VAR2CSA and protective immunity against pregnancy-associated Plasmodium falciparum malaria

SUMMARY People living in areas with stable transmission of P. falciparum parasites acquire protective immunity to malaria over a number of years and following multiple disease episodes. Immunity acquired this way is mediated by IgG with specificity for parasite-encoded, clonally variant surface antigens (VSA) on the surface of infected erythrocytes (IEs). However, women in endemic areas become susceptible to P. falciparum infection when they become pregnant, particularly for the first time, regardless of previously acquired protective immunity. This conundrum was resolved when it was observed that the selective placental accumulation of IEs that characterizes pregnancy-associated malaria (PAM) is caused by an immunologically and functionally unique subset of VSA (VSAPAM) that is only expressed by parasites infecting pregnant women, and that protective immunity to PAM is mediated by IgG with specificity for VSAPAM. In this review we summarize the research leading to the identification of the distinctly structured PfEMP1 variant VAR2CSA as the dominant PAM-type VSA and as the clinically most important target of the protective immune response to placental P. falciparum infection.

Plasmodium falciparum malaria is both the commonest and most serious form of malaria affecting humans, and is an infection of colossal medical and economic consequence (Greenwood and Mutabingwa, 2002 ;Miller et al. 2002 ;Sachs and Malaney, 2002). In areas of intense parasite transmission, severe P. falciparum malaria is concentrated among infants and young children, because acquisition of protective immunity causes malaria-related mortality and severe morbidity to decrease with increasing age. However, development of immunity is slow, and substantial protection is only achieved after a number of disease episodes. Even then, immunity is partial and though severe morbidity is rare among adults, sterile protection is probably never achieved. Several studies have documented the efficacy of passive immunization of malaria patients with IgG from clinically immune adults, showing that antibodies targeting asexual bloodstage antigens are an important component of protective immunity acquired following natural exposure to P. falciparum parasites (Cohen, McGregor and Carrington, 1961 ;McGregor, Carrington and Cohen, 1963 ;Sabchareon et al. 1991). Transplacental transfer of IgG also seems to be responsible for resistance to malaria for the first months of life in infants born to clinically immune mothers (Bruce-Chwatt, 1952). Antibodies with specificity for a number of antigens expressed by asexual blood-stage parasites are likely to contribute to this protection. However, parasite-encoded, clonally variant surface antigens (VSA) expressed on the surface of the infected erythrocytes (IEs) appear to be of particular importance (Marsh et al. 1989), and the intra-and inter-clonal diversity of VSA go a long way towards explaining the sluggish acquisition of immunological protection (reviewed by Hviid, 2005). The general rule of P. falciparum malaria as a childhood disease in areas of intense parasite transmission has one important exception that has attracted substantial attention in recent years : pregnancy-associated malaria (PAM).
It is a long-recognized fact that pregnant women are at increased risk of malaria, with adverse maternal and foetal consequences (reviewed by Duffy and Desowitz, 2001). This pregnancy-related susceptibility is apparent despite previously acquired protective immunity, and even in malaria-endemic areas, P. falciparum infection is therefore both more prevalent and more severe in pregnant women than in their non-pregnant peers (Walton, 1949). It has often been speculated that PAM is an unavoidable consequence of maternal immuno-suppression or -modulation to protect the foetal allograft from rejection (Menendez, 1995). However, pregnancyinduced immune modulation mainly affects the cellular arm of immunity, whereas humoral immunity (upon which protection from malaria hinges) is largely unaffected (reviewed by Guilbert, Abbasi and Mosmann, 2001). Furthermore, susceptibility to PAM is concentrated among women of low parity, in particular primigravidae (McGregor, 1984). This suggests that susceptibility to PAM is due to a specific absence of immunity to a particular form or subset of P. falciparum parasites that can only infect pregnant women, and that protection against PAM-specific parasites is acquired following exposure to them during pregnancy in a manner similar to acquisition of protection from malaria in general. The hypothesis of a pregnancy-specific subset of P. falciparum is further supported by the facts that PAM is characterized by a selective accumulation of IEs in the placenta (Blacklock and Gordon, 1925) and that pregnancy-associated parasitaemia generally resolves spontaneously shortly after expulsion of the placenta at delivery (Nguyen-Dinh et al. 1988).

PAM is caused by functionally and immunologically unique parasites
Erythrocytes infected by mature (trophozoite and schizont) stages of P. falciparum parasites can bind to a range of different receptors in the host vasculature. The first adhesion receptor identified was CD36 (Barnwell, Ockenhouse and Knowles, 1985 ;Ockenhouse et al. 1989), followed by a range of other molecules, including the proteoglycan chondroitin sulphate A (CSA) ( Robert et al. 1995 ;Rogerson et al. 1995). The first direct piece of evidence in favour of the hypothesis of a discrete parasite subset being responsible for PAM was the finding that IEs isolated from the placenta of women with PAM exclusively bind to CSA, a receptor that is rarely -if ever -exploited as an adhesion receptor by P. falciparum-IEs in non-PAM infections (Fried and Duffy, 1996). This key observation was followed by the similarly important observation that placental and CSA-adhering IEs are not only functionally but also immunologically distinct from all other IEs (Fried et al. 1998 ;Beeson et al. 1999 ;Ricke et al. 2000). As adhesion of IEs in general is mediated by VSA, these findings together pointed to a functionally and immunologically distinct subset of VSA that is exclusively expressed by placental and CSAadhering parasites ; a subset often referred to as VSA PAM .
Protective immunity to PAM is mediated by IgG with specificity for PAM-specific VSA It is generally suspected that protective immunity to malaria depends on VSA-specific IgG that can interfere with receptor-specific IE adhesion (David et al. 1983 ;Udeinya et al. 1983), and that the piecemeal acquisition of protective immunity to malaria reflects the need to acquire a broad repertoire of such antibodies with specificity for a multitude of antigenically distinct VSA (Marsh and Howard, 1986 ;Bull et al. 1998). The importance of VSA PAM in protective immunity against PAM was therefore reinforced when it was shown that serum antibodies can inhibit adhesion of IEs to CSA in a paritydependent manner (Fried et al. 1998 ;Ricke et al. 2000). In addition to inhibition of IE adhesion, opsonization and phagocytosis of IEs may also play an important role in protective immunity to PAM. The dominance of cytophilic subclasses among VSA PAM -specific IgG Megnekou et al. 2005) and the finding of impaired opsonization of VSA PAM -expressing IEs in HIVinfected women (Mount et al. 2004 ;Keen et al. 2007) support this possibility. In any case, levels of VSA PAM -specific IgG at delivery correlate inversely with placental parasitaemia (Staalsoe et al. 2001), and direct and compelling evidence of the clinical importance of CSA adhesion-inhibitory as well as VSA PAM -specific IgG have recently become available (Duffy and Fried, 2003 ;Staalsoe et al. 2004).

Molecular identification of VSA PAM
Most efforts to identify VSA PAM in molecular terms have focused on the best-described family of VSA, P. falciparum erythrocyte membrane protein 1 (PfEMP1) (Leech et al. 1984). These high-molecular weight antigens are encoded by the var multi-gene family (Baruch et al. 1995 ;Smith et al. 1995 ;Su et al. 1995). Scherf et al. (1998) observed transcription of a particular var gene, dubbed var CSA , in P. falciparum FCR3 selected for IE adhesion to CSA, and the Duffy-binding-like (DBL) 3-c domain of the encoded PfEMP1 variant was found to have affinity for CSA (Buffet et al. 1999). Most P. falciparum genomes contain a var gene very similar to var CSA , and these related genes were grouped together in the var1csa sub-family of var genes (Rowe et al. 2002 ;Salanti et al. 2002). These findings, and the observation that var1csa is often highly transcribed by placental parasites (Fried and Duffy, 2002), fitted the earlier prediction (Fried et al. 1998) that the antigen mediating IE adhesion to CSA, and the suspected target of PAM-specific immunity, was conserved between parasite clones. However, a number of other findings do not support the hypothesis of VAR1CSA as a PAM-type VSA, or its involvement as a target of protective immunity against PAM. Thus, high transcription of var1csa is not restricted to placental and CSA-adhering parasites (Rowe et al. 2002 ;Kyes et al. 2003), and levels of VAR1CSA-specific IgG do not correlate with gender or parity, in contrast to expectation regarding VSA PAM -specific IgG . Finally, the ability of IEs to adhere to CSA is maintained after disruption of var1csa (Andrews et al. 2003).

An unusually structured gene, var2csa, is selectively transcribed by CSA-adhering and placental parasites
The availability of the complete P. falciparum genome (Gardner et al. 2002) opened the possibility of more precise analysis of var gene transcription than had previously been possible with degenerate primers (Taylor et al. 2000). Taking this approach, Salanti et al. (2003) analyzed changes in var gene transcription in response to selection of NF54-IEs for adhesion to CSA, which resulted in acquisition of the VSA PAM -type gender-specific and paritydependent IgG recognition pattern of selected IEs (Ricke et al. 2000). The striking result was that the selection resulted in markedly increased and dominant transcription of a single, distinctly structured, var gene, PFL0030c . PFE1640w, which belongs to the var1csa sub-family previously implicated in adhesion of IE to CSA, is a pseudo-gene in NF54/3D7, but data from other parasite lines have shown that selection for IE adhesion to CSA does not increase the abundance of var1csa transcripts . Genes with high similarity to PFL0030c, and now grouped in the var2csa sub-family, appear to be present in all P. falciparum clones and to be selectively transcribed by CSA-selected and placental parasites alike Duffy et al. 2005 ;Tuikue Ndam et al. 2005). The findings that several VAR2CSA domains have affinity for CSA  and that disruption of var2csa interferes with the ability to acquire the CSA-adhering phenotype further supports the PAM-relevance of this sub-family (Viebig et al. 2005 ;Duffy et al. 2006). It is noteworthy that the var2csa response to CSA-selection had gone unnoticed in earlier studies because they used degenerate primers targeting DBL1-a or DBL1-c encoding sequences (Taylor et al. 2000 ;Fried and Duffy, 2002) ; domains that are not present in VAR2CSA.

VAR2CSA has the characteristics expected of VSA PAM
The distinct structure of var2csa compared to other var genes corresponded well with the expected functional and antigenic uniqueness of the PAMtype VSA it was assumed to encode Kraemer and Smith, 2003). Indeed, analysis of levels of VAR2CSA-specific IgG in P. falciparumexposed adults confirmed the female-restricted and parity-dependent pattern expected of VSA PAM (Salanti et al. 2004). Importantly, VAR2CSA is present on the surface of intact VSA PAM -type IEs and absent from the surface of IEs that do not have this phenotype (Salanti et al. 2004 ;Barfod et al. 2007). Finally, we have shown that high plasma levels of VAR2CSA-specific IgG correlate with protection from adverse clinical consequences of PAM (Salanti et al. 2004).
VAR2CSA appears to be the dominant target of the protective immune response to PAM The importance of VSA-specific IgG relative to IgG with other antigen specificities for clinical protection against P. falciparum malaria is unknown, but may be high (Marsh et al. 1989). A parallel issue is the importance of IgG with specificity for VAR2CSA relative to other VSA PAM specificities in protection against PAM. To address this latter question, we cloned memory B cells from recently pregnant, P. falciparum-exposed multigravidae, using a recently developed and highly efficient immortalization method (Traggiai et al. 2004). The clones were subsequently screened for production of VSA PAMspecific monoclonal IgG using a panel of IEs displaying the VSA PAM phenotype. All except one of the eight VSA PAM -specific monoclonals had specificity for either the DBL3-X or the DBL5-e domain of VAR2CSA (Barfod et al. 2007). The characteristics of the remaining antibody suggest that it is also VAR2CSA-specific. Although var2csa is composed of alternating stretches of substantial and restricted interclonal diversity Dahlbäck et al. 2006 ;Trimnell et al. 2006), all the human monoclonal IgG antibodies targeted polymorphic rather than conserved epitopes (Barfod et al. 2007). This finding suggests that domains of IE surface-expressed VAR2CSA that are accessible to protective antibodies are under selection pressure that favours polymorphism, in agreement with conclusions drawn from in silico analysis of var2csa (Dahlbäck et al. 2006 ;Trimnell et al. 2006). This pressure is most likely due to protective host immunity, supporting the clinical significance of the VAR2CSA-specific IgG response to PAM.
C O N C L U D I N G R E M A R K S VAR1CSA (as well as several non-PfEMP1 antigens) continues to be implicated in the pathogenesis as well as in the protective immune response to pregnancy-specific P. falciparum infection (Chia et al. 2005 ;Badaut et al. 2007), and is being explored as a candidate for development of vaccines against PAM (Gamain et al. 2004 ;Chia et al. 2005 ;Bir et al. 2006). However, from our perspective the bulk of current evidence identifies VAR2CSA as the main antigen mediating placental sequestration of P. falciparum-IEs and as the dominant and clinically most relevant target of the human IgG response to placental P. falciparum infection.