Assessing Malaria Vaccine Efficacy

Abstract

After many years of silence, eradication of malaria is, once again, one of the top priorities on the agenda of many international health and development agencies. To meet this idealistic goal, a combination of control tools is needed. From this armentarium, a malaria vaccine is central to prevent infection and/or disease. However, numerous malaria vaccine candidates have shown limited efficacy in Phase II and III studies. One reason for these failures has been that the assessment of efficacy in the context of malaria has been difficult to standardize. In this article, we have reviewed and discussed the different ways to assess the outcome of a malaria vaccination.

1. Introduction

Malaria remains one of the major infectious diseases with a huge burden, affecting a large fraction of the world population. Although most of the deaths, caused by Plasmodium falciparum, P. vivax and to a lesser extent P. malariae and P. ovale, occurred in Africa; significant morbidity is evident in South America and Asia [1]. Different control measures such as insecticide-treated bed nets, powerful drugs (i.e., artemisinin-based combination therapies) and early diagnostics have had a positive impact in reducing malaria mortality worldwide [2]. However, these methods have led to complete eradication of malaria in only a few countries in intertropical zones [3]. This is mainly due to increasing drug resistance of the parasites and the failure of vector control strategies resulting from the change in mosquito behavior and the emergence of insecticide resistance [4, 5]. An antimalarial vaccine is thus a necessity to achieve the goal of complete global malaria eradication [6, 7].

Vaccine development in malaria have employed a composite of rational and empirical approaches and depended on multiple epidemiological and experimental studies. Individuals living in endemic regions acquire immunity over time after repeated exposure to the parasites. Such immunity, also called premonition, is partial, species-specific and biphasic [8, 9]. In the first phase, the hosts still get infected but do not develop clinical symptoms. On the contrary, the second phase, which is the prevention or limitation of parasite multiplication, takes long to develop. This second phase is heavily dependent on parasite exposure - more the exposure the host gets, lesser the time this immunity takes to develop [9, 10]. Hence, the goal of vaccine strategies is to reduce the time needed to acquire protective immunity and to make the immunity long-lasting.

The use of experimental models is critical to vaccine development. Many researchers advocate the use of human parasites in human hosts as it is the optimal experimental model for malaria [11]. However, field studies are inherently limited by the inability to control multiple experimental parameters such as the number of infective mosquito bites, the number of parasite per infective dose, and the genetic background of the host and parasite. In addition, there are numerous ethical considerations, which restrict access to peripheral blood samples for antibody and T cell studies, important for investigating long-term protection. Thus, many researchers have turned to more controllable models, such as monkey or human Plasmodium in monkeys [12, 13, 14] or rodent Plasmodium in mice [15]. Using these models, there have been numerous vaccination studies using genetically-attenuated parasites [16], irradiated parasites [17, 18], chemically-attenuated parasites [19, 20], live parasites under drug prophylaxis [21, 22, 23], and defined antigenic formulations [24, 25, 26, 27, 28, 29, 30]. These studies have demonstrated that vaccination can reduce parasite development, prevent pathology in infected animals, prevent transmission to mosquitoes, and even induce sterile immunity. Another major advantage of these models is that the outcomes following vaccination is well-defined and easily measurable, such as development of sterile immunity, inhibition of parasite development in the liver or in the blood, and/or prevention of certain pre-defined clinical signs or of pathologies. So far, only whole parasite formulations using irradiated sporozoites [31, 32] or live parasite immunization under chloroquine [33, 34, 35], and a limited number of sporozoite antigen formulations, such as RTS,S [36], have been shown to induce sterile immunity in significant proportion of the human volunteers.

2. Vaccine developmental phases

Vaccine clinical testing in humans involves multiple phases. Phase IA involves a small group of naïve volunteers (<100), from non-endemic regions, with no previous experience of malaria, while Phase I involves malaria-exposed individuals from endemic regions. In both phases, vaccine safety and immunogenicity are assessed. Only after the vaccine has shown a good safety profile with encouraging immunogenicity data, phase IIa test study can be initiated with a larger set of volunteers (>100–1000) from non-endemic regions. In Phase IIa, vaccine efficacy is assessed by subjecting the volunteers to a challenge with mosquito bites or intravenous injections of infected red blood cells. Phase IIb involves assessing the vaccine efficacy in a larger set of volunteers from endemic regions. Promising Phase II results qualify moving the vaccine testing to Phase III, which comprises assessing vaccine safety (including potential side effects) and efficacy over a longer time period in a cohort consisting of thousands of volunteers from endemic regions. If sufficient safety and efficacy has been demonstrated in Phase III (2 to 5 years), the vaccine can then be licensed and marketed for human use, after which mass-deployment for endemic regions can be launched.

3. Testing malaria vaccines in the field

A malaria vaccine could potentially target many different stages of the infection. It could work by: (1) preventing de novo infection (either in the liver or the blood), (2) controlling parasite levels in the blood and duration of the blood infection, (3) preventing pathology induced by the infection and thus preventing or reducing morbidity or mortality, and (4) preventing or reducing transmission to mosquitoes. However, not all of these outcomes can be assessed accurately in the field. Only the first and fourth outcomes mentioned can be assessed accurately and experimentally with reproducible results, mainly due to the standard operating procedures that have been implemented over the years.

Outcome assessment in the field is complicated due to the nature of the infection itself. In endemic regions, malaria infections are usually chronic [37]. Low-level parasite persistence may affect immune reactivity by amplifying or down-regulating vaccine-induced immune responses. It may also confuse diagnostics such as fever detection [38]. Occurrence and extent of chronicity may vary according to age, endemicity and host genetics. Thus, in many trials, antimalarial treatments are applied to the tested cohorts to clear prior malarial infections to reduce confounding factors [39].

Presence of co-infections is another factor that makes outcome assessment in the field difficult. Often, endemic cohorts are also infected with other pathogens, such as worms, bacteria or viruses, without being overtly sick [40, 41, 42, 43, 44]. Immune responses to these pathogens may either potentiate or inhibit the development of the protective response induced by infection or vaccination [45]. Due to cost constraints, it is rarely possible to make a full analysis for all possible pathogens, but it is advisable to perform retrospective studies to assess their possible influence on the malaria vaccination outcomes.

Another major roadblock for malaria vaccine development is the absence of correlates or surrogate markers of protection. These markers are crucial as they would facilitate the testing of large sets of vaccine formulations and would reduce costs and organization constraints [46]. As an example, vaccine development against Hepatitis B was greatly simplified when it was shown that concentration of Hepatitis B S antigen antibodies over 10 UI/ML level was a surrogate marker of protection (for review Plotkin et al.) [47]. This greatly accelerated the testing of multiple new formulations in a limited number of volunteers and also helped in the development of subsequent improved formulations. There is clearly a gap in our knowledge of the immune correlates of protection against malaria. It is still not clearly known what defense mechanisms are crucial in humans for mediating protection against malaria. This severely handicaps our progress towards effective vaccine development [48].

To assess vaccine efficacy in the field, it is also critical to have epidemiological data concerning the vaccine site. The level of endemicity will have an impact on the surveillance time following the last immunization and also on the size of the cohort. Low endemic conditions will require longer follow-up and a larger cohort to obtain statistically significant results.

4. Vaccines targeting specific parasite stages

The malaria parasite has a complex life cycle, alternating between the human and mosquito host. In the human host, the malaria parasite transits across different body compartments and alternates between intracellular and extracellular locations (Figure 1). This developmental complexity of the malaria parasite has a profound impact on the study design of the malaria vaccine and assessment efficacy (Table 1).

Figure 1.

Plasmodium falciparum life cycle and vaccine strategies. The cycle in humans includes three stages: the pre-erythrocytic stage, which is asymptomatic; the asexual blood stage, which induces pathology; and the sexual stage, which is transmitted to Anopheles mosquitoes. At each of these stages, the parasite expresses various proteins that are targets of vaccine candidates. The different vaccine strategies for each stage are indicated.

Targets	Induced immunity	Mechanisms	Readout
Pre-erythrocytic vaccines
Sporozoite antigens; Liver stage antigens	Inhibition of parasite development and replication/survival	Antibodies against sporozoites; T cells against liver stage	Presence of parasites in the blood
Blood stage vaccines
Asexual blood stage antigens	Inhibition of erythrocyte invasion and parasite replication/survival	Antibodies; Antibody cell dependent inhibition (ADCI); Cellular immunity	Blood parasite load
Parasite derived toxins; Parasite adhesion ligands	Inhibition of pathogenesis	Antibodies neutralizing inflammatory factors; Antibodies inhibiting parasite/host interactions	Fever; Blood parasite load; Severe complications^*
Transmission-blocking vaccines
Sexual blood stage antigens; Mosquito stage antigens	Inhibition of parasite development in the mosquito	Antibodies blocking gamete mating, ookinete formation or oocyst maturation	Presence of parasites in the mosquito vector (midgut, salivary glands)

Table 1.

Targets and mechanisms for anti-malaria vaccines.

During the pre-erythrocytic stage, the parasites exist as the extracellular motile sporozoite upon injection by the mosquito during feeding, and the intracellular liver parasites. Vaccines developed to target the pre-erythrocytic stage aim at inducing antibodies that target mainly the sporozoites and/or inducing T cells that will eliminate intracellular hepatic forms, thus preventing or controlling the extent of the subsequent blood stage development.

Vaccines targeting the asexual blood stages of the parasites are divided into two categories. The vaccines can be anti-parasite, which aim to control and eliminate parasite development in the blood or anti-disease, which aim to prevent the pathologies induced by the parasite. These vaccines need to induce different types of immune responses targeting different phases of the asexual blood stage. Vaccines targeting the sexual stage parasites aim to prevent transmission of gametocytes to the mosquito and/or gamete mating and ookinete development in the mosquito midgut.

It is worth noting that while the parasite expresses different set of genes at different stages of its life cycle, there are also many antigens that are expressed across the different parasite stages. Vaccination against these shared antigens may have an effect at different phases of the life cycle [22, 49, 50, 51, 52], making them just as attractive for vaccine against malaria.

5. Anti-parasite vaccines

For anti-parasite vaccines targeting the pre-erythrocytic stage, the assessment of vaccine efficacy is relatively easy. Complete efficacy for this stage is defined as sterile protection, whereby no parasite can be detected in blood of immunized individuals after the sporozoite challenge. This is an all or none phenomenon, because a single sporozoite developing in the liver can lead to full-blown blood infections.

Intuitively, one would expect great success of pre-erythrocytic vaccines since the limited numbers of sporozoites (a mean of 5–50) injected by infected mosquitoes [53] would be easily eliminated by the different arms of the immune system induced by the vaccine. However, this has proved to be the contrary. To date, only one vaccine formulation, RTS,S, an hybrid molecule containing a large segment of the circumsporozoite protein and S antigen of the Hepatitis B virus mixed with the AS02 adjuvant, has been shown to induce sterile protection in a substantial proportion of the naïve volunteers [36, 54] but to a much lower extent in field trials [55, 56, 57, 58].

One reason that could contribute to the lack of success stories with pre-erythrocytic anti-parasite vaccines is the procedures implemented to assess protection. Immunized volunteers were subjected to five mosquito bites, a dose required to ensure that naïve control volunteers would develop patent parasitemia 7–14 days after challenge [59, 60]. Alternative protocols using purified sporozoites injected either intradermal or intravenously have been developed, and so far, have proven to be safe and reproducible [61, 62]. Detection of parasitemia is performed by microscopy on Giemsa-stained blood smears over a 20–25-day period. Once a positive blood film is confirmed, the volunteers were treated with blood schizonticides to eliminate blood parasites and prevent any blood stage parasite-induced pathologies [63]. One limitation of this method of detection is that the time taken to detect parasites in the blood can differ up to 7 days. Hence, a delay of parasitemia does not necessarily translate in reduction of liver load. Moreover, there might be other confounding factors affecting the ability to detect blood parasitemia that are not related to the vaccination. To address this problem, sensitive PCR methods have been developed to detect the first wave of released liver merozoites and to assess the efficacy of the vaccine against pre-erythrocytic parasites. Using elegant regression methods, quantitative PCR techniques [64, 65, 66, 67] allow an estimation of the reduction of the parasite liver load and an accurate measure of the effect on the growth rate of blood stage parasites.

Assessing pre-erythrocytic vaccine efficacy in the field is complicated due to factors mentioned earlier, such as the nature of the infection and presence of other co-infections. Evaluation of pre-erythrocytic vaccine in the field had mostly relied on microscopy and long follow-up (usually 6 to 24 months). As mentioned above, this assay may not be the most suitable to accurately assess the efficacy of any formulation targeting the pre-erythrocytic stage. In field conditions, many of the volunteers have been previously infected and, depending on age and exposure, may have developed some immunity against blood stage parasites. Thus, to eliminate possible confounding effects of a synergistic immunity of on-going blood stage infection with immunity induced by vaccination, it is important that volunteers are cleared by drug treatment of low-level parasitemia during immunization and before the surveillance period. It is also necessary that low-level blood infection occurrence be assessed by PCR. When implemented, this approach has reduced the follow-up time period to 1 month, saving costs and allowing the assessment of new formulations [39].

For vaccines against the pre-erythrocytic stage of P. vivax, efficacy assessment is further complicated by the fact that this species may produce non-replicating liver form called hypnozoites. These hypnozoites are responsible for relapse up to 18 months after a sporozoite injection [68], thus complicating analysis and may require longer follow-up to detect relapse. Up to now, few challenges with P. vivax sporozoite have been performed [69]. There are no standard protocols and many issues need to be addressed [70]. First, the production of P. vivax sporozoites is limited since it requires infected blood from infected patients or monkeys to feed mosquitoes. Second, contrary to P. falciparum, no P. vivax cloned lines are available. Most of the lines available are derived from infected patients [71] or have been maintained in monkeys [72]. These lines contain multiple clones, which are poorly characterized at the molecular level [73]. This makes it difficult to obtain reproducible infection profile after experimental infection with mosquito bites of naïve volunteers and to characterize hypnozoite relapse profile. Moreover, as with anti-malarial drug studies in the field, the absence of validated genetic or serologic tools to distinguish between reinfection and relapse [74, 75, 76, 77] may also prevent detecting strain-specific effect. For anti-parasite vaccines targeting the blood stage, efficacy is assessed after sporozoite or asexual blood stage parasite challenge. Sterile protection occurs when no parasite can be detected in blood of immunized individuals. Detection of parasitemia can be monitored either by microscopy or by PCR, the latter providing more information. Due to its higher sensitivity, it allows the detection of at least 3–5 parasite cycles even before the parasite is detectable by microscopy. PCR [78, 79, 80, 81] bar-coding methods [82] can also be applied to genotype blood parasites. This allows assessing multiplicity of infection and determines whether the vaccine efficacy observed is strain-specific [83, 84]. Strain-specific vaccines have little interest since they will select vaccine-resistant parasites.

To assess the vaccine efficacy of anti-parasite vaccines, a challenge is essential. As mentioned earlier, challenge can be performed using sporozoites or blood stage parasites. However, due to the limited availability of insectaries that can provide infected mosquitoes on a regular basis, and the absence of accepted surrogates of protection, there is a necessity for blood stage challenge in healthy volunteers. Contrary to murine or monkey models where direct challenge with blood stage parasites is common, challenge with blood stage parasites in human has only been performed in limited vaccine studies using naïve volunteers [85, 86]. Because of safety reasons, blood parasites used for challenge need to be fully characterized. For a long time, only 3D7, a clone of the NF54 line, has been used. This line is susceptible to a wide range of antimalarials. Other parasite lines have been recently developed [87, 88]. However, since most of blood stage candidates are polymorphic, it is of utmost importance to assess the effect of polymorphism to have an idea of potential vaccine coverage. In addition, blood cells used for blood stage parasite propagation need to be pre-screened for the presence of a wide range of potential pathogens [86].

Despite these limitations, studies have shown that blood stage challenge can be safe and may allow the assessment of anti-asexual blood stage vaccine efficacy [89, 90]. Moreover, as recently reported, blood stage growth in vivo could be quantified more accurately after challenge with asexual blood stage parasites than with sporozoites [91], highlighting the advantage of this procedure. However, as it is not possible for safety reasons to let the parasitemia develop to high levels, its application might be limited. Vaccine efficacy may depend on the development of additional immune responses by the host during infection, which requires more time to be active as shown in mouse model [92]. In addition, some immune mechanisms may need higher parasite challenge dose to be triggered. Antibody-dependent cell immunity or ADCI has been proposed to be effective with parasitemia approaching level detectable by microscopy [93].

In endemic settings, efficacy of blood stage vaccines has been assessed in natural conditions after challenge by mosquito bites. This type of challenge is perfectly suited if the vaccine can induce sterile immunity. However, it might not be the most appropriate when blood parasite multiplication rate must be measured. This rate depends on the numbers of liver merozoites released and timing of their release. As mentioned above, liver merozoite release is not a homogenous phenomenon in terms of quantity and timing. Thus, to obtain parasite growth curve suitable for comparative analysis between individuals and groups, blood sampling must be carefully planned. Indeed, this implies an active and close follow-up of the volunteers to obtain multiple time points. One possibility to obtain more homogenous results would be to perform the challenge with defined number of infected red blood cells at a same time of infection across all groups [86, 94]. However, this requires overcoming a series of hurdles such as the development of standardized inoculums with known number of parasites at the same stage of development and the availability of donor blood, which have to be heavily tested for the presence of any pathogens. In addition, the parasite in the challenge inoculums would also need to be fully characterized and clearly defined in term of parasite clonality.

To accurately assess the efficacy of anti-parasite blood stage vaccine, it is necessary to evaluate any pre-existing immune responses to the antigens in the blood stage vaccine. Individuals leaving in endemic areas acquire immunity over time. The time required to develop this immunity depends on the endemicity level and their genetic background. This immunity may influence growth rate of the parasite. Pre-existing immunity can synergize with the immunity induced by the vaccination. Vaccination may also boost pre-existing antigen-specific immune responses, which would be ideal for any vaccine formulations. On the contrary, pre-existing immune response may inhibit or mask the immune response induced by vaccination. It has been shown that the antibodies to the N-terminal of P. falciparum merozoite surface protein 1 can block the inhibitory activity of antibodies recognizing the C-terminal part [95]. Thus, if such an antibody interference mechanism exists for antigen(s) used in vaccine formulation, it would be necessary to evaluate carefully pre-existing immune responses to these antigens.

6. Anti-disease vaccines preventing or reducing morbidity and mortality

Anti-disease vaccines aim to prevent the pathologies induced by the parasite. Hence, to assess the efficacy of these vaccines, it is important to clearly define the symptoms. Symptomatic malaria infections are characterized by recurrent fever and if not treated could develop into more severe complications (i.e. anemia, multi-organ dysfunctions affecting the lungs, kidneys, liver and brain…), and ultimately leading to death. These different clinical occurrences can be considered as end-points when assessing vaccine efficacy. For safety and ethical reasons, these end-points are looked for in experimental clinical trials. However, they are not measured in many field trials. Active and passive case detections are undertaken to detect clinical malaria episodes and define rate of the first episode or all episodes. Criteria to define a malaria case include presence of fever (≥37.5°C) and detection of malarial parasites in peripheral blood. Careful clinical assessment of the origin of fever is needed to ensure the fever is due to the parasite but not due to concomitant bacterial or viral infection. It should be mandatory to prevent undermining the vaccine efficacy. It is also crucial to clear any asymptomatic infections prior to vaccine testing. Clearing asymptomatic parasitemia allows a better identification of malaria-attributable fever [38]. Assessment of the reduction of severe symptom occurrence and mortality is more difficult to use as end-point. Because of active intervention (drug treatment and patient management), severity and mortality occur only in small fraction of clinical cases. Thus, in order to have sufficient statistical power to assess the vaccine testing, very large cohort is required, resulting in huge cost. Moreover, there have been concerns that decreasing the level of exposure to the parasites might, in return, results in an increase in mortality in the long–term [96]. It has been suggested that reduced exposure prevents the development of naturally-acquired clinical immunity [93], which is thought to result from constant parasite exposure. Thus, for any vaccines entering in Phase III trials, these end-points need to be assessed.

Two types of vaccine strategies aimed at reducing specifically morbidity and mortality are being developed. Anti-sequestration vaccines are based on the assumption that cytoadherence of infected red blood cells leading to parasite sequestration in deep tissues is responsible for most of malaria pathologies. These vaccines are designed to target parasite ligands such as members of the var. multigene family encoding the proteins Plasmodium falciparum erythrocyte membrane protein-1 (PfEMP-1), which mediates cytoadherence [97, 98]. It has been proposed that parasites sequester to avoid splenic elimination [99]. The more clinically-advanced anti-sequestration vaccine candidate aims at preventing pregnancy-associated malaria [100]. Few var. genes, which encode PfEMP-1 binding to chondroitin sulfate A (CSA), have been implicated in placental sequestration, thus making them attractive vaccine candidates [101, 102]. Anti-sequestration vaccines are designed to produce antibodies, which prevent the interactions between infected red blood cells and their cognate host cells (endothelial cells, syncytiotrophoblast…). This will lead to an increase in the circulation of blood parasites at all development stages and hence their elimination by the spleen. Primary end-point measures for such vaccines are both parasitological and clinical. Efficacy of desequestration as measured by the number of mature blood forms can be evaluated simply by microscopical observation of Giemsa stained-blood smears. However, preventing sequestration may lead to rapid increase of parasitemia and possibly a faster development of fever episodes and faster treatment application. Thus, the time-window for monitoring parasite development might be limited. Ultimately, assessment of severity and mortality are the most relevant measures for desequestration vaccines. A large cohort is needed to assess efficacy, like any blood stage vaccine targeting parasite growth.

Another type of anti-disease vaccine is targeting parasite moieties behaving as toxins and inducing immune-mediated pathologies [28, 103]. As these vaccines may have no effect on the parasitemia, the assessment of their efficacies will require very large cohorts to assess clinical outcomes with both active and passive case detection. In terms of safety, a strict clinical follow-up starting as early as the last vaccine dose administration is needed since these may perturb the immune network and induce immunopathology.

7. Transmission blocking vaccines

Transmission blocking vaccines (TBV) are designed to prevent or reduce the development of the sporogonic stage inside the mosquito host. This leads to a reduction in the numbers of infected mosquitoes and hence lesser malaria transmission in the population. As such, experimental clinical trials cannot provide straightforward answers of vaccine efficacy since the effect of such vaccine is at the population level and not at individual level. Assessment of TBV efficacy is done using in vitro assays and the membrane-feeding assay. Mosquitoes are fed in vitro with Plasmodium gametocytes mixed with serums from immunized individuals and the level of sporogonic development is assessed by counting the oocysts in dissected mosquito midguts [104] or, more rarely, the sporozoites in the mosquito salivary glands. Although this assay has been instrumental in identifying target antigens, it remains to be seen whether it might help to define correlate of protections for TBV development.

Currently, none of the TBV has progressed to clinical trials in the field. As the principal outcome of TBV is to reduce the number of infected mosquitoes at the population level, methods for assessing their number in field conditions should be implemented in a timely manner. There are, to date, no standardized methods to estimate the number of infected mosquitoes in the field and estimation would require large sampling size. Moreover, it seems that infected mosquitoes can cluster in discrete locations [105], thus requiring extensive studies on the distribution of infected mosquitos before and after vaccine trials. Since the number of infected mosquitoes depend on the number of circulating gametocytes [106], defining the number of gametocyte carrier prior to vaccine implementation is also a pre-requisite. In addition, defining transmission intensity of the vaccine site is important since it may influence the outcome of the vaccination. This can be defined by seroepidemiology and geographical information system (GIS) applications. For the latter, GPS mapping of mosquitoes and infected humans needs be done. It must be noted that one major limitation of these trials is that they assume that the human and mosquito populations tested are not mobile, which is often not true. An influx of infected individuals can modify the outcome by creating new reservoirs, and an influx of external infected mosquitoes would maintain transmission. It has been suggested that TBV can be tested with accuracy only in enclaved locations such as islands. Ultimately, the main expected outcome is that TBV will reduce transmission and thus reduce morbidity and mortality. The effect of such vaccine is at the population level, a large and costly cohort will be needed to be assessed over a long period of time. However, recent advances in modeling might facilitate TBV assessment by identifying end-point measures, which may serve as correlates of protection [107, 108]. Different end-point measures have been developed to assess TBV efficacy. For TBV that targets gametocytes, numbers of gametocytes and duration of gametocytes are important measures since it is expected that reduction in the number of gametocyte-carriers (reservoir) will decrease transmission. Microscopic determination of gametocytes on Giemsa-stained smears has long been used but they must be complemented with PCR methods since they have shown to underestimate gametocyte load [109]. Gametocyte infectivity to mosquitoes can be measured with the membrane-feeding assay mentioned earlier or with feeding of mosquitoes directly on the skin of gametocyte carriers, which reproduces the natural situation. An honest correlation between the two assays has been described in few studies [110] but the membrane-feeding assay still awaits definitive validation [111]. In summary, it would be relevant for future TBV trials to perform feeding directly on gametocyte carriers using local mosquitoes. These mosquitoes would have to be raised in local insectaries and tested for the absence of any other human pathogens. Measures of TBV efficacy should not be limited to development of oocysts but also to salivary glands sporozoites since the latter are the infectious forms to humans. Hence, future studies should aim at measuring salivary gland sporozoite loads and sporozoite infectivity.

8. Concluding remarks

Here we discussed the different types of malaria vaccines and the different ways to access the vaccine efficacy. We also highlighted the limitations involved and the difficulties encountered by researchers aiming to develop an efficacious vaccine against a complex parasite such as Plasmodium. Despite decades of research efforts in vaccine development, no efficient malaria vaccine (i.e. with an efficacy >50%) has been developed. The most clinically-advanced RTS,S, which has been tested in Phase III, conferred at best 30–40% protection against clinical malaria [112]. Modeling studies have been proposed that, together with other malaria interventions, RTS,S vaccination may reduce the incidence of clinical malaria and deaths in many sub-Saharan African countries [113, 114]. Thus, this has led to the approval of licensure for the RTS,S vaccines by the European Union. However, the World Health Organization has not recommended its use in the extended program of immunization for children due to its discouraging vaccine efficacy data. Thus, the future of this vaccine for mass deployment remains uncertain. One of the major reasons of the limited efficacy of RTS,S vaccines and the discontinuation of various other vaccine development efforts is certainly due to antigen polymorphism [83]. In addition to antigen polymorphism, the malaria parasite utilizes many other immune escape mechanisms [115], which have severely hampered the development of malaria vaccines. With the renewed interest in malaria eradication, the development of an effective malaria vaccine is high on the agenda. Diverse strategies are being proposed to develop better vaccines: identification of new vaccine candidate [116], combinations of different antigens targeting the same stage or different stages [117]; new delivery systems and prime-boost strategies using different modalities [118]; and new adjuvants to induce stronger and longer lasting efficient immune responses [119, 120, 121, 122]. However, for all vaccine types described, the absence of validated surrogates of protection to help select and prioritize different vaccine formulations is a major roadblock, which should be given priority to accelerate vaccine testing.

Acknowledgments

This work was supported by the Agency for Science Technology and Research (A*STAR), Singapore.