Identification and serological responses to a novel Plasmodium vivax merozoite surface protein 1 (PvMSP-1) derived synthetic peptide: a putative biomarker for malaria exposure

Study population

This study enrolled serum samples from 156 adult P. vivax-infected patients aged 18 to 68 years (median age: 38 years). The majority of the samples (75%) were obtained from male subjects between February 2006 and January 2008. Patients were selected based on classic malaria symptoms (fever, chills, headache, myalgia, nausea, and vomiting), seeking medical attention at Hospital Universitário Júlio Muller in Cuiabá (HUJM-CB), Mato Grosso State, Brazil. Notably, active malaria transmission does not occur in Cuiabá. Patients reported short exposure to other areas in the Brazilian Amazon where malaria is endemic, suggesting potential infection acquisition during these visits. Indeed, the cumulative exposure in the endemic regions was limited to less than one year, exhibiting a median incidence of three previous malaria cases. Our analysis also included 15 patients with P. falciparum infection, selected based on identical criteria at HUJM-CB during the same period. All patients were diagnosed by thick blood smear microscopic observation and the infection was confirmed by a species-specific nested PCR amplification of the Plasmodium spp. 18S SSU rRNA gene (Scopel et al., 2004). As controls, 36 malaria-naïve individuals residing in a non-endemic area (Belo Horizonte, state of Minas Gerais, Brazil) with no prior exposure to malaria transmission were also included. Blood was collected by vein puncture into EDTA tubes, plasma samples were separated by centrifugation (520× g for 15 min) and stored at −20 °C until serological experiments were conducted. The sera tested comprise a Biorepository at the Malaria Laboratory/ Universidade Federal de Minas Gerais (UFMG), adhering to Ethics Committee of the National Information System on Research Ethics Involving Human Beings regulations (SISNEP–CAAE: 01496013.8.0000.5149). All individuals included in this study were anonymized and provided written informed consent for collecting samples and subsequent analysis. A detailed description of the characteristics of P. vivax-infected patients has been published elsewhere (Morais et al., 2005; Mourão et al., 2012; Mourão et al., 2016).

Peptide SPOT-synthesis and immunodetection of reactive peptides

A total of 580 overlapping 15-mer peptides, frame-shifted by three residues, covering the entire amino acid sequence of PvMSP-1 Sal I strain (Gene PVX 099980) (Putaporntip et al., 2002) were simultaneously synthesized using an automatic synthesizer ResPep SL (Intavis) according to the SPOT-synthesis technique (Frank, 2002). Briefly, the peptides were built on specific spots on a derivatized cellulose membrane (Intavis), using Fmoc chemistry. The Fmoc group was removed with 4-methylpiperidine 25% v/v in dimethylformamide (DMF; Merck, Rahway, NJ, USA). The amino acids were activated using diisopropylcarbodiimide (DIC; Sigma-Aldrich, St. Louis, MO, USA) and Oxyma Pure (Sigma-Aldrich) at 1.1 M (1:1) and deposited onto the membrane for coupling. After, non-reactive amine residues were acetylated with acetic anhydride (3% v/v in DMF). These steps were repeated until all the amino acids were added. At the end of the synthesis, the membrane was submerged for one hour in a cleavage solution composed of 95% (v/v) trifluoracetic acid (TFA; Sigma Aldrich) associated with 2.5% (v/v) water and 2,5% (v/v) triisopropylsilane (TIPS; Merck) to remove protective groups from the amino acids side chains. After, the membrane was washed 4 times with dichloromethane (DCM; Sigma-Aldrich), 4 times with DMF, and 2 times with ethanol (Sigma-Aldrich). The membrane was dried and the spots were checked with UV light.

Immunoblotting assays were performed using pools of plasma from P. vivax-infected individuals (n = 10), P. falciparum-infected patients (n = 10), and healthy donors (n = 10). Briefly, the membrane was blocked overnight with PBS containing 3% (w/v) bovine serum albumin (BSA; Sigma-Aldrich) and 5% (w/v) sucrose. The blocked solution was removed, and the membrane was washed three times with PBS with 0.1% Tween (PBST) with agitation for 10 min each. Antibodies from each pool diluted 1:1000 in PBST were added to the membrane, one at a time, for 2 h at room temperature, under vigorous stirring. After each incubation, the membrane was rewashed three times with PBST 0.1% for 10 min. Antibody binding was detected using HRP-conjugated anti-human IgG (Sigma-Aldrich) diluted 1:10000 on PBST, and was incubated for 1 h at room temperature (23 °C) following three washes with PBST for 10 min. Luminata (Millipore/Merck, Rahway, NJ, USA) was added to the membrane as a substrate. The membrane was automatically exposed, scanned, and digitalized by an ImageQuant LAS 500 chemiluminescence detector (GE Healthcare, Chicago, IL, USA).

Membrane analysis and spot quantification

The spot intensity of each peptide was quantitatively measured by densitometry scanning analysis using the software Image J–Protein Array Analyzer plug-in (http://image.bio.methods.free.fr/ImageJ/?Protein-Array-Analyzer-for-ImageJ.html) as described elsewhere (Reis-Cunha et al., 2022). Briefly, densitometry values were normalized and the mean densitometric value of the four less reactive peptides from each experiment was subtracted from the densitometric value of each spot of the same membrane. The cut-off value was defined based on the mean plus two times the standard deviation of the reactivity of all spots in the membrane incubated with non-infected pool sera.

In silico peptide validation

B cell linear epitopes were predicted in silico using the program BepiPred 2.0 (https://services.healthtech.dtu.dk/services/BepiPred-2.0/) as described before (Bueno et al., 2011). Briefly, based on a single FASTA sequence input each amino acid receives a prediction score based on a random forest algorithm trained on epitopes annotated from antibody-antigen protein structures. The cut-off value of 0.5 was used to validate a linear B cell epitope. Additionally, the amino acid sequence was analyzed using the Protein BLAST^® (https://blast.ncbi.nlm.nih.gov/Blast.cgi) to determine the specificity of those sequences for P. vivax.

Synthesis of soluble peptides

The 15-mer peptides selected by immunoblotting assays were synthesized on a 10 µmol scale by SPPS technique using the automatic synthesizer ResPep SL (Intavis). Briefly, the Fmoc protective group was removed using 25% 4-methylpiperidine (v/v in DMF), and amino acids were activated with a 1:1 solution of Oxyma Pure (Merck) and DIC (Sigma-Aldrich). The amino acids were incorporated into an H-Rink Amide ChemMatrix resin (Sigma-Aldrich) and non-reagent amino groups were acetylated with acetic anhydride (3% v/v in DMF). These steps were repeated until the synthesis of each peptide was fully completed. Peptides were released from the resin using a cleavage cocktail as previously described (Siqueira, 2023). Briefly, the peptides were precipitated by adding cold methyl tert-butyl ether and then lyophilized. We performed a MALDI-TOF spectrometry to check each synthesized peptide’s mass-to-charge ratio (m/z) H+ using a saturated matrix solution containing 10 mg/mL α-cyano-4-hydroxycinnamic. The peptide: matrix mixture (1:2) was applied on an MTP AnchorChip™ 600/384 plate (Bruker Daltonics) and maintained to dry at room temperature.

Serological assays

IgG and IgG subclass antibody responses against the selected peptides were detected by enzyme-linked immunosorbent assay (ELISA). Briefly, each well of a 96-well flat-bottomed polystyrene microplate (Corning Incorporation, Corning, NY, USA) was coated with 0.5 µg/µL of peptide in bicarbonate–carbonate buffer (pH 9.6; 0.1 M) and incubated overnight at 4 °C. Plates were blocked with BSA 3% (w/v) in PBS pH 7.4, for 1 h at 37 °C. Then, each plasma diluted 1:100 in PBST 0.05% containing 3% BSA was tested in duplicate and incubated for 2 h at 37 °C following four washes with PBST 0.05%. HRP-conjugated anti-human (Sigma-Aldrich) IgG (1:2000) or anti-human IgG1, IgG2, IgG3, and IgG4-biotin antibodies were diluted (1:500) in 0.05% PBST with 3% BSA and incubated for 90 min at 37 °C. For IgG subclass detection, HRP-conjugated streptavidin (RD systems, Minneapolis, MN, USA) (1:1000) diluted in 0.05% PBST was added for 60 min at 37 °C. Antibody reaction was revealed using 0.5 mg/mL o-phenylenediamine dihydrochloride (OPD) substrate (Sigma-Aldrich) diluted in 0.05 M phosphate-citrate buffer pH 5.0 and the reaction was stopped with 4M H₂SO₄. The optical density (OD) value at 492 nm was determined in a Multiskan GO Reader (Thermo Fisher Scientific, Waltham, MA, USA) microplate spectrophotometer. Levels of IgG and IgG subclasses were expressed as reactivity index (RI). To establish a cut-off, the mean OD was calculated by adding three standard deviations of samples from 10 naïve volunteers who had no previous exposure to malaria. The RI was calculated as a ratio between the mean OD of each patient’s duplicate sample by the cut-off value. Any reactivity above 1 was considered responsive.

Statistical analysis

Data were analyzed using GraphPad Prism 8.0 software (GraphPad Software, La Jolla, CA, USA). To assess if the data fit in a Gaussian distribution, we used the Shapiro–Wilk normality test. Comparisons between groups were analyzed by the Kruskal–Wallis test followed by Dunn multiple comparison tests. Receiver operating characteristic (ROC) analysis was performed to obtain sensibility, specificity, and accuracy values for IgG, IgG1, and IgG3 serological responses against p314. We used the Principal Component Analysis (PCA), an effective technique used to reduce the dimensions of a dataset by replacing the original variables with a smaller set of derived variables (principal components, PCs), to analyze the correlation between IgG subclass levels, malaria exposure, and parasitemia. To accomplish this, we assessed the proximity of position and value of loadings contribution to each variable (Zhang & Castelló, 2017; Beattie & Esmonde-White, 2021). This analysis was performed in the R software (4.3.0 version) using factoextra (https://cran.r-project.org/web/packages/factoextra/index.html), ggplot2 (https://cran.r-project.org/web/packages/ggplot2/index.html) and MASS (https://cran.r-project.org/web/packages/MASS/index.html) packages. A p-value <0.05 was considered statistically significant.

PvMSP-1 peptides are exclusively recognized by antibodies from P. vivax-infected patients

We employed epitope mapping to pinpoint the immunodominant epitopes within PvMSP-1 exclusively recognized by antibodies from P. vivax-infected patients. The peptide recognition patterns by antibodies from each examined group are illustrated in Fig. 1A, revealing that antibodies from P. vivax-infected individuals recognize a greater number of peptides in comparison to other groups. We further examined whether the antibody response exhibited variations against the peptides constituting the conserved and polymorphic blocks of PvMSP-1 following the categorization by Putaporntip et al. (2002). Antibodies from P. vivax-infected patients notably exhibited a higher reactivity to both conserved and polymorphic blocks than antibodies from individuals of other groups (Fig. S1). A Venn diagram (Fig. 1B) illustrates the peptide recognition by antibodies from each studied group, depicting 255 peptides (53.5%) recognized by antibodies from subjects of the three groups. Of the peptides recognized by a single group, P. vivax patients recognized 142 peptides (29.8%); non-infected individuals, 11 peptides (2.3%); and P. falciparum- infected patients, 8 peptides (1.7%). Figure 1C depicts the twelve most reactive peptides exclusively recognized by antibodies from P. vivax-infected patients. Among these, we selected two, p70 and p314, for further investigation due to their highest relative intensity values. Furthermore, in silico analysis revealed six/seven amino acids as B-cell epitopes, LKISDK, and DELDLLF for p70 and p314, respectively (Fig. 2). Concerning the homology between p314 and p70 and other FASTA sequences available on the NCBI website for Plasmodium spp., our analysis revealed that the p314 sequence exhibited alignment solely with the PvMSP-1 (Table S1), thereby confirming the specificity of this peptide. Conversely, the p70 sequence exhibited alignment not only with the PvMSP-1 but also with homologous proteins of P. coatneyi and P. cynomolgi, non-human primate species occurring in Southeast Asia (Eyles, Coatney & Getz, 1960; Ta et al., 2014; Zhang et al., 2016) (Table S1).

Peptide p314 is highly recognized by IgG from P. vivax-infected patients

IgG responses to p70 and p314 were assessed by ELISA and results were expressed as reactivity index. Antibody response against p70 exhibited median values below the cut-off value in all studied groups (Fig. 3A). On the other hand, P. vivax-infected individuals exhibited higher levels of IgG against p314 compared to healthy individuals (p = 0.001) and those infected with P. falciparum (p = 0.0001) (Fig. 3B). Based on these results, we further investigate IgG subclass responses only against p314.

Cytophilic IgG subclasses against p314 are higher in vivax malaria and correlate with parasitemia

Our findings indicate that among the IgG subclasses targeting p314 in plasma from P. vivax-infected individuals, IgG3 predominates, followed by IgG1 (Fig. 4). IgG1 and IgG3 levels were significantly three times higher in individuals infected with P. vivax compared to non-infected and P. falciparum-infected patients. Statistically significant differences were observed between subjects with vivax malaria and the other studied groups (IgG1 p < 0.0001, IgG3 p < 0.001). However, IgG2 and IgG4 exhibited median values for all groups below the positivity threshold (Fig. S2).

Aiming to pinpoint the predominant anti-p314 IgG subclass among 120 positive P. vivax patients, a Venn diagram was created. The distribution of anti-p314 IgG subclasses indicated that IgG1 was positive in 69 patients (57.5%), IgG2 in 52 (43.3%), IgG3 in 102 (85%), and IgG4 in 20 (16.7%) (Fig. 5A). Among those 120 patients, four (3.3%) individuals were positive only for IgG1, 11 (9.2%) only for IgG2, 24 (20%) only for IgG3, and one (0.8%) only for IgG4. Altogether, 80 patients (66.7%) displayed a combination of at least two IgG subclasses, with the most prevalent being a mixed response of IgG1 and IgG3 in 64 ones (53.3%). Next, we correlated the levels of IgG subclass responses with parasite burden and malaria exposure (number of previous malaria episodes). Our findings indicate that IgG3, closely trailed by IgG1, emerges as the predominant factor strongly linked to the parasitemia. Notably, these factors display contrasting trends compared to individuals without malaria infection (Fig. 5B).

The ROC analysis reveals a similar pattern between anti-p314 IgG and IgG3 responses

To assess the efficacy of anti-p314 antibodies, we analyzed Receiver Operating Characteristic (ROC) curves, evaluating the sensitivity, specificity, and accuracy of IgG, IgG1, and IgG3 immunoglobulins in our ELISA protocol (Fig. 6A). IgG anti-p314 exhibited a specificity of 76.47% and a sensitivity of 77.78% in distinguishing P. vivax-infected individuals from those never exposed to malaria parasites (Fig. 6B). For IgG1, we observed a specificity of 49.11% and a sensitivity of 48.39%. In the case of IgG3 anti-p314, the analysis indicates a sensitivity of 77.30% and a specificity of 78.57%. Considering these results, our findings highlight a well-defined antibody response against p314 characterized by IgG and IgG3 (Fig. 6B).