Structural vaccinology of malaria transmission-blocking vaccines

Abstract

Introduction:

The development of effective vaccines remains a major health priority to combat the global burden of malaria, a life-threatening disease caused by Plasmodium parasites. Transmission-blocking vaccines (TBVs) elicit antibodies that neutralize the sexual stages of the parasite in blood meals ingested by the Anopheles mosquito, interrupting parasite development in the vector host and preventing disease spread to other individuals.

Areas covered:

The P. falciparum gametocyte surface antigens Pfs230, Pfs48/45, and Pfs47, the parasite ookinete surface protein Pfs25, and the male gametocyte specific protein PfHAP2 are leading TBV candidates, some of which are in clinical development. The recent expansion of methodology to study monoclonal antibodies isolated directly from humans and animal models, coupled with effective measures for parasite neutralization, has provided unprecedented insight into TBV efficacy and development.

Expert opinion:

Available structural and functional data on antigen-monoclonal antibody (Ag-mAb) complexes, as well as epitope classification studies, have identified neutralizing epitopes that may aid vaccine development and improve protection. Here, we review the clinical prospects of TBV candidates, progress in the development of novel vaccine strategies for TBVs, and the impact of structural vaccinology in TBV design.

1. Introduction

Malaria is a life-threatening disease caused by five Plasmodium parasite species (P. falciparum, P. vivax, P. ovale, P. malariae and P. knowlesi) that infect humans through the bites of infected female Anopheles mosquitoes. The intensity of transmission depends on various factors related to the parasite, vector, human host, and environment. P. falciparum and P. vivax pose the greatest public health threat and cause almost all mortality and morbidity associated with the disease. Clinical symptoms of malaria are the result of multiple asexual cycles within human erythrocytes, followed by rupture and release of more blood stage parasites that invade and deplete more erythrocytes. In 2018, there were an estimated 405, 000 deaths and 228 million cases of malaria worldwide, with the highest burden of malaria cases (93%) and deaths (94%) in the WHO African region [1]. Children under five years of age are the most vulnerable group, accounting for 67% (272, 000) of all malaria deaths worldwide [1]. Increasing resistance to anti-malarials and insecticides highlights the need for alternative treatment strategies and preventive measures that include vaccines to fight this global disease.

Malaria vaccine development has been hindered by the complexity of the multi-stage parasite life cycle, as well as an incomplete understanding of host-pathogen interactions that should be targeted for effective vaccination [1]. RTS,S/AS01, the most clinically advanced malaria vaccine, is based on the parasite circumsporozoite protein (CSP) [1,2]. RTS,S/AS01 has been shown to significantly reduce malaria and life-threatening severe malaria in young African children, although its partial efficacy and longevity are suboptimal [1]. Many other potential vaccines currently in development can be broadly categorized as (1) pre-erythrocytic vaccines that aim to prevent infection such as RTS,S/AS01, (2) asexual blood stage vaccines that aim to prevent clinical manifestations of disease, and (3) sexual stage vaccines that aim to prevent parasite transmission by the mosquito host [3].

Transmission of parasites from one human to another relies on the generation of male and female gametocytes in the human host [4] (Figure 1(a)). These sexual stage parasites are ingested during a blood meal by an Anopheles mosquito and differentiate within the mosquito midgut to form microgametes and macrogametes. Microgametes then penetrate macrogametes to form zygotes, which then develop into motile and elongated ookinetes. Ookinetes invade the midgut wall of the mosquito where they develop into oocysts, which then grow, rupture, and release sporozoites. Sporozoites ultimately relocate to the mosquito’s salivary glands, enabling inoculation into the next human host to continue the malaria life cycle and transmission of disease (Figure 1(a)) (https://www.cdc.gov/malaria/about/biology/index.html).

Figure 1.

(a) Different life cycle stages of the P. falciparum parasite. It involves two hosts. In the exo-erythrocytic cycle, sporozoites infect hepatocytes and mature into schizonts inside them which rupture and release merozoites. These released merozoites infect RBCs and develop into trophozoites and schizonts in the erythrocytic cycle. Some parasites differentiate into gametocytes that develop inside erythrocytes and are ingested by female Anopheles mosquito during a blood meal. Parasite development in the mosquito is known as the sporogonic cycle (mosquito/transmission stages). The ingested gametocytes develop into zygotes and ookinetes which invade the midgut wall of the mosquito where they develop into oocysts. The oocysts rupture and release sporozoites that make their way to the salivary glands. Inoculation of the sporozoites into a new human host continues the malaria life cycle. TBVs elicit antibody responses against the target antigen of mosquito/transmission stages of the parasite, the gametes and ookinetes, to prevent infection in mosquito and further transmission of sporozoites into a new human host. This figure is created with BioRender.com. Overall view of the structure of (b) Schematic domain organization of all proteins described in this article. TBV candidates are colored in blue and the rest are in gray.

Transmission-blocking vaccines (TBVs) target these parasite sexual stages to prevent parasite transmission to mosquitoes [4–11], with the ultimate goal of effectively reducing malaria transmission at the population level toward malaria control and elimination [4]. Transmission-blocking activity induced by TBVs is expected to be primarily antibody-dependent, although T-cell mediated response may also play a role [3,12,13]. Preventing or reducing transmission will be critical for malaria eradication efforts and will complement existing efforts to prevent incidence of infection and/or clinical disease [4].

TBVs are unique in their mechanism of action as they do not directly prevent infection or clinical symptoms within a vaccinated individual, but rather act upon the parasite in the mosquito vector to prevent sporozoite development and transmission to another human [14]. Multiple gametocyte surface proteins have been identified by generating monoclonal antibodies against sexual stage parasites then screening these for transmission reducing activity [3,11,15–17]. TBV candidates include the ookinete surface protein Pfs25, and the gametocyte antigens Pfs48/45 and Pfs230 which play critical roles during sexual stage development. Monoclonal antibodies isolated from humans and rodents that target these P. falciparum TBV antigens have been recently described (Table 1). In addition to Pfs25, Pfs230 and Pfs48/45, two additional target immunogens of interest are Pfs47, involved in parasite immune evasion in the mosquito vector [18], and PfHAP2 which is expressed on the male gametocyte and microgamete, essential for membrane fusion during fertilization [19].

Table 1.

Structures of selected transmission blocking vaccine candidates with potent inhibitory mAbs.

Immunogen	PlasmoDB ID	PDB	Region	Description	Expression	Monoclonal antibody	Monoclonal Ab Source	Interface BSA (Å2)	CDRs	KD (nM)	Neutralization measure	Antigen residues making H-bonds	Antigen residues making salt bridges	Reference
Pfs230D1M	PF3D7_0209000	6OHG	542–736	# Portion of non-structured domain through domain 1	P. pastoris	4F12	Mice immunized with parasite derived Pfs230	1,500	H3, L1, L2, L3	23.8	42.2% reduction in oocysts at ~1.0 mg/ml 84.3% reduction in oocysts at 0.8 mg/ml	Ala583, Gln585, Thr596, Asp597, Gln598, Ser604, Lys609	-	[10,44]
				# Mutated to remove the N-glycosylation site		rh4F12 (recombinant chimeric 4F12 IgG with human IgG1 Fc)		-	-	-		-	-
Pfs48/45–6 C	PF3D7_1346700	6E62	291–428	# C-terminal 6-cys domain (Domain III) # With N-glycosylation sites intact	HEK293 mammalian cells	85RF45.1	Rats immunized with gametocytes	1,039	H1, H2, H3, L1, L2, L3	3.2 ± 3.1	1–3 µg/ml (IC80)	Asp347, Asp351, Gln355, Glu365, Ser367, Tyr371, Lys394, Asp415	Asp347, Lys413, Lys416	[5]
Pfs25	PF3D7_1031000	6B0A	23–193	# Mutated to remove the N-glycosylation site	HEK293F mammalian cells	1269	Humanized mice (Kymice, possessing human	1,129	H1, H2, H3, L1, L2, L3	3.7 ± 0.3	63 µg/ml (IC80)	Asn87, Glu88, Thr93, Lys108, Cys112, Gln125, Asn126, Asp102, Gln123		[6,9]
		6B0G				1245	immunoglobulin (Ig) repertoire)	803.1	H1, H2, H3, L1, L2	31.0 ± 5.6	65 µg/ml (IC80)	Glu22, Lys24, Asn27, Lys40, Gly95, Lys98, Asn115, Ala149, Asp151, Gly152, Asp157	Lys155
		6PHC				2544	Human	1,005	H2, H3, L1, L3	4.6 ± 1.2	16 µg/ml (IC80)	Ala1, Val3, Thr4, Val5, Asp6, Val8, Gln16, His20, Glu34, Glu35, Thr103, Lys135, Cys136, Ser137, Lys139, Cys140	Lys2
		6PHB				2530		797.1	H1, H2, H3, L2, L3	19.2 ± 1.9	65 µg/ml (IC80)	Asn27, Lys40, Leu42, Glu46, Lys51	Asp28, Glu39, Lys43, Lys47, Lys118, Asp131, Lys135
		6PHD				2586		824.5	H2, H3, L1, L2, L3	2.5 ± 1.2	96 µg/ml (IC80)	Gly117, Lys130, Asp131, Glu133, Cys136, Asp151, Gly152, Tyr154
Pvs25		1Z3G	23–195	-	S. cerevisiae	2A8	Mice	1,330	H1, H2, H3, and fourth hypervariable loop	Within 1–10 nM range	-	Cys57, Glu59, Pro61, Met68, Cys73	Lys44, Glu75	[32]

Transmission-reducing antibodies against Pfs48/45, Pfs230 and/or Pfs25, when present in blood meal [4,6,10,20], significantly inhibit oocyst development measured by an in vivo standard membrane feeding assay (SMFA) in preclinical studies [11,20–23]. The SMFA is considered one of the most reliable assays to assess the effectiveness of TBVs. This assay measures transmission-reducing activity (TRA) and transmission-blocking activity (TBA) defined as percent inhibition of mean oocyst count per mosquito (TRA) and percent inhibition of the prevalence of infected mosquitoes (TBA), respectively [8]. TRA and TBA have been extensively used to describe and compare SMFA results for diverse antigens [8].

In this review, we summarize structural studies of TBV candidates, examine the quality of antibody responses elicited by transmission-blocking antigens (Table 1), and discuss how available structural and immunological data may guide design for more potent vaccine immunogens. We focus on TBV candidates Pfs230, Pfs48/45 and Pfs25, which have been tested in clinical trials (Table 2) and have available structural data defining how potent neutralizing monoclonal antibodies function to block transmission. We also discuss the potential for Pfs47 and PfHAP2 as TBV candidates. Finally, we describe the importance of these findings to improve malaria vaccines using structure-based approaches.

Table 2.

Transmission blocking vaccine candidates currently under preclinical development or in clinical trials.

Immunogen		Vaccine type	Current status	Adjuvant	ClinicalTrials.gov ID	References
Pfs48/45	-	Subunit	Preclinical	-	-
Pfs230	Pfs230D1M-EPA	Subunit	Phase II	Alhydrogel®	NCT02334462	[2,10,44]
	Pfs230D1M-EPA			AS01	NCT02942277
	Pfs230D1M-EPA			AS01	NCT03917654
Pfs25	Pfs25M-EPA	Subunit	Phase I	Alhydrogel®	NCT02334462	[2,6,9]
	Plant-Derived Pfs25 VLP-FhCMB			Alhydrogel®	NCT02013687
	Pfs25M-EPA			AS01	NCT02942277

EPA, ExoProtein A; VLP, Virus Like Particle.

2. TBV antigens are members of conserved protein families

TBV antigens Pfs48/45, Pfs230 and Pfs47 are part of the 6-Cysteine (6-Cys) family of parasite proteins (Figure 1(b)); each contains the signature 6-Cys domain. Proteins of this family are expressed in all stages of the parasite life cycle [4,5,24–28]. The 6-cys domain is a β-sandwich formed by two β-sheets with a combination of parallel and antiparallel strands. The domain is approximately 120 amino acids in length and typically contains six positionally conserved cysteines which form three disulfide bonds (C1-C2, C3-C6, and C4-C5) (https://www.ebi.ac.uk/interpro/entry/InterPro/IPR010884/). C1-C2 and C3-C6 pin together the two β-sheets of the β-sandwich, whereas C4-C5 links an ancillary loop to β-strand of the core domain [24,26].

Pfs25/Pvs25 contain four epidermal growth factor (EGF)-like domains (Figure 1(b)). EGF-like domains are thirty to forty amino acids in length stabilized by three disulfide bridges [29–31]. This structurally conserved domain is present in a large number of diverse proteins usually found in the extracellular domain of membrane-bound proteins or in proteins known to be secreted. The EGF-like domain consists of two central β-sheets connected to two shorter β-sheets with six cysteines stabilizing the domain (https://www.ebi.ac.uk/inter pro/entry/InterPro/IPR000742/). The EGF-like domains of Pfs25/Pvs25 have six cysteines with C1-C3, C2-C4 and C5-C6 connectivity along with two central β-strands connected by a B loop [32]. EGF-like domain 1 interfaces with EGF-like domain 3 and 4, forming a triangle arrangement [32].

PfHAP2 belongs to the HAP2/GCS1 (Generative cell-specific protein 1) family of proteins [33] (Figure 1(b)) and is necessary for the fusion of the gametes in fertilization [34]. Fusion is thought to involve a two-phase mechanism where species-specific proteins bring male and female gamete membranes together initially in the prefusion attachment step. This is followed by fusion of the two membranes together and is facilitated by broadly conserved proteins with a HAP2-GCS1 domain expressed only in the male gametocyte and microgamete [34] (https://www.ebi.ac.uk/interpro/entry/InterPro/IPR040326/). The HAP2-GCS1 domain is conserved from plants to lower eukaryotes [34].

3. Pfs230

Pfs230 is a gametocyte and gamete surface antigen currently under clinical evaluation as a TBV candidate [35]. Pfs230 is a sexual stage protein that is composed of fourteen 6-cysteine-rich domains (Figure 2(a)) [10,11]. Upon expression in gametocytes, Pfs230 is processed during gamete emergence from red blood cells in the mosquito midgut [10,36]. Proteolysis occurs between the glutamate-rich motifs and the first cysteine motif domain to yield either 300 and 35 kDa or 307 and 47 kDa fragments [3,37]. Secreted Pfs230 is associated with the parasite plasma membrane, where it forms a stable membrane-bound complex with GPI-anchored Pfs48/45 [10,38,39]. Pfs230 mediates the binding of male emerging gametocytes to erythrocytes to form the exflagellation center [40]. Antibodies raised against Pfs230 were shown to reduce transmission [36], and a disruption of the gene leads to reduced fertilization and oocyst formation [40]. The presence of naturally acquired antibodies for Pfs230, generated by the natural course of infection in individuals from malaria-endemic regions, correlates with the TBA of their sera [41,42].

Figure 2.

Structural basis for transmission-reducing activity of mAb 4F12 against Pfs230D1M. (a) Domain organization of full-length Pfs230 and its truncations. SP signal peptide (b) Overview of the Pfs230D1M structure (PDB: 6OHG). Overlay of Pfs230 domain I with the Pfs48/45–6 C (PDB: 6E62), both domains of Pf12 (PDB: 2YMO) and Pf41 (PDB: 4YS4). Root mean square deviations (rmsd) between Pfs230 and the structures are indicated in brackets. Disulfide bridges are shown as sticks. (c) Binding of Fab fragment of 4F12 on the Pfs230D1M. The Pfs230D1M is represented as a gray surface and epitope is colored in blue. mAb 4F12 recognizes conformational epitope on Pfs230 DI by LCDR1–3, HCDR1 and HCDR3. (d) Polymorphic residues mapped onto the Pfs230 surface. The 4F12 epitope is colored in blue. Amino acid substitutions are shown in yellow and the observed substitutions are also indicated. This figure is modified and adapted from [10].

Pfs230 is a member of the 6-Cys protein family. A fragment of Pfs230 spanning the N-terminal pro-domain (Figure 2(a)) and first three 6-Cys domains has been shown to elicit potent transmission blocking antibodies [11]. Disulfide bridging with the three 6-cys domains appears critical for proper folding of each of these domains, as many Pfs230-specific transmission-blocking monoclonal antibodies did not recognize reduced Pfs230 [10,38].

In an attempt to identify the minimal Pfs230-domain involved in the generation of transmission-blocking antibodies, Pfs230 constructs containing the Pro (pro-domain residues 443 to 590), Pro+I (pro-domain and domain I, residues 443 to 715), Pro+I,II (pro-domain through domain II, residues 443 to 915), and Pro+I,II,III (pro-domain through domain III, residues 443 to 1132) domains (Figure 2(a)) were produced individually in the wheat germ cell-free system [43]. Interestingly, the N-terminal Pro domain, which lacks cysteines, was sufficient to elicit complement-dependent TBA in the SMFA, suggesting that transmission-blocking antibodies may also be directed against non-conformational epitopes [43]. Recombinant Pfs230 domain I construct (Pfs230D1M) (Figure 2(b)), corresponding to amino acid residues 542 to 736, was produced in Pichia pastoris as a properly folded protein and elicited transmission-blocking antibodies in rodents [44]. The primary amino acid sequence of this Pfs230 domain I encodes for one putative N-linked glycosylation site when expressed in Pichia pastoris: Asn585 that is placed near the N-terminus of the construct was changed to a Gln to avoid possible glycosylation [10,44].

The crystal structure of Pfs230D1M suggests that its overall fold (β-sandwich) is similar to those of other members of the 6-Cys family, such as Pf12 (PDB: 2YMO) [26], Pf41 (PDB: 4YS4) [45], and Pfs48/45–6C (PDB: 6E62) [5], despite containing only four cysteines (Figure 2(b)) [10]. The two disulfide bridges in Pfs230D1M closely overlay the disulfides in the other 6-cysteine domains (Figure 2(b)). Both domains of Pf12 and Pf41 are similar to Pfs230D1M (Figure 2(b)). Domain I and II of Pf12 overlay on Pfs230D1M with a root mean square deviation (rmsd) of 1.9 Å over 82 α-carbons and 2.4 Å over 86 α-carbons, respectively. Similarly, domain I and II of Pf41 closely superpose on Pfs230D1M with an rmsd of 1.7 Å over 82 α-carbons and 2.1 Å over 86 α-carbons, respectively. The Pf48/45–6C is also similar to the Pfs230D1M (Figure 2(b)) with an rmsd of 2.2 Å over 91 α-carbons. In general, loops connecting β-strands at one end of the beta-sandwich domain are shorter and closely overlay, while those at the other end are longer and significantly deviating (Figure 2(b)) [10].

The murine mAb 4F12 specific for Pfs230D1M is known to reduce transmission [44]. Complement-fixing activity is thought to be important for antibody targeting of Pfs230 and may subsequently trigger lysis of gametes in the infected mosquito [10,38]. Replacement of the Fc region of 4F12 with a human IgG1 Fc domain, which may fix complement, increased TRA significantly [10]. The transmission-reducing antibodies 4F12 and 5H1 recognize different epitopes on Pfs230D1M. In SMFA, mAbs 4F12 and rh4F12 (with a human IgG1 Fc domain) in combination with 5H1 enhance transmission-reducing activity [10]. Moreover, Pfs230D1M mAbs rh4F12 and 5H1, when used in combination with Pfs48/45-specific mAb 3E12 in SMFA, demonstrate enhanced transmission-reducing activity [10]. A combination TBV that includes both Pfs230D1M and Pfs48/45–6C domain or their fusion proteins elicits strong transmission-blocking antibody responses in mice [11]. Specifically, the fusion protein that includes a Pfs230 pro-domain (residues 443 to 590), domain I and Pfs48/45–6C elicited over three-fold excess transmission-blocking antibody responses than the single antigens alone [11].

4F12 binds to Pfs230D1M through complementarity-determining regions (CDRs) H3, L1, L2, and L3, and demonstrates an interacting buried surface area (BSA) of 1,500 Ų (PDB ID: 6OHG) (Figure 2(c)) [10]. CDRs are loops of the variable domains of antibodies that bind to the antigen and define antigen specificity. Each variable domain of the anti-body heavy and light chain contains three CDRs [46]. The CDR-L1 of 4F12 interacts with the residues Lys581, Tyr582, Ala583, Ser584, Gln585, and Asn586 in the first beta-strand of Pfs230D1M. The CDR-L3 interacts with the residue Asp594 in the subsequent second beta-strand. Further, CDR-L1, CDR-L2, and CDR-L3 interact with the residues Thr596, Asp597, Gln598, Lys600, Thr602, Glu603, Ser604, and Lys607 in the loop between the second and third beta-strand of Pfs230D1M. CDR-L1 also interacts with the residue Lys609 in the third beta-strand [10]. The disulfide bond (593–611) between two beta-sheets at the second and third beta strands seems to stabilize the 4F12 epitope on Pfs230D1M [10]. The CDR-H3 of 4F12 binds Pfs230D1M at the loop between the second and third beta-strands of Pfs230D1M, interacting with Pro601 and overlapping the light chain contacts to residues Asp597 and Lys600. Most interactions between mAb 4F12 and Pfs230D1M are mediated by hydrogen bonds from CDR loops of both heavy and light chains. Amino acid substitution at thirteen residue positions within Pfs230D1M have been identified in over 2618 P. falciparum sequences (https://www.malariagen.net/projects/Pf3k) and are excluded from the 4F12 epitope (Figure 2(d)) [10,44,47].

In the past few years, nanoparticle-based vaccine platforms have been explored to enhance immunogenicity and functional response to malaria vaccine antigens. Chemical conjugation of poorly immunogenic TBV antigens to carrier Exoprotein A (EPA) elicited enhanced immune responses [48]. The Phase 2 trial of a Pfs230D1-EPA conjugated nanoparticle formulated in AS01 or Alhydrogel^® is currently underway (clinicaltrials.gov IDs: NCT03917654; NCT02334462; NCT02942277). Recently, the Outer Membrane Protein Complex (OMPC), a membrane vesicle derived from Neisseria meningitidis, has also been evaluated as a carrier for Pfs230D1M [35]. Chemical conjugation of Pfs230 to OMPC enhanced immunogenicity and functional activity of the Pfs230D1M antigen. Pfs230D1-OMPC conjugates induced a strong Th1-biased response with increased levels of IgG2 compared to its EPA conjugates, which may enhance antibody functional activity in a complement-dependent manner [35]. Together, these studies support the strong clinical data establishing Pfs230 as the leading TBV candidate for malaria.

4. Pfs48/45

Pfs48/45 is another promising TBV candidate that is conserved across strains of P. falciparum, and belongs to a cysteine-rich structural family that includes Pfs230, Pfs47, Pf12, Pf41 and many others [25,49]. Pfs48/45 is expressed on the surface of gametocytes and gametes. The glycosylphosphatidylinositol (GPI)-anchored Pfs48/45 forms a complex with Pfs230 on the gametocyte surface and is required for gamete fusion [4,27,38,40]. P. falciparum parasites lacking Pfs48/45 expression are severely impaired in their ability to form oocysts in mosquitoes [27,40]. In the rodent malaria parasite P. berghei, male gametes lacking the Pfs48/45 orthologue Pbs48/45 are unable to penetrate female gametes and form zygotes [4,27]. Immunization of animals with Pfs48/45 results in sera containing antibodies that when present in a parasite-infected blood meal, block the sporogonic cycle of the parasite within the infected mosquito [4,15,20,22,50–53]. Pfs48/45 is expressed on the surface of gametocytes isolated from human blood, and the presence of Pfs48/45-specific antibodies in individuals from malaria-endemic regions correlates with the TBA of their sera [52,54–59] as assessed by SMFA [4,60].

The sequence diversity of Pfs48/45 is relatively low across P. falciparum strains [4,54,61] which may aid in development of a strain-transcending TBV [49]. Pfs48/45 contains two 6-Cys domains which are separated by a 4-Cys central domain (Figure 2(a)). Like many other Plasmodium surface antigens, full-length Pfs48/45 has been challenging to express and purify in a soluble homogenous recombinant form [49]. The development of stable, soluble recombinant Plasmodium antigens is therefore an area of intense research [4,5,49–51,62–65]. Recently, it has been shown that soluble, correctly folded full-length Pfs48/45 that does not contain GPI-anchor but retains the seven potential N-glycosylation sites intact, can be expressed in large quantities using a Drosophila Schneider-2 insect cell expression system [4]. Truncations of Pfs48/45 have also been evaluated alone and as fusions to GLURP [65,66]. Successful truncations include the central 4-Cys central domain and the 6-Cys C-terminal domain (Pfs48/45–10C), or the 6-Cys C-terminal domain alone (Pfs48/45–6C), and these have been fused to the N-terminal region of the glutamate-rich protein (GLURP) and produced in various expression systems [4,5,11,62,65]. N-linked glycosylation is also crucial for soluble homogeneous Pfs48/45 expression in HEK293 mammalian cells [5].

This available structure of Pfs48/45–6C expressed in mammalian cells provides a guide for structure-based design to improve expression and stability, and to elicit enhanced anti-body response through epitope focusing [5]. Pfs48/45–6C adopts a typical β-sandwich fold that comprises two β-sheets formed by a mix of parallel and anti-parallel β-strands including three disulfide bonds (Figure 3(b)) [4,5]. This fold is predicted to be adopted by at least 14 additional protein members of the 6-Cys family (or s48/45 family) of P. falciparum [5,24,67]. Domain II from the two other structurally characterized P. falciparum proteins Pf12 (PDB: 2YMO) [26] and Pf41 (PDB: 4YS4) [45] reveals a structurally conserved fold (Figure 3(b)). Pf12 and Pf41 may be essential for cell adhesion and immune evasion, though their exact functions remain unclear [5,67]. Both proteins are expressed on the merozoite surface [4].

Figure 3.

Structural basis for transmission-reducing activity of mAb 85RF45.1 against Pfs48/45. (a) Domain organization of Pfs48/45 immunogen. SP signal peptide, GPI glycosylphosphatidylinositol anchor. (b) Overview of the Pfs48/45–6C structure (PDB: 6E62). Overlay of Pfs48/45–6C with the domain II of Pf12 (PDB: 2YMO) and Pf41 (PDB: 4YS4). Root mean square deviations (rmsd) between Pfs48/45–6C and the structures are indicated in brackets. Disulfide bridges are shown as sticks. (c) Binding of Fab fragment of 85RF45.1 on the Pfs48/45–6C domain. The Pfs48/45–6C domain is represented as gray surface and epitope is colored in blue. mAb 85RF45.1 recognizes conformational epitope on Pfs48/45–6C by all CDRs (LCDR1–3 and HCDR1–3). (d) Polymorphism mapped onto the Pfs48/45 surface. The 85RF45.1 epitope is colored in blue. Amino acid substitutions are shown in yellow and the observed substitutions are also indicated. This figure is modified and adapted from [5].

Epitope mapping of Pfs48/45 has identified at least five different epitopes recognized by murine antibodies, with and without transmission reducing activity, that span its primary sequence [5,23,68]. Human, murine and rat mAbs have been identified that target epitope I, IIb, III, and V (Figure 3(a)). mAb 85RF45.1 is derived from rats immunized with whole gametocytes followed by fusion and screening by enzyme-linked immunosorbent assay (ELISA) for Pfs48/45 Abs [15,51,52,69]. mAb 85RF45.1 is a rat IgG1 that recognizes epitope I on Pfs48/45–6 C and has the highest reported TRA [52]. The two most potent, well-characterized mAbs 85RF45.1 [52] and 32F3 [15] bind to Pfs48/45–6C, suggesting the C-terminal domain of Pfs48/45 is a focus of the neutralizing immune response. However, the central domain of Pfs48/45 may also be targeted by transmission-blocking antibodies, and both domains could be considered for TBV design [4,15,62].

The structure of Pfs48/45–6C bound to the most potent transmission-blocking antibody 85RF45.1 has been solved (PDB: 6E62, 6H5N) (Figure 3(c)) [4,5]. mAb 85RF45.1 targets a conserved conformational epitope on Pfs48/45–6C that is bound by all the complementarity-determining regions (CDRs) of the mAb. The 85RF45.1 epitope lies on a face of Pfs48/45–6C that is opposite to the GPI-anchor (Figure 3(c)) and forms a negatively charged patch on Pfs48/45–6 C [4,5]. A total interacting BSA of 1039 Ų between the mAb and Pfs48/45–6C demonstrates an extensive binding surface [5]. Most interactions are mediated by hydrogen bonds or salt bridges between Pfs48/45–6C residues Gly346, Asp347, Asp351, Gln355, Glu365, Lys394, Lys413, Lys414, Asp415, Lys416 and residues from both heavy and light chains (Figure 3(c)) [5]. CDR loop H2 and H3 each present hydrophobic residues (Ile54 and Met102), which form van-der-Waals interactions with hydrophobic patches on the surface of Pfs48/45–6C [4].

mAb 85RF45.1 is the most effective TB antibody identified against Pfs48/45 to date, with an IC₈₀ of approximately 1–3 μg/mL in SMFA experiments [4,5]. 85RF45.1 binds to recombinant Pfs48/45–6C with an affinity of 3.2 ± 3.1 nM. The high inhibitory potential cannot be explained by enhanced affinity alone, as other mAbs such as 32F3 have similar affinity for Pfs48/45. It appears that the high efficacy of mAb 85RF45.1 is associated with epitope location in a functional segment of Pfs48/45, and may block the normal function of Pfs48/45 or disrupt interaction with its binding partner [4,5]. Sequence analysis of Pfs48/45–6C indicated that polymorphisms are rare within the mAb 85RF45.1 epitope. The Pfs48/45–6C epitope residues Asp347, Asp351, Glu365, Lys394, Lys413, Asp415, and Lys416 are conserved [5]. A few Pfs48/45–6C residues at the binding interface exhibit polymorphisms (Ile/Val349, Gln/Leu355, Lys/Glu414, and Lys/Asn416) (Figure 3(d)), but these polymorphisms had no impact on binding affinity [4,5].

Immunization of mice with Pfs48/45–6C generates antibodies that reduce transmission in the standard SMFA, but also generated antibodies with no observable transmission-blocking activity [4]. Designing an effective Pfs48/45–6C vaccine candidate that specifically elicits TBA will require a focused approach driven by a complete understanding of the spatial organization of Pfs48/45–6C domains and its key epitopes. Collectively, this suggests that a Pfs48/45 immunogen-based vaccine could generate broadly neutralizing strain-transcending antibodies that target a conserved functional region on Pfs48/45. Focused targeting of an essential component of the sexual life cycle should block parasite transmission to uninfected individuals and form the basis for a potent TBV.

5. Pfs25

Pfs25 has been tested extensively in clinical studies because of its low sequence variability within P. falciparum strains [6,9]. Blood meals that contain Pfs25-specific antibodies block parasite progression in the mosquito [9,70,71]. Pfs25 is a GPI-anchored protein, consists of four EGF-like domains characterized by a distinct disulfide-bridging pattern, and is expressed on the surface of ookinetes that develop from the fertilized zygote [3,9,32,72,73]. Pfs25 is essential for ookinete survival in the mosquito midgut, and may aid in the penetration of the midgut epithelium and maturation of the oocyst [9,74,75]. Similarly, the P. vivax orthologue Pvs25 is also expressed on the surface of mosquito stage ookinetes and is a leading malaria TBV candidate for P. vivax [6,9]. Immunization of rabbits and rhesus monkeys with Pfs25 and Pvs25 immunogens elicits antibody responses that exhibit potent TRA and TBA in SMFA [6,7,70,76–78]. The first human trial targeting P. vivax malaria transmission assessed a recombinant Pvs25 formulated with Alhydrogel^®. This vaccine developed antibodies in some volunteers with significant transmission-blocking activity, although antibody levels were too low for potent protection [70,79]. In a second human trial (ClinicalTrials.gov Identifier: NCT00295581), the immunogenicity of Pfs25 and Pvs25 was improved by using the adjuvant Montanide ISA 51 as water-in-oil emulsions. Volunteers who completed the lowest scheduled doses of Pfs25/ISA 51 showed significant antibody responses and serum anti-Pfs25 Ab levels corelated with TBA. However, these vaccines were unexpectedly reactogenic, precluding further development [70].

Pfs25 and Pvs25 both contain four epidermal growth factor (EGF)-like domains (Pfs25: residues 4–40, 45–87, 90–131, and 137–174; Pvs25: 7–41, 45–89, 92–134, 141–176) which are stabilized by a conserved pattern of eleven disulfide bridges and attain a similar flat triangular arrangement (Figures 4(a) and (b); 5b). Pfs25 (PDB: 6PHC) [6] and Pvs25 (PDB: 1Z3G) [32] are structurally similar, with an rmsd of 1.72 Å over aligned 153 residues, with a majority of the differences contributed by the loop regions of EGF-like domains and EGF-like domain 4 (Figure 5(c)). In contrast, residues that promote the unique triangular fold are highly conserved, including a salt-bridge swap of residues Glu/Lys22 and Lys/Glu98 between Pfs25 and Pvs25 (Figure 5(a) and (c)) [9]. Pfs25/Pvs25 form a triangular and flat architecture that may enable formation of a protective interlocking sheet with Pfs28 on the surface of the ookinete [3,80,81]. This model is consistent with a role in ookinete protection [3].

Figure 4.

Structural basis for transmission-reducing activity of mAbs 1269, 1245, and 2544 against Pfs25. (a) Domain organization of Pfs25 immunogen. SP signal peptide, GPI glycosylphosphatidylinositol anchor. (b) An overview of the Pfs25 structure (PDB: 6PHC) and its EGF-like domains 1 to 4 are colored in blue, green, orange, and firebrick, respectively. Disulfide bridges are shown as sticks. (c) Surface representation of Pfs25 displaying conformational epitopes of 1269 (site I) (PDB: 6B0A), 1245 (site II) (PDB: 6B0G), 2544 (near to site I) (PDB: 6PHC), 2530 (site III) (PDB: 6PHB) and 2586 (between sites I and II) (PDB: 6PHD). (d) Binding of Fab fragments of 1269, 1245, 2544, 2530 and 2586 on the Pfs25 protein. Pfs25 is represented as a gray surface and epitopes for 1269, 1245, 2544, 2530 and 2586 are colored in red, blue, green, pink and yellow orange, respectively. For each mAb, K_D (nM) and IC₈₀ (µg/ml) values are also indicated. mAbs 1269, 1245, 2544, 2530 and 2586 recognize different conformational epitopes on Pfs25. This figure is modified and adapted from [6,9].

Figure 5.

Sequence similarity, structure homology and molecular details of Pvs25–2A8 Fab structure. (a) Sequence alignment of Pfs25 with Pvs25. (b) An overview of the Pvs25 structure (PDB: 1Z3G) and its EGF-like domains 1–4 are colored in blue, green, orange, and firebrick, respectively. (c) Overlay of Pfs25 (PDB: 6PHC) with the Pvs25 (PDB: 1Z3G) structure. Inset shows conserved residues between Pfs25 and Pvs25 making H-bond contact that promote the triangular arrangement. Root mean square deviations (rmsd) between Pfs25 and Pvs25 is indicated in brackets. (d) Binding of Fab fragment of 2A8 on the Pvs25 protein. Pvs25 is represented as a gray surface and epitope is colored in blue. mAb 2A8 recognize conformational epitope on B loop and central β-strands of the EGF-like domain II by CDRs H1-H3 and hypervariable forth loop. Disulfide bridges are shown as sticks. This figure is modified and adapted from [9,32].

Two transmission-blocking mAbs 1269 and 1245 were derived from Kymouse human immunoglobulin (Ig) loci transgenic mice [9]. mAb 1269 targets EGF-like domain 3 and domain I of Pfs25 (Site I) (Figure 4(c) and (d)). The epitope for mAb 1269 broadly overlaps with the well-studied murine mAb 4B7 epitope in EGF-like domain 3 (PDB: 6B0A) [9]. The binding affinity between mAb 1269 and Pfs25 is 3.7 ± 0.3 nM, and all CDR loops of mAb 1269 contribute to binding. The majority of the contacts between mAb 1269 and Pfs25 are mediated by the heavy chain with seventeen hydrogen bonds, compared to the light chain with two hydrogen bonds. Overall contacts between mAb 1269 and Pfs25 contributes 1,129 Ų of BSA, demonstrating a large interaction surface. A less potent Site II-directed mAb is mAb 1245 which interacts with all four EGF-like domains (Site II) and demonstrates a smaller interacting interface with a BSA of 803.1 Ų compared to 1269. The binding of mAb 1245 with Pfs25 is mediated by the CDR loops H1, H2, H3, L1 and L2 (PDB: 6B0G) (Figure 4(c) and (d)) [9]. The binding affinity between mAb 1245 and Pfs25 is tenfold weaker than for mAb 1269, with an affinity of 31.0 ± 5.6 nM. Most interactions formed by mAb 1245 heavy chain with Pfs45 are contributed by CDR-H3, with eight out of twelve total hydrogen bond partners from mAb. None of these mAbs bind to the epitope recognized by Pvs25-reactive mAb 2A8 [9]. Combination of mAbs 1269 and 1245 exhibited additive inhibitory activity in SMFA [9].

Human mAbs 2530, 2586 and 2544 recognize novel epitopes beyond the previously identified distinct Sites I and II characterized for antibody recognition by transmission-blocking antibodies elicited in Kymice [6]. These novel epitopes include epitopes that combine Sites I and II, and another distinct epitope termed Site III. mAb 2530 interacts mostly with the EGF-like domain II with only a few interactions with EGF-like domain I (PDB: 6PHB) [6]; mAb 2586 contacts the EGF-like domains III and IV (PDB: 6PHD) [6]; and mAb 2544 interacts with the EGF-like domains I, III, and IV (PDB: 6PHC) [6].

mAb 2544 is the most potent mAb targeting Pfs25 with an IC₈₀ of 16 µg/mL [6]. mAb 2544 contacts Pfs25 near the previously recognized site I epitope, but on the opposing side of the B loop of domain III recognized by murine mAb 4B7 (Figure 4(c) and (d)) [6,9]. The detailed analysis of mAb 2544 Fab-Pfs25 crystal structure (Figure 4(c) and (d)) suggested that the antibody buries 1,005 Ų of surface area, and the majority of this interface is contributed by the heavy chain (713 Ų), in comparison to the light chain (292 Ų). All interactions formed by heavy chain with Pfs25 are contributed by CDR H3 that reaches into the pocket of EGF1 domain of Pfs25, and CDR H2 with 21 hydrogen bonds and one salt bridge. All interactions formed by the light chain with Pfs25 are mediated by CDR-L1 and CDR-L3, with four hydrogen bonds to the EGF4 domain of Pfs25. CDR-L2 does not interact with Pfs25 but plays a structural role to help shape the mAb 2544 paratope [6]. The short B loop of domain 3 is the other identified epitope recognized by transmission-blocking mAbs 4B7 [82] and 32F81 [83] which bind P. falciparum Pfs25 [32]. mAbs 2530 and 2586 have a lower potency compared to mAb 2544. 2530 binds to the recombinant Pfs25 with an affinity of 19.2 ± 1.9 nM and demonstrated an IC₈₀ of 65 µg/mL in SMFA [6]. mAb 2530 recognizes the Site III epitope on Pfs25 that is opposite to that of mAb 2586 (Figure 4(c) and (d)) [6]. The Site III-specific mAb 2530 interacts with Pfs25 in similar manner as mouse mAb 2A8 binding to Pvs25. mAb 2530 uses CDRs H1-H3, L2 and L3 to bind to Pfs25, resulting in interacting interface with a BSA of 797.1 Ų. mAb 2586 binds to recombinant Pfs25 with an affinity of 2.5 ± 1.2 nM and showed an IC₈₀ of 96 µg/mL in SMFA. Binding of mAb 2586 and Pfs25 is mediated by CDR-H3 and CDR-L3, each overlapping a portion of both Sites I and II and exhibits an interacting interface with a BSA of ~824.5 Ų (Figure 4(c) and (d)) [6].

Surprisingly, there is no correlation between mAb binding kinetics to recombinant Pfs25 and mAb potency in SMFA [6,9]. mAb 2544 has a binding affinity to recombinant Pfs25 in the low nanomolar range (K_D = 4.6 ± 1.2 nM) that is similar to other identified mAbs with much lower potency (1269: K_D = 3.7 ± 0.3 nM, IC₈₀ = 63 µg/mL; 2586: K_D = 2.5 ± 1.2 nM, IC₈₀ = 96 µg/mL) [6,9]. Available structural epitope mapping data indicated that most of the Pfs25 surface is accessible as displayed on the VLP, and elicits B cell responses after vaccination in humans [6]. Potency appears to be primarily dependent on epitope location and may be related to the angle of approach [6]. Two possible explanations for epitope location driving potency are site-specific antibody blockade of an important interaction motif in Pfs25 as observed in other parasite antigens [84–86], and/or epitope accessibility on the surface of the ookinete. Further insight into how epitope location drives potency awaits a clearer understanding of the role of Pfs25 in the ookinete.

Several monoclonal antibodies to Pvs25 have been developed. Murine mAbs 2A8, 1H10 and 1A5, when present in a P. vivax-infected blood meal, all significantly inhibit parasite growth [32]. The binding affinities (K_D) for 2A8, 1H10 and 1A5 to Pvs25 are in the range of 1–10 nM. The structure of mAb 2A8 in complex with yeast-produced Pvs25 (PDB: 1Z3G) [32] revealed that binding is mediated by the mAb heavy chain interacting with residues at a peak of EGF-like domain 2, making contacts with the B loop and central β-strands (Figure 5(d)) [32]. mAb 2A8 buries a large surface area of 1,330 Ų of Pvs25, as residues from all three CDR loops of the heavy chain variable domain and the hypervariable fourth loop make contacts with 19 Pvs25 residues [32]. Epitopes for 1H10 and 1A5 overlap with 2A8 and likely also bind the B loop of domain 2.

One challenge faced by TBVs is to elicit high enough antibody titers for robust TRA. Data from human clinical trials suggest that higher antibody titers will be required to block transmission in the mosquito midgut relative to levels observed in preclinical work [3,87]. Antibody titers in excess of 100 μg/ml against Pfs25 are expected to be required to achieve effective transmission-blocking immunity, and such high titers require potent immunogenic delivery platforms and/or adjuvants such as the RTS,S/AS01 formulation [3,88]. Efforts using Pfs25 as an immunogen elicited sub-optimal immune responses even when used with adjuvants expected to enhance the antibody response [9,70,89]. Evaluation of each of the four EGF-like domains of Pfs25 produced in Saccharomyces cerevisiae and boosted the responses with the analogous full-length Pfs25 (TBV25H: amino acids 22–193) revealed that EGF-like domain 2 induced the highest transmission-blocking response [3,78,90]. The EGF-like domain II/TBV25H vaccination strategy elicited antibody responses that block transmission in SMFA. In contrast, EGF-like domain 3 was shown to be the target of known Pfs25 transmission-blocking mAbs, including conformation-independent 4B7 and conformation-dependent 1D2 [3,78]. These studies suggest multiple neutralizing epitopes in Pfs25 exist and must be accounted for in further development of Pfs25-based vaccines.

In animal models, immunization with a S. cerevisiae-generated Pfs25 or Pvs25 in aluminum hydroxide adjuvant results in transmission-blocking antibody responses [3,91]. In contrast, S. cerevisiae-produced Pvs25 with the same adjuvant did not produce sufficiently high antibody titers in a Phase Ia clinical trial, confirming that more potent adjuvants are likely necessary for clinical development [3,79]. To overcome this issue, multimerization approaches in combination with various adjuvants have been examined in an attempt to improve the antibody response [6,9,89,92,93]. In mice, immunization with the Pfs25-EPA conjugated nanoparticle formulated in Alhydrogel^® had a 75 to 110 fold increase in Pfs25-specific antibodies compared to an unconjugated Pfs25/Alhydrogel^® formulation [89]. Recently, Scaria et al. evaluated Pfs25 and Pfs230 chemically conjugated to four different carriers: tetanus toxoid, a recombinant fragment of tetanus toxin heavy chain, and recombinant CRM197 produced in Pseudomonas fluorescens (CRM197) or in E. coli (Eco CRM^®). These conjugates were compared to EPA conjugates in mouse immunogenicity studies with Alhydrogel^® and GLA-LSQ. All four novel carriers showed higher functional activity compared to the EPA conjugate [92]. The plant-produced Pfs25 virus-like particle vaccine formulated in Alhydrogel^® elicited potent antibody responses with complete transmission-blocking activity [93] and has undergone a phase I clinical trial (ClinicalTrials.gov Identifier: NCT02013687) [94]. Pfs25 has demonstrated viability as a malaria vaccine candidate, and structural studies have begun to identify epitope targets that may improve the neutralizing antibody response for future Pfs25-based designs.

6. Pfs47

Pfs47 (PlasmoDB: PF3D7_1346800) is another 6-cys family protein present on the surface of P. falciparum gametocytes and female gametes, and mediates parasite evasion of the mosquito immune system [18,95]. Immune evasion is critical for efficient parasite transmission, as disruption of Pfs47 greatly reduced parasite survival in the mosquito [18]. Pfs47 consists of three domains: Domains I (Thr32-Asn154) and III (Asn268-Ala414) have six cysteine residues expected to contain a 6-Cys domain fold (s48/45), while domain II (Ser155 to Gln267) has only two cysteine residues [95]. Immunization of BALB/c mice elicited antibodies against E. coli produced full-length Pfs47 target domains I and III, but not domain II. However, neither polyclonal IgG nor mAbs isolated from this immunization with the full-length protein showed significant TBA. In contrast, mAbs that target the central region of domain II have been shown to have potent TBA [95]. Pfs47-mediated TBA is independent of human complement, suggesting a distinct mechanism from Pfs230 [95]. The protective antibody response elicited by Pfs47 appears independent of epitope conformation, and prevents fertilization by acting at an early stage [95]. There is no reported structural data for Pfs47 to date and these studies will likely inform the development of Pfs47 in the future.

7. PfHAP2

PfHAP2 (PlasmoDB: PF3D7_1014200) is a member of the HAP2/GCS1 family of proteins that are class II fusion proteins with a cysteine-rich extracellular region [33]. The HAP2/GCS1 family consists of conserved male-specific proteins expressed on the male gametocyte and microgamete, and are essential for mediating membrane fusion during fertilization [19]. Disruption of PbHAP2 in P. berghei affects the ability of male gametes to fuse with female gametes and interrupts the essential process of fertilization required for subsequent parasite stages in the mosquito [34,96]. The conserved fusion loop (cd loop) is comprised of 174 to 205 amino acid residues in PbHAP2 and from 178 to 207 amino acid residues in PfHAP2, and has been demonstrated to elicit transmission-blocking immunity [19]. This suggests that the cd loop may be essential for gamete fusion and is a potential TBV target [19]. PbHAP2 and PfHAP2 have shown to elicit functional antibodies in rodents [33,97]. Future structural definition of this essential protein family is expected to drive the development of PfHAP2 as a TBV candidate.

8. Expert opinion

TBVs hold great promise as essential malaria control tools and as central components for eradication efforts. Malaria vaccine development has been traditionally hindered by limited knowledge of the antigens expressed during various stages of the parasite life cycle, lack of structural and functional data for protective antigens, and poor definition of immune correlates of protection. Furthermore, the limitations of available experimental models for studying malaria in vitro and in vivo have complicated the assessment of the role and impact of antibodies within humans during an infection. Finally, translation of antibody potency in the lab to susceptible populations, particularly in endemic settings with preexisting immunity and genetically diverse parasite and human populations, have proven difficult in the past.

Recent advances in clinical studies coupled with comprehensive parasitology and structure-function analyses have begun to overcome some of these hurdles. The available data from recent clinical studies provide proof-of-concept that P. falciparum-targeted antibodies can be raised in humans and block transmission. These studies are also major step forward in defining the predictive value of preclinical assays and the potency of target epitopes for vaccine design. Furthermore, an explosion of recent reports has expanded insight into the structural definition of transmission-reducing mAbs bound to sexual stage parasite antigens. These available structural data define epitopes associated with potent transmission blocking activity and provide a guide to understand the antibody response in terms of titer and durability as well as target epitope specificity. Structural data allows for structure-based design to focus and enhance antibody responses to sites associated with potent inhibitory activity. Structure-based approaches have been successful in preferentially eliciting neutralizing antibody responses to structural epitopes on the HIV-1 Env [98,99], the Respiratory Syncytial Virus F protein [100,101], and the influenza hemagglutinin [102,103], among other examples. Immunogens can now be designed to elicit neutralizing antibody responses against conformational epitopes that are defined by neutralizing antibodies isolated from vaccine-elicited individuals or individuals with naturally acquired immunity against malaria.

Further challenges for TBVs include the need to elicit highly potent antibodies at high titers for long-term TBA. The successful soluble expression and purification of homogenous recombinant proteins, as well as their characterization and functional analysis, pose a second challenging step in malaria vaccine development. Another important open question to be addressed is whether a single antigen is sufficient for full transmission-blocking activity or whether combinations of multiple antigens will be required to increase the breadth of neutralizing antibody responses. The development of novel vaccine strategies and targeted rational design of immunogens are areas of active research and may alleviate these challenges. Finally, TBV development may also be informed by transcriptomic and proteomic data that identify novel immunogens from the sexual stage of parasite life cycle.

Improvements in the presentation of antigens to elicit improved antibody titers provide a strong foundation for another direction of active research that can improve TBVs. Nanoparticle-based platforms such as virus-like particles (VLPs), with carriers such as exoprotein A (EPA) and outer membrane protein complex (OMPC), have already been used to elicit enhanced immune responses against TBV candidates with great success [6,9,35,48,89,92]. TBV candidates could also be linked on the surface of self-assembled protein-based nanoparticles such as Ferritin, as has been successfully used for influenza H1 (clinicaltrials.gov ID: NCT03814720) [102,104,105]. Attaching structure-guided immunogens on such protein-based nanoparticle platforms should improve the quality of antibody responses resulting in optimal transmission-reducing potential. The structural insights on TBV cited in this review provide a guide for development of next generation potent TBVs.

Article Highlights.

• Transmission blocking vaccines (TBVs) elicit human antibodies that neutralize the sexual stages of the malaria parasite in the mosquito vector to prevent sporozoite development and transmission to another human.

• The P. falciparum gametocyte surface antigens Pfs230, Pfs48/45, and Pfs47, the parasite ookinete surface protein Pfs25, and the male gametocyte specific protein PfHAP2 are leading TBV candidates, some of which are in clinical development.

• TBV antigens contain conserved domains including the 6-Cys domain, EGF-like domain and HAP2/GCS1 domain.

• Available structure data for TBV candidates provide a guide for the development of potent novel vaccine strategies and structure-based immunogen design to focus and enhance antibody responses to sites associated with potent inhibitory activity.

• Clinical studies suggest that the nanoparticle-based platforms such as virus-like particles (VLPs) or with carriers such as exoprotein A (EPA) have elicited enhanced immune response against target antigens with great success.