Approaches to Malaria Vaccine Development Using the Retrospectroscope

Seventy-five years ago, J. Gordon Thomson addressed the Royal Society of Tropical Medicine and Hygiene on the topic of immunity in malaria (102). In introducing the subject, he made several points. (i) “The literature bearing on this one aspect (immunity in malaria) of the disease in man and other animals is considerable and, therefore, it may be useful to review it and attempt to analyze critically the epidemiological, clinical, experimental and pathological evidences of immunity.” (ii) “The study of malaria by competent workers in many parts of the world has directly and indirectly supplied a mass of facts which materially assist the immunologist to make useful observations and suggest further studies.” (iii) “The knowledge already accumulated leads one to conclude that the problems of immunity in protozoal diseases are probably related closely to those already extensively investigated by bacteriologists.” (iv) “Within recent years more accurate knowledge of immunity in malaria has gradually been acquired; but it is necessary to stress the fact that there are many unsolved problems of great importance. It is not too much to state that if the mechanism of immunity in malaria were entirely solved the control of the disease and its treatment would be nearer accomplishment.” Fifty years later, 25 years ago, the first gene encoding a malaria vaccine immunogen, the circumsporozoite protein (CSP), was molecularly cloned and sequenced (29, 40). This was followed closely by the molecular cloning of the first asexual erythrocytic-stage vaccine immunogen, merozoite surface protein 1 (MSP1) (49, 74). These two antigens still remain the focus of current vaccine efforts. In the intervening years, dozens of individual genes, and ultimately the complete genome, have been sequenced (43). Concurrent with elucidation of the complete primary genomic structure of Plasmodium falciparum, the development of transfection techniques allowed direct functional analysis of specific genes (28, 110, 111). Several genes encoding vaccine targets appear to be essential for stage-specific growth, whereas other genes appear to be nonessential, which has led some investigators to conclude that these latter targets may not be ideal vaccine candidates. Significant advances have also been made in describing the cellular immune responses induced during rodent malaria parasite infections (20, 59). This work has generated an additional “mass of facts” that now need to be reconciled with the immune responses of humans, including those so well described by Thomson and colleagues. Despite this immense progress, few, if any, competent workers who study human malaria would state that the mechanism of protective immunity is entirely solved or the complete control of the disease by active immunization is likely to be a near-term accomplishment.

When contemplating vaccine development for a specific pathogen, two approaches are often considered. One approach, referred to hereafter as semiempirical, is to incorporate, into a well-tolerated regimen, repeated exposures to presumed protective immunogens, the goal of which is to recapitulate naturally acquired clinical resilience, subclinical susceptibility, or ultimately clinical insusceptibility that was empirically described more than a century ago. Examples of semiempirical approaches include the whole killed virus vaccine for hepatitis A and the subunit-based vaccine for hepatitis B virus. An orthogonal approach, referred to hereafter as noetic (derived from the Greek noetikos, understanding or rational), has the goal of inducing protective immune responses that may not occur either qualitatively or quantitatively during natural infection. The subunit-based vaccine for Lyme disease is an example of a noetic approach, which was based on an understanding of the biology of the spirochete in the tick gut and not on evidence derived from empirical studies of naturally acquired immunity in humans.

Herein we revisit, three-quarters of a century later, the topics so eloquently reviewed in the Transactions of the Royal Society of Tropical Medicine and Hygiene. We do so by using a retrospectroscope (25) and triangulating from three perspectives (pure basic research, use-inspired research, and applied product development [96]). We attempt to briefly review and critically analyze some key epidemiological, clinical, experimental, and pathological evidence of immunity in malaria, which remains incompletely solved, and then to systematically explore the two approaches being considered in efforts to develop malaria vaccines.

NATURALLY ACQUIRED PROTECTIVE IMMUNITY TO MALARIA: KEY EPIDEMIOLOGICAL, CLINICAL, EXPERIMENTAL, AND PATHOLOGICAL EVIDENCE—SEMIEMPIRICAL VACCINE APPROACHES

Perhaps as good a place as any to start analyzing evidence of protective immunity to malaria is where Thomson started in 1933, that is, by acknowledging that “it is now regarded almost as axiomatic that (malaria is most acute and severe) during the first few years of residence in hyperendemic zones, while subsequent attacks become progressively less severe and more infrequent.” Recently, Marsh and Kinyanjui (63) nicely captured graphically the epidemiology-based notion of “naturally acquired immunity,” which is the often observed acquisition of increasing insusceptibility first to severe malaria and then mild malaria, and finally, after years of presumably repeated exposures, insusceptibility to parasitemia (Fig. 1).

FIG. 1.

The age pattern of severe, mild, and asymptomatic malaria. Representative data from a number of studies in Kilifi, Kenya, demonstrate the percentages of individuals in each age group with severe, mild, and asymptomatic parasitemia. The number of homologous infections is indicated in red below the x axis. The data can be divided into four categories of immunity, full clinical susceptibility (preinfection), clinical resilience (post one infection), subclinical susceptibility (post two or three infections), and clinical insusceptibility (more than three or four infections). Modified from reference 63 with permission of Blackwell Publishing Company.

Several other lines of epidemiological, clinical, and experimental evidence are often cited to support the existence of naturally acquired immunity, including: (i) experimental passive-transfer studies that demonstrated hyperimmune gamma globulin from adult or cord blood has an anti-P. falciparum effect in infected children (12, 21, 37, 65); (ii) experimental data from clinical studies in the 1940's to the 1960's of iatrogenically induced malaria, as part of immunotherapy, that demonstrated that a single prior exposure provides some clinical resilience (23) and even older data from similar clinical studies in the 1920's and 1930's that demonstrated that repeated inoculation of parasites leads to clinical insusceptibility (17); and (iii) epidemiologic studies of previously malaria-naïve individuals who migrated to regions where malaria is hyperendemic that indicated acquisition of clinical immunity within four infections (4).

HUMAN PASSIVE-TRANSFER EXPERIMENTS

The most compelling data for the possibility of an antibody-mediated vaccine that mimics natural immunity are the three independent passive-transfer experiments that clearly demonstrated the ability of the immunoglobulin G (IgG) fraction isolated from protected adults to effectively reduce asexual parasitemia in children (Fig. 2) (21, 65, 88). The decline in parasitemia was followed by a decrease in clinical symptoms such as fever, suggesting that resolution was in response to the lower parasitemia and not the IgG fraction directly. The IgM fraction was not tested; therefore, the role of nonpeptide, T-cell-independent antigens that induce primarily IgM responses remains unknown (79). One study also demonstrated that cord blood obtained during delivery could effectively decrease parasitemia (37). Interestingly, there was no effect on gametocytemia, suggesting differential acquisition of effective immunity against asexual and sexual erythrocytic stages, as discussed below. Additionally, postinfusion no mature asexual parasites were seen in the peripheral circulation, suggesting that the IgG fraction did not release sequestered parasites. However, the ability of the IgG fraction to block adhesion or to prevent infection or symptoms was not directly evaluated in vivo since all of the patients were parasitemic and symptomatic prior to administration. In vitro analyses of functional antibody activity are consistent with the IgG fraction decreasing parasitemia by enhancing monocyte clearance of parasites and, at least in one laboratory, were not associated with blocking of merozoite invasion or growth (12). These studies suggest that IgG molecules bind to exposed antigens on the surface of the infected red blood cell (RBC) or extraerythrocytic parasite stages, such as the merozoite, targeting them for phagocytosis. A number of merozoite surface antigens have already been identified and considered as vaccine candidates, but with limited success to date (www.who.int/vaccine_research/documents/en/, malaria vaccine, candidate malaria vaccines in clinical development, October 2008; see Table S1 in the supplemental material).

FIG. 2.

Peripheral parasitemia following passive transfer of IgG from malaria-resilient adults. Daily Giemsa-stained blood smears were used to follow the peripheral parasitemia in a child for a week after the first of three intramuscular injections of IgG purified from adults living in East Africa. The number of asexual parasites (trophozoites) and gametocytes in a cubic millimeter of blood is plotted on the y axis, and the arrows under the x axis indicate times of IgG administration. A total of 1.2 to 2.5 g of IgG was injected over the course of the experiment. Adapted from reference 21 by permission from Macmillan Publishers Ltd.: Nature (Cohen et al., 192:733-737), copyright 1961.

The role of nonhumoral immunity against asexual-stage parasites was suggested by the demonstration that three of four malaria-naïve volunteers were protected after three cycles of inoculation with 30 infected erythrocytes, followed by a drug cure, even though no parasite-specific antibodies could be detected (81). Parasite-induced lymphoprofileration, gamma interferon production, and nitric oxide synthase activity were observed. However, the subsequent realization that the drugs used had a longer half-life than expected may have compromised the challenge experiment (38).

IMMUNOLOGY

These lines of epidemiological, clinical, and experimental evidence are consistent with the supposition that repeated infections lead to naturally acquired immunity through conventional B- and T-cell immune responses to one or more protective immunogens. However, as noted by James and Watson during the discussion of Thomson's presentation in 1933, immunity to malaria is not akin to “that which we see in diseases like yellow fever, diphtheria, scarlet fever, or other bacterial diseases, where one attack confers freedom from subsequent attacks.” This difference, noted 75 years ago, suggests that alternative mechanisms may also contribute to clinical resilience; one such mechanism, which is as consistent with the data as acquisition of conventional immunity, is acquisition of regulatory immune responses that function to downregulate the strong proinflammatory responses associated with each of the three distinct, though often overlapping, clinical syndromes that constitute severe malaria (i.e., impaired consciousness, respiratory distress, and severe anemia [62]). From studies of a wide variety of pathogens, ample precedent exists for the induction of both stimulatory and inhibitory immune responses (26, 86, 95, 107). The resulting tachyphylaxis of a strong proinflammatory response may represent a negotiated compromise reached between the host and the pathogen, the consequences of which are enhanced survival of the host, persistence of infection, and transmission of the pathogen.

It is clear that the measurable humoral response induced by a single Plasmodium infection is inadequate to provide subclinical susceptibility by either a homologous or a heterologous second infection (23, 24); rather, as Koch reported more than a century ago, protective immunity to parasitosis requires repeated, life-long exposure to malaria infections. The continued susceptibility of clinically resilient adults to reinfection for >20 years indicates that naturally acquired clinical immunity is not sterilizing (78). To this latter point, anecdotal evidence is sometimes invoked to suggest that protective immunity is short-lived and requires ongoing parasite exposure and presumably occasional subclinical parasitosis. Although generally believed to wane within months to a few years of nonexposure, naturally acquired protection against severe malaria may persist for many years despite lack of parasite exposure (11, 22, 64, 99).

PATHOGENESIS

Evidence from studies of the pathogenesis of severe malaria is also cited to support the existence of naturally acquired immunity and/or the development of potential target immunogens for inclusion in a malaria vaccine. The key underlying premise is that the asexual erythrocytic cycle of Plasmodium is responsible for all of the morbidity associated with malaria. The molecular mechanisms involved in RBC invasion, sequestration of parasitized RBCs, clearance of uninfected RBCs from the periphery, and suppression of hematopoiesis are all seen as fertile areas for immunogen discovery.

With respect to RBC invasion, molecules on the merozoite surface (MSPs) adhere to RBCs seconds after release from the schizont. RBC attachment initiates the release of AMA1 from the apical organelle, and this facilitates the formation of a tight junction between the apical end of the parasite and the RBC and then active invasion. In P. vivax, PvDBP is required for junction formation and invasion (16, 67), while the more virulent species P. falciparum can use a number of alternative genes, including those that encode EBA-175, BAEBL/EBA-140, JESEBL/EBA-181, PfRh1, PfRh2b, and PfRh4 (27). All of these genes have been proposed as vaccine candidates, and those for PfAMA1, PfGLURP, PfMSP1, PfMSP2, PfMSP3, and PfRESA have advanced to phase 1 vaccine trials (www.who.int/vaccine_research/documents/en/, malaria vaccine, candidate malaria vaccines in clinical development, October 2008).

With respect to RBC sequestration, the ability of P. falciparum-infected RBCs to induce aggregation of uninfected RBCs and bind host cells lining blood vessels or the placenta is thought to contribute to respiratory distress, cerebral malaria, and pregnancy malaria. P. falciparum-specific PfEMP1 molecules appear to be responsible for this agglutination and adhesion and are encoded by polymorphic var genes that undergo antigenic variation (57). A single var gene, Pfvar2csa, has been associated with syncytiotrophoblast adherence and the development of immunity to pregnancy malaria, making it a prime vaccine candidate (90). However, there is still significant polymorphism between strains, complicating vaccine development (10, 55). Specific var genes have not yet been directly linked to either cerebral malaria or respiratory distress syndrome (69).

With respect to the clearance of uninfected RBCs from the periphery and suppression of hematopoiesis, P. falciparum-infected RBC lysis during schizont rupture has been proposed to contribute to both (48), although the precise mechanisms remain poorly understood. Severe anemia may result from subsequent cytokine release and/or other factors postlysis, whereas uninfected erythrocyte removal from the periphery may result from high plasma concentrations of parasite ligands that are shed during invasion or released upon lysis of schizont-infected RBCs and that subsequently adhere to uninfected erythrocytes, possibly resulting in immunoglobulin- or complement-mediated removal. The parasite-derived material released during rupture is also thought to stimulate tumor necrosis factor alpha secretion, which may contribute to the paroxysmal fever associated with malaria. In sum, as stated by Thomson in 1933, “…it is necessary to stress the fact that there are many unsolved problems of great importance.”

INTRAERYTHROCYTIC SEXUAL STAGES

It is probably prudent at this point to briefly mention acquired immunity to gametocytes, the sexual form of the parasite responsible for transmission from humans to mosquitoes. Similar to asexual parasitemia, gametocytemia also declines with age. Although the decrease is thought to be primarily due to decreased asexual parasitemia, gametocyte-specific antibodies are generated during natural infection (75). However, the response is distinct from immunity to asexual stages, as passively transferred IgG does not decrease gametocytemia (Fig. 2) (21, 88) and only a subpopulation of serum samples from adults living in areas where malaria is endemic can block the ability of parasites to be transmitted to mosquitoes (34, 45, 82). Despite the rather slow acquisition of naturally occurring transmission-blocking activity, inclusion of immunogens that induce similar activity might contribute substantially to the herd immunity effect of a malaria vaccine.

NOVEL PROTECTIVE IMMUNITY: EXPERIMENTAL EVIDENCE—NOETIC VACCINE APPROACHES

An alternative to replicating natural immunity is to develop vaccine strategies that protect better than exposure to the disease. For example, even though adults living in areas where malaria is endemic are protected against clinical disease and high parasitemia, they can still be infected by sporozoites; therefore, acquired resistance to sporozoite and/or liver stage infection represents a novel mechanism of protective immunity. Such a noetic approach was demonstrated more than a quarter century ago when Clyde et al. inoculated human subjects with irradiated P. falciparum sporozoites and induced sterilizing immunity in one of the three individuals tested (18). In subsequent studies, 10 of 11 volunteers immunized with >1,001 irradiated sporozoites were protected and in 4 of 5 volunteers given a second immunization, protection lasted for 23 to 42 weeks (51). Also more than a quarter century ago, several groups showed that antibodies generated by vaccinating a vertebrate host with mosquito midgut stage parasites blocked the production of infectious sporozoites after parasites are taken up in a blood meal by a mosquito (14, 47, 66). Since these late sexual stage antigens are presumably not expressed in the human host during a natural infection, an immune response is typically not generated. Both of these approaches generate a protective immune response that is not induced during a natural infection. Consequently, there has been little selective pressure to drive the development of immune evasion responses, such as strain-specific polymorphism. This should allow the production of a vaccine that is effective against multiple isolates. On the other hand, immunity may not be boosted by an anamnestic memory response during a subsequent natural infection. The relative effect these two factors have in the development of an effective vaccine has yet to be assessed: whether it will be easier to overcome the potential lack of a memory response with repeated booster immunizations than it will be to simultaneously induce long-lived protective immunity against a broad range of polymorphic antigens has to be determined empirically.

LIVER STAGE IMMUNITY

The intriguing induction of protective immunity in response to irradiated sporozoites, but not live or killed sporozoites, is still not well understood. Simply vaccinating with a preparation of killed parasites does not induce protection without the addition of strong adjuvants that stimulate significant local inflammation at the injection site (70, 92). However, inoculation with irradiated sporozoites was found to protect against a subsequent live-sporozoite challenge for a wide range of Plasmodium species (73). In humans, the induction of protective immunity against P. falciparum infection required 1,000 bites from an irradiated-sporozoite-infected mosquito and only lasted 42 weeks (19, 51). Of note, irradiated asexual P. falciparum parasites did not induce a similar protective response (89).

It appears that irradiated sporozoites are able to invade liver cells and initiate development, but DNA replication is blocked and the parasite fails to complete development. The disruption of any one of several genes in the rodent malaria parasite Plasmodium berghei (i.e, PbUIS3, UIS4, P52, or P36) also inhibits complete liver stage development after invasion and induces protective immunity against a subsequent challenge (58, 71). That aberrant parasites remain in the liver and chronically stimulate a protective immune response is one hypothesis (68). Another is that aborted development prevents the parasite from evading the immune response, possibly inhibiting the liver cell from undergoing apoptosis (93). The development of genetically attenuated P. falciparum sporozoites is being explored as an alternative to attenuation by irradiation (105). Another approach has been to use the sporozoite surface protein CSP as an immunogen (98).

The development of sporozoite and liver stage vaccines has also been markedly facilitated by the ability to conduct human phase Ib studies in which vaccinated subjects are challenged with sporozoites via infected-mosquito bites. As clinical insusceptibility (i.e., sterilizing immunity to asexual erythrocytic-stage infection) is the primary goal of liver stage vaccines, any appearance of erythrocytic-stage parasites indicates that the vaccine was not completely protective. The trial can then be ended immediately with drug treatment before the parasitemia progresses. Using this experimental medicine model, many synthetic and recombinant vaccine strategies have been tested in humans, the most promising being RTS,S, which includes CSP amino acids 207 to 395 and is described in more detailed in Current Challenges, Uncertainties, and Opportunities (ii) Noetic Vaccine Approaches. The partial protective effect of RTS,S, as well as insecticide-treated bed nets, does demonstrate that decreasing the number of infective sporozoites can reduce malaria morbidity and mortality, which is counter to previous dogma that sporozoite immunity had to be 100% effective to have any protective role.

TRANSMISSION-BLOCKING IMMUNITY

At the other end of the life cycle, murine monoclonal antibodies (MAbs) and/or polyclonal antibodies specific for several antigens exposed on the parasite surface in the mosquito midgut, Pfs48/45, Pfs230, Pfs25, and Pfs28 and P. vivax orthologs Pvs25 and Pvs28, have been shown to block transmission to mosquitoes and are being developed as vaccine candidates (85, 106). Recombinant Pfs25, Pfs28, Pvs25, and Pvs28 all induce antisera in mice that can significantly decrease oocyst production in a mosquito membrane feed (35, 53, 76, 108). Moreover, antibodies against recombinant Pvs25 and Pfs25 obtained from a phase 1 clinical trial with humans were also shown to decrease P. falciparum oocyst production in mosquitoes by 20 to 30% (61, 109).

FRAMING THE VACCINE APPROACHES

For either semiempirical or noetic vaccine approaches, the multistage life cycle of the malarial parasite provides a range of target immunogens or vaccine candidates (Table 1). The most proximal vaccine opportunity after the bite of an infected mosquito is to interfere with the deposition of the sporozoite and/or its migration and/or invasion of the liver parenchyma. The next potential vaccine target would be to interrupt the clinically latent liver stage as the parasite replicates in a hepatocyte. Subsequent targets include merozoites as they are released from liver cells and/or infect RBCs, the latter of which initiate and perpetuate the asexual erythrocyte cycle. Inhibiting the subsequent increase in asexual parasites or the progressive increase in clinical symptoms, such as fever, chills, vertigo, nausea, and metabolic disorders, consistent with a direct role for these stages in disease pathogenesis (72), is yet another target. The parasitized RBC surface receptors that appear to be involved in at least some of the specific molecular mechanisms associated with the clinical manifestations of severe malaria provide additional candidates. The most distal vaccine opportunity is to block the production and/or circulation of the gametocytes responsible for transmission of the parasite from humans to female mosquitoes and/or the sexual reproduction and further sporogonic development of the parasite in the infected mosquito.

TABLE 1.

Comparison of semiempirical and noetic vaccine approaches

Vaccine approach and target	Primary goal	Challenges and uncertainties	Advantages and opportunities
Semiempirical
Severe malaria and asexual erythrocytic stages	Prevent debilitating symptoms	Unknown mechanism of protective immunity, lack of clinically qualified functional immunoassay and/or challenge model	Passive immunization studies with MAbs; recapitulates natural immunity, so likely boosted/sustained by subsequent natural subclinical infections
Asexual erythrocytic stages	Prevent parasite replication and binding	Antigenic and/or allelic polymorphism, poor immunogenicity as monomeric subunit vaccine	Passive immunization studies with MAbs; recapitulates natural immunity, so likely boosted/sustained by subsequent natural subclinical infections
Pregnancy malaria	Prevent maternal and pre- and antenatal deaths	Target immunogen to be defined, unknown immunogenicity of subunit approach	Functional immunoassay; passive immunization studies with MAbs; recapitulates natural immunity, so likely boosted/sustained by subsequent natural subclinical infections
Sexual (early) stages	Prevent mosquito transmission	Poor immunogenicity as monomeric subunit vaccine, antigenic and/or allelic polymorphism	Membrane feeding assay, direct mosquito feeding
Noetic
Preerythrocytic (sporozoite and liver stages)	Sterile immunity, reduced liver stage burden, or reduced initial merozoite burden	Subunit, poor immunogenicity and/or contribution of CD8+ T-cell immunity, unknown durability and/or boosting following natural infection; attenuated, economically feasible manufacturability and delivery	Sporozoite challenge model
Mosquito stages (late sexual and sporogonic)	Prevent mosquito transmission	Subunit, poor immunogenicity as monomeric subunit vaccine, unknown durability and/or boosting following natural infection	Membrane feeding assay, direct mosquito feeding

An additional exercise in framing the vaccine approach(es) is to carefully consider the potential target product profile(s) (TPP; see, for example, http://www.who.int/immunization/sage/target_product_profile.pdf), specifically, the indication for use, the target efficacy, and the duration of protection (i.e., need for and timing of booster immunizations). With respect to these TPP attributes, two goals have been proposed: “by 2015, to develop and license a first-generation P. falciparum malaria vaccine with a protective efficacy against severe disease and death of more than 50% and which lasts longer than one year; and by 2025, to develop and license a malaria vaccine with a protective efficacy against clinical disease of more than 80% and which lasts longer than four years” (http://www.rollbackmalaria.org/gmap/2-4a.html). A third goal that has recently been proposed is eradication of malaria (http://www.gatesfoundation.org/speeches-commentary/Pages/bill-gates-malaria-forum.aspx).

Based on these goals and the observations captured in Fig. 1, at least four potential clinical endpoints could be contemplated. (i) Vaccines designed to induce immunity that mimics naturally acquired clinical resilience might have as a primary endpoint reduction of the incidence of the three clinical syndromes that constitute severe malaria (7) with a P. falciparum density of ≥2,500 parasites/μl. (ii) Vaccines designed to induce immunity that mimics naturally acquired subclinical susceptibility might have as a primary endpoint a reduction in the incidence of all clinically apparent episodes of malaria, defined as an axillary temperature of 37.5°C or higher with a P. falciparum density of >2,500 parasites/μl. (iii) Vaccines designed to induce immunity that mimics naturally acquired clinical insusceptibility might have as a primary endpoint a reduction in the incidence of any episodes of malaria parasitosis. (iv) Vaccines designed to induce immunity that, in the context of other sustainable control measures, provides sterilizing immunity and/or eliminates parasite transmission might have as a primary endpoint reduction in the incidence of any episodes of asexual parasitemia, as well as gamteocytemia and/or evidence of transmission-blocking immunity.

Depending upon the primary efficacy endpoint selected (as well as other factors, such as the local malaria epidemiology), the sample size and duration of the pivotal clinical trial(s) will vary substantially. Even though the exercise of starting a vaccine research-and-development program by thoughtfully and carefully planning backward by drafting, in the following order, (i) the desired TPP, (ii) the primary efficacy endpoint(s) for the pivotal registration trial(s), (iii) the primary endpoints for the proof-of-concept study, (iv) the safety and immunogenicity endpoints for phase I studies, and (v) the preclinical candidate attributes, as well as assigning clear go/no-go criteria at each milestone, may be considered best practice, it seems an exercise rarely completed.

CURRENT CHALLENGES, UNCERTAINTIES, AND OPPORTUNITIES (i) SEMIEMPIRICAL VACCINE APPROACHES

For the semiempirical vaccine approaches, malaria vaccine development has been complicated by a plethora of challenges and uncertainties (Table 1). Although many of the current asexual vaccine candidates were first identified in other Plasmodium species and then the orthologs cloned from P. falciparum or P. vivax (52), the unique biological properties of each Plasmodium species, especially with regard to host-parasite interaction, limit the usefulness of animal models in solving the challenges or reducing the uncertainties in developing vaccines for P. falciparum. As such, it remains imperative to move quickly to the clinic, which provides more reliable insights than animal testing.

A particularly vexing problem in vaccine development for the most lethal human malaria parasite, P. falciparum, is sequestration, presumptively mediated by PfEMP. Parasite-produced PfEMP1 is transported to the surface of the infected RBC and mediates tight binding to host cells, including endothelial and dendritic cells (6, 94, 100). Endothelial cell adhesion allows the parasite to be sequestered in the microvasculature and at high parasitemia is thought to cause circulatory obstruction and tissue damage. Parasites are also sequestered in the bone marrow and spleen, sites of B-cell maturation, selection, and activation. In addition to adhesion, P. falciparum-infected RBCs have been reported to have direct effects on endothelial cells and dendritic cells, which could contribute to the development of pathology and adversely affect induction of immunity (39, 44a, 104, 112). No other species, except for the closely related chimpanzee malaria parasite Plasmodium reichenowi, expresses a similar protein on the RBC surface (103). The lack of this key surface protein, coupled with species-specific differences in host cell invasion, complicates the extrapolation of pathogenesis and immunology data obtained with other Plasmodium species to P. falciparum and emphasizes the importance of studying human malaria directly in the natural host.

Beyond the absence of an animal model, the number of antigenically distinct PfEMP1 molecules, presumably expressed by the current population of circulating P. falciparum, coupled with the large size of PfEMP1, substantially complicates the development of an antiadhesion vaccine (57). One approach being considered comprises identifying and recreating a conserved functional epitope. Recently, the crystal structure of the CD36 binding domain of PfEMP1 was solved (54). Since CD36 binding is a common characteristic of many otherwise polymorphic PfEMP1 proteins, the crystal structure could facilitate identifying a conserved functional epitope. It remains to be determined whether a conserved functional epitope is involved in a protective immune response and ultimately whether a method to induce production of the appropriate Ig molecules can be established. The development of a functional adhesion assay(s) as a tool for developing pregnancy malaria vaccines appears to be paying off (42); investing in an analogous functional adhesion assay(s) as a tool for developing asexual erythrocytic-stage vaccines that protects by blocking sequestration or adhesion seems sensible.

Other challenges, particularly for vaccines that target other parasite surface proteins (e.g., MSP, AMA, etc.) described, include the lack of a clinically qualified functional immunoassay(s) and/or human challenge models, antigenic and/or allelic polymorphism, and poor immunogenicity as monomeric subunit vaccines. MAbs that inhibited parasite growth in vitro (9, 30, 41, 80) were used first to identify target antigens in the parasite and later to screen recombinant antigens for the presence of the corresponding “inhibitory” epitope (2, 13, 15, 36). Subsequent human vaccine trials have been disappointing, hampered by weak immunogenicity and strain-specific polymorphism (31, 44, 97, 101). New approaches using novel adjuvants or protein conjugation are being developed to increase the immune response, but it is still a work in progress (87).

In the absence of a clearer understanding of the mechanism(s) of protective immunity targeting parasite surface proteins, it is difficult to establish that the recombinant proteins being tested actually contain the epitopes required for protective natural immunity and/or whether in vitro growth inhibition is the best tool to guide further vaccine development. MAbs used to characterize these immunogens have not been evaluated for a protective effect in humans, and the reactivity of vaccine immunogens with these immunological reagents should be considered circumstantial evidence, at best, of appropriate antigenicity. MAb technology has advanced considerably since its introduction more than 2 decades ago and has yielded several approved products for the treatment of arthritis and cancer and preventive therapy for specific acute viral infection in the very young (60). It is technically, biologically, and ethically feasible to develop and evaluate corresponding recombinant MAbs, alone or in combinations, for investigational use in humans (80a). Vaccine candidates could be prioritized and fast tracked for vaccine development if recognized by recombinant antibodies that demonstrate the biological activity and clinical effect desired by passive immunization.

New approaches to identify novel vaccine candidates are needed in addition to validating the antigens currently in the pipeline. Possibilities include large-scale protein-protein microarray comparison of antibodies from genetically related protected and unprotected individuals or subtractive libraries of recombined Ig variable domains (5, 33, 46). The proteins screened by microarray could be derived from the Plasmodium genome or randomly generated mimotopes. Those proteins differentially recognized by serum from protected adults could be tested directly for immunogenicity in animals and the antibodies generated assayed for parasite reactivity and for inhibitory activity in vitro. Similarly, antibodies found preferentially in serum from protected individuals could be produced recombinantly and screened for parasite reactivity and inhibitory activity. The target antigens of inhibitory antibodies could be identified by mass spectroscopy of immunoprecipitated material, produced recombinantly, and tested as immunogens. Orthologs in rodent or primate malaria could also be tested for protective efficacy in vivo. Ultimately, antigens that induce inhibitory antibodies in either screen would be evaluated for pilot trials with humans.

Basic research characterizing the progression of the human immune response through the development of protective immunity could provide additional insight into the host-parasite interaction. As mentioned above, direct species-specific interaction between the parasite and host cells makes it imperative to evaluate this directly in humans. Important questions include the effect of parasites sequestered in bone marrow on B-cell development, selection, and tolerance and that of parasites sequestered in the spleen on the activation of both B and T cells (91).

(ii) NOETIC VACCINE APPROACHES

For the noetic vaccine approaches, facile human models, be they sporozoite challenge or direct mosquito feeding or functional immunoassays such as the standard membrane feeding assay, have been used to leverage vaccine development efforts. The most effective effort to date has been the RTS,S vaccine, which was recently shown to reduce malaria incidence by 62.2% and decrease severe malaria in children under 1 year of age by 20 to 60%, suggesting that an immune response induced by amino acids 207 to 395 of sporozoite surface protein CSP can play a role in providing some clinical resilience (3, 8) Although these preliminary findings are encouraging, matching the strong clinical resilience and decreased susceptibility to parasitemia induced by naturally acquired immunity has proven challenging. Attempts to improve the efficacy of RTS,S above 60% have not been successful, even though a more potent adjuvant increased anti-CSP titers 10-fold (1). The lack of correlation between anti-CSP antibody titer and protection is consistent with previous results for other liver stage vaccines (50). Further evaluation of the long-term effect of RTS,S and its mechanism of action may lead to methods to increase its efficacy. Meanwhile, RTS,S will begin phase III clinical trials in 2009. If approved, this vaccine could be used as part of an integrated malaria control program, but it is unlikely to be effective enough to be used as a control measure alone. Additional characterization of liver stage and erythrocytic-stage parasite development and immunity may identify new target antigens or vaccination strategies to enhance immunity (77).

As discussed above, an alternative to a subunit approach is to use whole irradiated or genetically attenuated sporozoites, the production of which needs to be scaled up for widespread use. Infectious sporozoites can currently be produced only in mosquitoes and therefore are unlikely to be a practical vaccine until an efficient in vitro system is developed (32) and/or immunogenicity improved such that only minute quantities of live attenuated sporozoites are required. Other vaccine quality attributes such as formulation, stability, and delivery of a live attenuated sporozoite vaccine also need to be addressed in a manner that provides an economical process for manufacturing, shipping, distributing, and administering the vaccine. Other than altering the dose, route of administration, and/or schedule, the immunogenicity, be it antibody or CD4⁺ or CD8⁺ T-cell immunity, of a specifically attenuated pathogen is largely immutable. Based on the irradiated-sporozoite data, the durability beyond 42 weeks of protective immunity produced by live attenuated sporozoite vaccination is uncertain, as is the propensity to boost immunity after a natural sporozoite infection.

With respect to transmission-blocking vaccines directed against the late sexual stages, and as found with other human malaria parasite antigens, monomeric recombinant protein vaccines are not very immunogenic alone and adjuvants are required. Conjugate vaccines and other technologies may be required to induce a sustained transmission-blocking immune response in humans (56, 84). The search for effective and safe adjuvants is ongoing (83, 109).

FUTURE STRATEGIES: (i) SEMIEMPIRICAL VACCINE APPROACHES

The ultimate goal of a semiempirical vaccine is to reproduce natural immunity generated by multiple natural exposures. If protection is mediated by classic B and T cells, akin to that seen in many acute bacterial and viral infections, then presumably immunization with a manageable number of protective target epitopes in a multivalent vaccine that recapitulates natural protective immunity may be feasible. To do so, new strategies may be needed to identify, validate, and recreate these protective epitopes, including those targeted by the protective IgG fraction isolated from adults living in regions where malaria is endemic. Methods to efficiently evaluate the potential efficacy of vaccine candidates are also required. Serious consideration should be given to pursuing the use of passive immunization with MAbs as an experimental medicine tool in identifying and validating vaccine candidate immunogens. Alternative means of producing polyvalent vaccines may also be considered, the extreme case being whole or split asexual erythrocytic-stage parasites.

(ii) NOETIC VACCINE APPROACHES

For the two major noetic vaccine approaches, fully harnessing the power of recombinant DNA technologies, be it by creating genetically attenuated sporozoites or multimeric-subunit sexual stage transmission-blocking vaccines, makes sense; however, whether an economical process for manufacturing, shipping, distributing, and administering a live whole parasite can be achieved remains uncertain. In both cases, existing efficacy assays should be fully leveraged early in development to prove the principle: in the case of preerythrocytic-vaccine approaches, the critical assay is the human sporozoite challenge model, while in the case of transmission-blocking vaccines, both the standard membrane feeding assay and the direct mosquito feed should be exploited.

SUMMARY

The start of the first phase III trial of a malaria vaccine in 2009 marks a significant advance in malaria control. However, the goal of a vaccine with at least 80% efficacy lasting 4 or more years continues to be elusive. To reiterate Thomson's statement in 1933, “… it is necessary to stress the fact that there are many unsolved problems of great importance. It is not too much to state that if the mechanism of immunity in malaria were entirely solved the control of the disease and its treatment would be nearer accomplishment.” Creative approaches are needed to better define the mechanisms of immunity that contribute to and inhibit protection against human malaria. Given the limitations imposed by the species specificities of host-parasite interactions, direct evaluation of the protective human immune response is paramount. Equally important to the development of active immunization strategies that effectively generate a protective response in humans by reproducing the protective response acquired through repeated natural infections or through novel mechanisms that do not occur in nature is the development of new experimental medicine approaches that safely permit elucidation of protective immune responses in humans, including passive immunization. In the end, it is highly likely that a multiantigen and multistage approach will be required to achieve the desired goal for malaria vaccines and should be planned accordingly.