**5.1 The role of MSP-1 in natural immunity**

Various biological factors influencing the function of MSP-1 specific antibodies have been reported in individuals with natural immunity:

(1) The role of MSP-1 specific antibody titers, isotype and the association with protection and/or reduction in morbidity:

A meta-analysis of 33 clinical studies revealed that the presence of MSP-1p19 specific antibodies is associated with a lower incidence rate of malaria (Fowkes et al., 2010). Moreover, high levels of anti-PfMSP-1p19 immunoglobulin G were associated with reduced malaria in an age-adjusted multivariate analysis (Perraut et al., 2005). In contrast, other reports failed to show any associations between MSP-1p19 (MSP-1) Abs and clinical outcome (Dodoo et al., 1999; Nebie et al., 2008). At this point we can only speculate about the cause of this discrepancy. As outlined above, the differences in the methodology and/or

The Impact of Immune Responses on the Asexual

Erythrocytic Stages of *Plasmodium* and the Implication for Vaccine Development 217

case of the single allele vaccine - moderate growth inhibitory activity was induced. The MSP-1p42 vaccines generated at WRAIR underwent further clinical evaluation in Phase Ib studies conducted in Kenya (Stoute et al., 2007) and Mali (Thera et al., 2006), where adults were immunized with the MSP-1p42 (3D7) adjuvanted in AS02A. In Kenya, although preexisting MSP-1p42 antibody titers in the participants were high, they could be boosted by the vaccine. In addition, Phase Ib (Withers et al., 2006) and Phase IIb (Ogutu et al., 2009) studies were conducted in Kenyan children (1-4 years old) using the same vaccine formulation. Results from this study indicated that younger children mounted stronger vaccine responses in terms of the magnitude of the antibody response. Sera from the Phase IIb study displayed strong growth inhibitory activity against the heterologous FVO parasites indicating that the predominant strain circulating at the time of natural exposure was different from the vaccine strain 3D7 (Angov et al., manuscript in preparation). Growth inhibitory activity to the 3D7 parasites was only observed in a small proportion of the study population with no significant difference between the rabies control group (Rabipur) and the malaria vaccine group. Some 3D7 specific GIA activity was reversible by antigen add-back confirming that some of the activity in the sera was due to MSP-1 specific antibodies. The major conclusion from this trial was that although the vaccine was safe and immunogenic, in the context of the heterologous exposure, the vaccine did not induce sufficiently cross-reactive responses. Future studies should include MSP-1p42 allele(s) that are better matched to the dominant circulating parasites. Another important consideration is to clear parasitemia in the study participants prior to and during the course of the vaccination. The presence of parasites during vaccination

could lead to competitive immune responses thus curtailing the vaccine's potential.

parasites, and showed no activity in growth inhibition assays.

Some vaccine approaches have focused on the MSP-1p19 fragment rather than the fulllength MSP-1p42 since this region of the molecule contains the highly conserved functionally important B cell dominant epitopes. In one particular construct, helper T cell epitopes from tetanus toxoid were used to generate a chimeric P30P2-MSP-1p19 protein (Keitel et al., 1999). Others approached the lack of T cell help by generating a chimeric AMA-1/MSP-1p19 vaccine, which was tested in a Phase Ia trial (Malkin et al., 2008). This vaccine induced antibodies to both antigens; however the antibodies reacted primarily to the recombinant antigens by ELISA, less well by immunofluorescence assay (IFA) on whole

Therefore, with regards to further characterizing of MSP-1-based vaccine approaches, there is a need for reliable readout preclinical methods to enable prediction of protection and to facilitate the down-selection of vaccine candidates. One such attempt was the development of a transgenic rodent malaria parasite (*P. berghei*) expressing the *P. falciparum* MSP-1p19 transgene (De Koning-Ward et al., 2003). These parasites may be useful in the down-selection of vaccine candidates by either immunization strategies in the murine model (which does not necessarily address immunogenicity in humans) or passive transfer strategies to characterize the immune potential of antibodies induced in vaccinated individuals. Moreover, a recent report suggests that some anti-MSP-1 specific antibodies may mediate ADCI. Due to the incomplete compatibility between human Abs and murine Fc receptors, current mouse models are unable to use this pathway and may lead to false negative results. There are some experimental alternatives such as transgenic mice that express one of the human Fc receptors (CD32 or CD64) or humanized mice. Humanization refers to irradiated mice that have been reconstituted with human leukocytes (Badell et al., 2000; Pleass et al., 2003) or hematopoietic

stem cells which leads to the production of an array of human blood cells in the mice.

choice of plate antigen may be responsible for some of these issues. It is, however, interesting to note that full length MSP-1p42 was only used in one study as the plate antigen and this study reported a reduced risk of malaria (Al-Yaman et al., 1996).

(2) The role of antibody isotype and functional activity against the parasite:

Longitudinal studies have demonstrated an association between the IgM and IgG responses to MSP-1p19 and the degree of clinical disease and anemia in infants and pregnant women (Branch et al., 1998). Similarly, high antibody levels of MSP-1 specific IgG1 were associated with reduced morbidity (Riley et al., 1992; Al-Yaman et al., 1996) with protection against high-level parasitemia (Fowkes et al., 2010) and clinical disease (Egan et al., 1996; Cavanagh et al., 2004; Soe et al., 2004).

(3) The role of fine specificity of antibodies and association with protection:

IgG derived from sera obtained from Kenyan residents were tested for their impact on parasite viability and growth. The results demonstrated that the invasion inhibitory antibodies were specific for the C-terminal MSP-1p19 (John et al., 2004). Overall, there was a lack of association of total IgG or IgG subclass Abs to MSP-1p19 measured by ELISA with either invasion-inhibitory activity or protection against infection. In contrast, a study analyzing sera from children in West Africa (Sierra Leone and Gambia) demonstrated a strong association between antibody titers to the C-terminus of MSP-1 (MSP-1p19) and protection against clinical malaria and high level parasitemia (Egan et al., 1996). Thus, the fine specificity (*i.e.*, epitope specificity) of the MSP-1 specific antibodies appears to be important and testing only for antibody titers to the total molecule or a fragment may result in the loss of an association with a clinical response (Corran et al., 2004; Okech et al., 2004). The C-terminus of the MSP-1 is comprised of two EGF-like domains and depending on which of the domains the antibodies recognize results in either growth inhibition or no functional activity against the parasite (Chappel et al., 1994; Darko et al., 2005).

#### **5.2 Efficacy of MSP-1 based vaccines in naïve and malaria-exposed individuals**

Several clinical trials have been conducted testing either MSP-1p42 or MSP-1p19 as vaccine candidates. The objective of using the larger subunit, MSP-1p42, was to ensure that potential helper epitopes, which can induce antibodies are present in the immunogen. Moreover, the N-terminus p33 portion of the molecule contains most of the known T cell epitopes. The nature of the C-terminus, *i.e.*, due to several disulfide bridges associated with the EGF-like domains, renders the structure rigid and thus resistant to processing by antigen presenting cells, and therefore does not contain any dominant T cell epitopes. Thus, using the full length p42 fragment, which is expressed on pRBC starting at the trophozoite stage, would allow antibodies to bind to the parasites even inside the pRBC thereby potentially preventing the rupture of schizonts. This fact motivated two institutions (National Institute of Health (NIH) and Walter Reed Army Institute of Research (WRAIR)) to proceed with two independent MSP-1p42 based vaccines. At the NIH, a mixture of MSP-1p42 (FVO) and MSP-1p42 (3D7) was adjuvanted with Alhydrogel together with or without the addition of CpG7909 (Ellis et al., 2010) while the WRAIR vaccine consisted of single-allele vaccines, MSP-1p42 (3D7) (Ockenhouse et al., 2006) or MSP-1p42 (FVO) (Spring et al., manuscript in preparation), adjuvanted with GSK's adjuvant system, AS02A or AS01B, respectively. Both were tested in US naïve individuals. All studies reported good immunogenicity and - in the

choice of plate antigen may be responsible for some of these issues. It is, however, interesting to note that full length MSP-1p42 was only used in one study as the plate antigen

Longitudinal studies have demonstrated an association between the IgM and IgG responses to MSP-1p19 and the degree of clinical disease and anemia in infants and pregnant women (Branch et al., 1998). Similarly, high antibody levels of MSP-1 specific IgG1 were associated with reduced morbidity (Riley et al., 1992; Al-Yaman et al., 1996) with protection against high-level parasitemia (Fowkes et al., 2010) and clinical disease (Egan et al., 1996; Cavanagh

IgG derived from sera obtained from Kenyan residents were tested for their impact on parasite viability and growth. The results demonstrated that the invasion inhibitory antibodies were specific for the C-terminal MSP-1p19 (John et al., 2004). Overall, there was a lack of association of total IgG or IgG subclass Abs to MSP-1p19 measured by ELISA with either invasion-inhibitory activity or protection against infection. In contrast, a study analyzing sera from children in West Africa (Sierra Leone and Gambia) demonstrated a strong association between antibody titers to the C-terminus of MSP-1 (MSP-1p19) and protection against clinical malaria and high level parasitemia (Egan et al., 1996). Thus, the fine specificity (*i.e.*, epitope specificity) of the MSP-1 specific antibodies appears to be important and testing only for antibody titers to the total molecule or a fragment may result in the loss of an association with a clinical response (Corran et al., 2004; Okech et al., 2004). The C-terminus of the MSP-1 is comprised of two EGF-like domains and depending on which of the domains the antibodies recognize results in either growth inhibition or no

and this study reported a reduced risk of malaria (Al-Yaman et al., 1996). (2) The role of antibody isotype and functional activity against the parasite:

(3) The role of fine specificity of antibodies and association with protection:

functional activity against the parasite (Chappel et al., 1994; Darko et al., 2005).

**5.2 Efficacy of MSP-1 based vaccines in naïve and malaria-exposed individuals** 

Several clinical trials have been conducted testing either MSP-1p42 or MSP-1p19 as vaccine candidates. The objective of using the larger subunit, MSP-1p42, was to ensure that potential helper epitopes, which can induce antibodies are present in the immunogen. Moreover, the N-terminus p33 portion of the molecule contains most of the known T cell epitopes. The nature of the C-terminus, *i.e.*, due to several disulfide bridges associated with the EGF-like domains, renders the structure rigid and thus resistant to processing by antigen presenting cells, and therefore does not contain any dominant T cell epitopes. Thus, using the full length p42 fragment, which is expressed on pRBC starting at the trophozoite stage, would allow antibodies to bind to the parasites even inside the pRBC thereby potentially preventing the rupture of schizonts. This fact motivated two institutions (National Institute of Health (NIH) and Walter Reed Army Institute of Research (WRAIR)) to proceed with two independent MSP-1p42 based vaccines. At the NIH, a mixture of MSP-1p42 (FVO) and MSP-1p42 (3D7) was adjuvanted with Alhydrogel together with or without the addition of CpG7909 (Ellis et al., 2010) while the WRAIR vaccine consisted of single-allele vaccines, MSP-1p42 (3D7) (Ockenhouse et al., 2006) or MSP-1p42 (FVO) (Spring et al., manuscript in preparation), adjuvanted with GSK's adjuvant system, AS02A or AS01B, respectively. Both were tested in US naïve individuals. All studies reported good immunogenicity and - in the

et al., 2004; Soe et al., 2004).

case of the single allele vaccine - moderate growth inhibitory activity was induced. The MSP-1p42 vaccines generated at WRAIR underwent further clinical evaluation in Phase Ib studies conducted in Kenya (Stoute et al., 2007) and Mali (Thera et al., 2006), where adults were immunized with the MSP-1p42 (3D7) adjuvanted in AS02A. In Kenya, although preexisting MSP-1p42 antibody titers in the participants were high, they could be boosted by the vaccine. In addition, Phase Ib (Withers et al., 2006) and Phase IIb (Ogutu et al., 2009) studies were conducted in Kenyan children (1-4 years old) using the same vaccine formulation. Results from this study indicated that younger children mounted stronger vaccine responses in terms of the magnitude of the antibody response. Sera from the Phase IIb study displayed strong growth inhibitory activity against the heterologous FVO parasites indicating that the predominant strain circulating at the time of natural exposure was different from the vaccine strain 3D7 (Angov et al., manuscript in preparation). Growth inhibitory activity to the 3D7 parasites was only observed in a small proportion of the study population with no significant difference between the rabies control group (Rabipur) and the malaria vaccine group. Some 3D7 specific GIA activity was reversible by antigen add-back confirming that some of the activity in the sera was due to MSP-1 specific antibodies. The major conclusion from this trial was that although the vaccine was safe and immunogenic, in the context of the heterologous exposure, the vaccine did not induce sufficiently cross-reactive responses. Future studies should include MSP-1p42 allele(s) that are better matched to the dominant circulating parasites. Another important consideration is to clear parasitemia in the study participants prior to and during the course of the vaccination. The presence of parasites during vaccination could lead to competitive immune responses thus curtailing the vaccine's potential.

Some vaccine approaches have focused on the MSP-1p19 fragment rather than the fulllength MSP-1p42 since this region of the molecule contains the highly conserved functionally important B cell dominant epitopes. In one particular construct, helper T cell epitopes from tetanus toxoid were used to generate a chimeric P30P2-MSP-1p19 protein (Keitel et al., 1999). Others approached the lack of T cell help by generating a chimeric AMA-1/MSP-1p19 vaccine, which was tested in a Phase Ia trial (Malkin et al., 2008). This vaccine induced antibodies to both antigens; however the antibodies reacted primarily to the recombinant antigens by ELISA, less well by immunofluorescence assay (IFA) on whole parasites, and showed no activity in growth inhibition assays.

Therefore, with regards to further characterizing of MSP-1-based vaccine approaches, there is a need for reliable readout preclinical methods to enable prediction of protection and to facilitate the down-selection of vaccine candidates. One such attempt was the development of a transgenic rodent malaria parasite (*P. berghei*) expressing the *P. falciparum* MSP-1p19 transgene (De Koning-Ward et al., 2003). These parasites may be useful in the down-selection of vaccine candidates by either immunization strategies in the murine model (which does not necessarily address immunogenicity in humans) or passive transfer strategies to characterize the immune potential of antibodies induced in vaccinated individuals. Moreover, a recent report suggests that some anti-MSP-1 specific antibodies may mediate ADCI. Due to the incomplete compatibility between human Abs and murine Fc receptors, current mouse models are unable to use this pathway and may lead to false negative results. There are some experimental alternatives such as transgenic mice that express one of the human Fc receptors (CD32 or CD64) or humanized mice. Humanization refers to irradiated mice that have been reconstituted with human leukocytes (Badell et al., 2000; Pleass et al., 2003) or hematopoietic stem cells which leads to the production of an array of human blood cells in the mice.

The Impact of Immune Responses on the Asexual

**6. Future direction** 

(Draper et al., 2009).

Erythrocytic Stages of *Plasmodium* and the Implication for Vaccine Development 219

One biological function known to be displayed by some MSP-1 specific antibodies is the inhibition of the secondary processing of the MSP-1 molecule into the p33 and the p19 portion (Blackman et al., 1994). This activity was found in sera from individuals with acquired natural immunity (Patino et al., 1997). In this assay, merozoites are prepared from synchronized blood stage cultures and their ability to process MSP-1p42 into p33 and p19 in the presence or absence of immune antibodies is evaluated. This method is qualitative at best and due to its nature not designed for high-throughput testing. Thus, very few laboratories use the technique for the evaluation of functional activity in immune sera.

The current body of literature clearly supports the development of a blood stage vaccine. Although such a vaccine would not prevent infection, it can reduce morbidity and mortality associated with malaria infection and therefore such a vaccine would save the lives of many residents of malaria-endemic areas. The experience and knowledge gained from these studies should be used to rationally design new vaccine formulations and future clinical

*(1) Vaccine platform.* Most erythrocytic vaccines tested so far are based on recombinant, soluble proteins. When using recombinant proteins it is paramount to assure proper, thus native-like, protein folding. Sera from malaria-endemic areas are a useful tool to establish the degree of cross-reactivity between the recombinant vaccine and the "native" antigen. Alternative vaccine platforms to those primarily described here, soluble proteins plus adjuvant, are particle-based approaches. When using this approach it is important to assure proper orientation of the protein on the particle. Some proteins that are inherently unstructured such as MSP-2 could benefit from particle formation because the particle provides a stabilizing scaffold. Moreover, the distance between the epitopes and the density may be crucial in order to induce proper immune responses. Such particle presentation could be achieved by using either recombinantly expressed antigen on whole-killed bacteria or viruses. For example, mouse studies using recombinant adenovirus encoding MSP-1 demonstrated "protection" (defined by the authors as delayed and lower parasitemia)

*(2) Development of preclinical and clinical models that better predict human anti-malarial responses.* When using preclinical animal models, the parasite growth kinetic is frequently different between those *Plasmodium* species that are suitable for the respective animal model and *P. falciparum* thus failing to simulate the clinical situation. Testing vaccine candidates in animals can be challenging when testing *P. falciparum* antigens: except for one confirmed antigen, *Pf*CelTOS, (Bergmann-Leitner et al., 2010), malaria antigens are relatively speciesspecific, *i.e.,* immunization with *P. falciparum* antigens does not confer protection against a heterologous *Plasmodium* species. Thus, investigators often make their decisions based solely on immunogenicity in the animal model (cellular and/or humoral responses notionally thought to be important). Alternatively, investigators have searched for orthologs of the *P. falciparum* antigen in the respective *Plasmodium* species relevant for their preclinical model to conduct immunization and challenge studies. A caveat of this approach is that the ortholog may have a different function than the *P. falciparum* antigen in human malaria or there simply may not be a valid ortholog (*e.g*. *P. falciparum* LSA-1 does not have an ortholog in

trials. Factors that need to be considered for their success are:

#### **5.3 Biological effect of MSP-1 specific antibodies on parasite growth and function**

Characterizing immune responses induced by MSP-1 vaccines revealed that several factors impact the immunogenicity and functional activity of the induced antibodies (vaccine platforms will be discussed in detail in Section 6): (1) the expression system used for the production of the recombinant protein (*i.e., E. coli*, baculovirus, or yeast) (Arnot et al., 2008; Reed et al., 2009), (2) the amino-acid sequence used for vaccine development (*i.e.* full length gp195, MSP-1p42 or MSP-1p19) (Stowers et al., 2001; Woehlbier et al., 2006) and (3) the vaccine platform used to deliver the MSP vaccine (*i.e*, recombinant protein, recombinant viral vectors or DNA vaccines).

Our laboratory has intensively studied the anti-parasite effects induced by MSP-1p42 specific antibodies. Early observations indicated that MSP-1 specific antibodies impacted the various parasite strains differently depending on their classification as MAD20 or Wellcome/K1-like. These two alleles differ markedly in their p33 fragments while only by four amino acids (E-TSR *vs.* Q-KNG, respectively) in the p19 portion of the molecule. To this end, MSP-1 specific antibodies were able to significantly delay the intra-erythrocytic development of the 3D7, but not the FVO parasite clone (Bergmann-Leitner et al. 2009). In the case of FVO parasites, anti-MSP-1p42 antibodies prevented schizont rupturing by stalling or arresting intra-erythrocytic parasite development likely through direct interactions with intra-erythrocytic parasites within the parasitophorous vacuole, which is putatively connected to the surface of the pRBC by the parasitophorous duct. This duct gives antibodies, but not larger immune components access to the parasite inside the vacuole (Bergmann-Leitner et al., 2009). In contrast, the same antisera tested on the 3D7 parasite clone were unable to interfere with the release of the merozoites. These antibodies were still able to agglutinate merozoites and interfere with invasion. We expanded our analysis to the parasite clone CAMP/FUP that has a p33 and a p19 EGF-like domain 1 identical to the sequence of the 3D7 parasite clone, and an EGF-like domain 2 identical to the FVO parasite clone. We observed the same response pattern as reported for the FVO parasite indicating that antibodies within the EGF-like domain might be responsible for stalling the rupture of the schizonts. In contrast, no significant growth inhibition was observed following successful invasion indicating that EGF-like domain 1 specific antibodies may be mediating this particular biological effect. To test this working hypothesis, the activity of affinity purified antibodies specific to the entire p19 or each of the EGF-like domains were compared to the source material (antibodies induced by immunization with the MSP-1p42 vaccine representing either the FVO or the 3D7 allele). We concluded that only antibodies that bound to regions within the p42 or to the p19 subunit, but not the EGF-like domain 1 or 2 subunits, displayed growth inhibitory activities. This was surprising given previous observations suggesting that responses directed to the p19 were associated with reduced parasite density or clinical disease (see above). However, the proper folding of the recombinant fragments used, which represent the two EGF-like domains has not yet been confirmed. This result may indicate that protective epitopes may depend on proper tertiary structure of the molecule. In support of this theory, it has been reported that inhibitory anti-MSP-1 specific antibodies map to epitopes formed through the "properly" folded p19 subunit and not to its sub-domains (McBride and Heidrich, 1987; Uthaipibull et al., 2001).

One biological function known to be displayed by some MSP-1 specific antibodies is the inhibition of the secondary processing of the MSP-1 molecule into the p33 and the p19 portion (Blackman et al., 1994). This activity was found in sera from individuals with acquired natural immunity (Patino et al., 1997). In this assay, merozoites are prepared from synchronized blood stage cultures and their ability to process MSP-1p42 into p33 and p19 in the presence or absence of immune antibodies is evaluated. This method is qualitative at best and due to its nature not designed for high-throughput testing. Thus, very few laboratories use the technique for the evaluation of functional activity in immune sera.
