete: values remained above the 397 MFI-bg assay cut-off threshold in the presence of 10 μg/ml P. malariae competitor protein. Given the very high P. malariae MSP119 control antibody responses for these samples (> 27,000 MFI-bg units), it is possible that the competitor protein concentration was insufficient for complete antibody blocking. As previously demonstrated, addition of GST control protein to the competition assay at a concentration of 10 μg/ml had no effect on the antibody responses.

One sample (Mozambique 1, Table 4) demonstrated a partial loss (40% reduction) of anti-P. malariae MSP119 antibody response in the presence of P. falciparum competitor protein, but antibodies to the other 3 species antigens were specific. Four samples, represented by Mozambique 16 in Table 4, demonstrated some combination of incomplete homologous response suppression and heterologous assay inhibition. In the case of Mozambique 16, the P. falciparum competitor protein partially inhibited the heterologous P. malariae MSP119 antibody response (68% reduction), and the P. vivax competitor protein had a major impact on the P. ovale heterologous response (99% reduction), but, in the same reaction, this competitor did not completely block the homologous P. vivax MSP119 antibody response (90% reduction). Reciprocal heterologous competition was only observed between P. ovale and P. vivax MSP119 antigens and only in two donors represented by Mozambique 19 (Table 4). Heterologous competition assays leading to response values below the cut-off were observed for both P. malariae and P. vivax MSP119 proteins and the P. ovale antigen partially reduced the P. vivax antibody response.

As shown in Table 5, what appeared to be a true, pan-malaria MSP119 antibody response was observed in blood spot elutions from two 20–30 years old Mozambique donors, numbers 3 and 12. In contrast to assays described above, control response values to MSP119 proteins for 3 of the malaria species (P. malariae, P. ovale, P. vivax) were reduced by about 50% upon addition of GST control protein. The reason for the observed signal suppression by GST is not understood, but it was probably not related to the presence of the GST component of the MSP119-GST fusion proteins since no antibody bound to the GST control bead and there was no decrease in the PfCSP peptide-GST response. For both donors, the response to the P. falciparum MSP119 antibody response was less affected by the addition of the GST control protein, and residual P. falciparum MSP119 antibody signal (6–9% of control) was observed in the presence of each of the heterologous competitor proteins. Antibody responses to the other 3 MSP119 proteins in the presence of heterologous competitor proteins were either below or very near the cut-off values for the respective assays (Table 5).

Table 5 Mozambique blood spot elutions that demonstrate multiple cross-reacting MSP119 antibody responses

Affinity purification of MSP119 antibodies

Another potential method to detect antibody cross-reactivity is to affinity purify antibody using an MSP119 protein from a single malaria species and then assess the reactivity of the eluted antibody using MSP119 coated beads from all 4 species. Tetanus toxoid, a protein lacking GST, was included in the multiplex panel as an additional assay control.

Plasmodium vivax MSP119-GST-coated magnetic beads were used to affinity purify antibodies from two Mozambique samples: sample 13, previously shown to have specific responses to all 4 MSP119 proteins; and sample 16, previously shown to have significant cross-reactivity between the P. vivax protein and the P. malariae antibody response (Table 4). In the case of Mozambique sample 13, only the antibody response to the P. vivax MSP119 decreased in the serum dilution after exposure to the magnetic beads (Table 6), and antibodies eluted from the magnetic bead only reacted with the P. vivax bead in the multiplex assay. In a separate experiment with sample 13, captured P. malariae antibodies were eluted from beads coated with the homologous antigen (Additional file 4). For sample 16, antibody responses to both the P. ovale and the P. vivax MSP119 proteins decreased upon exposure of the serum dilution to P. vivax protein-coated magnetic beads, and eluted antibodies reacted to both P. vivax and P. ovale beads in the multiplex assay. Thus, the cross-reactivity previously observed in the competition assays was confirmed for this sample.

Table 6 Affinity purification of MSP119 binding antibodies using magnetic bead capture

Plasmodium falciparum MSP119-GST-coated magnetic beads were used to affinity purify antibodies from two additional Mozambique samples: sample 15, previously shown to have specific responses to all 4 MSP119 proteins (Table 4); and sample 12, previously shown to have a pan-malaria cross-reactive response (Table 5). As expected for a species-specific antibody response, only the P. falciparum MSP119 antibody response decreased in the serum dilution following exposure to magnetic beads, and the elution only contained antibodies that recognized the P. falciparum protein in the multiplex assay (Table 6). Incubation of Mozambique sample 12 with the P. falciparum MSP119-GST-coated magnetic beads drastically decreased the antibody responses to proteins from all 4 species in the post-treatment serum dilution, but positive responses were not observed in the MBA analysis of the eluted antibodies. Plasmodium ovale MSP119 –GST-coated magnetic beads were also used for antibody capture from sample 12 with results similar to those described above (Additional file 4).

Sub-class and avidity studies

The inability to affinity purify and recover antibodies from a highly cross-reactive sample suggested that the antibody response in Mozambique 12 might have some unique features relative to responses that were species specific or weakly cross-reactive. Table 7 shows that, while the Mozambique sample 13 anti-MSP119 antibody responses were predominantly of the IgG1 sub-class, Mozambique sample 16 and 15 responses were a combination mainly of IgG1, IgG2, and IgG3 sub-classes. Of particular interest, the P. falciparum and P. malariae MSP119 responses for both donors had strong IgG3 components, but the P. ovale and P. vivax responses for sample 16 were largely of the IgG2 sub-class. Further, the MSP119 antibody responses observed in Mozambique 13, 16 and 15 samples appeared to be a mixture of high avidity and low avidity antibodies as determined by the 6 M urea treatment, with responses to the P. falciparum MSP119 having a high avidity (≥ 0.98) and responses to the P. vivax MSP119 being mainly low avidity (ratio ≤ 0.12). This low avidity, however, did not prevent antibody capture and elution using P. vivax MSP119-coated beads shown in Table 6.

Table 7 IgG sub-class and antibody avidity index for samples from Mozambique

The pan-malaria cross-reactive response of Mozambique 12 was completely different. The response to MSP119 proteins from all four malaria species was almost exclusively of the IgG3 sub-class, and the entire IgG response was low avidity with avidity index values of 0.01–0.03 (Table 7). The donor was clearly capable of making high avidity IgG antibodies of other sub-classes as evidenced by the responses to the P. falciparum CSP and tetanus toxoid (Table 7). Unfortunately, this observation could not be confirmed using the other highly cross-reactive sample (Mozambique 3) as no additional antibody eluate was available for testing.

Discussion

Serologic antibody responses to malaria MSP119 antigens are increasingly used to map geographic distributions and transmission intensities of malaria infection, but questions about the specificity of the responses remain incompletely explored [17,18,19]. Genomic sequence analysis demonstrates limited allelic variability within species (often only 2–3 amino acids), but considerable sequence heterogeneity between species (this work [35, 38, 65, 66, 67]). In a recent serologic IgG antibody survey of two communities in northern Mozambique [41], a non-trivial 2–4% prevalence for IgG antibodies to P. vivax MSP119 antigen was observed in a population that is expected to be ≥ 95% negative for the Duffy marker used for RBC invasion [10,11,12, 70]. The current study was undertaken to determine whether these unexpected responses represented antibody cross-reactivity resulting from the transmission of P. malariae and P. ovale in the context of intense P. falciparum infection or whether they represented true P. vivax infections [41].

First, monoplex bead assays using a single MSP119 antigen were compared to multiplex bead assays that included beads coated with antigens from all 4 species as well as GST control and PfCSP peptide coupled to GST. The monoplex versus multiplex results for all 4 MSP119 antigens using a panel of 2 sera and 4 blood spot elutions with a range of antibody response values were virtually identical, and no response dilution effect was detected. Similar results were previously reported by Kerkhof et al. [31] using the P. falciparum and P. vivax MSP119 antigens and 3 different positive control serum dilutions. However, the observation that a two-fold increase in the number of beads used per assay well had only marginal effects on the measured P. falciparum and P. vivax MSP119 antibody responses [31] suggests that this technique may not be a sensitive method for identifying partial cross-reactivity.

Second, banked sera from chimpanzees infected with a single species of malaria in a controlled laboratory setting were tested by MBA. While all 4 animals had homologous antibody responses to the laboratory-administered parasite infection, 2 of the animals also had weak heterologous antibody responses to the P. vivax antigen and strong heterologous responses to the P. falciparum antigen despite the fact that they had never been infected with either of these parasites in the laboratory. Responses to the PfCSP antigen were also observed despite the lack of laboratory exposure to P. falciparum sporozoites. The presence of the P. falciparum MSP119 and CSP responses strongly suggested that infections had occurred in the wild, perhaps with one of the Laveranian great ape malaria species that are genetically very similar to P. falciparum [69]. Muerhoff et al. came to the same conclusion regarding a P. falciparum MSP119 response observed in chimpanzee sera by ELISA [36]. The absence of pre-exposure baseline sera for the chimpanzees meant that it was impossible to discriminate between cross-reactive responses resulting from the laboratory infections and pre-existing responses from infections acquired in the wild prior to capture.

Suppression of antibody binding by pre-absorption with excess heterologous or homologous MSP119 antigen should be a sensitive method for the identification of cross-reactive antibody responses in MBA. Amanfo et al. [42] used 2 sera with a competition ELISA technique to demonstrate a lack of cross-reactivity between P. falciparum, P. ovale and P. malariae MSP119 antigens (the P. vivax antigen was not included in their analysis). In the third part of this current study, 12 samples from low transmission areas in Haiti and Cambodia and 20 samples from a high transmission area in Mozambique were used to assess cross-reactivity by antigen competition MBA. Eleven of 12 sera from residents of the low transmission areas had MSP119 antibody responses that were completely species specific. Only one individual had a heterologous competition response decrease that met the > 30% reduction definition. In the Mozambique sample set, antibody responses for 12 of the 20 samples tested were totally species specific, and 6 of these 12 samples were positive for antibodies to all 4 malaria parasite species. For 6 additional Mozambique samples, the P. falciparum MSP119 responses were species specific, but various levels of incomplete heterologous competition were observed for the non-P. falciparum assays ranging from a 31 to 99% response reduction. The high specificity of the P. falciparum assay may reflect the affinity maturation of the immune response upon repeated infection with P. falciparum in the high intensity transmission setting of Mozambique. Most heterologous competition was non-reciprocal, suggesting that infection with one malaria species elicited both specific and cross-reactive antibodies while infection with the other malaria species resulted in only specific antibodies. Most commonly, P. malariae responses cross-reacted with P. falciparum antigen. Two examples of reciprocal heterologous competition were also identified, and both of these involved P. vivax and P. ovale responses. Whether higher concentrations of competitor MSP119-GST protein (> 10 μg/ml) might have resulted in more complete heterologous competition of these responses was not determined.

Two individuals were identified who had very high responses to all 4 MSP119 antigens (> 9000 MFI-bg units) and who appeared to have pan-malaria MSP119 antibody responses by antigen competition MBA. However, perhaps because of the very high levels of antibodies generated by intense levels of P. falciparum transmission, heterologous antigens could only partially suppress the P. falciparum antibody response. These 2 samples represent only 15% of the 13 samples that were positive for antibodies to all four malaria species in the high transmission area sample set, and it should be noted that the sample set was not randomly selected from the overall Mozambique bed net study population. In fact, samples with high responses to the non-P. falciparum MSP119 antigens were intentionally chosen in an attempt to identify those ‘worst case scenario’ samples where cross-reactivity might be observed. Because only 40 samples from the overall Mozambique bed net study set (N = 2408) had responses above the cut-off values to all 4 MSP119 antibodies [41], the number of potential pan-malaria reactive individuals in Mozambique is likely quite low (< 0.3%).

Finally, an antibody capture/elution technique was used with the MBA to confirm the results of the competition assays described above. While species specific and partially cross-reactive MSP119 antibodies could be eluted from capture beads, appreciable quantities of captured antibodies could not be recovered from the pan-malaria responsive DBS elution despite repeated attempts with multiple capture antigens. Further analysis of the samples from the species specific and partially cross-reactive donors revealed IgG responses of varying avidity dominated by the IgG1 and IgG3 sub-classes. Others have reported that exposure to P. falciparum MSP119 elicits a mixed pattern of IgG1 and IgG3 antibodies and that repeated infection causes a shift towards an IgG1-dominated response [35, 71,72,73,74,75,76,77,78]. The species specific and partially cross-reactive results presented here are consistent with those reports. In contrast, the pan-malaria response from Mozambique sample 12 exhibited very low avidity binding to MSP119 antigens from all 4 malaria species and was skewed entirely to the IgG3 sub-class. Low avidity responses to the P. falciparum MSP119 are relatively rare [74], and only one previously reported example of a mixed IgG1/IgG3 response that skewed almost entirely to an IgG3 response upon repeat infection with P. falciparum was found in the literature [78]. At present, it cannot be determined whether these observations result from host-specific factors or are a universal feature of pan-malaria responses, nor can any definitive conclusions be drawn about the impact of such responses on malaria immunity or potential malaria pathology.

Previous studies on allele-specific antibody responses to P. falciparum MSP119 and apical membrane antigen 1 (AMA1) suggested that children develop allele-specific responses upon primary infection and that the prevalence of cross-reactive antibodies to conserved epitopes increases with age and increasing experience of infection [79, 80]. The number of samples in this study was too small to definitively address the issue of age as a proxy for infection experience and the development of cross-reactive antibody responses. However, two of the partially cross-reactive samples from Mozambique were from 5-years old donors, and 5 of the donors with specific antibody responses to all 4 malaria species were > 50 years of age. Simultaneous infection with multiple malaria species, which is known to occur in Mozambique [81], might play a larger role in the development of antibody responses against shared MPS119 epitopes than total infection experience [82]. Thus, even in a high transmission setting with multiple co-endemic malaria species, most antibody responses to P. falciparum, P. malariae, P. ovale, and P. vivax MSP119 antigens are likely indicative of previous infection history with those parasite species.

Conclusions

Globally, most areas of malaria transmission are seldom mono-specific. In sub-Saharan Africa, P. falciparum is the most prevalent infection with the highest intensity of transmission, but significant transmission attributable to P. malariae, P. ovale, and, in some areas, P. vivax occurs. In South and Central America, P. malariae, P. falciparum, and P. vivax are transmitted endemically whereas in Asia all four human malaria species can be transmitted. MSP-119 is a major antigen recognized by the IgG antibody response of a majority of exposed individuals in an endemic population. Malaria control efforts would likely benefit from being able to rapidly and easily monitor immune responses not only for the main targeted species such as P. falciparum and P. vivax, but also the lesser species, P. malariae and P. ovale. The analyses presented in this work indicate that the multi-species MSP-119 multiplex bead assay will be a useful tool in future malaria epidemiologic surveillance and control program studies.

Footnotes

Use of trade names is for identification only and does not imply endorsement by the Public Health Service or by the US Department of Health and Human Services. The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.