**4. Immune effector mechanisms in malaria immunity**

Although both cellular and humoral immunity are thought to be involved in malaria immunity, the relative importance of each in protection against malaria is not yet well established. In particular much of the data on cellular immunity comes from animal models. However, many of these animals are poor models of human malaria. Furthermore, data from different animal species and between different strains of same species often vary considerably making it difficult to generate definitive conclusion regarding immune effector mechanisms in malaria.

#### **4.1 CD8+ T-cells (CTL)**

CD8+ T cells are also referred to as cytotoxic T-cells (CTL) because they can kill infected cells directly by various cytotoxicity mechanisms. Because hepatocytes express class 1 HLA, the receptor for CD8+ T cells, the liver stage of malaria parasites is thought to be capable of inducing CTL responses. The role of CTL in the protection against malaria was first demonstrated in the classical experiments involving the immunization of animals and human with irradiated sporozoites. Such immunization resulted in complete, though short-

The Immunology of Malaria 181

Subsequent to the innate responses that mediate early resistance to malaria infection, the adaptive immunity takes over with CD4+ T-cells becoming the main producers of cytokines. Traditionally mature CD4+ T-cells are placed in two groups that are associated with distinct cytokine profiles. Production of interferon alpha/gamma (INF-α/γ), lymphotoxin-α (TNFβ), interleukin-12 (IL-12) defines type 1 helper cells (Th1) and is associated with a strong cell-mediated immunity while production of IL-4, 5, 6, 9, 10 and 13 define type 2 (Th2) which are associated with antibody production. However, because some T-cells and non-Tcells can produce both Th1 and Th2 cytokines, it may be more appropriate to talk of a type 1 (TR1) or a type 2 response (TR2) (Clerici and Shearer 1994). In malaria, the TR1/TR2 dichotomy is most evident in the mouse-*P. chabaudi* model (Langhorne *et al* 1989, Taylor-Robinson and Phillips 1993). In this model, TR1 dominates the early response of mice to acute *P. chabaudi* infection and parasite killing is mediated by INF-γ, tumour necrosis factor (TNF-α) and nitric oxide (NO) secreted by activated Th1 CD4+, macrophages, and natural killer cells. In *P. berghei* and *P. yoelii* models, TR1 response induced through sporozoites vaccination have been shown to provide strong protection against challenge infections (Oliveira *et al* 2008, Purcell *et al* 2008). On the other hand, a shift towards TR2 leads to less symptomatic chronic infections (Clerici and Shearer 1994). Along with inhibiting both INF-γ and TNF-α, type 2 cytokines also stimulate B-cells to secrete of antibodies (Fell and Smith 1998, Pretolani and Goldman 1997, Taylor-Robinson 1995). The dual anti-parasite/ pathogenetic nature of TR1 is also evident in *P. berghei* infections (Hirunpetcharat *et al* 1999, Rudin *et al* 1997). Other murine-malaria models display variable tendencies towards either type of responses during acute and chronic infections (Taylor-Robinson and Smith 1999).

The distinction between type 1 and 2 responses is less clear in human malaria. Increased IFN-γ is associated with the resolution of parasitaemia in acute malaria episodes (Winkler *et al* 1998) and a delay in re–infection (Luty *et al* 1999) while reduced levels accompany hyperparasitaemia in children (Winkler *et al* 1999). Similarly, levels of type 1 response were lower among Malawian malaria patients than among patients of other disease, with a reverse trend being observed for the type 2 responses (Jason *et al* 2001). INF-γ levels were found to be higher in pregnant women who did not have placental malaria than in those who did (Moore *et al* 1999). These observations argue for a possible anti-parasite role of TR1 in humans. Furthermore CD4+ secreted IL-2 and TNF−α are associated with the protection provided by the experimental vaccine RTS'S (Lumsden *et al* 2011). On the other hand, IL-10 and IL-4, both type 2 cytokines, have been associated with protection against malarial anaemia (Biemba *et al* 2000, Kurtzhal *et al* 1998). Although reduced secretion of INF-γ by immune T-cells in response to malaria led to the conclusion that reduced pathology in immune individuals may be attributable to down-regulation of TR1 cytokines (Chizzolini *et al* 1990), Winkler et al (1999) observed a striking increase in type 1 cytokines in immune adults (Winkler *et al* 1999). It is likely that efficient immunity to malaria requires a balance in

There is now an increasing recognition of the role played by a third population of CD4+ Tcells in malaria immunity (Walther *et al* 2009). This cells, designated regulatory T-cells (Tregs), additionally express CD25 and FOXP3 cellular markers and mediate their actions through immunomodulatory cytokines IL-10 and TGF-β. Studies in mouse models suggest

between TR1 and TR2.

**4.3 Regulatory T cells** 

lived, immunity (Clyde 1975, Nussenzweig *et al* 1967, Rieckmann *et al* 1974). Adoptive transfer of CTL and CTL depletion experiments in animals showed that although high levels of anti-sporozoite antibodies were observed in the immunized subjects, the protection observed was mediated by CTL (Schofield *et al* 1987, Suss *et al* 1988, Weiss *et al* 1988). The fact that adoptively transferred CTL pre-primed with *P. berghei* failed to protect against infection by *P. yoelii* indicates that protection by CTL is species-specific (Romero *et al* 1989). Later studies showed that CTL mediate their protection by preventing parasite development in the liver (Rodrigues 1991). Although CTL can kill parasites by perforin-mediated lysis and FASinduced apoptosis of infected cells (Kagi *et al* 1994, Lowin *et al* 1994), depletion of FasL and perforin did not affect CD8+ mediated-protection against *P yoelii* infection in mice suggesting that the protection is probably cytokine-mediated (Morrot and Zavala 2004). Indeed, The importance of the cytokine pathway in which IFN-γ stimulates the host cell to kill the parasites through nitric oxide (NO) production has been demonstrated in mice (Schofield *et al* 1987).

Indirect evidence for CTL protection against malaria in humans is borne in the association between some class 1 HLA alleles and protection against malaria (Hill *et al* 1991). Over 30 peptides on the sporozoites and liver stage antigens of malaria parasites have now been identified as epitopes for human CTL (Aidoo and Udhayakumar 2000, Aidoo *et al* 1995, Bottius *et al* 1996) Some of these epitopes exhibit extensive polymorphism generated by nonsynonymous mutations, an indication that they are under some sort of selection possibly by host immunity (Hughes and Hughes 1995, Lockyer *et al* 1989, Schofield 1989).

#### **4.2 Cytokine response to malaria – Role of innate immunity and CD4+ T-cells**

Malaria disease is characterised by production of a wide range of cytokines. Studies suggest that these come from both the innate arm and the adaptive arm of the immune system. Because parasites multiply very rapidly, it is likely that the innate arm mediates early cytokines responses against malaria. An early interferon gamma (IFN-γ) response has been shown to be important in protecting against development of severe disease symptoms (Cabantous *et al* 2005, D'Ombrain *et al* 2008, Perlaza *et al* 2011) as has the ability of one's cells to produce TNF-α or INF-γ upon stimulation with live parasites in vitro (Robinson *et al* 2009). Natural killer cells (NK) have been implicated as the source of early proinflammatory responses such as IFN-γ and TNF-α against malaria parasites (Artavanis-Tsakonas and Riley 2002, Korbel *et al* 2005) while other studies point to γδΤ−cells and αβT-cells (D'Ombrain *et al* 2008, Horowitz *et al* 2010). However, the relative contribution of each type of cells is debatable. Activation of innate immunity depends on broad recognition of pathogens. This recognition is driven by receptors that recognise pathogen associated molecular patterns (PAMPs). Among the best characterised of these pattern recognition receptors are Toll-like receptors (TLR), of which ten have been described in man so far. Upon recognition of PAMPs, TRLs induce a signalling cascade leading to secretion of proinflammatory cytokines, chemokines, and interferons. Malaria parasite glycophosphoinostol (GPI) has been shown to interact with TLR2 and to some extent TLR4 (Franklin *et al* 2009, Gowda 2007, Krishnegowda *et al* 2005). While some studies suggest that haemozoin, a product of haemoglobin digestion by malaria parasites interacts with TLR9 (Coban *et al* 2005), other studies suggest that haemozoin is immunologically inert and it's the parasite DNA that it complexes with that interact with TRLs to induce a proinflammatory response (Parroche *et al* 2007). It is also possible that malaria parasites can induce the innate immune system through interaction between other non-TLR receptors and AT-rich parasite DNA fragments (Sharma *et al* 2011).

lived, immunity (Clyde 1975, Nussenzweig *et al* 1967, Rieckmann *et al* 1974). Adoptive transfer of CTL and CTL depletion experiments in animals showed that although high levels of anti-sporozoite antibodies were observed in the immunized subjects, the protection observed was mediated by CTL (Schofield *et al* 1987, Suss *et al* 1988, Weiss *et al* 1988). The fact that adoptively transferred CTL pre-primed with *P. berghei* failed to protect against infection by *P. yoelii* indicates that protection by CTL is species-specific (Romero *et al* 1989). Later studies showed that CTL mediate their protection by preventing parasite development in the liver (Rodrigues 1991). Although CTL can kill parasites by perforin-mediated lysis and FASinduced apoptosis of infected cells (Kagi *et al* 1994, Lowin *et al* 1994), depletion of FasL and perforin did not affect CD8+ mediated-protection against *P yoelii* infection in mice suggesting that the protection is probably cytokine-mediated (Morrot and Zavala 2004). Indeed, The importance of the cytokine pathway in which IFN-γ stimulates the host cell to kill the parasites through nitric oxide (NO) production has been demonstrated in mice (Schofield *et al* 1987).

Indirect evidence for CTL protection against malaria in humans is borne in the association between some class 1 HLA alleles and protection against malaria (Hill *et al* 1991). Over 30 peptides on the sporozoites and liver stage antigens of malaria parasites have now been identified as epitopes for human CTL (Aidoo and Udhayakumar 2000, Aidoo *et al* 1995, Bottius *et al* 1996) Some of these epitopes exhibit extensive polymorphism generated by nonsynonymous mutations, an indication that they are under some sort of selection possibly by

Malaria disease is characterised by production of a wide range of cytokines. Studies suggest that these come from both the innate arm and the adaptive arm of the immune system. Because parasites multiply very rapidly, it is likely that the innate arm mediates early cytokines responses against malaria. An early interferon gamma (IFN-γ) response has been shown to be important in protecting against development of severe disease symptoms (Cabantous *et al* 2005, D'Ombrain *et al* 2008, Perlaza *et al* 2011) as has the ability of one's cells to produce TNF-α or INF-γ upon stimulation with live parasites in vitro (Robinson *et al* 2009). Natural killer cells (NK) have been implicated as the source of early proinflammatory responses such as IFN-γ and TNF-α against malaria parasites (Artavanis-Tsakonas and Riley 2002, Korbel *et al* 2005) while other studies point to γδΤ−cells and αβT-cells (D'Ombrain *et al* 2008, Horowitz *et al* 2010). However, the relative contribution of each type of cells is debatable. Activation of innate immunity depends on broad recognition of pathogens. This recognition is driven by receptors that recognise pathogen associated molecular patterns (PAMPs). Among the best characterised of these pattern recognition receptors are Toll-like receptors (TLR), of which ten have been described in man so far. Upon recognition of PAMPs, TRLs induce a signalling cascade leading to secretion of proinflammatory cytokines, chemokines, and interferons. Malaria parasite glycophosphoinostol (GPI) has been shown to interact with TLR2 and to some extent TLR4 (Franklin *et al* 2009, Gowda 2007, Krishnegowda *et al* 2005). While some studies suggest that haemozoin, a product of haemoglobin digestion by malaria parasites interacts with TLR9 (Coban *et al* 2005), other studies suggest that haemozoin is immunologically inert and it's the parasite DNA that it complexes with that interact with TRLs to induce a proinflammatory response (Parroche *et al* 2007). It is also possible that malaria parasites can induce the innate immune system through interaction between other

host immunity (Hughes and Hughes 1995, Lockyer *et al* 1989, Schofield 1989).

non-TLR receptors and AT-rich parasite DNA fragments (Sharma *et al* 2011).

**4.2 Cytokine response to malaria – Role of innate immunity and CD4+ T-cells** 

Subsequent to the innate responses that mediate early resistance to malaria infection, the adaptive immunity takes over with CD4+ T-cells becoming the main producers of cytokines. Traditionally mature CD4+ T-cells are placed in two groups that are associated with distinct cytokine profiles. Production of interferon alpha/gamma (INF-α/γ), lymphotoxin-α (TNFβ), interleukin-12 (IL-12) defines type 1 helper cells (Th1) and is associated with a strong cell-mediated immunity while production of IL-4, 5, 6, 9, 10 and 13 define type 2 (Th2) which are associated with antibody production. However, because some T-cells and non-Tcells can produce both Th1 and Th2 cytokines, it may be more appropriate to talk of a type 1 (TR1) or a type 2 response (TR2) (Clerici and Shearer 1994). In malaria, the TR1/TR2 dichotomy is most evident in the mouse-*P. chabaudi* model (Langhorne *et al* 1989, Taylor-Robinson and Phillips 1993). In this model, TR1 dominates the early response of mice to acute *P. chabaudi* infection and parasite killing is mediated by INF-γ, tumour necrosis factor (TNF-α) and nitric oxide (NO) secreted by activated Th1 CD4+, macrophages, and natural killer cells. In *P. berghei* and *P. yoelii* models, TR1 response induced through sporozoites vaccination have been shown to provide strong protection against challenge infections (Oliveira *et al* 2008, Purcell *et al* 2008). On the other hand, a shift towards TR2 leads to less symptomatic chronic infections (Clerici and Shearer 1994). Along with inhibiting both INF-γ and TNF-α, type 2 cytokines also stimulate B-cells to secrete of antibodies (Fell and Smith 1998, Pretolani and Goldman 1997, Taylor-Robinson 1995). The dual anti-parasite/ pathogenetic nature of TR1 is also evident in *P. berghei* infections (Hirunpetcharat *et al* 1999, Rudin *et al* 1997). Other murine-malaria models display variable tendencies towards either type of responses during acute and chronic infections (Taylor-Robinson and Smith 1999).

The distinction between type 1 and 2 responses is less clear in human malaria. Increased IFN-γ is associated with the resolution of parasitaemia in acute malaria episodes (Winkler *et al* 1998) and a delay in re–infection (Luty *et al* 1999) while reduced levels accompany hyperparasitaemia in children (Winkler *et al* 1999). Similarly, levels of type 1 response were lower among Malawian malaria patients than among patients of other disease, with a reverse trend being observed for the type 2 responses (Jason *et al* 2001). INF-γ levels were found to be higher in pregnant women who did not have placental malaria than in those who did (Moore *et al* 1999). These observations argue for a possible anti-parasite role of TR1 in humans. Furthermore CD4+ secreted IL-2 and TNF−α are associated with the protection provided by the experimental vaccine RTS'S (Lumsden *et al* 2011). On the other hand, IL-10 and IL-4, both type 2 cytokines, have been associated with protection against malarial anaemia (Biemba *et al* 2000, Kurtzhal *et al* 1998). Although reduced secretion of INF-γ by immune T-cells in response to malaria led to the conclusion that reduced pathology in immune individuals may be attributable to down-regulation of TR1 cytokines (Chizzolini *et al* 1990), Winkler et al (1999) observed a striking increase in type 1 cytokines in immune adults (Winkler *et al* 1999). It is likely that efficient immunity to malaria requires a balance in between TR1 and TR2.

#### **4.3 Regulatory T cells**

There is now an increasing recognition of the role played by a third population of CD4+ Tcells in malaria immunity (Walther *et al* 2009). This cells, designated regulatory T-cells (Tregs), additionally express CD25 and FOXP3 cellular markers and mediate their actions through immunomodulatory cytokines IL-10 and TGF-β. Studies in mouse models suggest

The Immunology of Malaria 183

merozoites invasion of red blood cells (Haynes *et al* 2002, Tham *et al* 2009, Vande Waa *et al* 1984), depress parasite growth (Crompton *et al* 2010, Dent *et al* 2008, McCallum *et al* 2008, Wilson *et al* 2010), and promote parasite phagocytosis by macrophages (Druilhe and Khusmith 1987, Groux *et al* 1990). Furthermore, immune serum can disrupt rosettes (Carlson *et al* 1994, Vigan-Womas *et al* 2010) and the binding of infected erythrocytes to endothelial cell ligands (Iqbal *et al* 1993, Ricke *et al* 2000, Udeinya *et al* 1983), two phenomena which, as

However, it is not clear how well in vitro antibody activities correlate with effector mechanisms in vivo. Malaria literature is replete with reports of lack of a correlation between antimalarial antibody titres measured in vitro and malaria protection (Erunkulu *et al* 1992, Marsh *et al* 1989, Thelu *et al* 1991) This is because the majority of malaria antibodies are probably directed against cellular debris released when schizonts burst and are of little

Fig. 2. Malaria antigens associated with various stages of the parasite that are thought to be

earlier indicated, have been implicated in the pathogenesis of severe malaria.

consequence with regard to protection.

targets for immune responses.

that these cytokines help reduce immunopathology by suppressing proinflammatory cytokines (Nie *et al* 2007) although if induced too early in an infection they may suppress the protective effects of the proinflammatory cytokines and allow the parasite to multiply uncontrollably (Walther *et al* 2005). In humans, TGF−β, which appears to interact with Tregs, is associated with increased risk of clinical disease and a high parasite growth in vivo (Todryk *et al* 2008). As such a fine balance between the symptom-suppressing effects of Tregs and the parasite-suppressing effect of symptom-inducing cytokines is needed for an optimal outcome following malaria infection (Berretta *et al* 2011).

#### **5. Humoral responses in malaria**

#### **5.1 Evidence for involvement in protection against malaria**

There is no doubt that humoral responses are important in protection against malaria. Direct evidence for this comes from passive transfer experiments both in animal models (Groux and Gysin 1990) and humans. In a series of experiments carried out in the early 1960s by Cohen, Macgregor, and Carrington intra-muscular administration of purified IgG from malaria immune African adults into Gambian and East African children suffering from clinical malaria caused a marked drop in parasitaemia within five days. IgG from Europeans without prior exposure to malaria did not show this parasitocidal effect; indicating that the antibodies from Africans were malaria-specific (Cohen *et al* 1961, McGregor 1963). Similar results were obtained in more recent transfer experiments that used African immune serum to treat Thai malaria patients (Druilhe and al 1997, Sabchareon *et al* 1991). In addition, Edozien et al (1962) showed that antibodies that protected against malaria could be obtained from cord blood thus demonstrating that the passively acquired maternal immunity against malaria in infant, is at least in part, antibody–mediated (Edozien *et al* 1962).

Strong, albeit indirect, evidence for the protective efficacy of antimalarial antibodies comes from classical longitudinal studies where a person's history of malaria disease during a follow-up period is assessed for association with levels of antibodies to various malaria antigens at the beginning of the follow-up period. Using this approach a number of antibody responses to various antigens, some of which listed in figure 2, have been shown in to be associated with protection against malaria. Two shortcomings of this approach are the assumptions that all those who did not get an malaria episode during the follow-up are immune and that the levels of antibodies measured at the beginning of the follow-up period last through the period and any failure to see protection reflects lack of protection by the antibodies. (Bejon *et al* 2010, Kinyanjui *et al* 2009). However, the presence of variation of exposure even within limited geographic region (Bejon *et al* 2010) means that some people may fail to get an episode during follow-up simply because they were not exposed rather than because they are immune. In addition, humoral responses to malaria have been found to be short-lived (Kinyanjui *et al* 2003) as such even people with protective levels at the beginning of a follow-up might have non-protective levels by the time they encounters the next infection.

#### **5.2 Mechanisms by which antibodies protect against malaria**

In vitro, antibodies from immune individuals have been shown to inhibit sporozoites invasion of hepatocytes (Dent *et al* 2008, Fidock *et al* 1997, Pasquetto *et al* 1997), prevent

that these cytokines help reduce immunopathology by suppressing proinflammatory cytokines (Nie *et al* 2007) although if induced too early in an infection they may suppress the protective effects of the proinflammatory cytokines and allow the parasite to multiply uncontrollably (Walther *et al* 2005). In humans, TGF−β, which appears to interact with Tregs, is associated with increased risk of clinical disease and a high parasite growth in vivo (Todryk *et al* 2008). As such a fine balance between the symptom-suppressing effects of Tregs and the parasite-suppressing effect of symptom-inducing cytokines is needed for an

There is no doubt that humoral responses are important in protection against malaria. Direct evidence for this comes from passive transfer experiments both in animal models (Groux and Gysin 1990) and humans. In a series of experiments carried out in the early 1960s by Cohen, Macgregor, and Carrington intra-muscular administration of purified IgG from malaria immune African adults into Gambian and East African children suffering from clinical malaria caused a marked drop in parasitaemia within five days. IgG from Europeans without prior exposure to malaria did not show this parasitocidal effect; indicating that the antibodies from Africans were malaria-specific (Cohen *et al* 1961, McGregor 1963). Similar results were obtained in more recent transfer experiments that used African immune serum to treat Thai malaria patients (Druilhe and al 1997, Sabchareon *et al* 1991). In addition, Edozien et al (1962) showed that antibodies that protected against malaria could be obtained from cord blood thus demonstrating that the passively acquired maternal immunity against

Strong, albeit indirect, evidence for the protective efficacy of antimalarial antibodies comes from classical longitudinal studies where a person's history of malaria disease during a follow-up period is assessed for association with levels of antibodies to various malaria antigens at the beginning of the follow-up period. Using this approach a number of antibody responses to various antigens, some of which listed in figure 2, have been shown in to be associated with protection against malaria. Two shortcomings of this approach are the assumptions that all those who did not get an malaria episode during the follow-up are immune and that the levels of antibodies measured at the beginning of the follow-up period last through the period and any failure to see protection reflects lack of protection by the antibodies. (Bejon *et al* 2010, Kinyanjui *et al* 2009). However, the presence of variation of exposure even within limited geographic region (Bejon *et al* 2010) means that some people may fail to get an episode during follow-up simply because they were not exposed rather than because they are immune. In addition, humoral responses to malaria have been found to be short-lived (Kinyanjui *et al* 2003) as such even people with protective levels at the beginning of a follow-up might have non-protective levels by the time they encounters the

In vitro, antibodies from immune individuals have been shown to inhibit sporozoites invasion of hepatocytes (Dent *et al* 2008, Fidock *et al* 1997, Pasquetto *et al* 1997), prevent

optimal outcome following malaria infection (Berretta *et al* 2011).

**5.1 Evidence for involvement in protection against malaria** 

malaria in infant, is at least in part, antibody–mediated (Edozien *et al* 1962).

**5.2 Mechanisms by which antibodies protect against malaria** 

**5. Humoral responses in malaria** 

next infection.

merozoites invasion of red blood cells (Haynes *et al* 2002, Tham *et al* 2009, Vande Waa *et al* 1984), depress parasite growth (Crompton *et al* 2010, Dent *et al* 2008, McCallum *et al* 2008, Wilson *et al* 2010), and promote parasite phagocytosis by macrophages (Druilhe and Khusmith 1987, Groux *et al* 1990). Furthermore, immune serum can disrupt rosettes (Carlson *et al* 1994, Vigan-Womas *et al* 2010) and the binding of infected erythrocytes to endothelial cell ligands (Iqbal *et al* 1993, Ricke *et al* 2000, Udeinya *et al* 1983), two phenomena which, as earlier indicated, have been implicated in the pathogenesis of severe malaria.

However, it is not clear how well in vitro antibody activities correlate with effector mechanisms in vivo. Malaria literature is replete with reports of lack of a correlation between antimalarial antibody titres measured in vitro and malaria protection (Erunkulu *et al* 1992, Marsh *et al* 1989, Thelu *et al* 1991) This is because the majority of malaria antibodies are probably directed against cellular debris released when schizonts burst and are of little consequence with regard to protection.

Fig. 2. Malaria antigens associated with various stages of the parasite that are thought to be targets for immune responses.

The Immunology of Malaria 185

select for particular parasite variants. However, the huge in drop parasitaemia seen in the transfer experiment is more consistent with a parasitocidal rather than the parasitostatic effect implied by ADCI and other antibody-mediated mechanisms cannot be excluded.

Among people living in endemic areas, levels of antibodies to many malaria antigens vary with the seasonality of malaria transmission, often being higher during periods of high malaria transmission than at the end of a low transmission season (Cavanagh *et al* 1998, Giha *et al* 1998, Nebie *et al* 2008). Second, levels of antibodies to malaria antigens often tend to be higher in individuals who also have malaria parasites at the time when their antibodies are measured than in those without parasites (al-Yaman *et al* 1995, Bull *et al* 2002, Kinyanjui *et al* 2004). These phenomena are typically seen in young children, probably because adults typically have much higher antibody levels that take longer to decay appreciably even in the absence of an infection [(Fruh *et al* 1991, Riley *et al* 1993, Taylor *et al* 1998). These observations and those from other longitudinal studies where malaria antibodies fell from relatively high levels to low levels within a few weeks of treatment of a clinical episode (Branch *et al* 1998, Fonjungo *et al* 1999, Fruh *et al* 1991) suggest that antibody responses to many malaria antigens are relatively short-lived. The preponderance of IgG3 subclass, which has a shorter half-life than the other IgG subclasses, might, in part, explain the brevity of antimalarial antibody responses. However, detailed kinetics studies on the decay of antimalarial antibodies suggest that even the other subclasses decline at a rate that is faster than can be explained by normal catabolic decay (Kinyanjui *et al* 2003). This brevity of circulating antibody responses might explain the rapid re-infection seen among

Like other parasites, malaria parasites are not passive partners in the interaction with the host immune system. The immune system exerts a strong selective pressure on malaria parasites. They have therefore over time evolved a number of mechanisms to evade the

Polymorphism is a common feature of many malaria antigens and is generated through recombination during fertilization or clonal antigenic variation (Anders and Smythe 1989, Borst *et al* 1995). The circumsporozoite protein (CSP) (Dame *et al* 1984, Lockyer *et al* 1989) and thrombospondin–related adhesive protein (TRAP) (Robson *et al* 1998) on the surface of sporozoites all have regions of extensive polymorphism as does the major merozoite antigens; merozoite surface proteins (MSP-2 & MSP-2) (Cooper 1993, Felger *et al* 1994), ring stage erythrocyte surface antigen (RESA) (Perlmann *et al* 1984) and the apical membrane

The antigens inserted by mature parasites on to the surface of the host red cells, which include PfEMP1, rifins, and STEVOR, not only exhibit extensive polymorphism between isolates from different patients, they also under go clonal variation so that each new generation of parasites exhibits different variant from the previous one (Bachmann *et al* 2011, Baruch *et al* 1995, Blythe *et al* 2004, Chen *et al* 1998, Cheng *et al* 1998, Niang *et al* 2009). In many instances, there is little immunological cross reactivity between different

**5.4 Longevity of antibody responses to malaria antigens** 

individuals living in endemic areas after malaria treatment.

antigen-1 (AMA-1) (Verra and Hughes 1999).

immune system.

**6. Mechanisms of immune evasion by malaria parasites** 

Under a variety of in vitro situations, malaria antibodies are often ineffective against parasites in the absence of effector cells and may even promote parasite growth (Galamo *et al* 2009, Shi *et al* 1999). Despite exhibiting potent anti-parasitic activity in vivo, the antibodies used in the transfer experiments in Thailand showed no activity in vitro except in presence of monocytes (Bouharoun-Tayoun *et al* 1990, Sabchareon *et al* 1991). Conversely, antibodies that do not protect in vivo were unable to interact with monocytes in vitro (Groux and Gysin 1990). Thus it has been suggested that the ability of antibodies to cooperate with effector cells may be more important than their quantity (Bouharoun Tayoun and Druilhe 1992). It has been noted that humoral responses to malaria show pronounced skewing towards cytophilic antibodies IgG1 and IgG3 subclasses, unlike responses to other pathogens where IgG1 and IgG2 dominate (Ferrante and Rzepczyk 1997). This bias has been reported severally in responses against ring-infected erythrocyte surface antigen (RESA) (Beck *et al* 1995, Dubois *et al* 1993), merozoites surface antigens (MSA1/2) (Rzepczyk *et al* 1997, Taylor *et al* 1995) and schizont antigens (Nguer *et al* 1997, Piper *et al* 1999, Thelu *et al* 1991).

This skew towards cytophilic antibodies, which need to bind to effector cells before they can mediate any action against antigens, could explain the failure of malaria antibodies to exert anti-parasitic activity on their own. In vitro work has shown that while cytophilic antibodies cooperate with monocytes in inhibiting parasites, non-cytophilic subclasses antagonise this cooperation (Bouharoun Tayoun and Druilhe 1992). Data from field studies indicate that young children and non-immune adults have a high proportion of non-cytophilic antibodies (Wahlgren *et al* 1983), while cytophilic antibodies are associated with protection against infection (Aribot *et al* 1996, Ferreira *et al* 1996, Salimonu *et al* 1982) and better prognosis during acute malaria episodes (Sarthou *et al* 1997). Taken together, these data suggests that acquisition of immunity to malaria may involve a shift in responses from non-cytophilic to cytophilic antibodies (Bouharoun Tayoun and Druilhe 1992).

#### **5.3 Antibody dependent cellular inhibition (ADCI)**

An interesting observation in the transfer experiments was the failure of passively transferred antibodies to completely eradicate all the parasites. This may have parallels in the failure of otherwise highly immune individuals to eliminate chronic low-grade infections. One proposal is that the parasites that escaped the transferred immunity comprised "strains" of parasites against which the antibodies lacked specificity. Two arguments against this are that the antibodies from immune African adults are expected to be directed against multiple antigens, which should help overcome restriction by the strainspecificity of responses to individual antigens, and more importantly, the same antibodies were subsequently shown to be effective against the breakthrough parasites. A density dependent mechanism designated antibody dependent cellular inhibition (ADCI), has been proposed to explain the interaction between cytophilic antibody and monocytes (Druilhe and Perignon 1997). This interaction causes the monocytes to release mediators that reversibly inhibit the growth of parasite ring stages. The amount of inhibiting mediators released is proportional to the ratio of merozoites to monocytes, which explains why the drop in parasitaemia following injection of immune IgG was proportional to the initial parasitaemia. Decline of either antibody levels, or numbers of monocytes or merozoites reverses inhibition and the parasite population flares up. A further implication of the hypothesis is that since the inhibiting mediators are non-specific, this mechanism does not

Under a variety of in vitro situations, malaria antibodies are often ineffective against parasites in the absence of effector cells and may even promote parasite growth (Galamo *et al* 2009, Shi *et al* 1999). Despite exhibiting potent anti-parasitic activity in vivo, the antibodies used in the transfer experiments in Thailand showed no activity in vitro except in presence of monocytes (Bouharoun-Tayoun *et al* 1990, Sabchareon *et al* 1991). Conversely, antibodies that do not protect in vivo were unable to interact with monocytes in vitro (Groux and Gysin 1990). Thus it has been suggested that the ability of antibodies to cooperate with effector cells may be more important than their quantity (Bouharoun Tayoun and Druilhe 1992). It has been noted that humoral responses to malaria show pronounced skewing towards cytophilic antibodies IgG1 and IgG3 subclasses, unlike responses to other pathogens where IgG1 and IgG2 dominate (Ferrante and Rzepczyk 1997). This bias has been reported severally in responses against ring-infected erythrocyte surface antigen (RESA) (Beck *et al* 1995, Dubois *et al* 1993), merozoites surface antigens (MSA1/2) (Rzepczyk *et al* 1997, Taylor *et al* 1995) and schizont antigens (Nguer *et al* 1997, Piper *et al* 1999, Thelu *et al*

This skew towards cytophilic antibodies, which need to bind to effector cells before they can mediate any action against antigens, could explain the failure of malaria antibodies to exert anti-parasitic activity on their own. In vitro work has shown that while cytophilic antibodies cooperate with monocytes in inhibiting parasites, non-cytophilic subclasses antagonise this cooperation (Bouharoun Tayoun and Druilhe 1992). Data from field studies indicate that young children and non-immune adults have a high proportion of non-cytophilic antibodies (Wahlgren *et al* 1983), while cytophilic antibodies are associated with protection against infection (Aribot *et al* 1996, Ferreira *et al* 1996, Salimonu *et al* 1982) and better prognosis during acute malaria episodes (Sarthou *et al* 1997). Taken together, these data suggests that acquisition of immunity to malaria may involve a shift in responses from non-cytophilic to

An interesting observation in the transfer experiments was the failure of passively transferred antibodies to completely eradicate all the parasites. This may have parallels in the failure of otherwise highly immune individuals to eliminate chronic low-grade infections. One proposal is that the parasites that escaped the transferred immunity comprised "strains" of parasites against which the antibodies lacked specificity. Two arguments against this are that the antibodies from immune African adults are expected to be directed against multiple antigens, which should help overcome restriction by the strainspecificity of responses to individual antigens, and more importantly, the same antibodies were subsequently shown to be effective against the breakthrough parasites. A density dependent mechanism designated antibody dependent cellular inhibition (ADCI), has been proposed to explain the interaction between cytophilic antibody and monocytes (Druilhe and Perignon 1997). This interaction causes the monocytes to release mediators that reversibly inhibit the growth of parasite ring stages. The amount of inhibiting mediators released is proportional to the ratio of merozoites to monocytes, which explains why the drop in parasitaemia following injection of immune IgG was proportional to the initial parasitaemia. Decline of either antibody levels, or numbers of monocytes or merozoites reverses inhibition and the parasite population flares up. A further implication of the hypothesis is that since the inhibiting mediators are non-specific, this mechanism does not

cytophilic antibodies (Bouharoun Tayoun and Druilhe 1992).

**5.3 Antibody dependent cellular inhibition (ADCI)** 

1991).

select for particular parasite variants. However, the huge in drop parasitaemia seen in the transfer experiment is more consistent with a parasitocidal rather than the parasitostatic effect implied by ADCI and other antibody-mediated mechanisms cannot be excluded.
