Pfs47 as a Malaria Transmission-Blocking Vaccine Target

INTRODUCTION

The past two decades saw major advancements in alleviating malaria burden, primarily by reducing disease transmission through massive distribution of insecticide-treated nets and deployment of residual insecticide spraying. However, progress has recently stalled because of the emergence of insecticide resistance by malaria vectors and parasite resistance to antimalarial drugs,¹ making more evident the need for new strategies to curb this devastating parasitic disease. Transmission-blocking vaccines (TBVs) are a potential new tool to assist in malaria elimination² by eliciting a human immune response that prevents parasite development in the mosquito vector.³

Malaria TBVs take advantage of the Plasmodium parasite’s vulnerability as it emerges from red blood cells in the mosquito gut to engage in a complex developmental cycle that results in a natural life cycle bottleneck. Malaria transmission begins when a mosquito vector takes a blood meal containing Plasmodium gametocytes. Within minutes, these sexual-stage parasites mature into gametes, leave the red blood cell, and fuse, forming a fertilized zygote that matures into a motile ookinete. Usually, five or fewer ookinetes traverse the mosquito gut successfully and develop into the oocyst stage. The parasite multiplies continuously during the oocyst stage, and mature oocysts rupture and release thousands of sporozoites that invade the mosquito salivary gland. Sporozoites are infectious to humans, and they are injected into the skin of a new host when an infected mosquito takes an additional blood meal.⁴ It is therefore attractive to stop Plasmodium infection in the mosquito before the parasite multiplies during the oocyst stage.

Pfs47 STRUCTURAL ORGANIZATION

Plasmodium falciparum sexual-stage surface protein 47 (Pfs47) is a predicted glycosylphosphatidylinositol (GPI) anchored protein of the Plasmodium falciparum six-cysteine (6-Cys) family, which includes two leading TBV targets: Pfs48/45 and Pfs230. This family comprises proteins with 1 to 14 cysteine-rich domains. The current 14 members of the 6-Cys family include secreted or membrane-anchored proteins that are involved in cell-to-cell interactions during fertilization, liver invasion, or mosquito immune evasion.⁵ Pfs48/45 and Pfs230 are expressed in gametocytes and on the surface of gametes, and are critical for male fertility.^6–8 They both were initially recognized as TBV targets using transmission-blocking monoclonal antibodies (mAbs) raised against gametes.⁹ Pfs230 has 14 protein domains, each with a conserved pattern of cysteine residues.¹⁰ The other 6-Cys family members were later identified searching the canonical cysteine pattern on novel protein sequences of the Plasmodium genome. A correct prediction of the cysteine pairs that form the disulfide bonds and of the folding in the 6-Cys domain¹¹ was followed by accurate computational modeling¹² and experimental confirmation.¹³

Based on their similar gene structure and contiguous genomic location, Pfs47 and Pfs48/45 appear to have arisen from a gene duplication event, although they have low amino acid (aa) sequence identity (27%). Both Pfs47 and Pfs48/45 contain three protein domains: D1, D2, and D3. In Pfs47, D1 and D3 have the canonical 6-Cys pattern, whereas D2 is a shorter and degenerate s48/45 domain with only two cysteines (Figure 1A).¹³

Pfs47 protein structure and the identified transmission-blocking vaccine (TBV) antigen Pfs47-Del2. (A) Schematic of Pfs47 drawn to scale shows Pfs47 three domains (D1, D2, D3), predicted disulfide bonds (S-S), and TBV antigen Pfs47-Del2 location (178–229 amino acid residue [aa], in yellow). Pfs47 D1 and D3 have the canonical cysteine arrangement of a six-cysteine domain, whereas D2 has only two cysteines. The approximate beginning of each protein domain is indicated by its aa number. The location of the aa polymorphisms that are known to determine Plasmodium falciparum compatibility to vectors is indicated with red stars. (B) Amino acid composition of TBV antigen Pfs47-Del2. The properties of the aa side chains are indicated by color: polar (black), nonpolar (gray), negatively charged (red) or positively charged (blue), and glycine (green). Predicted β-strands (–––) and disordered segments (- - - -) are indicated. The AlphaFold2-predicted Pfs47 three-dimensional structure is presented in a side view (C) and top views (D, E) without the signal peptide. The surface representation (E) shows negative (red) and positive (blue) charges. Pfs47-Del2 appears in yellow and the aa polymorphisms that determine P. falciparum compatibility to vectors are shown in red (C, D). GPI ω-site = predicted glycosylphosphatidylinositol anchoring site; SP = signal peptide.

Citation: The American Journal of Tropical Medicine and Hygiene 107, 3_Suppl; 10.4269/ajtmh.21-1325

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Pfs47 protein structure and the identified transmission-blocking vaccine (TBV) antigen Pfs47-Del2. (A) Schematic of Pfs47 drawn to scale shows Pfs47 three domains (D1, D2, D3), predicted disulfide bonds (S-S), and TBV antigen Pfs47-Del2 location (178–229 amino acid residue [aa], in yellow). Pfs47 D1 and D3 have the canonical cysteine arrangement of a six-cysteine domain, whereas D2 has only two cysteines. The approximate beginning of each protein domain is indicated by its aa number. The location of the aa polymorphisms that are known to determine Plasmodium falciparum compatibility to vectors is indicated with red stars. (B) Amino acid composition of TBV antigen Pfs47-Del2. The properties of the aa side chains are indicated by color: polar (black), nonpolar (gray), negatively charged (red) or positively charged (blue), and glycine (green). Predicted β-strands (–––) and disordered segments (- - - -) are indicated. The AlphaFold2-predicted Pfs47 three-dimensional structure is presented in a side view (C) and top views (D, E) without the signal peptide. The surface representation (E) shows negative (red) and positive (blue) charges. Pfs47-Del2 appears in yellow and the aa polymorphisms that determine P. falciparum compatibility to vectors are shown in red (C, D). GPI ω-site = predicted glycosylphosphatidylinositol anchoring site; SP = signal peptide.

Citation: The American Journal of Tropical Medicine and Hygiene 107, 3_Suppl; 10.4269/ajtmh.21-1325

• Download Figure
• Download figure as PowerPoint slide

Pfs47 protein structure and the identified transmission-blocking vaccine (TBV) antigen Pfs47-Del2. (A) Schematic of Pfs47 drawn to scale shows Pfs47 three domains (D1, D2, D3), predicted disulfide bonds (S-S), and TBV antigen Pfs47-Del2 location (178–229 amino acid residue [aa], in yellow). Pfs47 D1 and D3 have the canonical cysteine arrangement of a six-cysteine domain, whereas D2 has only two cysteines. The approximate beginning of each protein domain is indicated by its aa number. The location of the aa polymorphisms that are known to determine Plasmodium falciparum compatibility to vectors is indicated with red stars. (B) Amino acid composition of TBV antigen Pfs47-Del2. The properties of the aa side chains are indicated by color: polar (black), nonpolar (gray), negatively charged (red) or positively charged (blue), and glycine (green). Predicted β-strands (–––) and disordered segments (- - - -) are indicated. The AlphaFold2-predicted Pfs47 three-dimensional structure is presented in a side view (C) and top views (D, E) without the signal peptide. The surface representation (E) shows negative (red) and positive (blue) charges. Pfs47-Del2 appears in yellow and the aa polymorphisms that determine P. falciparum compatibility to vectors are shown in red (C, D). GPI ω-site = predicted glycosylphosphatidylinositol anchoring site; SP = signal peptide.

Citation: The American Journal of Tropical Medicine and Hygiene 107, 3_Suppl; 10.4269/ajtmh.21-1325

• Download Figure
• Download figure as PowerPoint slide

The structure of Pfs47 was recently predicted using AlphaFold2,¹⁴ protein structure modeling software, and became publicly available (https://alphafold.ebi.ac.uk/entry/Q8IDN0). The software predicted most of the structure with confidence or high confidence (predicted local distance test, > 70 or > 90; predicted alignment error, 0.727). It predicts the presence of three domains with the canonical s48/45 domain conformation consisting of a β-sandwich formed by antiparallel and parallel β-sheets (Figure 1C and D).¹³ The approximate boundaries of the Pfs47 domains in the predicted structure are D1, aa 27 to 182; D2, aa 183 to 282; and D3, aa 283 to 415. The three Pfs47 domains are adjacent to each other, forming a triangular or “three-petal flower” arrangement (Figure 1D). The predicted GPI ω-site is toward the C-terminal of Pfs47 D3 (Figure 1C) presumably facing the cell membrane. Furthermore, D1 and D3 have very few negative charges (Figure 1E) on the surface facing away from the cell membrane, whereas this surface has many negative charges in D2 (Figure 1E, shown in red). The C-terminal of Pfs47 D2, containing the polymorphisms responsible for mosquito compatibility, is at the lower edge of the predicted β-sandwich (Figure 1C, in red). Multiple efforts to elucidate the structure of Pfs47 experimentally have not succeeded so far (M. Boulanger, personal communication).

ROLE OF Pfs47 AND ITS ORTHOLOGUES IN PLASMODIUM FERTILIZATION AND MOSQUITO IMMUNE EVASION

Several functions have been attributed to Pfs47 and its orthologues. Pfs47 is expressed within Plasmodium female gametocytes, and on the surface of female gametes, and ookinetes.^15–17 In vivo studies with Plasmodium berghei, a mouse malaria model, in which the Pbs47 gene (the ortholog of Pfs47) was disrupted, revealed that Pbs47 is required for optimal female gamete fertility, as Pb47 null parasites had lower gamete fusion and reduced infectivity in Anopheles stephensi Nijmegen (SDA 500) mosquitoes.⁸ However, laboratory infections with in vitro cultured gametocytes from a P. falciparum-Pfs47 null line revealed that Pfs47 is not essential for female gamete fertility nor for infectivity in An. stephensi Nijmegen (SDA 500).¹⁷ Of note, this widely used laboratory mosquito strain is of Indian origin and was genetically selected to be highly susceptible to P. falciparum infection.¹⁸

Analysis of a genetic cross between two P. falciparum lines (GB4 and 7G8) that differ in their survival in Anopheles gambiae L3-5, led to the discovery that Pfs47 allows the malaria parasite to evade the mosquito complement-like system, a major final effector of mosquito anti-plasmodial immunity that targets the ookinete stage.¹⁹ Genetic mapping, linkage group selection,²⁰ and functional genetics studies revealed that four aa polymorphisms between the two cysteines in Pfs47-D2 are key determinants of mosquito infectivity.¹⁶ Later studies showed that Pfs47 is required for efficient malaria transmission by several dominant anopheline vectors, including An. gambiae (Africa), Anopheles dirus (Southeast Asia) and Anopheles albimanus (New World).^16,21 Pfs47 disrupts activation of a c-Jun N-terminal kinase/caspase-mediated apoptosis that involves a strong nitration response in the invaded midgut cell, preventing nitration of the subjacent basal lamina.^22,23 Circulating mosquito hemocytes are attracted to the midgut by prostaglandin E2 release by the midgut in response to ookinete invasion.²⁴ If hemocytes patrolling the midgut surface encounter a nitrated basal lamina, they undergo apoptosis and release microvesicles.²⁵ Local release of hemocyte-derived microvesicles is required for effective activation of the mosquito complement-like system and for parasite elimination.²⁵ Pfs47 allows Plasmodium immune evasion by disrupting midgut epithelial nitration, thus preventing local hemocyte-derived microvesicle release and rendering the mosquito complement system ineffective against the parasite. Pbs47 was as well found to be essential for ookinete evasion of the An. gambiae complement-like response.²⁶

Several studies have shown that Pfs47 is highly polymorphic in Pfs47 D2 and exhibits a striking geographic population structure.^21,27,28 Furthermore, direct experimental evidence showed that polymorphisms in Pfs47 D2 determine the compatibility of P. falciparum to An. gambiae, An. dirus, and An. albimanus, suggesting that selection of Pfs47 haplotypes was involved in the adaptation of P. falciparum to phylogenetically distant mosquito vectors.²¹ Pfs47 interacts with a receptor protein in the mosquito midgut, and the Pfs47 haplotype of a given parasite must be compatible with the receptor of the local vector, as a lock and key, for effective immune evasion that facilitates malaria transmission.²⁹

Pfs47 AS A TBV TARGET

Initial attempts to assess the potential of Pfs47 and its orthologue in Plasmodium vivax (Pvs47) as TBV targets had variable success. Three rat mAbs against Pfs47 had no significant effect on P. falciparum infection in standard membrane feeding assays (SMFAs) using in vitro cultured P. falciparum gametocytes to infect An. stephensi Nijmegen females.¹⁷ In contrast, DNA vaccine immunization of mice against full-length Pvs47 generated antisera with significant transmission blocking in An. dirus. This activity required human complement, in a direct membrane feeding assay with P. vivax gametocytes obtained directly from patients with malaria.³⁰

The discovery of the key role of Pfs47 in the evasion of mosquito immunity, and the genetic evidence of the relevance of Pfs47 D2,^16,21 led to a more systematic evaluation of Pfs47 as a TBV target. Soluble, full-length recombinant Pfs47, was expressed successfully as a fusion protein with thioredoxin (T-Pfs47) in Escherichia coli, and was immunogenic in mice. Pfs47 polyclonal antibodies recognized gametocytes and female gametes, but together with 14 Pfs47 mAbs had modest (< 65%) or no transmission-reducing activity (TRA) in SMFAs with An. gambiae G3 females—an unselected laboratory strain susceptible to P. falciparum infection. Interestingly, binding specificity analysis showed that antibodies after immunization with T-Pfs47 mostly recognized Pfs47 D1 or D3, but not D2.¹⁵ A similar study, in which mice were immunized with recombinant full-length Pfs47 ectodomains expressed in HEK293E cells, led to polyclonal antibodies with no gamete recognition by indirect fluorescent antibody assay and no transmission-blocking activity.³¹

In contrast, immunization with a Pfs47 D2 recombinant peptide (aa 155–264), in which the two cysteines were mutated to alanine to facilitate expression, generated polyclonal and mAbs that recognized gametocytes and female gametes, and had significant transmission blocking (78%–99% TRA). mAbs and deletion analysis identified a 52-aa region mostly in D2 (Pfs47-Del2, aa 178–229), immediately preceding the predicted Pfs47 D2 disulfide bond, where antibody binding disrupts malaria transmission (Figure 1A, yellow). This target region is highly conserved (96%–98% aa identity) in parasites from different geographic regions, with only seven unique haplotypes that mostly differ by a single aa.¹⁵ Further studies are needed to test with heterologous parasites the cross-reactivity and transmission-blocking activity of sera developed against the most frequent Pfs47 haplotype in Africa (GB4).

The higher immunogenicity of Pfs47 D1 and D3, which generate non-transmission-blocking antibodies, suggests that they may act as antigenic decoys to prevent antibody generation against functional regions in Pfs47 D2. The predicted three-dimensional structure of Pfs47 (Figure 1B–D) indicates that Pfs47 D2 is accessible to antibodies. The lower immunogenicity of Pfs47 D2 coincides with higher negative surface charge, and a reduced number of cysteines (one pair), compared with Pfs47 D1 and D3 (three pairs), which could make Pfs47 D2 conformation more flexible and less immunogenic, and may also explain the difficulty in obtaining a Pfs47 crystal structure.

Transmission-blocking mAb IB2 reduced the number of ookinetes in the mosquito midgut lumen independent of human complement, suggesting that IB2 blocks gamete fertilization or ookinete development by steric hindrance. Although Pfs47 appears not to be essential for fertilization, since Pfs47 null gametocytes infect An. stephensi Nijmegen females.¹⁷ However, mosquitoes are fed large numbers of gametocytes in SMFAs, so that even if the process was less efficient for the Pfs47 null parasite, enough fertilization events may still occur. IB2 has strong TRA in both An. gambiae G3 and An. stephensi Nijmegen, and disruption of the mosquito complement system in An. gambiae G3 did not affect the TRA, suggesting that it is also independent of mosquito complement.¹⁵ In contrast, mAb JH11, which binds to Pfs47 D1 close to the boundary with Pfs47 D2, increased consistently and significantly the number of midgut oocysts in SMFA.¹⁵ The effect of mAb JH11 may be a result of stabilization of a conformation that facilitates the interaction between gametes during fertilization or between Pfs47 with its mosquito receptor.

Interestingly, potent transmission-blocking antibodies against Pfs48/45 and Pfs230 are directed mainly to a specific domain—similar to observations with Pfs47. Although some Pfs48/45 mAbs that bind to epitopes in D2 and D3 reduce transmission,^32,33 antibodies with the strongest transmission blocking activity target Pfs48/45 D3,³⁴ a domain with six cysteines. To date, only antibodies that target Pfs230 D1 (aa 589–730, with four cysteines) are effective.³⁵

However, there are important differences in the architecture of their target domains and the mechanism of transmission blocking of antibodies against these three 6-Cys proteins. The TRA of both Pfs48/45 and Pfs47 antibodies does not require human complement. In contrast, human complement greatly enhances the TRA of most antibodies that target Pfs230.³⁶ Pfs47 mAbs identified so far with strong transmission-blocking activity recognize reduced Pfs47 protein, indicating that they bind to linear epitopes¹⁵; however, most transmission-blocking antibodies that target Pfs48/45 and Pfs230 only recognize their target proteins when their disulfide bonds are intact (non-reduced state), suggesting they bind to conformational epitopes.^9,32,36,37 Pfs230 mAb 4F12 is an interesting exception because it has been shown to recognize both native and denatured Pfs230, and it has effective transmission blocking in the absence of human complement.³⁸ The antibodies that do not require human complement for transmission blocking may disrupt domain structure and/or block access to interacting proteins.

The Pfs47-Del2 antigen, which generates antibodies that block transmission, forms a disordered segment toward its N-terminal region (E190-T202) followed by three β-strands of the β-sandwich in Pfs47 D2, according to the AlphaFold2-predicted Pfs47 structure (Figure 1B–D). Positively charged aa’s predominate in the disordered region, which is followed by a region where non-polar and negatively charged aa’s predominate (Figure 1B).

Individuals in malaria-endemic areas generate antibodies against Pfs47³⁹; however, no association has been established between transmission-blocking activity and natural anti-Pfs47 antibody levels.⁴⁰ This is not surprising, as we have found that antibodies to the most immunogenic Pfs47 domains (D1 and D3) do not block disease transmission.

The efficacy of a P47-based TBV was tested in vivo with P. berghei. Full-length recombinant Pbs47 was expressed in E. coli as a fusion protein with thioredoxin. Immunization with full-length Pbs47 generated antibodies that mostly recognized Pbs47 D1 and Pbs47 D3, but not Pbs47 D2, and had no significant transmission blocking in direct mosquito feeding on mice—similar to vaccinating with full-length Pfs47. Furthermore, a region of Pbs47 D2 similar to Pfs47-Del2 (Pbs47-Del1, with the two cysteines substituted for alanine) generated polyclonal antibodies with significant TRA in both active and passive immunization assays, with IgG concentrations as low as 50 μg/mL in mouse serum. Affinity-purified Pbs47-Del1rabbit antibodies had strong TRA (> 80%) after passive immunization in mice with 1 μg/mL IgG,⁴¹ suggesting that increasing the immunogenicity of the Pfs47 antigen may lead to high transmission blocking in vivo.

OPTIMIZATION OF A Pfs47-BASED TBV

In early experiments, immunization with Pfs47-Del2 peptide together with a cytosine phosphoguanine (CpG)-based adjuvant (Magic Mouse^©, Creative Diagnostics, Shirley, NY) generated antibodies with significant and reproducible transmission-blocking activity (> 89% TRA) at a concentration of 200 μg/μL IgG.¹⁵ A 58-aa Pfs47 target antigen (aa 178–235) that includes Pfs47-Del2 was then linked to a virus-like particle (VLP) using the AP205-SpyCatcher:SpyTag platform system to enhance immune recognition and increase antibody titers.⁴² Of the different antigen combinations tested, sera obtained after an initial immunization with VLP-Pfs47 followed by a boost with Pfs47 peptide gave the highest transmission blocking, requiring antibody concentrations as low as 5 μg/μL IgG in SMFA.⁴³ However, binding specificity tests indicated that the majority of the antibodies produced were directed to the virus portion of the VLP and not the Pfs47 antigen⁴³; thus, other platforms are being tested to optimize Pfs47 antigen presentation further (unpublished results).

The potential of delivering the relatively small (52–58 aa) Pfs47 antigen using microneedle technology was explored, because this vaccine delivery platform offers substantial advantages.⁴⁴ Antigen-loaded dissolving microneedles (DMNs) are micrometer-size structures that can be applied to the skin in a Band-Aid-like fashion. This avoids the need for a needle injection and could, in principle, be kept at room temperature for an extended time, simplifying vaccine storage and delivery under field conditions. Gelatin-based microneedles containing Pfs47 antigen, and CpG (toll-like receptor 9 agonist) as an adjuvant, were manufactured successfully and tested for stability and solubility. The DMNs punctured the mouse ear skin and dissolved efficiently. The Pfs47 antigen in DMNs showed no signs of degradation after 1 week of storage at room temperature. The CpG adjuvant in the Pfs47 DMNs was also active, and stimulated toll-like receptor 9 signaling in reporter cells and splenic dendritic cells.⁴⁵ The in vivo immunogenicity of Pfs47 DMNs remains to be established.

CONCLUSION

Further development and clinical testing of a Pfs47-based malaria TBV is called for, based on the strong experimental evidence that antibody binding to a specific region of Pfs47 blocks malaria transmission and that this target region is immunogenic. At present, it is not known whether naturally acquired anti-Pfs47 antibodies are protective nor whether antibodies to immuno-dominant domains, which do not block transmission, limit the access to Pfs47 D2 epitopes where antibody binding could disrupt transmission. It is also not clear whether a Pfs47 D2-based vaccine would be boosted by natural malaria infections. Because the transmission blocking mechanism of the Pfs47 target antigen appears to be independent from that of other TBV targets, testing whether this antigen could synergize when combined with other vaccine antigens is warranted.