Evaluation of transmission-blocking potential of Pv22 using clinical Plasmodium vivax infections and transgenic Plasmodium berghei

Abstract

Antigens expressed during the sexual development of malaria parasites are transmission-blocking vaccine (TBV) targets. Pb22, a protein expressed and localized to the plasma membrane of gametes and ookinetes in Plasmodium berghei, is an excellent TBV candidate. Here, we evaluated the TB potential of the Plasmodium vivax ortholog Pv22 using a transgenic P. berghei parasite line and P. vivax clinical isolates. The full-length recombinant Pv22 (rPv22) protein was produced and used to immunize mice and rabbits to obtain antibodies. We generated a transgenic P. berghei line (TrPv22Pb) by inserting the pv22 gene into the pb22 locus and showed that Pv22 expression completely rescued the defects in male gametogenesis of the pb22 deletion parasite. Since Pv22 in the transgenic parasite showed similar expression and localization patterns to Pb22, we used the TrPv22Pb parasite as a surrogate to evaluate the TB potential of Pv22. In mosquito feeding assays, mosquitoes feeding on rPv22-immunized mice infected with TrPv22Pb parasites showed a 49.3–53.3% reduction in the oocyst density compared to the control group. In vitro assays showed that the rPv22 immune sera significantly inhibited exflagellation and ookinete formation of the TrPv22Pb parasites. In a direct membrane feeding assay using three clinical P. vivax isolates, the rabbit anti-rPv22 antibodies also significantly decreased the oocyst density by 53.7, 30.2, and 26.2%, respectively. This study demonstrated the feasibility of using transgenic P. berghei parasites expressing P. vivax antigens as a potential tool to evaluate TBV candidates. However, the much weaker TB activity of Pv22 obtained from two complementary assays suggest that Pv22 may not be a promising TBV candidate for P. vivax.

1. Introduction

Plasmodium vivax is widespread throughout Southeast Asia and is recognized as an important source of malaria morbidity [1]. Although it is generally associated with milder symptoms than Plasmodium falciparum, P. vivax infection can also result in severe illness and death [1, 2]. P. vivax forms dormant hypnozoites in the liver, which provoke recurrent attacks over the months and prolong the period of transmission [3, 4]. Gemetocytogenesis occurs early in P. vivax before the appearance of clinical symptoms, which favors transmission to mosquitoes. In addition, the emergence of P. vivax resistance to frontline treatment also poses a significant challenge to managing P. vivax cases [5, 6]. In light of the increased predominance of P. vivax in many elimination scenarios, it is imperative to develop innovative strategies to disrupt parasite transmission, including transmission-blocking vaccines (TBVs).

TBVs typically target antigens expressed during the sexual stages of the parasite, which are obligative for the transmission of the parasite to mosquito vectors [7, 8]. In principle, antibodies against the sexual stage antigens, once ingested by mosquitoes, could interrupt transmission by inhibiting single or multiple steps of parasite development, including gametogenesis, fertilization, ookinete development, and midgut invasion. Depending on the expression time, TBV antigens are divided into two broad classes – pre-fertilization and post-fertilization antigens. Pre-fertilization antigens, such as P48/45 and P230, are expressed in gametocytes and gametes [9–11], whereas post-fertilization antigens such as P25 and P28 are expressed on the surface of zygotes and ookinetes [12]. Although TBV development for P. vivax has received much attention, only a limited number of candidates have been identified [13, 14], including the pre-fertilization antigens Pvs230 [15], Pvs48/45 [16, 17], and PvHAP2 [18], and post-fertilization antigens Pvs25 and Pvs28 [19]. The development of vaccines for P. vivax has been hampered, partly owing to the lack of a continuous in vitro culture system to propagate P. vivax [20]. Thus, direct membrane feeding assay (DMFA) is usually used to assess the transmission-blocking activity (TBA) of antibodies against P. vivax antigens [21]. However, DMFA for P. vivax can be performed in a few laboratories in the world since it requires access to P. vivax patients. Furthermore, different sources of gametocytes also contain numerous uncontrolled factors, which introduce large variability among assays [22]. More recently, transgenic rodent parasites expressing human parasite proteins have proved to be a valuable tool to evaluate malaria vaccine candidates, including TBV antigens [23].

To discover new TBV antigens, we identified Pb22 as a conserved protein across the genus Plasmodium. Pb22 was expressed intracellularly in schizonts and gametocytes, but became associated with the plasma membranes of gametes and ookinetes [24]. In P. falciparum, Pf22 was also expressed in sexual stages and the Pf22 protein was primarily detected in female gametocytes [25, 26]. Its ortholog in P. vivax, Pv22, was detected through transcriptomic analysis in asexual blood stages, liver stage and salivary gland sporozoites [27–29]. The functions of P22 orthologs in human malaria parasites have not been reported. In P. berghei, Pb22 is involved in male gametogenesis, and antibodies against Pb22 showed strong TBA in immunized mice. Given the conservation of P22 in Plasmodium and its excellent TBA in P. berghei, we selected Pv22 to determine its TB potential against this human malaria parasite. Using a transgenic P. berghei parasite expressing Pv22 and clinical P. vivax isolates, we conducted in vivo mosquito feeding and DMFA and demonstrated that antibodies against Pv22 had substantial transmission-reducing activity (TRA).

2. Materials and methods

2.1. Mice, parasites and mosquitoes

Female BALB/c mice and New Zealand white rabbits were purchased from Beijing Animal Institute. The P. berghei ANKA strain 2.34 was maintained by serial passage and used for challenge infection as described previously [30]. The Δpb22 parasite used for generating a transgenic parasite expressing Pv22 was from an earlier study [24]. Anopheles stephensi (Hor strain) and Anopheles dirus mosquitoes were fed on a 10% (w/v) glucose solution and kept in an insectary under 25°C and 50–80% relative humidity. All animal procedures were carried out per the welfare and ethical review standards of China Medical University.

2.2. Sequence analysis

The Pb22 (PBANKA_0305900) protein sequence and its ortholog in P. vivax (PVX_003895) were retrieved from PlasmoDB (http://plasmodb.org), aligned using ClustalW, and visualized in the BioEdit sequence editor program. Signal peptide and low-complexity region were predicted by the SMART program (http://smart.embl-heidelberg.de/).

2.3. Expression of recombinant Pv22 (rPv22) and immunization

A 196 amino acid (aa) fragment of Pv22 encoding aa 21–216 was synthesized (Genescipt, China) and cloned into pET28a for expression in Escherichia coli. The rPv22 protein was induced with 1 mM isopropyl-β-D-thiogalactopyranoside (Sigma) at 19°C for 8 h. The N-terminus of rPv22 was fused with a 6×His tag. The recombinant protein was purified on the Ni-NTA column (Millipore), followed by dialysis in phosphate-buffered saline (PBS, pH 7.4) at 4°C overnight. The purified protein was analyzed on a 10% SDS-PAGE gel and quantified by a BCA Protein Assay Kit (Beyotime). The glutathione S-transferase (GST) protein was produced as described previously [18] and used as a negative control for immunization.

To generate polyclonal antibodies against rPv22 and GST proteins, two rabbits were immunized subcutaneously, each with 500 μg of purified proteins emulsified with complete Freund’s adjuvant (Sigma). Two booster immunizations with 250 μg proteins in incomplete Freund’s adjuvant were performed at a three-week interval. Antisera were collected 14 days after the last immunization and pooled, while total antibodies were purified using Protein A columns. Antibody concentrations were determined using the BCA Protein Assay Kit. The IgGs were adjusted to 4.4 μg/μL and used for the mosquito membrane feeding assays.

2.4. Enzyme-linked immunosorbent assay (ELISA)

Antibody titers were assessed by ELISA as described previously [31]. Briefly, 96-well plates were coated overnight at 4°C with the purified rPv22 (5 μg/ml) in 0.05 M sodium carbonate buffer (pH 9.6). The plates were washed three times with PBS-T (0.05% Tween-20 in 0.1 M PBS, pH 7.4) and blocked with 1% bovine serum albumin (BSA, Sigma) for 1 h at 37°C. The sera were first diluted at 1:1000 in PBS with 1% BSA, then serially diluted in a 96-well plate. The primary antibodies were incubated at 37°C for 2 h. After three washes, horse-radish peroxidase (HRP)-conjugated goat anti-rabbit IgG antibodies (Invitrogen) at 1:5000 were added and incubated at 37°C for 1 h. After the final six washes, color development was performed using 100 μL of tetramethyl-benzidine (Amresco) in the dark for 10 min, then absorbance was detected at 490 nm wavelength. Endpoint titer was defined as above the cut-off value of the control antisera + 3×standard deviation (SD).

2.5. Generation of transgenic P. berghei expressing Pv22

To generate a transgenic P. berghei expressing Pv22 with a 3×HA tag referred to as TrPv22Pb, the complete pv22 open reading frame flanked by the 5’ and 3’ UTR of pb22 was inserted into the pUC57 vector at the HindIII and NotI sites. The plasmid was linearized and electroporated into purified Δpb22 schizonts [24]. Twenty-four hours after transfection, the parasites were injected intraperitoneally into a mouse and selected by 5-fluorocytosine (Sigma) for 4 days [32]. The TrPv22Pb parasites were subsequently cloned by limiting dilution. Diagnostic PCR was used to confirm the correct integration of pv22 into the Δpb22 genome at the pb22 locus and simultaneous removal of the hdhfr∷yfcu selection cassette. All the primer sequences are listed in Table S1.

2.6. Phenotypic analysis of the TrPv22Pb parasites

For phenotypic comparison, two groups of mice (3/group) were infected with 5×10⁶ infected RBCs with either the wild-type (WT) P. berghei or the TrPv22Pb parasites. Giemsa-stained blood smears were used to monitor levels of daily parasitemia, gametocytemia (mature gametocytes per 10⁴ RBCs) and gametocyte sex ratio (female/male). Exflagellation centers, macrogamete and ookinete numbers were determined as previously described [11]. In short, 10 μL of infected blood were added to 40 μL ookinete culture medium. After incubation at 25°C for 15 min, 1 μL of the culture was spotted onto a multi-well slide (Matsunami Glass Ind., Ltd., Japan), and exflagellation centers were observed under a light microscope at 400× magnification. To count the macrogametes, 10 μL of infected blood were mixed with 90 μL ookinete culture medium and incubated at 25°C for 15 min. Then, 0.5 μL of the mixture was placed on a slide, and macrogametes were identified by positive labeling with the anti-Pbs21 sera (1:500) but negative with the anti-Ter119 antibody. The culture was further incubated at 19°C for 24 h, and the ookinete numbers in 0.5 μL mixture were counted after staining with the anti-Pbs21 sera.

A mosquito feeding experiment was conducted to study the development of TrPv22Pb parasites in mosquitoes. Female A. stephensi mosquitoes starved for 8 h were allowed to feed on an anesthetized mouse with a parasitemia of 5–7% for 30 min. Engorged mosquitoes were kept for 12 days, and 30 mosquitoes in each group were dissected, and midguts were stained with 0.5% mercurochrome. The number of oocysts per infected mosquito was counted to evaluate the prevalence and intensity of infection. For parasites transmission from mosquito to mouse, infected mosquitoes (10/mouse) at 21 days post blood meal were allowed to feed on three naive mice for 30 min, and parasitemia was monitored for 14 days post mosquito bite.

2.7. Purification of Plasmodium at different stages

TrPv22Pb parasites of different developmental stages were purified on Nycodenz gradients as previously described [33]. For purifying schizonts, infected mouse blood (~5% parasitemia) was collected and cultured in a complete schizont culture buffer (RPMI 1640, 50 mg/L penicillin, 50 mg/L streptomycin, 100 mg/L neomycin, 25% [v/v] FBS, 6 U/mL heparin) for 16 h at 37°C. Schizonts were purified on a 55% (v/v) Nycodenz (Axis-Shield). To purify gametocytes, infected mice (~15% parasitemia) were treated with 20 mg/L sulfadiazine (Sigma) in drinking water for 48 h to kill off asexual blood stages. Gametocytes were fractionated on a 48% (v/v) Nycodenz gradient at 4°C to avoid activation of gametocytes [34]. Early asexual stage parasites (ring-trophozoite) were obtained after removal of gametocytes and schizonts using 48% (v/v) and 55% (v/v) Nycodenz gradients. Above parasites were released from RBCs by 0.15% saponin lysis for 10 min on ice and washed twice with PBS. For ookinetes, purified gametocytes were cultured in a complete ookinete culture medium for 24 h at 19°C. Ookinetes were purified on a 62% (v/v) Nycodenz gradient.

2.8. Western blot

Purified parasites were resuspended in the lysis buffer containing 2% SDS and protease inhibitor to extract the total proteins. Equal amounts of parasite lysates (15 μg/lane) were separated on a 10% SDS-PAGE gel under reducing conditions and transferred to PVDF membranes. Membranes were blocked with 5% non-fat milk in TBST (Tris-buffered saline with 0.1% Tween-20) for 2 h, then probed with anti-rPv22 sera (1:200) or anti-HA mAb (1:1000, Invitrogen). The anti-Hsp70 sera (1:1000) was used as a loading control. After three washes with TBST, membranes were incubated with HRP-conjugated anti-rabbit IgG antibody (1:5000, Invitrogen), and the blot was visualized using an ECL Western blot Kit (Thermo Fisher Scientific).

2.9. Indirect immunofluorescence assay (IFA)

The localization of Pv22 in the TrPv22Pb parasites was analyzed by IFA [35]. Briefly, TrPv22Pb parasites were fixed with 4% paraformaldehyde and 0.0075% glutaraldehyde in PBS for 30 min at room temperature. After washing once with PBS, cells were permeabilized with 0.1% Triton X-100 for 10 min and blocked with 3% BSA for 1 h. Subsequently, samples were stained with primary antibodies (rabbit anti-rPv22 sera at 1:200 or anti-HA mAb at 1:1000) for 1 h at 37°C. Parasites were co-incubated with the mouse anti-α-tubulin II (1:500), Pbs47 (1:500), SET (1:500) and Pbs21 (1:500) as stage-specific markers for male gametocytes/gametes, female gametocytes, nucleus and zygotes/ookinetes, respectively [36–39]. After three washes, parasites were incubated with Alexa Fluor 488-conjugated goat anti-rabbit IgG antibodies (1:500, Invitrogen) and Alexa Fluor 555-conjugated goat anti-mouse IgG antibodies (1:500, Abcam) for 1 h at 37°C. Asexual stages (merozoite, ring, trophozoite, and schizont) were only incubated with anti-rPv22 sera or the anti-HA mAb. Hoechst 33258 (1:1000, Invitrogen) was used to stain nuclear DNA. As negative controls, TrPv22Pb ookinetes were incubated with only the secondary antibodies or the control sera from the rabbit immunized with GST. Images were acquired using a Nikon C2 fluorescence confocal laser scanning microscope (Japan).

IFA was performed to detect Pv22 in the parasite-enriched fraction of a P. vivax patient. After obtaining informed consent, 5 ml of blood were collected and mixed with 10 ml of suspended animation buffer (10 mM Tris, pH 7.4, 170 mM NaCl, 10 mM glucose) at 37°C to avoid exflagellation. Then, the parasites were isolated on a 48% (v/v) Nycodenz gradient and placed on glass slides for IFA as described above.

2.10. In vivo and in vitro quantification of TBA

The TBA of Pv22 was assessed using an in vivo assay using the transgenic parasite TrPv22Pb. BALB/c mice (n = 6) were subcutaneously immunized with the purified rPv22 (50 μg/mouse) emulsified in complete Freund’s adjuvant, followed by two boosters of 25 μg/mouse emulsified in incomplete Freund’s adjuvant at a two-week interval. The same immunization procedure was followed for the control group of mice using the GST protein. Serum was collected from the tail vein of each mouse on the day before each immunization and at 10 days after the final immunization, and antibody titers were measured by ELISA. Subsequently, mice were infected with 5×10⁶ TrPv22Pb parasites. Parasitemia was measured daily, while gametocytemia and gametocyte sex ratio were monitored on day 3 post-infection. The percentage of schizonts in infected RBCs was measured by Giemsa counting from days 3 to 5 post-infection. Exflagellation centers, macrogametes and ookinetes were counted as described above. Starved female A. stephensi were allowed to feed on the infected mice, and infected mosquitoes and oocyst numbers were determined 12 days after feeding.

To investigate the effect of immune sera on parasites invasion, we used an in vitro invasion assay with isolated TrPv22Pb merozoites [40]. Briefly, purified mature schizonts were forcefully filtrated via 1.2 μm pore syringe filter. Merozoites obtained from 5×10⁸ schizonts in 500 μL of complete medium were mixed with 50 μL of pre-warmed (37°C) mouse RBCs (~5×10⁷ RBCs) with the GST or rPv22-immune sera at final concentrations of 50 μg/mL. The invasion assay was performed in 1.5 mL tubes with shaking at 1000 rpm for 1 h. Then, the tubes were centrifuged at 2000 × g for 1 min, supernatant removed, and the pellet washed once to remove the antisera. The pellet was re-suspended in fresh complete medium and in vitro cultured at 37°C in a 48-well plate. Parasitemia was evaluated after 18 h of culture.

An in vitro assay was carried out to determine the TBA of the immune sera on ookinete formation. Ookinete cultures were set as described above with the culture medium containing the anti-rPv22 sera or the control sera at final dilutions of 1:5, 1:10, and 1:50. Exflagellation centers were counted at 15 min, and ookinete numbers were estimated at 24 h.

2.11. Evaluation of TBA by direct membrane feeding assay (DMFA)

Patients visiting malaria clinics at the Thai-Myanmar border were diagnosed with P. vivax infection by microscopy. Written informed consent was obtained from three volunteers. Direct membrane feeding assay (DMFA) was performed using a published protocol [21]. In brief, purified IgGs from rabbits immunized with rPv22 or GST protein were diluted with normal human AB⁺ serum at the ratio of 1:1 in a total volume of 180 μl, which were mixed with RBCs from P. vivax patients (1:1, v/v). Each reconstituted blood sample was introduced into a membrane feeder after incubation at 37°C for 15 min. Female A. dirus mosquitoes were allowed to feed on infected blood for 30 min. After removing unfed mosquitoes, the remaining mosquitoes were maintained in an insectary for 7 days. Twenty mosquitoes from each group were dissected, and oocysts in the midguts were counted by microscopy.

2.12. Analysis of genetic polymorphisms

DNA from P. vivax isolates used in the DMFA was extracted using a QIAamp DNA Blood Mini kit (Qiagen, Germany). The pv22 DNA fragment encoding aa 21–216 was amplified by PCR with primers designed based on Sal-I sequence: Pv22-F (5’-TGCAACAAAAACATAATCGAGCTG-3’) and Pv22-R (5’-CCTGCCCCCCCCCTGCTGCATAC-3’). The purified PCR products were sequenced using the ABI Prism^® BigDye^™ cycle sequencing kit (Applied Biosystems, Thermo Fisher Scientific).

2.13. Statistical analysis

Statistical analysis was carried out using SPSS software, version 22.0. Antibody titers, parasitemia, gametocytemia, exflagellation and ookinete numbers among groups were compared by Student’s t test. The prevalence of infection was analyzed by Fisher’s exact test, while the intensity of infection (oocysts/midgut) was performed by Mann-Whitney U test. All data were from three independent experiments.

3. Results

3.1. Production of rabbit antibodies against Pv22

Having determined the immunogenicity and TB potential of the sexual stage antigen Pb22 in P. berghei [24], we sought to explore the P. vivax ortholog Pv22 as a TBV candidate for the human malaria parasite. Pv22 encodes a protein of 216 aa with a predicted molecular weight of 22 kDa. Pv22 exhibited a 52% aa identity with Pb22 (Fig. S1). The rPv22 without the signal peptide was expressed in E. coli and purified by affinity chromatography. SDS-PAGE analysis of the purified rPv22 showed a relatively homogenous band of ~22 kDa (Fig. 1A). Rabbits were immunized with purified rPv22 to obtain polyclonal antibodies, and the antisera after the final booster showed a titer of 1:512000, as determined by ELISA (Fig. 1B). GST served as the immunization control [18]. Antibody concentrations after immunization with rPv22 and GST, determined from antibodies purified using protein A column, were 4.4 and 6.3 μg/μL, respectively.

Fig. 1.

Purification of rPv22 and antibody titers of anti-rPv22. (A) Purified rPv22 protein was separated by 10% SDS-PAGE and stained with Coomassie brilliant blue. (B) Rabbits were immunized with rPv22 or GST control protein. Antibody titers after the last immunization were analyzed by ELISA. Error bars indicate mean ± SEM. * P < 0.05, ** P < 0.01

3.2. Generation and phenotypic analysis of the TrPv22Pb parasite

Genetically modified transgenic P. berghei parasites expressing Pv22, designated as the TrPv22Pb lines, were generated by the gene insertion-marker out (GIMO) approach (Fig. 2A) [32]. After transfection and negative selection with 5-fluorocytosine, marker-free transgenic parasites were examined by diagnostic PCR to verify successful genomic integration (Fig. 2B). This was illustrated by the prominent band of 1.46 kb amplified only in the TrPv22Pb parasite with the primer pair P1xP4, while the presence of less prominent bands with other primer pairs (P1xP3 and P5xP6) was probably due to non-specific amplifications.

Fig. 2.

Generation of TrPv22Pb parasites and genotypic analysis. (A) Schematic representation of the transfection construct integrated into the Δpb22 locus by GIMO. (B) Successful integration of pv22 into the Δpb22 genome and removal of the hdhfr∷yfcu selection cassette was confirmed by PCR amplification with primers 1–6. Lane 1: primers 1+2 (1449 bp); Lane 2: primers 1+3 (1182 bp); Lane 3: primers 1+4 (1460 bp); Lane 4: primers 5+6 (2504 bp).

Deletion of pb22 resulted in a significant defect in male gametogenesis [24]. Phenotypic characterization of TrPv22Pb parasite showed no differences in asexual blood-stage growth, gametocyte production, sex ratio, and female gametogenesis from the WT parasites (Fig. S2A–D). The expression of Pv22 fully rescued the defects of Δpb22 in male gametogenesis; similar numbers of exflagellation centers were observed for both WT and TrPv22Pb (Fig. S2E). Further, the TrPv22Pb parasite formed a similar number of ookinetes from in vitro culture as the WT (Fig. S2F). Additionally, the proportions of infected mosquitoes and oocyst densities from mosquito feeding assays were comparable between the WT and TrPv22Pb parasites (Fig. S2G, I). In a bite-back experiment, the infected mosquitoes could also infect 3/3 mice (Fig. S2I). These observations collectively demonstrated that pv22 fully complemented the aberrant phenotypes of the Δpb22 parasite, and the TrPv22Pb parasite was equally competent as the WT parasite in completing the entire life cycle.

3.3. Pv22 shows similar stage-specific expression and localization as Pb22

To study Pv22 expression in the transgenic parasite, we purified ring-trophozoite stages, schizonts, gametocytes and ookinetes of the TrPv22Pb parasites, as well as gametocytes of the WT P. berghei. Western blot analysis of protein extracts from purified parasites using both the anti-rPv22 sera and anti-HA mAb specifically recognized the ~22 kDa band in the TrPv22Pb parasites, but not the WT parasites (Fig. 3A). Pv22 expression was low in early asexual stages but similarly high in schizonts, gametocytes, and ookinetes.

Fig. 3.

Expression and localization of Pv22 in TrPv22Pb parasites. (A) Western blot analysis of TrPv22Pb parasite lysates from ring-trophozoite stages (R-T), schizonts (Sch), gametocytes (GC) and ookinetes (Ook), as well as gametocytes of P. berghei (WT). Proteins were probed with anti-rPv22 sera diluted 1:200 (top panel), anti-HA mAb diluted 1:1000 (middle panel), or anti-Hsp70 sera diluted 1:1000 (bottom panel). (B) Representative IFA images of cells with membrane permeabilization (+Triton X-100). TrPv22Pb parasites were incubated with anti-rPv22 sera diluted 1:200 as the primary antibodies. The parasites were also labeled with antibodies against the marker proteins for different stages (α-tubulin II for male gametocytes and gametes, Pbs47 for female gametocytes, SET for nucleus of gametocytes and Pbs21 for zygotes/ookinetes). Alexa Fluor 488 (AF 488)-conjugated goat anti-rabbit IgG antibodies and Alexa Fluor 555 (AF 555)-conjugated goat anti-mouse IgG antibodies were used as the secondary antibodies. The nucleus was stained with Hoechst 33258 (blue). TrPv22Pb ookinetes incubated with anti-GST sera were used as a negative control. DIC, differential interference contrast microscopy. Scale bar, 5 μm.

IFA using the anti-rPv22 sera and anti-HA mAb in the TrPv22Pb parasites detected a similar localization pattern of Pv22 as the Pb22 in WT parasites (Fig. 3B, S3 and S4). Pv22 was expressed in asexual blood stages, localized in the cytosol ring and trophozoite stages, but in schizonts it appeared to be associated with the merozoite surface (Fig. S3). In sexual stages, Pv22 was detected in both male and female gametocytes, gametes and ookinetes (Fig. S4). Of particular relevance is the peripheral localization of Pv22 in gametes, zygotes and ookinetes, which suggests the adequacy of the transgenic parasite TrPv22Pb for in vivo evaluation of Pv22’s TB potential.

3.4. Evaluation of Pv22’s TB potential using transgenic parasite

To determine the the TB potential of Pv22, we immunized two groups of mice (n=6 per group) with either rPv22 or GST using immunization scheme shown in Fig. S5A. Anti-Pv22 antibody titers determined by ELISA showed that the rPv22 protein elicited an effective antibody response during the immunization period (Fig. S5B), with the antibody titer against rPv22 reaching 1:128000 after the last booster (P < 0.01 compared to the control group; Fig. S5C).

We used the TrPv22Pb parasites as a surrogate of P. vivax for studying Pv22’s TB potential. TrPv22Pb exhibited normal asexual blood stages and gametocytes in mice immunized with either rPv22 or GST control (Fig. S6). In rPv22 immunized mice, the daily parasitemia and schizont development were similar to those in the GST control group (Fig. S6A,B). We also found that the presence of GST and rPv22-immune sera did not affect the merozoite’s invasion efficiency (Fig. S6C). Besides, the gametocytemia, sex ratio, and macrogamete numbers were also similar between the two immunization groups (Fig. S6D–F). However, rPv22 immunization led to a 53.5% decrease in the number of exflagellation centers (P < 0.01; Fig. 4A), and a 52.8% reduction in ookinete development in vitro (P < 0.01; Fig. 4B). In the mosquito feeding assays, the oocyst density in mosquitoes feeding on the rPv22-immunized mice had a 49.3–53.3% reduction compared to those that fed on the GST-immunization control mice (P < 0.01; Fig. 4C and Table 1), although there was no reduction in the prevalence of infected mosquitoes.

Fig. 4.

In vivo transmission-blocking activity of rPv22 protein. BALB/c mice (3/group) immunized with the rPv22 fusion protein or GST control protein were infected with 5×10⁶ TrPv22Pb parasites 10 days after the third immunization. (A) Exflagellation centers/10 fields. (B) Ookinete numbers in 0.01 μL infected blood. (C) Oocyst numbers per midgut at 12 days after feeding in the two groups. Data of mosquito feeding assays are shown in Table 1. Error bars indicate mean ± SEM. ** P < 0.01

Table 1.

In vivo evaluation of transmission-blocking effects of Pv22 in mosquito feeding experiments.

Experiment	Group	Oocyst density Mean ± SEMa	% reduction in oocyst densityb	P valuec	% infection prevalence (infected/dissected)d	% reduction in prevalencee	P valuef
1	GST	111.4 ± 14.5			93.3 (28/30)
	Pv22	52.0 ± 8.7	53.3	0.0010	86.7 (26/30)	6.6	0.6714
2	GST	105.3 ± 12.6			96.6 (28/29)
	Pv22	53.4 ± 7.8	49.3	0.0011	86.7 (26/30)	9.9	0.3526
3	GST	96.7 ± 13.1			93.3 (28/30)
	Pv22	46.6 ± 8.1	51.8	0.0031	82.1 (23/28)	11.2	0.2463

^aMean oocyst numbers and standard error of mean.

^bReduction in oocyst density was calculated as (mean _GST – mean _Pv22)/mean _GST × 100%.

^cThe mean number of oocyst was statistically analyzed by Mann-Whitney U test and P-values less than 0.05 were considered statistically significant.

^dThe prevalence of infection was calculated by the number of mosquitoes with oocysts/number of mosquitoes dissected in each group × 100%.

^eReduction in prevalence was calculated as % prevalence _GST – % prevalence _Pv22.

^fThe prevalence of infection was statistically analyzed by Fisher’s exact test and P-values less than 0.05 were considered statistically significant.

Subsequently, we examined the TB potential of rPv22-immune sera using an in vitro assay. Incubation of the TrPv22Pb parasites with the rabbit anti-rPv22 sera at dilutions of 1:5 and 1:10 reduced the number of exflagellation centers by 33.6 and 27.6%, respectively (P < 0.05; Fig. S7A). Similarly, the ookinete numbers in cultures incubated with anti-rPv22 sera at 1:5 and 1:10 dilutions were reduced by 41.1 and 31.1%, respectively (P < 0.05; Fig. S7B). The inhibitory effects of the anti-rPv22 antibodies on exflagellation and in vitro ookinete development were attenuated at the 1:50 dilution (Fig. S7). The control sera, irrespective of the dilution, showed no noticeable effects on exflagellation and ookinete formation (Fig. S7).

3.5. Evaluation of Pv22’s TB potential using DMFA

Transcriptomic analyses of P. vivax showed that the Pv22 transcripts were detected throughout the entire life cycle [27–29]. IFA was conducted to investigate the expression and localization of Pv22 in P. vivax gametocytes. The rabbit anti-rPv22 sera specifically reacted with male and female gametocytes in a clinical isolate (Fig. 5A), and the cytoplasmic localization patterns were similar to Pb22 in P. berghei gametocytes [24]. In contrast, the control anti-GST sera did not stain P. vivax gametocytes.

Fig. 5.

Detection of Pv22 expression in P. vivax gametocytes and TRA of antibodies against rPv22 in DMFA. (A) Gametocytes were stained with the anti-rPv22 sera (1:200) and Alexa Fluor 488-conjugated anti-rabbit IgG antibodies. Antisera against GST were used as the negative control. Nuclei were stained with Hoechst 33258 (blue). BF, bright field; AF488, Alexa Fluor 488; Merge, AF488 + Hoechst. (B) DMFA was performed using three P. vivax isolates with purified IgGs mixed with heat-inactivated (complement minus) AB⁺ human serum in the ratio of 1:1. Numbers of oocysts in mosquito midguts were shown as scatter dot plots. The red horizontal bar indicates the mean number of oocysts in each group. Statistical difference of the mean number of oocysts between GST and Pv22 groups was analyzed by the Mann-Whitney U test (* P < 0.05, ** P < 0.01). Data of DMFA are shown in Table 2.

The TB potential of Pv22 was finally evaluated using three P. vivax clinical samples by DMFA with An. dirus mosquitoes. Although the mean oocyst density resulting from the three P. vivax isolates varied, the rabbit anti-rPv22 antibodies possessed significant TRA and reduced the midgut oocyst density by 53.7, 30.2, and 26.2%, respectively, compared with the GST control group (P < 0.05, Fig. 5B and Table 2). The anti-rPv22 antibodies did not decrease the infection prevalence in mosquitoes.

Table 2.

Infection prevalence and oocyst density in mosquitoes from DMFA with three clinical P. vivax isolates.

P. vivax isolates	Antibodies	Oocyst density Mean ± SEMa	% reduction in oocyst densityb	P valuec	% infection prevalence (infected/dissected)d	% reduction in prevalencee	P valuef
Case 1	GST	8.2 ± 1.2			95.0 (19/20)
	Pv22	3.8 ± 0.7	53.7	0.0028	90.0 (18/20)	5.0	1.0000
Case 2	GST	4.3 ± 0.4			100.0 (20/20)
	Pv22	3.0 ± 0.6	30.2	0.0312	85.0 (17/20)	15.0	0.2308
Case 3	GST	69.4 ± 8.2			95.0 (19/20)
	Pv22	51.2 ± 3.9	26.2	0.0214	100.0 (20/20)	-	-

^aMean oocyst numbers and standard error of mean.

^bReduction in oocyst density was calculated as (mean _GST – mean _Pv22)/mean _GST × 100%.

^cThe mean number of oocyst was statistically analyzed by Mann-Whitney U test.

^dInfection prevalence was calculated by number of oocyst-infected mosquitoes per 20 mosquitoes dissected in each group.

^eReduction in prevalence was calculated as % prevalence _GST – % prevalence _Pv22.

^fInfection prevalence was statistically analyzed by Fisher’s exact test.

To determine whether the variation of TRA among the different isolates might be attributed to genetic polymorphisms of the Pv22 gene, we sequenced the DNA fragment of Pv22 in the three P. vivax isolates used in DMFA. These samples had identical amino acid sequences with the Sal-I strain (data not shown).

4. Discussion

Evaluation of the TBA of P. vivax antigens critically depends on DMFA, which is inconvenient and laborious to perform because of the need to access volunteers naturally infected with P. vivax. The lack of a continuous culture for P. vivax also means that blood samples from multiple P. vivax patients need to be evaluated, given the large variations among the clinical samples. For example, variations in mosquito infectivity have been commonly observed in ex vivo mosquito feeding assays [15, 18, 41]. Thus, a suitable small-animal model would be ideal for supporting preclinical TBV discovery. The availability of an animal malaria model would substitute for the artificial MFA, especially in places where it is not currently feasible to culture and produce infectious gametocytes. The use of transgenic rodent parasites expressing antigens of the human malaria parasites opens new avenues for malaria vaccine research [23]. Several transgenic rodent malaria parasites have been generated, where the endogenous P. berghei genes were replaced with its human malaria parasite orthologs, such as CSP [42–44], PvCelTOS [45], Pvs48/45 [46], and P25 [47, 48]. For the TBV candidates, the transgenic parasites expressing Pvs25 or Pvs48/45 were transmission-competent, but mosquito infectivity (oocyst density in infected mosquitoes) was reduced [46, 47]. This study found that Pv22 was fully functional in P. berghei, although Pv22 and Pb22 only shared 52% sequence identity. The Δpb22 parasite displayed significant defects in male gametogenesis and a complete loss of mosquito infectivity [24]. In contrast, replacing the endogenous pb22 gene with pv22 completely rescued these defects in sexual development. The transgenic parasites TrPv22Pb showed comparable mosquito infectivity as the WT parasites (in both prevalence of infection and oocyst density). Moreover, Pv22 and Pb22 also showed similar patterns of expression and localization. However, in the specific case of Pv22, the suitability of the TrPv22Pb chimeric parasite for evaluating the TB potential of Pv22 requires further detailed studies of Pv22 expression and localization in P. vivax gametocytes and ookinetes.

The excellent TBA of Pb22 from mosquito feeding assays (>80% reduction in infection prevalence and >90% reduction in oocyst density) with the rodent parasite prompted us to explore its feasibility as a TBV candidate in the human malaria parasite P. vivax [24]. We found obvious TRA of the Pv22 antisera using both transgenic TrPv22Pb and DMFA with clinical P. vivax isolates. Using the TrPv22Pb parasites, we found that the Pv22 antisera significantly inhibited exflagellation and ookinete formation from in vitro assays, consistent with the findings for Pb22. Subsequently, immunization with rPv22 impaired the transmission of parasites to mosquitoes, resulting in a substantial reduction in oocyst density. These results were further corroborated by DMFA with clinical P. vivax isolates from three patients, which showed a 26.2–53.7% reduction in oocyst density, despite that mosquito infectivity varied considerably among the three feeding experiments. Such wide variations are typical of DMFA experiments using P. vivax samples from patients [15, 18, 41]. Of note, the anti-rPv22 sera did not lead to a significant reduction in the prevalence of infected mosquitoes compared to the control antisera in both the in vivo and DMFA studies. These TRA results of Pv22 were much weaker than those for Pb22 [24]. Since the antibody titers in both studies were high after immunization with relatively high doses of rP22 proteins and Freund’s adjuvants [24], such differences in TBA may not be due to the difference in immunogenicity between the two proteins. This may reflect the functional difference between the two orthologs and suggest that Pv22 plays a less important role in fertilization and ookinete development than Pb22. Therefore, despite the excellent TBA result for Pb22 in the rodent malaria parasite, the less significant TBA results obtained from DMFA with clinical P. vivax isolates and direct feeding with the transgenic chimeric parasite indicate that Pv22 is not a promising transmission-blocking vaccine candidate.

Highlights.

• We generated transgenic P. berghei parasites expressing Pv22.

• We used the transgenic parasites to evaluate Pv22’s transmission-blocking potential.

• Clinical P. vivax isolates were used in membrane feeding.

• Antibodies against rPv22 significantly reduced the oocyst density in mosquitoes.

• Feeding assay results disfavor Pv22 as a promising TBV candidate.