**3.1.1 Circumsporozoite Protein (CSP)**

CSP, which is found in the outer sporozoite membrane, was shown to be a target of antibodies that prevented hepatocyte infection, thereby interrupting an obligatory stage in the parasite's lifecycle [15]. Cloning of the CSP gene (amongst other antigens) opened the doors to mass production and it was hoped, a malaria vaccine [53]. CSP contains a central (NANP) repeat region containing B-cell epitopes flanked by non-repetitive sequences containing T-cell epitopes [53-55]. In 2004, a considerable protective effect was observed with a CSP-based vaccine known as RTS,S with 49% efficacy against severe malaria for a period of 18 months in children aged 1-4 years [44]. This vaccine is composed of the central repeat and the C-terminal regions of CSP isolated from the reference strain, 3D7, combined with a Hepatitis B surface antigen (which provides protection against Hepatitis B and therefore satisfies ethical requirements). Recently published data from a phase III trial including 6000 African children has shown a 50% reduction in clinical episodes and 35% reduction in severe disease in children aged 5-18 months [25]. Therefore, RTS,S is likely to be the first licensed malaria vaccine and should be available within the next few years [56]. Most would agree though that it is an important proof of principle rather than an end to the hunt for a vaccine.

#### **3.1.2 Thrombospondin Related Adhesion Protein (TRAP)**

TRAP also known as Sporozoite Surface Protein (SSP) [18, 57] is present within the micronemes of the sporozoite and is essential for gliding motility and host cell invasion in the mosquito and human pre-erythrocyte stages [58-60]. TRAP is also expressed in the blood stages though its function in this stage is not well understood [18, 57]. TRAP is a Type I transmembrane protein consisting of multiple adhesive domains and a central repeat region flanked by unique sequence. Anti-repeat monoclonal antibodies have been shown to at least partially inhibit sporozoite invasion *in vitro* [61] and potent T-cell responses have been identified in individuals vaccinated with irradiated sporozoites [62, 63] as well as people

Using Population Genetics to Guide Malaria Vaccine Design 237

Another approach for AMA1 that is being tested is a diversity-covering vaccine (PfAMA1 DiCo) whereby 355 published sequences were aligned and common polymorphisms (i.e. those found in >16% of sequences) were used to design three synthetic constructs, which together encompass 97% of the sequence diversity [99]. So far there have not been any clinical trials however broader parasite growth inhibitory activity was observed in animals vaccinated with the DiCo mix than those vaccinated with native AMA1 from the

MSP1 is the most abundant and best studied of the *P. falciparum* merozoite surface antigens. The protein is synthesised as a 190-kDa-precursor protein that is cleaved into four fragments of 83, 30, 38, and 42 kDa [100]. These exist as a non-covalently associated complex tethered by a glycosylphosphatidylinositol (GPI) anchor at the C-terminal [101]. During erythrocyte invasion, the N-terminal fragments are shed when MSP142 undergoes a secondary processing event generating a further 33 kDa (MSP133) and 19 kDa fragment (MSP119) [102- 104] which facilitates parasite entry into the erythrocyte [105]. The MSP119 fragment is

The C-terminal end of MSP1 is a leading malaria vaccine candidate. The MSP133 sequence is dimorphic, with diverse alleles clustering within the two families known as K1 and Mad20 [107], however MSP119 is relatively conserved across the two allele families with six nonsynonymous SNPs that are commonly used to describe MSP119 haplotypes [108-113]. Within each of the allele-families, MSP133 also contains several single nucleotide

The FMP1/ASO2A vaccine, which is formulated with the 3D7 allele of MSP142, has been tested in Phase II trials. The vaccine initially showed promise in safety and immunogenicity trials in 40 Malian adults, with responses generated to parasite clones carrying diverse MSP142 alleles (FVO and Camp/FUP) [114]. Phase II efficacy trials including 400 children in Western Kenya did not protect against infection or lower parasite densities, nor did it reduce clinical episodes [115]. Though the authors stated that the vaccine was no longer a promising vaccine candidate, this was without having investigated how many vaccinees were infected with the vaccine strain. In addition, high throughput genotyping studies in Mali have demonstrated that the 3D7 allele had a prevalence of only 16% [113]. A vaccine containing both MSP142 3D7 and FVO alleles fused together with conserved regions of MSP1

MSP2, a 45-52 kDa glycoprotein is tethered to the membrane a glycosylphosphatidylinosital (GPI) anchor and is the second most abundant protein (based on copy numbers) on the merozoite surface [49]. The protein consists of highly conserved N- and C-terminal ends flanking a highly polymorphic central repeat region. MSP2 sequences fall into two distinct allelic families namely FC27 and 3D7 (IC-1) [117-120]. Within these allele-families, the central repeats vary in length, number and sequence among isolates. Allele-specific and length polymorphism in MSP2 has been used as the basis for high-resolution genotyping of

FVO strain [99].

**3.1.4 Merozoite Surface Protein 1 (MSP1)** 

retained on the surface of the invading merozoite [106].

polymorphisms (SNPs) and a 3 bp deletion [108, 111].

is currently being tested [116].

*P. falciparum* isolates [121, 122].

**3.1.5 Merozoite Surface Protein 2 (MSP2)**

who are naturally exposed to malaria [64]. TRAP is highly polymorphic [65] and has also been shown to be under strong diversifying selection suggesting that it is a natural immune target [66].

Mapping of inhibitory epitopes has led to the development of the multiepitope (ME)-TRAP DNA vaccine, which contains 15 TRAP T-cell epitopes. The vaccine has been delivered using a prime-boost regime involving three doses of naked DNA or DNA fused to a viral vector, to produce potent T-cell responses in both animals and humans [67]. While safety and immunogenicity trials in the UK have shown promising results [68, 69], Phase II efficacy trials conducted in semi-immune adults in The Gambia and children in Kenya provided no evidence of protective efficacy [70-73].

### **3.1.3 Apical Membrane Antigen 1 (AMA1)**

AMA1 is an 83 kDa integral membrane protein expressed late in the asexual stages and found in the micronemes at the apical tip of merozoites [74]. Prior to schizont rupture the AMA1 prodomain is cleaved to reveal a 66 kDa protein containing three subdomains (domains I, II and III) defined by their disulfide bonds [24]. AMA1 then relocates to the parasite surface [75] where further processing during invasion leads to shedding of two fragments of 44 and 48 kDa whilst the 22kDa cytoplasmic tail remains in the membrane and is carried into the invaded erythrocyte [76, 77]. AMA1 is now known to function in formation of the tight junction through interactions with rhoptry neck proteins [78, 79]. This molecule is also expressed in sporozoites where it is involved in hepatocyte invasion [80].

Antibodies against AMA1 are found in people living in malaria endemic areas [81-83], and these have been shown to block the invasion process [84-86]. The potential of AMA1 as a malaria vaccine candidate has been demonstrated in rodent models of malaria with both strain-specific [87-89] and cross-protective [90] protection observed. High levels of polymorphism in AMA1 [8, 91, 92] due to strong balancing selection [93] has resulted in hundreds of distinct AMA1 haplotypes; this might indicate that the development of a broadly effective AMA1-based malaria vaccine will be difficult. However, little is known about the antigenic diversity of AMA1 and recent studies suggest that immunization with a small number of different alleles might give broad reactivity [94, 95]. The availability of a 3D structural model for AMA1 has greatly advanced our understanding of antibody inhibition of AMA1 function by demonstrating that several polymorphisms are found on the edge of a hydrophobic pocket within which it is thought the receptor binds [96, 97]. A cluster of polymorphisms in this region, known as the "C1L cluster" contributes to immune escape [97], but the importance of residues outside this cluster remains unclear [92].

Vaccine candidates comprising full length AMA1 originating from the 3D7 strain alone (FMP2.1, [46] and in combination with FVO (AMA1-c1, [98]), have had variable success in Phase II efficacy trials. FMP2.1 demonstrated no efficacy using the primary endpoint of a clinical malaria episode, however when the secondary endpoint of clinical infections carrying the vaccine allele was defined, a substantial efficacy of 68% above that of the control was revealed [46]. The alternative vaccine candidate, AMA1-C1, showed no allelespecific efficacy however the prevalence of the vaccine allele amongst the samples was so low that sample numbers precluded a proper assessment of any effect [98]. Additionally, the adjuvant used was alum, which is less potent.

who are naturally exposed to malaria [64]. TRAP is highly polymorphic [65] and has also been shown to be under strong diversifying selection suggesting that it is a natural immune

Mapping of inhibitory epitopes has led to the development of the multiepitope (ME)-TRAP DNA vaccine, which contains 15 TRAP T-cell epitopes. The vaccine has been delivered using a prime-boost regime involving three doses of naked DNA or DNA fused to a viral vector, to produce potent T-cell responses in both animals and humans [67]. While safety and immunogenicity trials in the UK have shown promising results [68, 69], Phase II efficacy trials conducted in semi-immune adults in The Gambia and children in Kenya provided no

AMA1 is an 83 kDa integral membrane protein expressed late in the asexual stages and found in the micronemes at the apical tip of merozoites [74]. Prior to schizont rupture the AMA1 prodomain is cleaved to reveal a 66 kDa protein containing three subdomains (domains I, II and III) defined by their disulfide bonds [24]. AMA1 then relocates to the parasite surface [75] where further processing during invasion leads to shedding of two fragments of 44 and 48 kDa whilst the 22kDa cytoplasmic tail remains in the membrane and is carried into the invaded erythrocyte [76, 77]. AMA1 is now known to function in formation of the tight junction through interactions with rhoptry neck proteins [78, 79]. This molecule is also expressed in sporozoites where it is involved in hepatocyte invasion [80].

Antibodies against AMA1 are found in people living in malaria endemic areas [81-83], and these have been shown to block the invasion process [84-86]. The potential of AMA1 as a malaria vaccine candidate has been demonstrated in rodent models of malaria with both strain-specific [87-89] and cross-protective [90] protection observed. High levels of polymorphism in AMA1 [8, 91, 92] due to strong balancing selection [93] has resulted in hundreds of distinct AMA1 haplotypes; this might indicate that the development of a broadly effective AMA1-based malaria vaccine will be difficult. However, little is known about the antigenic diversity of AMA1 and recent studies suggest that immunization with a small number of different alleles might give broad reactivity [94, 95]. The availability of a 3D structural model for AMA1 has greatly advanced our understanding of antibody inhibition of AMA1 function by demonstrating that several polymorphisms are found on the edge of a hydrophobic pocket within which it is thought the receptor binds [96, 97]. A cluster of polymorphisms in this region, known as the "C1L cluster" contributes to immune escape

Vaccine candidates comprising full length AMA1 originating from the 3D7 strain alone (FMP2.1, [46] and in combination with FVO (AMA1-c1, [98]), have had variable success in Phase II efficacy trials. FMP2.1 demonstrated no efficacy using the primary endpoint of a clinical malaria episode, however when the secondary endpoint of clinical infections carrying the vaccine allele was defined, a substantial efficacy of 68% above that of the control was revealed [46]. The alternative vaccine candidate, AMA1-C1, showed no allelespecific efficacy however the prevalence of the vaccine allele amongst the samples was so low that sample numbers precluded a proper assessment of any effect [98]. Additionally, the

[97], but the importance of residues outside this cluster remains unclear [92].

target [66].

evidence of protective efficacy [70-73].

**3.1.3 Apical Membrane Antigen 1 (AMA1)** 

adjuvant used was alum, which is less potent.

Another approach for AMA1 that is being tested is a diversity-covering vaccine (PfAMA1 DiCo) whereby 355 published sequences were aligned and common polymorphisms (i.e. those found in >16% of sequences) were used to design three synthetic constructs, which together encompass 97% of the sequence diversity [99]. So far there have not been any clinical trials however broader parasite growth inhibitory activity was observed in animals vaccinated with the DiCo mix than those vaccinated with native AMA1 from the FVO strain [99].
