Fc-receptors and immunity to malaria: from models to vaccines

SUMMARY

The complexity and number of antigens (Ags) seen during an immune response has hampered the development of malaria vaccines. Antibodies (Abs) play an important role in immunity to malaria and their passive administration is effective at controlling the disease. Abs represent approximately 25% of all proteins undergoing clinical trials, and these ‘smart biologicals’ have undergone a major revival with the realization that Abs lie at the interface between innate and adaptive immunity. At least 18 Abs have FDA approval for clinical use and approximately 150 are in clinical trials, the majority for the treatment of cancer, allograft rejection or autoimmune disease. Despite these triumphs none are in development for malaria, principally because they are perceived as being too expensive for a disease mainly afflicting poor and marginalized populations. Although unlikely, at least in the foreseeable future, that Ab-based prophylaxis will be made available to the millions of people at risk from malaria, they may be incorporated into current vaccine approaches, since Abs act as correlates of protection in studies aimed at defining the best Ags to include in vaccines. Abs may also form the basis for novel vaccination strategies by targeting Ags to appropriate antigen presenting cells. Therefore, to develop the most efficacious vaccines it will be necessary to fully understand which Abs and Fc-receptors (FcRs) are best engaged for a positive outcome.

A ROLE FOR FC-RECEPTOR (FcR) TRANSGENIC MODELS IN UNDERSTANDING MALARIA IMMUNITY

The means by which antibodies (Abs) transfer protective immunity is not completely understood, but inhibition of merozoite invasion of erythrocytes and/or a role for Ab/complement mediated phagocytosis via Fc-receptors (FcRs) are the probable modes, although which of the large family of human FcRs are optimally involved remains unclear (1,2). Most FcRs have evolved as part of the immunoglobulin gene superfamily (IgSF), and important examples include receptors for IgG (FcγRI, FcγRII, FcγRIII, FcRn), IgE (FcεRI), IgA (FcαRI, Fcα/μR, pIgR) and IgM (Fcα/μR) (Figure 1). By contrast, the low affinity IgE receptor (FcεRII, CD23) is a C-type lectin. The IgSF FcRs each possess a unique ligand-binding chain (α-subunit) with a transmembrane region often complexed with signalling chains (β and γ subunits; Figure 1). A notable exception is FcγRIIIB, which is GPI-anchored to the lipid bilayer of neutrophils. Positive or negative effector mechanisms are triggered through immunoreceptor tyrosine-based activation motifs (ITAMs) or immunoreceptor tyrosine-based inhibition motifs (ITIMs), which, when phosphorylated following multimeric ligation, serve as sites either promoting (ITAM) or negatively regulating (ITIM) activation of cytoplasmic proteins into signalling complexes. However, the recent discovery that numerous ITAMs can induce inhibitory signals (designated ITAM_i), particularly when engaged by monovalent ligands, has disrupted this convenient paradigm (3). Furthermore, receptors that were once considered to mediate entirely negative signals via ITIMs have recently been shown to signal tyrosine phosphorylation, an event usually associated with cellular activation (4).

Figure 1.

The human (a) and mouse (b) leukocyte Fc-receptors (adapted from Refs. [2,21]). *Relative affinities of various ligands for each receptor are indicated in decreasing order, starting with the isotype with the highest affinity. Arrowheads and equal signs are used to show the differences in affinity. The Ig domains are colour coded according to their subunit homology. Whereas some receptors signal directly through activatory (green rectangles) or inhibitory motifs (orange rectangles) in their ligand binding α-chain, others depend on membrane association with the Fc common γ-chain to allow signalling through the γ-chain ITAM. Basos, basophils; Eos, eosinophils; Langs, Langerhans cells; Macs, macrophages; Monos, monocytes; Neuts, neutrophils; DC, dendritic cell; FDC, follicular dendritic cell; GPI, glycosylphosphatidylinositol; ITAM, immunoreceptor tyrosine-based activation motif; ITIM, immunoreceptor tyrosine-based inhibitory motif; ND, not determined; NK, natural killer; S-IgA, secretory IgA, E^b, IgE b allotype; h3, human IgG3; m2a, mouse IgG2a. Human FcγRIIIA_V158 has recently been shown to bind monomeric human IgG3 and must therefore be considered a high affinity FcγR in common with FcγRI.

A recent development with potentially major implications for malaria is the finding that innate pentraxins, including serum amyloid P component (SAP) and C-reactive protein (CRP), two acute-phase proteins expressed during malaria infections and that bind malaria parasite molecules, have been found to bind with high affinity to FcγRs (5). Binding of pentraxins to FcγRs (in particular FcγRI) was found to induce FcγR-mediated phagocytosis and cytokine secretion, in particular of TNFα. The co-crystal structure allowed the authors to show that the shared binding site for SAP and IgG results in competition for FcγR binding and the inhibition of immune-complex-mediated phagocytosis by soluble pentraxins. These results establish Ab-like functions for pentraxins in the FcγR pathway, suggest evolutionary overlap between innate and adaptive immunity, and have therapeutic implications for passive transfer and vaccination studies with Abs in malaria. Although polymorphisms at the CRP protein locus have been shown to predispose to numerous clinical conditions, including systemic lupus erythematosus (SLE), associations with malaria severity or susceptibility are yet to be made, although such studies suggest that this will be a very fertile area of future research (6).

The importance of FcR subunits in malaria immunity has been studied in animals with FcR deletions. Although informative, these gene-deficient mouse models may not always mimic the human immune system, due to differences in FcR biology and an apparent lack of true homologues (7). The γ-chain, a subunit common to FcγRI, FcγRIIIA, FcεRI and FcαRI, is required for efficient cell surface expression and signal transduction (Figure 1). Consequently FcRγ^−/− mice are unable to elicit phagocytosis or Ab-dependent cell-mediated cytotoxicity (ADCC) reactions through these receptors. Two studies with rodent malarias in the FcRγ^−/− have proved controversial, with one study showing a crucial role for FcR-mediated Ab-dependent phagocytosis in host resistance to blood-stage Plasmodium berghei XAT infection (8), and another study with P. yoelii concluding that the protective effects of Ab probably arise through FcR-independent mechanisms (9). However, these studies ignore two important possibilities. First, there might be other, as yet unidentified FcRs, involved in the observed response and second, the α-chain of many FcR are known to associate with signalling proteins other than the common γ-chain. With this in mind, it is interesting to note that mouse IgG3-opsonized Cryptococcus neoformans can still be phagocytosed by macrophages from these same FcRγ^−/− mice (10). This effect is probably mediated via an undefined FcR without requiring γ-chain for function, because of the known FcR, only murine FcγRI binds mouse IgG3, as demonstrated by transfection studies (11). More problematic is that FcR γ-chain deficient mice were found to express partially functional FcγRI in more recent mouse knockouts (12,13). It is now known that the α-chain of FcγRI can mediate MHC class II Ag presentation without active γ-chain signalling (14), and that the α-chain can interact with Periplakin to control receptor endocytosis and IgG binding capacity (15).

These potential drawbacks to the rodent FcγRI knockout and clear data showing that in humans the protective effects of Ab are mediated through interaction with FcRs (although the exact FcRs involved remain unknown) led us to investigate the possibility of using human FcR transgenic mice to investigate Ab function with relation to malaria (16). To achieve this goal we cloned Ab variable domains against P. falciparum MSP1₁₉ from Ab repertoire phage display libraries (generated from immune Gambian adults) and linked them to human IgG1, IgG3 and IgA1 (and more recently IgM) constant domains in suitable expression vectors to generate fully human IgG1 recognizing P. falciparum MSP1₁₉. We also provided a solution to the lack of an in vivo model for P. falciparum with which to test the efficacy of these engineered human Abs by using rodent parasites (P. berghei) transgenic for P. falciparum MSP1₁₉ in mice also transgenic for human FcRs. The model allowed us to show that human IgG1 was not only completely protective but was crucially dependent on human FcγRI for this effect, in a manner that was independent of interference with MSP1 processing (16). The availability of this rodent malaria model now provides an alternative to non-human primates for assessing and monitoring P. falciparum MSP1₁₉ based human Abs. By using this double transgenic approach we were able to show a specific Fc dependent mechanism of action in vivo, currently not possible in humans with P. falciparum malaria. Hence, recombinant human Abs engineered as described will be useful in correlating particular epitopes on MSP1₁₉ with protective immunity, as an aid to vaccine design, and will form the bases of effective in vivo assays before clinical trials in humans. For example, we could rapidly screen monoclonal human Abs or sera from vaccine trials to highlight Ags/Abs that are the most promising at inducing human FcγRI mediated protection. This would allow the malaria community to exclude Ags or Ab types that are unlikely to be effective for incorporation into vaccine constructs (17). This is vitally important since the complexity of the parasite proteome, especially the bewildering number of Ags seen during an immune response has so far greatly hampered the discovery of effective vaccines. We have recently shown that the model has applicability to vaccine trials since ten fold diluted sera from Gambians with moderate Ab titres to MSP1₁₉, protected in the rodent model, and validating MSP1₁₉ as a major vaccine candidate (R. McIntosh, P. Corran, E. Riley, R. J. Pleass, unpublished data from one experiment). This also suggested that the model could be used when only very small volumes of sera are available. Therefore, Abs from volunteer vaccinees could be assessed for protection using this mouse model, allowing vaccination regimens to be optimized rapidly before clinical trials in humans. It will be fascinating to determine if the double transgenic model may be usefully applied to other important malaria antigens for which transgenic parasites are already available (e.g. CSP), or are being generated (e.g. AMA1, MSP3 and MSP4). In parallel it will clearly be necessary to generate recombinant human Abs to these antigens. Although P. berghei has been used extensively for such studies, they should also be extended to P. yoelii YM, which unlike P. berghei is not restricted to growth in reticulocytes and is therefore more similar to P. falciparum in this regard.

Numerous interesting questions now arise from this work with FcγRI. For example: (i) Which FcγRI mediated signalling pathway leads to protection from malaria? It is known that the α-chain of FcγRI can mediate protective signalling events either through the common γ-chain or via Periplakin. (ii) Can human IgG3 mediate equal or better protection through FcγRI, and is this in turn mediated via the common γ-chain or Periplakin? We have recently generated human IgG3 versions of these IgG1’s, but have been limited to doing in vitro work, having yet to test them in vivo because of problems generating sufficient quantities of Ab. Mammalian expression systems appear to have difficulty with the extended hinge region of IgG3, which is also susceptible to proteolysis. (iii) Can anti-MSP1₁₉ IgG1 Abs mediate equally effective protection via other human FcγRs? Because there are no known murine equivalents of human FcγRIIA, FcγRIIC and FcγRIIIB, these receptors would be of particular interest to investigate using the transgenic approach (Figure 1). FcγRIIA is fascinating since the α-chain of this human receptor has its own unique ITAM and is therefore capable of signal transduction and phagocytosis in the absence of γ-chain or associated subunits (Figure 1). Naturally acquired anti-malarial Abs from clinically immune individuals are predominantly human IgG1 and IgG3 (18), subclasses that can trigger FcγRIIA (Figure 1). Importantly, polymorphisms in FcγRIIA have recently been implicated in susceptibility to severe malaria (7,19). In contrast with FcγRI, this receptor is expressed by neutrophils as their dominant activating FcγR (7). Since neutrophils are abundant in blood and have been shown to kill merozoites (20), comparing parasite clearance of FcγRIIA with FcγRI by the two IgG subclasses would be important. Although both IgG1 and IgG3 can bind equally well to FcγRI, recent analysis of specificity and affinity interactions of the two human subclasses with their FcγRs implies that human IgG1 is best at engaging FcγRIIA, whereas human IgG3 binds more efficiently to FcγRIIIA (Figure 1) (21). This suggests that anti-malarial IgG3 can more efficiently activate FcγRIIIA-expressing NK cells, monocytes and macrophages than IgG1, and there is certainly some support for the greater efficacy of IgG3 from the malaria literature.

The model has a number of advantages over both primate systems and in vitro studies. For example, a commonly used in vitro assay (inhibition of MSP1 processing) would not have predicted that the human IgG1 we engineered was protective because this in vitro assay does not replicate the important in vivo interaction between the Ab and the FcγRs. This important consideration also holds true for the controversial Ab-dependent cellular inhibition assay (ADCI), where immune Ab is believed to cooperate with human monocytes to inhibit parasite growth through a non-phagocytic mechanism, involving FcγRIIA/FcγRIIIA co-triggered release of unidentified toxins from primed monocytes (22-24). The ADCI assay would not have predicted a role for FcγRI in controlling malaria parasites, and since the mechanism of action is unknown, whether ADCI actually operates in an in vivo setting remains to be determined. It should be noted that FcγRI cannot be excluded from playing a role in ADCI since these assays make use of purified human monocytes whose FcγRI has a very high-affinity for monomeric IgG1, IgG3 and IgG4 (approximately 10⁹ M⁻¹) and therefore would be fully occupied with IgG on entry into the assay (21). Such problems may be partially circumvented by stripping off pre-bound Ab or by using THP-1 cell lines, although the latter do not accurately reflect typical human monocytes since THP-1 cells do not express FcγRIII, even in the presence of cytokines and pharmacologic agents (25). We therefore believe the solution to the current paucity of suitable in vitro assays of protection is to further refine the FcR humanized mouse model.

The P. berghei mouse transgenic combination is a reasonable model for P. falciparum in humans since mouse IgG2a binds with a similar order of affinity as human IgG1 to human FcγRI, indicating that the transgenic receptor in these mice would be fully occupied with mouse IgG2a, and therefore accurately reflecting the situation in humans with IgG1 (26). Although complete protection could be obtained in the human FcγRI transgenic animals with as little as 0·75 mg of Ab, it is likely that the presence of native mouse FcγRs may bind and interfere with the function of much of the injected Ab thereby decreasing its potential efficacy. Interestingly, although human IgG1 is known to bind mouse FcγRs (including FcγRIV) with good affinity, it was unable to protect in the non-transgenic mouse in both the P. berghei and P. yoelii systems (16,27). Possibly, recruitment of mouse FcγRs did occur but preferentially triggered an inhibitory signal rather than an activatory one. A similar lack of effect for human IgG1 in BALB/c mice has been seen with C. neoformans (28), although human IgG1 was shown to be protective in A/J mice (29), which are C5 deficient, suggesting that complement may play a role in protecting the parasite. That C5 deficiency has recently been shown to protect against murine cerebral malaria suggests that Ab immune complexes may also be involved in this process (30). Without further dissection of the molecules involved or the impact of mouse strain background, it is difficult to reconcile the reasons for the lack of protection offered by IgG1 in non-transgenic animals, although our results do confirm (for this epitope at least), that mere blocking of MSP1₁₉ function by some form of steric hindrance is insufficient to bring about protection (16,27), and that endogenous mouse FcγRs are not involved.

In order to refine this model system we are currently crossing the human FcγRI transgenic to mice deficient in FcγRI, FcγRII and FcγRIII (courtesy of Dr Sjef Verbeek, University of Leiden, The Nertherlands). FcγRIV has recently been reported in mice but its function can be blocked with mAb (31). Because the outcome of Ab-mediated immunity in malaria may depend on the interaction of IgG with multiple activating or inhibitory FcγR, developing fully humanized mice expressing all FcγRs on a mouse total FcγR knockout background would be very useful. Although extremely challenging, and arguably technically not feasible, these mice are nonetheless planned for testing HIV vaccines as part of a Bill and Melinda Gates Grand Challenge for Global Health initiative (www.gcgh.org/about/Pages/FcRHumanizedMice.aspx). Problems that can be envisaged in their development include: a requirement to cross mice with human Ab transgenic animals to avoid interference from mouse Abs (the same would apply for other ligands of human FcγRs absent or different in mice, e.g. signal transduction proteins, complement, pentraxins, etc.), issues of copy number and polymorphic variation present within human FcγRs that are absent in mice, and the role of mouse genetic background (1,2,7).

Natural rodent models for FcR mediated immunity to human malaria Ags

Although we have shown that transgenic systems are very useful for investigating FcγR function, particularly where there are no suitable animal models for P. falciparum infection in humans, an argument can be made that they are somewhat contrived, and therefore should be compared in parallel experiments in FcγR knockout models using rodent malarias. Among IgG subclasses, IgG2a and IgG2b are considered to be the most potent activators and dominate in successful passive transfer experiments in rodent malaria systems (32). Such functional distinctions have been attributed to differences in their capacity to fix complement (33), and/or interference with key Ags involved in erythrocyte invasion (34). However, previous studies from our group and experiments in complement deficient mice (C4, C3 or CR1/2 strains) have challenged this assumption for malaria and have focused attention on the FcγRs, as the primary mediators of IgG effector responses (27,35). Intriguingly, mice deficient in the α-chain of FcγRI, and thus with macrophages that cannot bind IgG3, have been shown to be protected from P. yoelii infection following passive transfer of an anti-MSP1₁₉-specific murine IgG3 mAb (36), suggesting that in mice at least, other FcRs may be important in controlling malaria. IgG3 differs from the other mouse IgG subclasses in that it forms aggregates and binds FcγRI but not FcγRII, FcγRIII or FcγRIV (11,31). Because of structural and functional differences between murine and human Abs (in particular with respect to IgG3), and to resolve existing controversies as to whether FcRs are indeed even required for protection from malaria in the mouse, we are generating a panel of recombinant mouse IgG1, IgG2a, IgG2b and IgG3 targeting the identical epitope on P. falciparum MSP1₁₉ to investigate the anti-malarial properties of the mouse subclasses in FcR knockout animals using the P. berghei PfMSP1₁₉ transgenic parasite. These reagents will allow us to determine activatory/inhibitory (A/I) ratios for each of the mouse IgG sublasses in the context of malaria, as has been done recently to predict which subclasses are optimal for clearing tumours (37). The A/I ratios of the mouse IgG subclasses in these experiments differed markedly, with values of 69 for IgG2a, 7 for IgG2b and 0·1 for IgG1. Interestingly, IgG3 showed no detectable binding to the receptors tested, so no A/I ratio was assigned. It may therefore be that the anti-MSP1₁₉ IgG3’s described by Vukovic et al. (36) and by Rotman et al. (9) do not work by FcR engagement but via alternative mechanisms not involving FcγRs. Whether the mAb IgG3 s used in these two studies can interfere with MSP1₁₉ processing was not investigated to determine if this is their mode of action in the absence of FcγR recruitment.

Some interesting predictions can be made for clearance of rodent malaria parasites based on these studies in mice: (i) IgG2a and IgG2b mediated activities will be dependent on FcγRIV and not FcγRI, as they are for humans, (ii) Mouse IgG1 exhibits the greatest (and IgG2a the least) dependence in its in vivo activity on FcγRIIB expression displaying significantly enhanced activity in mice lacking the inhibitory receptor. (iii) Removal or blocking of the inhibitory FcγRIIB may be a novel way to obtain greater therapeutic efficacy for passively transferred Abs. It has recently been shown that FcγRIIB deficient mice have increased clearance of P. chaubaudi malaria, develop less severe disease, and that polymorphic variants of human FcγRIIB resulting in loss of function are common in African individuals, who also show enhanced phagocytosis of parasites (38).

Do rodent malaria models translate to human studies?

How might these rodent studies translate to the human system with P. falciparum? Could we select IgG subclasses for optimal in vivo effect? We do not know, but important differences between rodents and humans may predict difficulties. First, mouse and human IgG subclasses are not directly comparable, for example, human IgG3 has an elongated hinge region, which may account for some of its enhanced biological activity. Second, in contrast to mouse Abs, human IgG2 can form covalent dimers in vivo (39), and the IgG4 subclass can exchange Fab arms in vivo, generating bispecific molecules that, in a therapeutic context, may have unanticipated effects (40). Third, human IgG1 (4 allotypes), IgG2 (5) and IgG3 (13 allotypes) subclasses are extremely polymorphic, and these allotypic differences are known to impact significantly on FcγR interactions (41,42). A recent paper comparing all four human IgG subclasses epitope matched for C. neoformans gives further evidence that there are no hard and fast rules (28). Contrary to expectations that cytophilic subclasses are better at controlling disseminated fungi, human IgG2 or IgG4 were shown to be optimal, albeit in mice, whereas human IgG1 or IgG3 were detrimental. Clearly it will be necessary to undertake similar experiments for malaria. The availability of mice transgenic for human FcRs together with rodent malarias transgenic for P. falciparum Ags will facilitate this process. Fourth, polymorphisms of the immunoglobulin genes are further complicated by polymorphisms in the FcRs (43) and complement components (44) with which they interact, and so it is not yet possible to make clear statements about the role of any given Ab. The H131R alleles of FcγRIIA are differentially distributed in Caucasians and Asians and have an impact on malaria susceptibility (7,19). Two alleles of the gene encoding FcγRIIIB generate two variants differing at four positions, NA1 (R³⁶, N⁶⁵, D⁸², V¹⁰⁶) and NS2 (S³⁶, S⁶⁵, N⁸², I¹⁰⁶) and an additional mutation in the NS2 allele (A78D) generates a further variant named SH (see Figure 1), although these have not been investigated thoroughly in relation to malaria. Furthermore, copy number variation in FcγRIIIB may allow a single individual to express all three FcγRIIIB variants complicating the situation for malaria still further. It has recently been shown that FcγRIIB deficient mice have increased clearance of P. chabaudi malaria and develop less severe disease, and that polymorphic variants of human FcγRIIB which result in loss of function are common in African individuals who also show enhanced phagocytosis of parasites (38). Finally, mouse FcγRIV can bind mouse IgE of the ‘b’ allotype as well as IgG2a and IgG2b, and ligation of FcγRIV by antigen-IgE^b immune complexes promoted macrophage-mediated phagocytosis, presentation of antigen to T cells, production of proinflammatory cytokines and the late phase of cutaneous allergic reaction (45). Intriguingly, murine IgE has been shown to confer protection from P. berghei in C57BL/6 mice, animals that produce Abs of the IgE^b allotype (46). Therefore murine IgE may influence in rodent malaria models with strain-dependent differences in parasite susceptibility. Thus, although there are significant differences in biology between the mouse and human receptor systems, we believe that by examining both transgenic and knockout models the principles that have emerged from these mouse studies are likely to apply to human Abs as well as their respective FcRs. Such considerations may prove important in the design of Ab-based therapies and active vaccination protocols for the treatment of malaria.

Unexplored FcRs and Ab classes in malaria

The role of IgA in human malaria research has largely been neglected principally because experimentation has been driven by murine systems. Mouse IgA is dimeric, and there is no murine equivalent of the 1–5 mg/mL monomeric IgA present in human plasma (1,2,7). In addition, mice do not possess an orthologue of FcαR (CD89), which in humans is constitutively expressed on blood neutrophils, and has been shown to be very effective at eliminating human blood pathogens (1,7). In humans, high titres of naturally occurring Plasmodium-specific IgA has been reported in serum (47) and breast milk (48), although we have never been able to detect MSP1₁₉-specific IgA in plasma, despite accounting for competition for binding from IgG (McIntosh R., Lazarou M., Corran P., Holder A., Riley E. and Pleass R.J., unpublished multiple ELISA observations from both groups). To address the role of human IgA in malaria, we generated recombinant versions recognizing MSP1₁₉ that were epitope matched for the human IgGs shown to be effective in the FcγRI transgenic model (16). Although these human IgAs bound with equal affinity to MSP1₁₉ as the IgGs, and were fully functional in in vitro assays with human neutrophils, they were unable to control parasitemias in human FcαRI (CD89) transgenic mice (McIntosh R., Lazarou M., Corran P., Holder A., Riley E. and Pleass R.J., unpublished observations from a single experiment). Recombinant human IgA1 directed against P. yoelii MSP1₁₉ was also without effect in normal wild-type mice (27).

A receptor for IgM (and IgA), the Fcα/μR that is closely related to the polymeric Ig receptor (pIgR) in its ligand-binding domain has been identified (49). The human Fcα/μR is expressed on a sub-population of pre-germinal B cells (IgD⁺/CD38⁺) and on follicular dendritic cells in secondary lymphoid organs (50), where it is believed to play a role in trapping IgM or IgA ICs and in presentation of intact Ags to B cells in germinal centres. We recently localized the amino acid residues within the Cμ4 domain of polymeric IgM that mediate PfEMP1 binding by rosetting parasites via the DBL4β domain, including two loops predicted to lie on the surface of the Cμ4 domain (51). Ongoing work in our laboratories has shown that mAbs that inhibit binding to PfEMP1 also prevented binding of IgM to the Fcα/μR (51,52), suggesting that the binding site for the two ligands lie close to each other on IgM. However, the interaction with Fcα/μR does require additional and unique contacts, since the domain swap Ab γ/L309C-Cμ4 did bind to PfEMP1, but did not bind Fcα/μR; and mouse IgM, shown not to bind PfEMP was capable of binding Fcα/μR (51,52). Furthermore, the extracellular portion of the pIgR (free secretory component; SC) previously shown to bind the Fc of IgM failed to prevent IgM binding to PfEMP1. However, free SC did compete out binding of IgM to Fcα/μR, supporting our notion that unique contacts are involved for binding of IgM to either Fcα/μR or the pIgR (51,52). Intriguingly, the PfEMP1 binding site on IgM is conserved with those on IgA and IgG bound by numerous viruses and bacteria (53,54). Given the overlapping nature of the binding sites, it could be argued that this region is important for IgM function, and that it is beneficial for the infected erythrocyte to block it. More work will be required to determine if this is indeed the case for Fcα/μR.

Of potentially greater relevance to malaria is the finding that CD300LG (also known as CLM9, TREM4 or NEP-MUCIN) shares 35% identity with the Fcα/μR, and is expressed on capillary endothelium particularly in placental tissues (55). Although CD300LG is a functional receptor for l-selectin mediated lymphocyte rolling (56), it has also been shown to bind IgM and IgA (55). Future experiments will address the possibility that IgM bound to PfEMP1 can cross-bridge with functionally important host receptors, such as Fcα/μR and CD300LG. Intriguingly, co-ligation of ICAM1 (CD54) and membrane IgM has been shown to negatively regulate B cell receptor signalling (57). Whether DBL domains of PfEMP1 can interact directly with the Cμ4 domain of the IgM B-cell receptor (BCR) to activate polyclonal B cell activation, as has been shown with the CIDR1α domain remains to be investigated (58).

Given these considerations, it is surprising that so little work has been done looking at the role of IgM in malaria. Only 21 published articles were found in Pubmed having IgM and Plasmodium in the title! Parasite-specific IgM can limit parasite replication, prime memory cell generation, and is a more potent adjuvant than Bordetella pertussis in murine malarias (59,60). Natural IgM is also an endogenous adjuvant for vaccine-induced protective CD8⁺ T cell responses against intracellular parasites (61). As a consequence, we are also generating MSP1₁₉-specific human IgM epitope matched for the earlier IgGs (16). Because these IgM’s have the advantage of increased binding strength through increased avidity it will be fascinating to determine from a biophysical standpoint, if they can more efficiently interfere with MSP1 processing and erythrocyte invasion, than the IgGs from which they were generated. It is known that MSP1₁₉ specific IgG mAbs that do not interfere with processing can be carried into the red cell on invasion (62). Whether the larger size of IgM (970 vs. 150 kDa for IgG) precludes such Trojan horse activity will be interesting to test for P. falciparum.

Targeting FcRs in vaccination

Abs can act like adjuvants by concentrating antigen at sites where lymphocytes are exposed to it, the ‘depot effect’ (63,64). They can potently enhance Ab and T-cell responses to the very antigen they are specific for. The adjuvant properties of Abs have been reviewed at great length (63), and it is generally believed that the aggregation of Ab by Ag into immune-complexes (ICs) potentiates this interaction with professional antigen presenting cells (APCs) expressing FcRs. We have recently reviewed the role played by ICs in malaria (65). FcγRs trigger activatory and/or inhibitory signalling pathways that set thresholds for cell activation and culminate in a well-balanced immune response (66). Activating and inhibitory FcRs are widely expressed throughout the haematopoetic system but particularly on professional APCs (66). For example in humans, FcγRI is constitutively expressed by blood myeloid dendritic cells (DCs) and FcγRII has been detected on every DC subset examined to date, whereas the expression of FcγRI, FcγRIIB and FcγRIII dominate on murine DCs (43,67). FcγRs also play a pre-eminent role in antigen presentation and IC mediated maturation of DCs, and in regulation of B-cell activation and plasma cell survival (66). Moreover, by regulating DC activity, FcγRs control whether an immunogenic or tolerogenic response is initiated after the recognition of antigenic peptides presented on the surface of DCs to cytotoxic T cells, T helper cells, and regulatory T cells. FcγRs also co-operate with Toll-like receptors (TLRs) in controlling levels of the important regulatory cytokines, IL-12 and IL-10 (68). Thus, FcγRs are involved in regulating innate and adaptive immune responses, which makes them attractive targets for the development of novel vaccination approaches (1,2,7). Most FcγRs (with the exception of FcγRI) can only interact with high affinity to Abs presented to the immune system as ICs. Several studies have shown that ICs are potent activators of DCs and can prime stronger immune responses than antigen alone (69-72). Importantly, FcγR-dependent IC internalization not only results in MHC-class-II-restricted antigen presentation but also in cross-presentation on MHC class I molecules, thereby priming both CD4⁺ and CD8⁺ T-cell responses (70). The magnitude of this response is controlled by the inhibitory FcγRIIB, as DCs derived from FcγRIIB-knockout mice generate stronger and longer-lasting immune responses in vitro and in vivo (73,74). More importantly, FcγRIIB-deficient DCs or DCs incubated with a mAb that blocks IC binding to FcγRIIB showed a spontaneous maturation (75,76). This suggests that the inhibitory FcγRIIB not only regulates the magnitude of cell activation but also actively prevents spontaneous DC maturation under non-inflammatory steady-state conditions. Indeed, low levels of ICs can be seen in the serum of healthy donors, emphasizing the importance of regulatory mechanisms that prevent unwanted DC activation (76).

The structurally unrelated neonatal Fc receptor for IgG (FcRn) is also found on professional APCs, including monocyte-derived DCs (77), and plays a role in mediating the presentation of IC derived Ags to LAMP1 positive lysosomes where antigen processing and loading take place (78). Importantly, multimeric ICs that are disabled in their ability to bind FcRn and monomeric IgG-Ag complexes were without effect. These differences were crucially dependent on FcRn function and were observed in the presence of the other FcγRs (78). These findings would suggest that delivering malaria Ags to DCs in the form of artificial ICs or via DC specific mAbs would hold great promise for novel vaccination strategies, especially if they can be engineered for high affinity binding to the activating FcγRs. FcγRI also represents the only FcγR with a well-documented capacity to facilitate immunological memory in vivo and is constitutively expressed by professional APCs including monocytes, macrophages, mast cells, and blood myeloid DCs (79,80). This will be crucial in malaria where the inability to induce or maintain long-term memory responses is likely to pose problems for developing effective vaccines (81). Such approaches have shown promise for model Ags (82). A recent method for targeting Pneumocystis carinii β-glucan with a Dectin-Fc fusion protein enhanced host recognition and clearance of P. carinii via FcγRs (83). Similar approaches may be possible for malaria, for example by fusing pattern recognition molecules like scavenger receptors (e.g. CD36 or CD31) to the Fc, for optimal clearance of P. falciparum infected erythrocytes. Given the ability of pentraxin to bind human FcγRs (5), fusions of malaria antigens directly to CRP or SAP may prove interesting alternatives to Abs as Ag delivery vehicles. Recent computational screens have predicted human proteins that may interact directly with PfEMP1, suggesting that these host proteins may also be partnered with Fc for the targeted removal of infected erythrocytes (84).

The ability of Abs to trigger innate effectors may make them potentially dangerous molecules. For example, over activation may lead to a pro-inflammatory cytokine storm with life-threatening complications. How are such possibilities evaluated for Fc based approaches? As described above, in vitro systems fail to recapitulate the diversity and specificity of Fc-FcγR interactions. Animal models, be they rodent or non-human primate models, are inadequate, as exemplified by the recent disastrous Northwick Park clinical trial with a CD28-specific super-agonist (85). In the absence of animal models with fully human immune systems, we believe FcγR humanized mice to be feasible alternatives for use as preclinical, in vivo platforms for the evaluation of Ab-based vaccines for malaria.