Alternative Invasion Mechanisms and Host Immune Response to Plasmodium vivax Malaria: Trends and Future Directions

Plasmodium vivax malaria is a neglected tropical disease, despite being more geographically widespread than any other form of malaria. The documentation of P. vivax infections in different parts of Africa where Duffy-negative individuals are predominant suggested that there are alternative pathways for P. vivax to invade human erythrocytes. Duffy-negative individuals may be just as fit as Duffy-positive individuals and are no longer resistant to P. vivax malaria. In this review, we describe the complexity of P. vivax malaria, characterize pathogenesis and candidate invasion genes of P. vivax, and host immune responses to P. vivax infections. We provide a comprehensive review on parasite ligands in several Plasmodium species that further justify candidate genes in P. vivax. We also summarize previous genomic and transcriptomic studies related to the identification of ligand and receptor proteins in P. vivax erythrocyte invasion. Finally, we identify topics that remain unclear and propose future studies that will greatly contribute to our knowledge of P. vivax.


Introduction
Plasmodium vivax malaria is a neglected tropical disease, despite being more geographically widespread than any other form of malaria [1], and causes 132-391 million clinical infections each year [2]. Compared to P. falciparum, P. vivax has a broader temperature tolerance and an earlier onset of gametocyte development, enabling the parasites to spread through diverse climates [3] and making them more difficult to control and eliminate [4]. Currently, there is no vaccine available for P. vivax, though several preventative medications have been shown to be effective [5,6]. The epidemiology of P. vivax malaria is further complicated by the pathogen's unique ability to form dormant-stage hypnozoites in the host liver cells, giving rise to recurrent relapse infections from weeks/months to years later [7,8]. Relapse infections have substantially impacted progress in malaria control, especially in countries that are approaching elimination [9][10][11].
Plasmodium vivax was previously thought to be rare or absent in Africa because people of African descent often lack the expression of a Duffy blood group antigen, known as the Duffy antigen-chemokine receptor (DARC). It is believed that the fixation of the Duffy negativity trait, and the rarity of P. vivax infection in Africa supports that Duffy-negative individuals are refractory to P. vivax. Unlike P. falciparum, which utilizes multiple erythrocyte receptors for invasion and has merozoite proteins with overlapping and redundant receptor-binding functions, invasion of erythrocyte by P. vivax merozoites exclusively relies on the interaction between PvDBP and DARC expressed on the surface of erythrocytes and reticulocytes. DARC is a glycoprotein on the surface of red blood cells (RBCs) that allows been proposed but remain impractical due to a large initial and continuous investment of labor and infrastructure [41]. The successes of short-term culture utilizing young reticulocytes from placental blood [40,42] and indefinite culture in Saimiri boliviensis and Aotus nancymae monkeys [43,44] shed light on pathogenesis in humans and potential ligands for invasion [39,44]; however, several unanswered questions remain.
While mature asexual P. vivax and its transmissive gametocytes occur in peripheral blood, histological analyses of P. vivax in Aotus and Saimiri monkeys have shown immature gametocytes and few asexual schizonts present in the parenchyma of bone marrow [45]. Asexual schizonts appear to be more concentrated in the sinusoids of the liver [45], suggesting that bone marrow could be a critical reservoir for P. vivax gametocyte development and proliferation. Indeed, the bone marrow reservoirs may suggest that microscopic detection is not ideal for active case detection and treatment of P. vivax until bone marrow samples are accessible. As P. vivax requires reticulocytes for growth [46][47][48][49], the general low proportion of reticulocytes (that make up only 1% of the total number of host erythrocytes) may explain low parasite loads in symptomatic patients [50][51][52] and a lack of observable schizonts in blood circulation [52,53]. Additionally, pathological analyses of S. boliviensis tissues showed that P. vivax infections also affect the lungs and kidneys, both of which had mononuclear infiltrates, higher macrophage levels, alveolar wall thickening, collagen deposition, and type II pneumocyte hyperplasia [44]. The level of tissue damage is parasite-load dependent and determined by the amount of by-product, namely hemozoin, being produced [44]. These findings may imply a large number of asymptomatic P. vivax carriers in the general populations. It is well-known that P. vivax has the ability to relapse from dormant liver-stage hypnozoites, from weeks to years after the clearance of the primary blood-stage infection, and this is a major obstacle to its control and elimination [20,54]. The liver and bone marrow have been shown to be major parasite reservoirs for P. vivax hypnozoites in Saimiri monkey models [45,55], but mechanisms of hypnozoite development remain largely unknown and are difficult to study due to a lack of long-term in vitro culture. Moreover, relapse varies systemically by geographic region and/or seasonal changes in the environment [54]. In regions where P. vivax transmission is intense and stable, relapse is common and enhances local transmission [20,54], whereas, in Africa, P. vivax transmission is relatively low and usually seasonal and unstable [56][57][58]. The rate of relapse is largely unknown. There is, as of yet, no information on the frequency and clinical impacts of relapse in Duffy-negative P. vivax infections, nor reliable biomarkers for relapse detection, due to limited technologies and substantial knowledge gaps in the biology of P. vivax hypnozoites and relapse. Future investigations employing a longitudinal study design that monitors the dynamics and consequences of relapse infections in both Duffy-positive and Duffy-negative individuals will offer deep insights into the epidemiology and biology of P. vivax infections.

Plasmodium falciparum
Our current knowledge of the molecular mechanisms of erythrocyte invasion in several Plasmodium species offers a reference model on candidate invasion ligands in P. vivax. Plasmodium falciparum invades a wide range of red blood cells, from young reticulocytes to mature normocytes. One of the main binding protein ligands is the erythrocyte binding ligand (EBL) family, which includes multiple members, such as EBA-175, EBA-140, EBL-1, and EBA-181. EBA-175 binds to the sialic acid-containing structure on human erythrocyte receptor glycophorin A (GpA) during invasion [59]. The role of the EBA-175 protein has been shown to be critical for erythrocyte invasion, as antibodies raised against EBA-175 prevent binding to GpA in vitro [60,61]. EBA-175 triggers changes in the erythrocyte membrane [62,63], and the shedding of EBA-175 causes uninfected red blood cells to cluster or form rosette, which allows for immune evasion [64]. The host immune responses may explain the polymorphisms and diversifying selection observed in EBA-175 [65]. Other ligands, such as EBA-140 and EBL-1, are known to bind to glycophorin C (GpC) [66] and glycophorin B (GpB), respectively, on the erythrocytes. Unlike GpA and GpC, the GpB exhibits high levels of polymorphisms, particularly in people of African ancestry, suggesting that a strong selective pressure may have provided an evolutionary advantage to parasite invasion [67]. For example, the S-s-U-and Dantu GpB phenotypes both showed moderate protection against invasion; however, this does not hold true for all GpB phenotypes [67][68][69][70]. To the best of our knowledge, the specific receptor for EBA-181 is chymotrypsin-sensitive, trypsin-resistant, and neuraminidase-sensitive to erythrocytic treatment [71,72], although it remains to be identified.
Another important binding protein family of P. falciparum is the reticulocyte-binding homologue (PfRh) that includes PfRh1, PfRh2a, PfRh2b, PfRh4, and PfRh5. PfRh1 binds to an unidentified receptor "Y", which has been characterized to be trypsin-and chymotrypsinresistant and neuraminidase-sensitive [73,74]. PfRh1 is necessary for sialic acid-dependent invasion of human red blood cells [74]. Antibodies raised against PfRh1 have been shown to block invasion by inhibiting calcium signaling in the merozoite [75]. PfRh2a and PfRh2b are identical for much of the N-terminus region, but each has a unique 500 C-terminus region [76] and differential expressions in various P. falciparum lines, including deletions, such as a deletion of PfRh2b in P. falciparum D10 [76,77]. The loss of PfRh2b does not appear to impact invasion or growth of the parasites and suggests compensatory mechanisms for the loss of PfRh2b [78]. PfRh2a binds to more than one receptor on erythrocytes, but these receptors have yet to be identified [79,80]. PfRh2b has been shown to be involved in merozoite calcium signaling [80]. It binds to an unknown receptor "Z" on erythrocytes, which is neuraminidase-and trypsin-resistant and chymotrypsin-sensitive [81]. PfRh4 has been shown to have sialic acid-independent binding activity with the complement receptor type I (CR1) on erythrocytes [82,83]. The PfRh5 complex is composed of PfRh5, Ripr, CyRPA, and Pf113, which collectively promote successful merozoite invasion of erythrocytes by binding to basigin (BSG, CD147) [84,85]. A BSG variant on erythrocytes, known as Ok a-, has been shown to reduce merozoite binding affinities and invasion efficiencies [86]. This variant was reported so far only from people of Japanese ancestry [87]. Previous knockout or double-knockout experiments have indicated that the EBL and PfRh gene families work cooperatively or can functionally compensate for the loss of each other [88,89]. For example, a loss of EBA-175 can activate PfRh4 [88,89]. When EBA-181 expression was disrupted, PfRH2b was no longer functional [89]. When EBA-181 and EBA-140 genes were disrupted, the parasite deleted the PfRh2b gene [89]. Further study is needed to gain a deeper understanding of how they may work synergistically to promote invasion and immune evasion.

Plasmodium knowlesi
Until recently, P. knowlesi was considered primarily a simian malaria that infects Macaca fascicularis, Macaca nemestrina, and Presbystis melalophos [90]. P. knowlesi is now confirmed to cause malarial infections in humans [91]. P. knowlesi has been shown to use different ligands to invade macaques and human erythrocytes [90]. Two gene families, DBL and RBP, are responsible for erythrocyte binding. The DBL gene family comprises PkDBP-α, PkDBP-β, and PkDBP-γ. In humans, the parasite ligand responsible for erythrocyte invasion is PkDBP-α, which binds to the DARC receptor. The other two Duffy-binding proteins, PkDBP-β and PkDBP-γ, bind only to macaque but not human erythrocytes [14]. The normocyte-binding protein Xa (NBPXa) is required for binding in human erythrocytes, but it is not necessary for invasion of Macaca mulatta erythrocytes [92]. Variation in PkNBPXa has been shown to be linked with parasite virulence and severity of disease [93]. The receptors for NBPXa and NBPXb necessary for invasion for either human or M. mulatta erythrocytes have yet to be identified [90]. Unlike P. vivax, both P. falciparum and P. knowlesi can be maintained in long-term culture, making them ideal systems for studying invasion mechanisms [94,95].

Plasmodium cynomolgi
P. cynomolgi is a vivax-like simian malaria that shares many genomic and phenotypic characteristics with P. vivax and has been often used as a reference model of P. vivax [96]. Two gene families, erythrocyte binding-like (EBL) and reticulocyte binding-like (RBL), are responsible for erythrocyte binding and invasion in P. cynomolgi [97][98][99]. The EBL gene family encodes PcyDBP-1 and PcyDBP-2, similar to PkDBP, which binds to the complementary DARC receptor on Duffy-positive erythrocytes. PcyDBP-1 is an ortholog for PkDBP-α, while PcyDBP-2 has no known orthologs with other Plasmodium DBPs [100]. Previous studies have shown no variation in gene copy number of either PcyDBP-1 or PcyDBP-2 among P. cynomolgi laboratory strains [101]. Studies of field isolates of both P. cynomolgi and P. knowlesi have shown that PcyDBP-1 exhibits high levels of nucleotide diversity, as compared to PcyDBP2 or PkDBPs [102]. The RBL gene family is composed of PcyRBP1, PcyRBP1a, PcyRBP1b, PcyRBP2a, PcyRBP2b, PcyRBP2c, PcyRBP2d, PcyRBP2e, PcyRBP2f, and PcyRBP3, most of which are responsible for mediating parasite invasion into reticulocytes [101]. Functional studies of PcyRBPs are further complicated, as different strains of P. cynomolgi have a different set of RBL genes. For example, PcyRBP2a is present in the P. cynomolgi B and P. cynomolgi Cambodian strains but absent in the P. cynomolgi Berok strain. Similarly, PcyRBP1b is present in the P. cynomolgi Berok and Gombak strain but absent in the P. cynomolgi B, Cambodian and Rossan strain [96]. While it is possible that the loss of PcyRBP1b can be compensated by the presence of PcyRBP2a [101], the relative role of PcyRBP1b and PcyRBP2a in RBC invasion requires further investigations [103].

Plasmodium vivax Ligand Proteins and Host Receptors for Erythrocyte Invasion
The Duffy-binding protein of P. vivax (PvDBP) and P. knowlesi (PkDBP-α) interact with DARC on erythrocytes [104]. A 140 kD region II of PvDBP (amino acids at site 198-522; Table 1) was identified as the key binding site to human erythrocytes [14]. In humans, there are two major codominant alleles for DARC, Fy a and Fy b , which differ by a single nucleotide substitution at amino acid position 42 with glycine and aspartic acid, respectively. Individuals who are Fy a+bhave been shown to be at reduced risk for clinical P. vivax in comparison to Fy a-b+ individuals [105,106]. Individuals with a single point mutation c.1-67T>C (rs2814778) in the GATA-1 box of the DARC gene are considered Duffy-negative (Fy -), as erythrocytic expression of DARC is abolished. Duffy-negative (Fy -Fy -) individuals were previously thought to be resistant to infection by P. vivax and P. knowlesi due to the parasites' inability to infect erythrocytes [107], but several recent studies have shown that P. vivax can infect Duffy-negative individuals, potentially utilizing another invasion ligand(s) [108]. The P. vivax erythrocyte binding protein (PvEBP), which is similar to PcyM DBP2 sequences in P. cynomolgi and contains a Duffy-binding like domain, was discovered in 2013 by de novo genome assembly of field isolates from Cambodia [109]. Binding assay of PvEBP region II (amino acids at site 171-484; Table 1) showed that, unlike PvDBP region II, PvEBP is able to moderately bind to Duffy-negative erythrocytes [17], lending support to the hypothesis of an alternative invasion pathway in P. vivax.
The comparisons of genomic sequences between P. vivax and other Plasmodium species have identified several members of the P. vivax reticulocyte binding protein (RBP) gene family [47,110,111]. The P. vivax RBP gene family comprises several full-length genes, including PvRBP1a, PvRBP1b, PvRBP2a, PvRBP2b, and PvRBP2c; partial genes, including PvRBP1p1, PvRBP2p1, and PvRBP2p2; and pseudogenes, including PvRBP2d, PvRBP2e, and PvRBP3. PvRBP1 comprises PvRBP1a and PvRBP1b and shares homologous regions with P. falciparum PfRh4 [112]. The binding regions of PvRBP1a and PvRBP1b are homologous to that of PfRh4, and the amino acids at site~339-599 were confirmed to interact with human reticulocytes [83,113] (Table 1). PvRBP1a is orthologous to the pseudogene of PkNBP1, but no orthologous region of PvRBP1b was detected in P. knowlesi [96]. While the host receptors of both PvRBP1a and PvRBP1b proteins are unclear, it has been shown that the receptors are neuraminidase resistant [83]. PvRBP1p1 contains a fragment which has 95% similarity to a C-terminal sequence in PvRBP1b [114]. Several members of PvRBP2 (PvRBP2a, PvRBP2b, PvRBP2c, PvRBP2d, PvRBP2e, PvRBP2p1, and PvRBP2p2) are orthologous to PfRH2a and PfRH2b [83,110], as well as PcyRBP2 [96]. Some of them, such as PvRBP2a and PfRh5, also share high structural similarity [115]. PvRBP2b and PvRBP2c are orthologous to PcyRBP2b and PcyRBP2c, respectively [96]. Recently, the receptor for PvRBP2b has been identified as transferrin receptor 1 (TfR1) and the PvRBP2b-TfR1 interaction plays a critical role in reticulocyte invasion in Duffy-positive infections [116]. While PvRBP2p1 has been identified in all human P. vivax infections, PvRBP2p2 was only present in certain lineages of P. vivax [117]. PvRBP2d, PvRBP2e, and PvRBP3 are pseudogenes that share homology with other PvRBPs but encode for nonfunctional proteins [117]. In addition, PvRBP2e is present in the Cambodian field isolates but not in the P. vivax Salvador I [109]. The role of these PvRBP genes in erythrocyte invasion remains unclear.
Tryptophan-rich antigens (TRAgs) are a family of antigens found on human and rodent malaria parasites. In P. vivax, the number of encoded tryptophan-rich proteins is much higher than that of P. falciparum. In the latter case, some of these proteins have been shown to play important role in red-cell invasion and thus are proposed as potential vaccine candidates against P. falciparum malaria [18,19]. Although the reason as to why such a large number of tryptophan-rich proteins are being expressed by P. vivax is unknown, the stage-specific expression of these genes is indicative of their different roles in the parasite's life cycle [20,118]. The PvTRAg family contains 36 members, each with a positionally conserved tryptophan-rich domain in the C-terminus region [119,120]. Of the 36 PvTRAgs, 33 transcribe differently during the ring, trophozoite, and schizont stages of P. vivax, indicating their involvement in blood-stage development [35]. A majority of transcription take place during the schizont-ring transition [121], suggesting their role in erythrocyte invasion. The proportion of non-synonymous SNPs in PvTRAg genes was shown to be significantly high, suggesting the effect of diversifying selection related to antigenic function [122]. Nine PvTRAgs have been shown to bind to human erythrocytes, including PvTRAg33.5, PvTRAg35.2, PvTRAg69.4, PvTRAg34, PvTRAg38, PvTRAg36, PvTRAg74, PvTRAg26.3, and PvTRAg36.6. These proteins possess erythrocyte-binding activity with predicted protein localization to be during the schizont stage [123,124], and each protein recognizes multiple erythrocyte receptors [124]. A comparison of P. vivax transcriptomes between Aotus and Saimiri monkeys indicated the expression of six PvTRAg genes in Saimiri P. vivax was 37-fold higher than in the Aotus monkey strains [39]. Five of these highly expressed PvTRAg genes were previously shown to bind to human erythrocytes [38,123]. Although most TRAg receptors remain poorly characterized and unnamed, the receptor of PvTRAg38 has been identified as Band 3 that binds to amino acid positions 197-214 [125] ( Table 1). The 10 PvTRAg ligands cross-compete with one other and each receptor is shared by more than one TRAg ligand, e.g., PvTRAg38 and PvTRAg74 share a common chymotrypsin-sensitive erythrocyte receptor [123,124]. In addition, the expression of the 10 PvTRAgs varies and is stage-specific, suggesting their differential roles in parasite growth and development. For example, PvTRAg, PvTRAg26.3, PvTRAg36.6, and PvTRAg69.4 are all expressed at the ring and early trophozoite stages of the parasites, indicative of an important role in rosetting; PvTRAg35.2, PvTRAg38, PvTRAg36, and PvTRAg34 are expressed during the schizont and merozoite stages, indicative of invasion properties [124]. Further, PvTRAg36.6 interacts with early transcribed membrane protein (PvETRAMP) to form a protein complex that is apically localized in merozoites, suggesting that this protein is critical for development or maintenance of the parasitophorous vacuole membrane [38]. Recent studies have shown that the PvTRAg35.2 gene sequences were highly conserved in the parasites, and amino acid residues 155-190 and 263-283 were involved in erythrocyte binding [126] (Table 1). PvTRAg35.2 competes with PvTRAg 33.5 and PvTRAg28 and may play a redundant role in erythrocyte invasion [126]. Orthologs of Pv-fam-a are present in P. knowlesi, P. falciparum, and P. yoelii, suggesting that these genes may play a critical role for invasion throughout Plasmodium evolution [111]. Other PvTRAgs, such as PvTRAg56.6 and PvTRAg56.2, interact with PvMSP1 and PvMSP7, to stabilize surface protein complexes on merozoites, and are likely not involved in erythrocyte binding [38]. Additionally, PvTRAgs elicit a strong IgG antibody immune response in P. vivax-infected individuals, with memory lasting up to 5-12 years after being infected [35]. Nine TRAgs (PvTRAg3, PvTRAg7, PvTRAg13, PvTRAg14, PvTRAg15, PvTRAg18, PvTRAg20, PvTRAg26, and PvTRAg35) showed high IgG positivity and conserved IgG reactivity in three Asian countries with low malaria endemicity [35], demonstrating the universal antigenicity of these TRAg proteins.
Merozoite Surface Proteins (MSPs) are a large family of genes found on the surface of merozoites, and a few members are suggested to be involved in non-DBP1 erythrocyte invasion pathways [59]. MSP1 is a 200 kDa highly conserved antigen that undergoes several cleavage events as invasion occurs [78,87]. MSP1 shows a strong binding affinity between 20 and 150 nM at the 42 and 19 kDa fragment cleavage sites, with high-activity binding peptides (HABPs) clustered close to these two fragments at positions 280-719 and 1060-1599, respectively [59], suggesting its critical role in erythrocyte invasion. MSP1 has the potential to be a vaccine target due to its strong immunogenicity [127], but further research is needed [128,129]. MSP3 transcribes during the trophozoite and schizont stages of P. vivax [130] and is highly expressed in Saimiri infections [39]. Although it is yet unclear whether MSP3 binds to human erythrocytes, MSP3.3 and MSP3.5 were expressed on the surface of mature schizonts and interact with MSP1, MSP7, and MSP9 [39]. The MSP3 gene family contained RNA expression of 11 members during the trophozoite and schizont stages, hinting at the MSP3 family playing an important role in both maturation and binding [130]. Interestingly, MSP3.7 was detected at the apical end of merozoites, which differentiates probable roles from other MSP3 family members [130]. Additionally, MSP3.11 transcripts were present, but with no corresponding protein being detected, questioning the exact role of this protein [130]. Although the MSP7 gene family has not been shown to bind to erythrocytes, it forms a complex with PvTRAg36.6 and PvTRAg56.2 and localizes on the surface, likely assisting in stabilization of the protein complex at the merozoite surface [38]. Several MSP7 genes, including MSP7C, MSP7H, and MSP7I, are strong antibody targets and contain high genetic diversity due to frequent positive selection [131]. The MSP9 family also undergoes frequent selection and recombination and forms a co-ligand complex with the 19 kDa fragment of MSP1, but no erythrocytic binding activity was observed [132]. It is possible that MSP9 assists MSP1 in binding to erythrocytes. MSP9 is highly immunogenic at conserved regions 795-808, making it a good vaccine candidate [133,134]. Erythrocytes 263-283 [126] Erythrocytes 280-719 [59] Erythrocytes 1060-1599 [59] Erythrocytes The empty cells represent the same gene above.

Humoral Immune Response against P. vivax and Vaccine Targets
The severity of malaria infection during the erythrocyte stage of Plasmodium depends on various factors, including the following: the location of parasitized red blood cells in the target organs; the local and systemic action of the parasite's bioactive products; pro-inflammatory cytokine production, as well as innate and adaptive immune system at the cellular levels that involves cytokine and chemokine regulators; and the activation, recruiting, and infiltration of inflammatory cells [136]. During P. vivax infection, some individuals can acquire immunity naturally. Such immunity consists of humoral IgG antibodies, cellular cytokines, and proteolytic enzymes production as part of the host response to the pathogen [137,138]. Apart from the invasion capability of P. vivax, host immune response to the pathogen is also a key factor for determining parasitemia and pathology. For example, patients with moderate parasitemia in endemic regions of Colombia were shown with cellular immune responses including high IFN-γ and TNF-α levels and a pro-inflammatory cytokine profile in unstable transmission regions [139]. The balance in interleukin (IL)-10/TNF-α rate could prevent increased parasitemia and host pathology [140]. To date, DBP is the primary target antigen for P. vivax vaccine development. Serological studies have shown that region II of PvDBP, which P. vivax uses to bind to human erythrocytes, induces antibodies against DBP and is naturally immunogenic in people residing in endemic regions, through repeated exposure to the infection [141,142]. However, the naturally acquired neutralizing antibodies against DBP are short-lived, increasing with acute infection, and are strain specific [143,144]. Antibodies from plasma from naturally exposed people and from animals immunized with recombinant Duffy binding protein (rDBP) have blocked the specific interaction between the PvDBP ligand domain in vitro and its receptor on erythrocyte surface [143,144]. Such inhibitory activity was correlated with antibody titers. The DBP1 binding domain is polymorphic, tending to compromise the efficacy of any vaccine associated with strain-specific immunity [145]. While almost all mutations in polymorphic residues did not alter RBC binding [146], such polymorphism has a synergic effect on the antigenic nature of DBP [147]. However, polymorphic SNP variants in the binding domain of PvDBP1 had no effect on the degree of inhibition by anti-DBP monoclonal antibodies. On the contrary, a higher PvDBP gene copy number was shown to reduce susceptibility to anti-PvDBP antibody response [148], but not for better invasion of FyA/FyA and FyA/FyB reticulocytes.
Apart from DBP, the MSP family is responsible of the interaction between merozoites and reticulocytes during the erythrocyte phase. PvMSP1, PvMSP3, and PvMSP9 are potential vaccine candidates, since they are exposed to the immune system and are recognized by antibodies from naturally infected individuals [149]. Cytoadherence assays demonstrated that MSP1 19 is an essential adhesion molecule used in P. vivax invasion to erythrocytes [150] and is immunogenic in people living in areas of unstable malaria transmission in Southeast Asia, Papua New Guinea, and Brazil [151]. Two recombinant polypeptides, rPvMSP 114 and rPvMSP 120 , from the MSP1 C-terminal region show high binding activity to reticulocytes, but no antibodies against these peptides were detected in immunized Aotus monkeys [87]. MSP3 is an abundant ligand on the merozoite surface that is essential for reticulocyte invasion. PvMSP3α block II and the C-terminal region were shown to be highly immunogenic. Individuals living in an endemic region with a high number of previous episodes of malaria were shown to have increased IgG1 and IgG3 anti-PvMSP3α [149,152,153]. Likewise, the PvMSP9 C-terminal and NT domains have also been shown to induce memory T-cell response (higher IFN-γ and IL-4 cytokine production) in individuals living in P. vivax-endemic regions of the Brazilian Amazon and Papua New Guinea [153,154] and are specific targets of P. vivax vaccine. It is unclear if Duffy-negative individuals acquire high levels of antibodies against the 19-kDa C-terminal region of the P. vivax PvMSP1, PvMSP3, and PvMSP9, resulting in a low susceptibility to P. vivax infection [106].
PvRBP2b was recently shown to bind transferrin receptor 1 of the reticulocytes, through the apical domain and the protease-like domain of TfR1 and the N-terminal region of Tf [116]. RBP2P1 protein was found to be expressed in schizonts and localized at the apical end of the merozoite, and it preferentially binds reticulocytes over normocytes. Monoclonal antibodies raised against PvRBP2b prevent reticulocyte binding and reduce P. vivax invasion [116]. P. vivax malaria patients had higher IgG levels against rRBP2P1 than did naive individuals. Human antibodies to this protein also exhibit erythrocyte binding inhibition and are associated with lower parasitemia [155]. PvRBP1a and PvRBP1b are highly transcribed during the parasite schizont stage [118]. PvRBP1a-34 and PvRBP1b-32 proteins bounded specifically to reticulocytes and showed significantly higher reticulocyte binding activity than normocyte-binding activity [83]. Clinical assays have indicated that PvRBP1 435-777 is poorly immunogenic, likely because PvRBP1 had multiple promiscuous T-cell epitopes, which did not induce specific genetic restriction [156]. IgG prevalence against the NT region (including the most polymorphic region) of the PvRBP1a and b was intermediate in a population from Thailand [117], but IgG prevalence against the PvRBP1a-34 and PvRBP1b-32 proteins was significantly higher in P. vivax patients than healthy individuals in the Republic of Korea [83]. Moreover, the highly conserved region III (between amino acids 1941-2229), with the greatest amount of high-affinity reticulocytebinding peptides and high binding affinity, was shown to induce high antibody titers in Aotus monkeys and be able to recognize the full PvRBP1 in parasite lysate [157]. The expression of PvRBPs in the African P. vivax and the antibody response against PvRBP in Duffy-negatives will provide important implications to the usefulness of a future vaccine in vivax malaria control in Africa.
While the circumsporozoite protein (CSP) is one of the most important proteins described in hepatocyte invasion by Plasmodium sporozoites, previous studies involving individuals residing in P. vivax malaria-endemic regions in Brazil showed low responses for antibodies directed against the repeat and C-terminal regions [158]. On the other hand, members of the TRAg gene family that were highly expressed in non-DARC Saimiri-infected P. vivax have also shown to induce antibody response in people from vivax-endemic regions. A recent study of 383 children in Papua New Guinea showed that antibodies against PvFAM-D2 were significantly more common in children with active P. vivax infections [159]. The coexpression of PvFAM-A2 and PvFAM-D2 proteins in infected reticulocytes is spleendependent, based on the Aotus monkey model [160]. These proteins were recognized by a high percentage of sera and are highly immunogenic targets of naturally acquired immune responses [160]. These results are in agreement with members of these families being highly expressed in transcriptional analysis of parasite isolates [39,161]. Moreover, PVX_108770 (VIR14) of the multi-gene VIR family largely located at the subtelomeric regions also presented high sero-positivity, despite the role of its conserved globular domains in eliciting cross-reacting antibodies being unclear [162,163]. Other proteins, such as HYP, which is 100% conserved among P. vivax isolates from Mauritania, North Korea, India, and Brazil, and had antibodies significantly associated with protection against clinical P. vivax episodes in children [159], are a potential target of blood-stage vaccine. Taken together, it is possible that individuals with low-to-no DARC expression have lower susceptibility to infection than individuals having high DARC expression by eliciting high frequency and magnitude of anti-DBP, anti-MSP, anti-RBP, and anti-FAM antibody response against P. vivax during the blood stage. This may imply that one of P. vivax's primary mechanisms for evading host immunity works through indirect, negative regulation of DARC, influencing the humoral response against erythrocyte invasion and parasite development.

Conclusions
The documentation of P. vivax infections in different parts of Africa where Duffynegative individuals are predominant [164][165][166][167][168][169][170][171] suggested that there are alternative pathways for erythrocyte invasion. It is apparent that Duffy-negative individuals are no longer resistant to P. vivax malaria. The increased risk of P. vivax infection and the growing clinical burden across Africa, as well as in Duffy-negative individuals, certainly highlight the public health concern of P. vivax malaria. Future studies should clarify the function and immunogenicity of various candidate parasite ligand proteins and identify their corresponding receptors involved in alternative Duffy-independent erythrocyte invasion, critically examine the host antibody response with respect to P. vivax proteins across wide ethic groups, provide a rigid comparison and analysis of asymptomatic reservoirs and transmission mechanisms of P. vivax in Duffy-negative populations, and unveil the biological features of relapse infections in Africa. Furthermore, future studies should investigate the biology of relapse infections, including the development and characteristics of hypnozoites in P. vivax, which remain largely unclear, despite the overwhelming importance with regard to the impact of relapses on genetic variation of the parasites, transmission, and antimalarial treatment of P. vivax cases. Last but not least, technical studies examining feasible methods for indefinite in vitro culture of P. vivax are vital for further understanding beyond the "omics" level of the parasites.
Funding: This research was funded by NIH/NIAID, grant number R15 AI138002.