Malaria invasion ligand RH5 and its prime candidacy in blood-stage malaria vaccine design

Abstract

With drug resistance to available therapeutics continuing to develop against Plasmodium falciparum malaria, the development of an effective vaccine candidate remains a major research goal. Successful interruption of invasion of parasites into erythrocytes during the blood stage of infection will prevent the severe clinical symptoms and complications associated with malaria. Previously studied blood stage antigens have highlighted the hurdles that are inherent to this life-cycle stage, namely that highly immunogenic antigens are also globally diverse, resulting in protection only against the vaccine strain, or that naturally acquired immunity to blood stage antigens do not always correlate with actual protection. The blood stage antigen reticulocyte binding homolog RH5 is essential for parasite viability, has globally limited diversity, and is associated with protection from disease. Here we summarize available information on this invasion ligand and recent findings that highlight its candidacy for inclusion in a blood-stage malaria vaccine.

Keywords:

• invasion
• malaria vaccine
• Plasmodium falciparum
• RH5
• vaccine development

Introduction

Malaria, caused by protozoan parasites of the order Apicomplexa, remains one of the most important vector-borne human diseases to this day. Plasmodium falciparum infections alone account for over a million deaths annually,¹ and P. falciparum has had a profound impact on infants and children in sub-Saharan Africa, where the emergence of drug-resistant strains of the parasite have spread across the continent, rendering affordable chemotherapy such as chloroquine and sulfadoxine-pyrimethamine ineffective, and is threatening the effectiveness of artemisinin-based anti-malarials.² Malaria also wreaks havoc in many other epidemiological groups and population settings. It is a major international public health problem, dramatically undermining worker productivity and draining national budgets. In recent years there has been a shift from managing this disease and reducing severe symptomatic cases back toward eradication and elimination. A major goal in this approach is the development of new prophylactic agents such as drugs or vaccines. The spread of drug-resistant P. falciparum has made vaccine research all the more urgent as vaccines hold the greatest potential for reducing malaria-associated morbidity and mortality in areas with the most intense transmission, as well as preventing malaria among travelers to those regions.

The symptoms of malaria disease occur during the erythrocytic stage of the parasite, dominated by anaemia and associated complications, and are caused by the cyclical invasion, multiplication and release of merozoites from red blood cells (RBCs). Producing a successful blood stage vaccine that interrupts this cycle will reduce clinical disease, and many blood stage antigens have been identified as potential vaccine candidates, the most studied being MSP1 and AMA1. Both MSP1 and AMA1 are genetically diverse antigens, with multiple non-synonymous mutations, but are immunogenic and antibodies to these antigens in individuals from malaria endemic regions have been associated with naturally acquired immunity,^3-8 suggesting they could be potent vaccine candidates. However, early clinical trials have shown that although vaccine-induced anti-AMA1or anti-MSP1 antibodies are produced after immunization, they are not associated with protection against disease, or protection has been only toward the homologous (vaccine) strain and they do not elicit strain-transcending immunity.^9,10 Thus, the main barrier to blood stage malaria vaccine development is the identification of an antigen able to provoke a strong immune response which is also able to neutralize a wide range of parasite variants. An ideal blood stage vaccine antigen would be highly conserved across a broad spectrum of strains to increase the ability for successful heterologous challenge, and would be essential to parasite viability or reproduction so resistance could not be easily acquired by mutation, or by simply switching off expression of that antigen in favor of an alternative. RH5, a member of the reticulocyte binding homolog family is the latest blood stage antigen to be considered as a vaccine candidate, and is fast becoming a front runner as it appears to be essential to parasite invasion and limited diversity has been observed by sequencing naturally circulating, globally diverse parasite populations, with only 12 non-synonymous mutations currently identified.^11,12

Parasite Invasion

Invasion of free merozoites into new RBCs is a critical pinch-point in the parasite life cycle as the parasite is exposed to the peripheral blood stream, including immune cells and antibody, while they interact with and invade new erythrocytes, yet invasion is accomplished within about a minute.¹³ However, invasion is a complicated process that is not fully understood or delineated, and requires a series of steps at the molecular level, starting with the initial contact and recognition between merozoites and erythrocyte. The merozoite then reorientates itself so that the apical end of the parasite, where the micronemes, rhoptries and dense granules are situated, is closest to the erythrocyte surface. Some of these released proteins bind to RBC surface receptors and directly contribute to the formation of a dynamic tight junction, which moves across the merozoites surface from fore to aft. Invasion finally concludes in the resealing of the erythrocyte membrane and simultaneous formation of a parasitophorous vacuole encasing the parasite inside the host cell.^13,14 It is known that accumulation of intracellular calcium (Ca²⁺) is involved in the signaling pathways of invasion of multiple intracellular parasites.¹⁶ One study suggests that the trigger for the Ca²⁺ accumulation and subsequent release of invasion proteins from the micronemes is associated with exposure to low potassium (K⁺) concentrations in the external micro-environment,¹⁵ but actual commitment to invasion by proteins released after those in the micronemes may rely on triggers originating from actual contact with the RBC surface.¹⁶ However, it has been shown that P. falciparum can thrive in vitro even when exposed to a broad range of K⁺ (as well as Na⁺ Cl^-, and H⁺) concentrations in the culture medium, suggesting K⁺ concentration may not be the absolute trigger for Ca²⁺ accumulation in merozoites.¹⁷ Further, the trigger for Ca²⁺ accumulation in Toxoplamsa gondii has been shown to be controlled by abscisic acid.¹⁸ Erythrocyte attachment and entry pathways vary by parasite species, and even strain, with the most diversity observed so far in P. falciparum, which utilize multiple parasite ligands and red cell receptors.¹⁹ The multiple ligand-receptor parings and subsequent redundancy observed within P. falciparum are likely an adaptive response to the heterogeneous nature of human RBC populations.^20-22 Given this redundancy within the invasion ligands, it is also possible that there is a redundancy in the triggers for egress leading to merozoite maturation and the release of invasion ligands.

The two main protein families have been identified as having crucial involvement during the invasion process are the Erythrocyte Binding-Like (EBL) proteins, homologous to the P. vivax Duffy Binding Protein (PvDBP), and the Reticulocyte binding homologues (RHs), homologous to the P. vivax reticulocyte binding proteins (PvRBP1 and PvRBP2). The EBL family consists of EBA-175,²³ EBA-140,^24,25 EBA-181,^26,27 and EBL-1,²⁸ which are located in the micronemes prior to invasion. Most members of the RH family were first recognized for their homology to the reticulocyte binding proteins of P. vivax²⁹ and P. yoelli³⁰. In P. falciparum, the RH family comprises RH1,^31,32 RH2a, RH2b, RH4,³³ and RH5.^11,34,35 RH2a and 2b are highly identical and are encoded by 2 genes that appear to have arisen from a duplication event as both genes are located adjacently on chromosome 13 in a head to head orientation.^20,36-38 EBA-165 and RH3 are pseudogenes within each family that are transcribed but not translated.^39,40

Not all parasite receptors on the RBC surface have been identified, but 2 major invasion pathways have been characterized: ligands from the EBL family largely bind to receptors which contain sialic acid residues and thus govern the sialic acid (SA) dependent pathway of invasion, and the RH family, except for RH1, mediate SA independent invasion as they bind to RBC receptors which do not contain sialic acid residues (Fig. 1). The glycophorins (GP) comprising GPA,⁴¹ GPB,^42,43 and GPC⁴⁴ are the major sialylated proteins on the RBC surface and have been identified as the receptors to which EBA-175, EBL-1 and EBA-140 bind, respectively. Neuraminidase, chymotrypsin and trypsin are enzymes that respectively cleave sialic acid residues, specific peptide bonds, and general proteins, and are a useful tool in defining the characteristics of RBC receptors. The receptor to which EBA-181 binds has yet to be identified, but has been found to be neuraminidase and chymotrypsin sensitive, and trypsin resistant (“E”),^26,45 and it does also bind to band 4.1, a protein located inside the erythrocyte.⁴⁶ RH2a, RH2b, RH4 and RH5 all utilize the sialic acid independent pathway. RH1 is an exception as it binds to an as-yet unidentified receptor which is neuraminidase-sensitive, chymotrypsin- and trypsin-resistant receptor (“Y”),^31,32 thus utilizing the sialic acid dependent pathway (Fig. 1). The receptor(s) to which RH2a and RH2b bind have not yet been identified, but RH2b recognizes a neuraminidase and trypsin resistant but chymotrypsin sensitive (“Z”).^{20,47, 48} RH4 binds to complement receptor 1,^49-51 and RH5 binds to basigin.⁵² The release of ligands for invasion follows a specific sequence with the proteins of the micronemes (EBL and AMA proteins) being released first followed by release of the rhoptry proteins, including the RHs,^15,16 although the exact timing and function of RH5 within the invasion process has yet to be fully elucidated. RH5 lacks a transmembrane domain or membrane anchor but is indirectly linked to the membrane via a complex with P. falciparum RH5 interacting protein (PfRipr), a protein released from the micronemes,⁵³ and with Cysteine-rich protective antigen (CyRPA), a GPI-linked protein also localized to the micronemes.⁵⁴ The expression of the RH proteins has been shown to correlate with the invasion phenotypes of different parasite lines. The laboratory strains Dd2, MCamp and FCR3 express higher levels of RH1 and low levels of RH2a/2b and as such are sialic-acid dependent strains,^55-57 while 3D7, HB3, 7G8 are sialic acid independent^55-57 as they express higher levels of RH2a/2b and lower levels of RH1.^20,32,58 The ability to switch between invasion pathways due to selective pressure is most clearly observed in the laboratory isolate Dd2, which switches to the non-sialic acid pathway in the presence of neuraminidase-treated RBCs,⁵⁹ and is associated with an up-regulation of RH4.^60,61

Figure 1. Schematic of invasion ligands strictly from the Erythrocyte Binding Like (EBL) and Reticulocyte binding Homologue (RH) families, and the erythrocyte receptors to which they bind, as far as they are known, visually organized to show whether they invade RBCs by the sialic acid dependent or independent pathways. Invasion ligands from the EBL and RH families are located at the apical end of merozoites prior to invasion within the micronemes and rhoptries, respectively. The only ligand essential for invasion identified so far is RH5, which is not anchored directly itself to the merozoite surface and binds to basigin (circled). GP – glycpophorin; ‘E’, ‘Y’, and ‘Z’ are unidentified receptors but their enzymatic profiles have been established; CR1 – complement receptor 1; BSG - basigin.

RH5

RH5 is located in a subtelomeric region of chromosome 4, upstream from RH4⁶¹ that contains species-specific genes involved in host-parasite interactions.⁶² It lacks a C-terminal transmembrane domain and the full-length translated protein is much smaller than the other RH proteins at ∼63 kDa.^11,34,35 A processed fragment of 45 kDa product have been observed in native parasite lysates,^11,34,35 thought to represent the C-terminus portion, with the N-terminus fragment being lost upon cleavage prior to binding to basigin,³⁵ but the exact sites of cleavage and the timing at which the protein is processed is not clear, and products of 28 and 31 kDa^11,63 were observed in the absence of the 45 kDa fragment, suggesting further cleavage events occur in RH5 at some point during the invasion process. Attempts at generating a RH5 knock-out have not been successful,^11,35 suggesting this protein pays an essential function in the overall invasion process. Although RH5 is located in the rhoptries at the apical end of merozoites,^11,34,35 it does not co-localize with other rhoptry proteins, like the PfRhopH2 and PfRAP1, which are known to reside in the rhoptry bulb,^11,34,35,64 or with those proteins in the rhoptry neck^63,65 but the processed 45 kDa fragment does form a complex with PfRipr in late-stage schizonts and in merozoites, and this RH5/PfRipr complex interacts with CyRPA at the apical end of merozoites during invasion.^53,54,63 If co-localization with PfRipr then CyRPA, occurs at the formation of the moving junction, it would only be accessible to antibodies for a short time of about 30 seconds as the merozoites makes contact and re-orientates to form the junction.⁶⁶ If RH5 is exposed between reorientation and junction formation only, the time for antibody to access the ligand is even shorter, closer to 11 seconds.⁶⁶

RH5 mediates invasion through the sialic acid independent pathway and binds to basigin on the erythrocyte surface.⁵² Basigin, also known as CD147, EMMPRIN and M6, is a member of the Ok blood group and has many biological functions including embryo implantation, spermatogenesis,⁶⁷ and retinal development.⁶⁸ It has been shown that all P. falciparum strains have the same dependency on basigin for invasion as anti-basigin inhibited parasite invasion in both lab and freshly isolated strains.⁵² Although the basigin sequence is highly conserved among humans and other higher apes, such as chimpanzee (Pan troglodytes) and gorilla (Gorilla gorilla), it has been shown that the human ortholog bound RH5 with a fold10- greater affinity than the non-human basigins, which bound poorly or did not bind at all,⁶⁹ suggesting only a few residues are responsible for binding between RH5 and basigin. Further, all non-human primate basigins encode a glutamine at residue 191; in humans, this is a lysine residue. Mutation of the chimpanzee basigin to contain a lysine at this position allowed RH5 to bind, highlighting the importance of this residue within the RH5 binding specificity to human RBCs. Such a mutation in basigin that prevents RH5 binding but does not prevent basigin function could be a target for anti-malarial therapeutics. The crystalline structure of RH5 alone⁷⁰ and of RH5 with basigin⁷¹ have been determined. RH5 forms a flat ‘kite-shaped’ or ‘elliptical’ structure, with basigin (which exists on the surface of erythrocytes in its short isoforms containing 2 domains⁵²) binding at one tip of RH5 with predominantly one domain only, creating an asymmetrical complex.⁷¹ However, as it appears there is a limited involvement of the basigin side chains for binding, the potential for mutations within basigin to prevent RH5 binding and subsequent invasion is also limited.⁷¹ The residues from each protein which are directly involved in binding have also been identified, including K191 of basigin as predicted, though none of the polymorphisms of RH5 observed from infections or cultures in human erythrocytes appear to be directly involved in binding with human basigin. Earlier work from Hayton et al.^11,72 showed that RH5 polymorphisms at residues I204K and I407V enable P. falciparum parasites to switch their ligand-receptor interactions and adapt to different RBC receptors, and the observed ability of strains from natural human infections in Africa to invade Aotus (New World) monkeys is likely from the diversity of ligand-receptor interactions exploited by P. falciparum in its various invasion pathways. The polymorphisms seen in the I204K and I407V codons may influence the binding to Aotus-specific RBC receptors, but they do not appear to be directly involved in binding to human basigin.⁷¹

Anti-RH5 Antibodies from Natural Infection

The prevalence of anti-RH5 antibodies in individuals naturally exposed to malaria is not very high, suggesting this antigen is not a primary target of naturally acquired immunity^65,73 but native RH5 is vulnerable to naturally acquired antibody.⁷⁴ Antibody responses to RH proteins in individuals from Peru with malaria were significantly higher in asymptomatic vs. symptomatic individuals, suggesting an association with the development of clinical immunity, but only 11% of parasitaemic individuals had anti-RH5 antibodies, despite being naturally exposed malaria.⁷³ Others have shown a similarly low prevalence of RH5 antibodies in individuals from malaria endemic regions. In Kenya, only 16% of individuals with malaria had anti-RH5 antibodies present.⁶⁵ Anti-RH5 responses from low and high endemicity areas in Senegal showed there was a significant age-association with RH5 positivity in high but not low endemicity areas, suggesting repeated exposure does lead to production of anti-RH5 antibodies,⁷⁵ and anti-RH5 responses have been shown to be associated with protection from high density parasitaemia during re-infection after treatment in Papua New Guinea,⁷⁶ and has been associated with an extended time to re-infection in Malian populations.⁷⁴ This poor immunogenicity in natural infections may well be explained by the supposition that RH5 is exposed only transiently during cell attachment and entry.⁵³ Although these data show that naturally acquired antibodies are associated with protection from various indicators of disease, they are not sufficient to show that anti-RH5 antibodies specifically are protective, and despite these associations, it is possible that anti-RH5 antibodies are not significant causal contributors to protection. However, it is notable that RH5 is analogous to the cryptic neutralizing epitopes of some important viral agents, such as HPV L2,⁷⁷ HIV env,⁷⁸ Flu HA,⁷⁹ the epitopes of which are being intensively studied as the potential bases of vaccines against those agents. Like them, RH5 is intimately involved with the processes of cellular entry and even if it is only transiently exposed to the immune system during this time, it is nevertheless susceptible to neutralization by antibody. As with the broadly neutralizing antibodies of HIV, only a small percentage of patients produce a significant anti-RH5 response, and then only after prolonged chronic exposure, yet this does still correlate with relative resistance to disease progression.^73,75 A clear link between anti-RH5 titers and protection needs to be determined to establish the vaccine candidacy of RH5 further, but available data shows promise.

RH5 Recombinant Antigens

Large scale production of pure, correctly conformed antigens are required for successful in vitro and in vivo characterization of vaccine candidates, yet producing recombinant malarial antigens is not straight forward due to the AT-rich malarial genome and overall difficulty of expressing plasmodial proteins in heterologous expression systems.^80-83 Partial recombinant RH5 proteins, covering residues Asn-31 to Val-174³⁴ or Asn-191 to His-359³⁵ both produced by E. coli expression systems, did bind to erythrocytes, but anti-RH5 sera against these fragments did not inhibit parasite invasion in vivo, suggesting although these regions of RH5 do each contain epitopes required for binding to basigin, they do not contain the neutralizing epitopes targeted by anti-RH5 antibodies, or that truncated RH5 antigens are not able to fold correctly thus unable to produce specific antibodies. Full-length recombinant RH5 antigens have successfully been produced in the mammalian cell expression systems,^52,84 viral-vectored vaccine constructs,⁶⁵ wheat-germ cell free^64,85 and E. coli expression systems,⁸⁶ and Drosophila S2 cells.⁷¹ These recombinants are all based on the 3D7 sequence and are all cable of producing neutralizing antibodies, suggesting they are each fully conformed and functional antigens, and clearly showing that RH5 is both immunogenic and anti-RH5 sera does mediate potent inhibition of invasion: in vitro assays with purified rabbit IgG from, RH5 synthesized via the viral-vectored vaccine construct E. coli, mammalian, and Drosophila expression systems all showed ∼75% −90% (inferred) inhibition at 10 mg/mL.^65,71,84,86 The wheat germ cell free system showed a neutralization effect of ∼70% inhibition of invasion in vitro at 0.5 mg/mL with purified IgG from mouse sera.⁸⁵ Further, production of monoclonal antibodies from mice immunized with the wheat germ recombinant RH5 that completely abrogate RBC invasion have been developed,⁶⁴ showing that there are critical epitopes within RH5 that are able to elicit specific antibodies with a high neutralizing activity.

Neutralizing epitopes on RH5 and potential mechanisms of action of inhibitory antibodies

Although it has been shown that antibody against RH5 can effectively neutralize parasites, the details of the exact mechanisms by which anti-RH5 antibody prevents parasite invasion have yet to be characterized fully, though the use of anti-RH5 mAbs, produced from mice immunized with a viral-vectored full-length recombinant RH5, have been employed to elucidate this process.^63,65 Monoclonal antibodies that were shown to be capable of blocking the RH5-basigin interaction completely in the avidity-based extracellular interaction screen (AVEXIS),^52,87 were not very potent inhibitors of invasion (∼40–50% inhibition). Conversely, potent inhibition of invasion was observed by anti-basigin mAbs that did not block the RH5-basigin interaction in the AVEXIS assay,⁶³ showing that absence of the ability to interrupt the RH5-basigin interaction was not a predictor of neutralizing ability in growth inhibition assays and does not fully reflect the in vivo dynamics of the invasion process. Thus it is possible that anti-RH5 mAbs that potently inhibit invasion may function by mechanisms other than direct inhibition of the RH5-basigin interaction. Mapping the targets of these mAbs, QA5 and 9AD4, revealed 2 linear epitopes within the RH5 antigen sequence: YGK(C/Y)IAVDAFIKKI (residues 200–213), and TNGIR(Y/F)HYD (residues 353–361), where C/Y and Y/F represent the location of 2 of the known non-synonymous changes in the RH5 antigen, suggesting immune pressure may be acting at these sites, though the presence of these changes did not affect binding of these mAbs to heterologous epitopes.⁶³ Although these epitopes are separated on the linear RH5 antigen, competitive binding assays suggest they are spatially related in the context of the fully confirmed RH5 protein, and likely close to the basigin binding site. Crystal structures of 9AD4 and QA1 in complex with RH5 provided further information the inhibitory epitopes targeted by these antibodies.⁷¹ QA1, a mAb shown to block the RH5-basigin interaction and parasite growth, binds to residues on the RH5 tip, overlapping with sites where the N-terminus domain of basigin binds. However, the 9AD4 mAb, which does not block the RH5-basigin interaction but is a high inhibitor of parasite growth, actually binds to different epitopes, according to the crystal structure, compared to those identified in peptide binding assays. The epitopes are close to but not overlapping with the basigin binding epitopes, suggesting the high neutralizing activity of this mAb is likely due to steric hindrance preventing the RH5-basigin interaction.⁷¹

Studies from our laboratory identified 3 anti-RH5 mAbs produced from mice immunized with the wheat germ recombinant RH5 antigen, 2E11, 5A03, and 5A08, that were found to completely block parasite invasion.⁶⁴ To identify the epitope to which the mAb 5A08 targets, a virus-like particle strategy was employed.^88–91 VLPs utilize the intrinsic ability of viral structural proteins to self-assemble into non-infective particles while retaining their antigenic similarity to the authentic virion, thus making VLPs an excellent vaccine candidate against the virus from which they were originally derived. However, the coat protein of the bacteriophage MS2 has the added advantage that it has surface-exposed loop in its coat protein that tolerates insertions without disruption of protein folding or VLP assembly, and it encapsids the nucleic acid encoding the coat protein and any guest peptide it displays within that loop.^89,90 When the guest peptides expressed in the surface of the MS2-VLP are random sequences of varying lengths, and when expressed in bacteria, each VLP that is assembled displays a different guest peptide on its surface, creating a large VLP-library to select against.

Although there is no definite limitation on the peptide size that can be accommodated within the AB-loop, the percentage of random peptide sequences that can support proper coat protein folding and assembly drops as the peptide gets longer,⁹⁰ and the sizes of insert tested so far are in the 6- to 10-mer range. The anti-RH5 mAb itself was then used against a VLP-library containing 6-, 7-, 8- and 10-mer inserts to affinity-select for its own target epitope,⁸⁹ identifying the 8-mer insert SAIKKPVT, with the RH5-spevcific epitope identified as being the central 4 amino acids, AIKK, correlating to residues 28–31, a linear tetrapeptide present at the N-terminus of the protein.⁶⁴ Sera from immunization with this epitope, again expressed in the context of a virus-like particle (VLP), was able to inhibit invasion to the same degree as the parent mAb, demonstrating that even a single epitope of RH5 could serve as an effective vaccine.⁶⁴ Thus, a systematic and comprehensive analysis of the entire RH5 protein must be carried out to identify other neutralization sensitive epitopes.

The main advantage of the VLP system is the high immunogenicity of peptides displayed multivalently on the MS2 surface permits the use of the affinity-selected VLP itself as a vaccine immunogen, thus integrating both epitope discovery and immunization functions into a single platform. An epitope-specific VLP-based vaccine may have advantages over a more traditional subunit vaccine as clinical trials conducted with VLP-based biologics have demonstrated that these products have good safety profiles,^92-95 and even at low doses, VLPs elicit high-titer and remarkably durable antibody responses, potentially obviating the need for adjuvants,⁹⁶ although the VLP immunogenicity does depend on high valency epitope display.^88-90,96 Further, as the VLPs can be produced at high levels in bacteria, they remain easy and cheap to manufacture, and the particle itself is relatively stable. These features are particularly important in the developing world, where the burden of malaria is the highest.

RH5 as a Vaccine Candidate

RH5 shows limited sequence diversity with only 12 non-synonymous amino acids changes (of 526 residues) identified across multiple laboratory strains¹¹ and naturally circulating isolates.¹² The prevalence of these non-synonymous amino acid changes show only 5 (Y147H, H148D, S197Y, C203Y, and I410M) are considered as the most common variants in global parasite populations, being present at more than 10% frequency in any given population.⁸⁴ Interestingly, only 2 of the 12 identified polymorphisms, Y148H and H148D, are from field isolates and are not observed in any laboratory strain sequenced so far, which suggests any as-yet unidentified polymorphisms would be at very low frequency. Although the level of sequence polymorphism is not the only factor that predicts whether a candidate antigen will prove to be a successful vaccine component, it is a significant contributor. As well having limited diversity, the support of RH5 as a promising vaccine candidate is strengthened with the fact that no inverse relationship between RH5 genotype and invasion inhibition from anti-RH5 antibodies has been observed, and anti-RH5 antibodies against 3D7 elicit potent strain-transcending parasite-neutralizing antibodies in vitro.^71,84,86 Although these data are extremely promising, there is currently no absolute in vitro assay that can accurately predict human in vivo protection and the contrasting data available for other blood stage antigens raises uncertainties about the use of these inhibition assays to predict vaccine outcomes in vivo. AMA1 was originally identified as it was shown to be the target of parasite neutralizing antibodies in invasion assays,⁹⁷ and even though purified IgG from malaria-naïve adults immunized with various AMA1 vaccine formulations in a Phase 1/2a trial was inhibitory to homologous (but not heterologous) parasites, the inhibition observed was not correlated with in vivo protection after parasite challenge.⁹⁸ However, it has also been shown that antibody titer (determined by ELISA) and inhibition of parasites in vitro with purified IgG from animals immunized with the carboxyl terminal of MSP-1 did correlate with in vivo protection in the Aotus animal model against homologous and heterologous challenge.⁹⁹

It is not clear exactly how much antibody is required to protect individuals from disease and the invasion inhibition assay is currently the most commonly used tool used to directly determine the effect of antibody on parasites, especially in the selection process of vaccine candidates.^99,100 Some reassurance of using invasion assays as an endpoint to assess RH5 is provided by data suggesting that the EC₅₀ of anti-RH5 antibody in invasion assays is lower than that of anti-AMA1 or anti-MSP1, both for vaccine-induced antibody¹⁰¹ and for naturally acquired antibody,⁷⁴ and that in combination with other antigens, such as EBA-175⁸⁵ or RH2 and RH4,¹⁰¹ a synergistic effect is observed, where the ability to inhibit parasite invasion is observed at lower antibody concentrations. As RH5 interacts with PfRipr⁵³ and CyRRA⁵⁴, neither of which can be knocked out, and antibodies against both are able to inhibit invasion, it is likely these proteins are under strict functional constraints preventing maintenance of multiple mutations, thus it may be difficult for the parasite to develop resistance to an effective RH5 vaccine by genetic drift driven by host immune responses.

Members of the P. falciparum EBL and RH families were first identified based on homologs on other malaria species, and one of the earliest examples of a protective blood-stage vaccine was actually purified parasite-derived Py235,¹⁰² and anti-Py235 mAbs are protective against P. yoelii.¹⁰³ It is possible that all Plasmodium spp. have at least one member of these protein families in their invasion armory, even if as-yet unidentified or characterized. However, there are currently no known homologues to RH5 or alternative “RH5-like” antigens in rodent malarias that can be utilized to further characterize this antigen as a vaccine candidate in rodent animal models, leaving the non-human primate (NHP) model as the only opportunity to study in vivo RH5 efficacy before conducting human clinical trials. Previous NHP studies assessing blood stage antigens have shown limited protection was only achieved when adjuvants unsuitable for human administration were used, and protection was toward homologous strains only. However recent data on in vivo efficacy of RH5 in the NHP model not only supports a causal link between invasion inhibition and actual protection for this antigen but significantly show heterologous protection was achieved when immunized using a viral-vectored platform known to be compatible with humans, and suggest that a human RH5 vaccine does have the capacity to deliver protection that is strain-transcending.¹⁰⁴ A Phase 1a trial of RH5 using the viral-vectored delivery platform used in the NHP studies, boosted with modified vaccinia Ankara is underway (https://clinicaltrials.gov/ct2/show/NCT02181088), which will show whether the promising in vitro and NHP data will translate into the human system.

Future Challenges or Closing Remarks

Despite the desperate need for alternative therapeutics due to the emergence and spread of drug resistance, only one malaria vaccine, RTS,S AS01, from a portfolio of about 40¹⁰⁵ has achieved the necessary level of efficacy to reach Phase III clinical trials, and highlights the need to assess all vaccine candidates as thoroughly as possible. RTS,S, which targets the pre-erythrocytic stages, resulted in only 50% protection from clinical malaria in children aged 5–17 months, and only about 30% protection in children aged 6–12 weeks,^106,107 and 3 y after vaccination, no vaccine efficacy could be detected.¹⁰⁸ However, pre-erythrocytic stage vaccines such as RTS,S aim to prevent clinical disease by eliminating parasites before they can enter the cyclical erythrocytic stage. If they are not 100% effective, parasites that have ‘escaped’ will enter the peripheral blood stream, able to cause the full range of clinical complications associated with this stage of the parasite life-cycle. An efficacious blood stage vaccine would prevent these clinical symptoms, and reduction in parasitemia would be beneficial to individuals and the general public health. The main challenges to a blood stage vaccine development include, but are not limited to, selecting an antigen that is both essential and elicits a strong, protective immune response, yet is not so diverse that immunity to only the vaccine strain would develop, and presenting the antigen to the host immune system with a sufficiently immunogenic but safe vaccine adjuvant, the availability and suitability of animal models, as well as gaps in our knowledge of the nature of antimalarial immunity and how it is acquired.^109,110 The malaria antigen RH5 described in this review may well live up to this challenge.

Disclosure if Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

We are indebted to Drs. Gabriel Gutierrez, Amy Noe, David Peabody and Annie Mo for the many valuable discussions during the course of this project.

Funding

Work done in the authors' lab cited in this review was funded by NIAID/DMID contract AI-N01–045210 to Leidos.