Main

Despite control efforts, malaria continues to kill 2–3 million people each year, 75% of whom are children, and the latest estimates suggest that malaria still kills more people than HIV1. The important human malaria parasite Plasmodium falciparum has developed resistance to cheap and previously reliable drugs such as chloroquine, and few new drugs are in development. Although a vaccine against malaria is the ultimate goal, its development has been beset with problems, and the search for alternative therapies has now begun to assume greater importance.

The recent publication of the P. falciparum genome sequence has given researchers the ability to begin to analyse which molecules are important targets of protective immunity2. Immunoglobulins (Igs) have an important role in immunity to malaria, and the passive administration of human Igs from hyperimmune sera can be effective at controlling the disease3,4. However, the difficulty in finding immune donors and the risks associated with the use of human blood products have necessitated the introduction of recombinant techniques for the production of antibodies against important parasite molecules for use in prototypic therapeutic and vaccine studies.

As the direction of malaria research begins to shift towards questions of protein structure, function and interaction — in particular at the host–parasite interface — antibodies can also provide important insights into the structure and function of key antigens, for example by identifying epitopes that are best suited for incorporation into vaccines5. Importantly, because there are no good animal models for P. falciparum malaria, antibodies can offer an inexhaustible supply of reproducible standards for use in neutralization assays to provide in vitro correlates of protection for the development of vaccines. Until now, work on human malaria parasites has often used antibodies purified from immune sera in these neutralization tests. However, sera contain a mixture of antibodies, some with inappropriate specificities such as BLOCKING ANTIBODIES5,6, or can contain ineffective antibody classes with the potential to trigger ineffectual killing responses. These drawbacks can be averted with recombinant engineered human MONOCLONAL ANTIBODIES (mAbs), which are homogeneous with respect to class and target epitope.

Antibodies and effector functions

There are five classes of human Igs, divided into nine subclasses (IgM, IgG1, IgG2, IgG3, IgG4, IgA1, IgA2, IgE and IgD) and all produced by B cells. Each antibody is composed of two pairs of identical heavy (μ, γ1, γ2, γ3, γ4, α1, α2, ε or δ) and light (κ or λ) chains, both containing VARIABLE DOMAINS and CONSTANT DOMAINS (Fig. 1). The variable domains contain three hypervariable or COMPLEMENTARITY-DETERMINING REGIONS, which interact with antigen. The heavy and light chains are held together by covalent and non-covalent bonds, forming a molecule with bilateral symmetry (Fig. 1). The two functional FAB DOMAINS have identical antigen specificity. The primary function of an antibody is to bind antigen, which in malaria can have a direct effect, for example by neutralizing malaria toxin or by preventing the parasite from invading erythrocytes. However, binding to antigen can also be ineffective unless secondary effector functions are activated. This is achieved through the FC REGION, which recruits effector function by binding FC RECEPTORS (FcRs), and/or complement.

Figure 1: Antibody formats.
figure 1

Schematic representation of an intact antibody (a); a Fab fragment (b); a single chain variable format (scFv) molecule (c), which comprises two variable regions and which is the smallest antigen-recognition unit that can be produced in eukaryotic or prokaryotic expression systems; and a bi-specific F(ab′)2 heterodimer (d). Both the scFv and F(ab′)2 formats can be used to generate recombinant therapeutic molecules. C, constant; H, heavy; L, light; V, variable.

There are FcRs for all classes of human antibodies. These receptors are widely distributed on cells of the immune system and, by binding to the Fc region, can couple the humoral immune response and cellular immune response7,8. FcRs can trigger or inhibit various cellular processes, including phagocytosis, antioxidant production, ANTIBODY-DEPENDENT CELLULAR CYTOTOXICITY and the release of cytokines, chemokines and other inflammatory mediators7. Most FcRs are structurally related and comprise a unique ligand-binding chain (α-chain), which is often complexed in the transmembrane region with a dimer of the FcR γ-chain, which is common to many FcRs. Effector mechanisms are signalled through immunoreceptor tyrosine-based activation or inhibitory motifs present in the cytoplasmic regions of either the α-chain or γ-chain.

In addition, it is now known that antibodies can be effective against microorganisms without mediating a direct effect9. For example, antibodies can be effective in controlling malaria by reducing the damage that results from excessive host inflammatory responses10. Recent work showing that AUTOANTIBODIES AND IMMUNE COMPLEXES (common in malaria) drive B-cell responses through the pattern-recognition receptor (PRR) Toll-like receptor 9 (TLR9) supports this concept for malaria, in which immune-complex deposition in tissues causes pathology11. Understanding the biology of the effector mechanisms recruited by antibodies, including FcRs, complement and PRRs, will be an important adjunct to developing optimal antibody-based therapies for malaria8,12.

Antibody format. Genetic engineering and mammalian expression systems allow the production of human mAbs of a defined class and specificity, to target protective or neutralizing epitopes on the parasite. However, passive administration of mAbs might not always be appropriate, for several important reasons. Whole antibody can potentially bind more than one type of FcR and drive inappropriate immune responses, for example by triggering inhibitory FcRs8. A mAb also has to compete with circulating antibodies for FcR binding. This could be important in malaria, as infection induces the production of abundant low-affinity polyclonal IgG and HYPERGAMMAGLOBULINAEMIA13. To circumvent these problems, researchers have been assessing the potential of antibody-based fragments produced by manipulating the variable regions of antibody genes (Fig. 1). Single-chain variable fragments (scFvs) that can be formed from two variable regions represent the smallest antigen-recognition unit that can be produced in eukaryotic or prokaryotic expression systems. By linking parasite-specific scFvs to scFvs that recruit human immune effectors, many potential therapeutic molecules can be generated (Fig. 1). In this article, we will explore the various antibody formats and effector functions that might be useful in controlling the malaria parasite.

Antibodies and immunity to malaria

Acquired immunity to malaria occurs only after many years of recurring infection and, although protective against severe clinical disease, it fails to provide sterile immunity. Consequently, in regions where malaria is endemic and transmission intensities are high, young children bear the brunt of the disease. One of the reasons that acquired immunity to malaria is slow to manifest is that clinical immunity results only after continued exposure from multiple infections with many parasites over time. As a result, high-affinity protective antibodies recognizing invariant epitopes of otherwise polymorphic antigens (arising as a result of ANTIGENIC VARIATION) are slow to develop.

However, the importance of antibodies in protection against malaria has been clearly shown. The most convincing evidence for their role in humans comes from clinical studies in which hyper-immune serum or purified Igs from adults substantially modified the course of infection, reducing peak parasitaemias and leading to spontaneous resolution when administered to sick children3,4. Maternally derived antibodies can protect neonates and infants from P. falciparum and Plasmodium vivax infection14. In animal models, passive immunization with mAbs15,16,17, ADOPTIVE TRANSFER of immunity with B cells18 and increased susceptibility of B-cell-deficient hosts have all highlighted the crucial role of antibodies19. In clinically immune individuals, the naturally acquired antibodies are mainly from the IgG1 and IgG3 subclasses20; however, there is little correlation between the levels of these antibodies and resistance to infection or disease. This is believed to be due in part to antigenic variation and the time taken by individuals to develop acquired immunity. This gradual development of acquired immunity indicates that proteins that undergo antigenic variation (for example, erythrocyte membrane protein 1 (EMP1)) could be vitally important21. The role of antibodies in protective immunity is not completely understood, but various mechanisms, including inhibition of merozoite invasion of erythrocytes16, antibody-mediated phagocytosis through FcR and complement22 and ANTIBODY-DEPENDENT CELLULAR INHIBITION, are possible4,20.

Designing therapeutic antibodies

Given the evidence for a role of IgG1 and IgG3 in immunity to malaria, it would seem logical to design therapeutic antibodies belonging to these subclasses. However, both IgG1 and IgG3 can bind FcγRs on cells that do not directly kill P. falciparum, including platelets, B cells, endothelial cells and even placental tissue7,23,24, resulting in the triggering of inhibitory FcγRIIb. This might benefit the parasite, analogous to the effect on tumour cell growth, in which passively administered anti-tumour antibodies enhanced tumour cell growth through FcγRIIb24. Another drawback of using IgG for malaria therapy is the high level of parasite-induced IgG in hypergammaglobulinaemia13. These antibodies might compete with the passively administered IgG for FcR occupancy, and therefore phagocyte recruitment. The presence of this pre-existing IgG could explain why large doses of antibody have been required to neutralize parasites both in vitro and in vivo25.

Other Ig classes should also be considered in the design of therapeutic antibodies. Elevated levels of parasite-specific IgE in asymptomatic individuals are associated with a reduced risk of subsequent clinical malaria26. However, IgE also has a role in malaria pathogenesis, and therefore might not be appropriate for use in immunotherapy27. Although non-specific binding of IgM is implicated in pathology, such as ROSETTE FORMATION by infected erythrocytes28, parasite-specific IgM is a potent adjuvant in murine malaria vaccination studies29. Immune complexes containing IgM serve as innate microbial recognition receptors, and have recently been shown to be powerful adjuvants, stimulating the development of acquired T-cell-mediated immunity, probably through their interaction with Fcα/μR30. Therefore, recombinant parasite-specific IgM might have potential in IgM-adjuvant-based malaria vaccines.

We believe that targeting the FcαR (CD89) with IgA could offer the greatest potential for controlling malaria with therapeutic antibodies8. Unlike IgM, IgG and IgE, which are all implicated in the pathology associated with rosette formation28, placental malaria23 and severe malaria27, IgA has not been implicated in malaria pathology, arguing for its consideration in antibody therapy. Although a direct role for IgA in killing malaria parasites has not been fully investigated, Plasmodium-specific IgA has been detected at high levels in serum31,32 and breast milk32,33 in humans from endemic areas. Ligation of the myeloid FcαR induces cytokine release and can stimulate a respiratory burst22,34, and FcαR is better than the FcγRs at triggering lysis of antibody-targeted tumours as well as phagocytosis of pathogens coated with antibodies, both in humans and mice22,35. Human FcαR is readily available for targeting with therapeutic IgA, as it is constitutively expressed on monocytes, neutrophils and eosinophils as well as KUPFFER CELLS and dendritic cells, making it a good target for antigen presentation34. The recent finding that FcαR is a discrete modulator of the immune system that can mediate both anti- and proinflammatory functions indicates that further exploration of the role of IgA in malaria could be important36.

Potential target molecules

The Plasmodium parasite has a complex life cycle (Fig. 2). The problem for researchers interested in using antibodies, either for therapeutic or vaccination purposes, is which life-cycle stage, and which antigens from the bewildering assortment available, to target. Although this question has not been fully addressed, the obvious target for therapeutic antibodies is the asexual blood stage (ABS), which is responsible for the pathology of malaria. Parasites in the ABS express many surface and secreted proteins that have been proposed as good vaccine candidates37. Three of these ABS antigens are considered below, together with the likely mechanisms of action of specific therapeutic antibodies. Although we focus solely on the ABS, other parasite stages might also express suitable antigens for targeting by recombinant antibodies.

Figure 2: The life cycle of Plasmodium falciparum.
figure 2

The life cycle of Plasmodium falciparum, showing the parasite stages and antigens that are amenable to antibody-based targeting. Of particular interest are merozoite surface protein 1 (MSP1), apical membrane antigen 1 (AMA1) and erythrocyte membrane protein 1 (EMP1).

Merozoite surface protein 1 (MSP1). MSP1 is an attractive target for antibody therapy. Animals actively or passively immunized against MSP1 can be protected against parasite challenge, and mAbs to MSP1 can block parasite invasion of erythrocytes in vitro5,6,15,38. MSP1 undergoes dual proteolytic processing. It is initially cleaved (primary processing) into multiple fragments that form a protein complex on the merozoite surface. Then, at erythrocyte invasion, the protein is cleaved again (secondary processing) and shed from the surface, except for a C-terminal 19-kDa polypeptide (MSP119) that comprises two epidermal growth factor (EGF)-like domains and which is carried into the newly invaded erythrocyte. Importantly, although MSP1 is polymorphic, sequence variation in MSP119 between P. falciparum isolates is limited39, possibly because the structure is constrained by its function40. Naturally exposed individuals have antibodies specific for MSP119 that correlate with resistance to clinical malaria41 and are the major erythrocyte-invasion-inhibitory factor in vitro42. Novel genetic approaches, including LINKAGE-GROUP SELECTION, have also identified MSP1 as an important target of immunity43. The fine specificity of the antibodies binding to MSP119 is important for their function, and only some MSP119-specific mAbs inhibit invasion, which they do by inhibiting the secondary processing of MSP1 (Ref. 44). The conformational epitopes recognized by such antibodies have been defined both broadly and in detail5,45,46,47. Such antibodies, in addition to causing direct elimination of the parasite, should also present conserved MSP119 epitopes to antigen-presenting cells, stimulating cellular immunity and limiting parasite proliferation.

To explore further the potential of MSP119-targeted antibodies, recombinant human chimaeric IgG1 and IgA1 against MSP119 from the rodent malaria parasite Plasmodium yoelii 22 have been generated (Fig. 1). These reagents, in particular the IgA1, triggered efficient in vitro phagocytosis of merozoites by human neutrophils. However, in contrast to the parental mouse IgG2b (from which both the chimaeric antibodies were derived), both recombinant IgG1 and IgA1 failed to protect against parasite challenge in vivo when used to passively immunize mice, despite possessing comparable affinity for MSP119 (Ref. 22). This failure of the engineered antibodies to protect in vivo probably stems from an inability of their human Fc regions to trigger murine effector mechanisms, and indicates that blocking MSP119 function by steric hindrance alone is insufficient to provide protection. These reagents will permit evaluation of human effector mechanisms in vivo, by using mice transgenic for human FcR and complement genes. Such in vivo experiments are not possible with human malaria parasites, including P. falciparum, which are adapted for life in their human hosts.

Apical membrane antigen 1 (AMA1). AMA1 is a promising vaccine candidate that has induced protection in both rodent and non-human primate models of malaria48. Antibodies to AMA1 have also been shown to inhibit erythrocyte invasion and parasite growth in vitro49, indicating that AMA1 is also an attractive target for therapeutic antibodies. As AMA1 is expressed by infectious sporozoites from the mosquito as well as merozoites, targeting AMA1 has the additional advantage of attacking two stages of the parasite's life cycle50 (Fig. 2).

Erythrocyte membrane protein 1 (EMP1). EMP1 is a highly polymorphic, parasite-derived antigen that is expressed on the surface of infected erythrocytes and undergoes antigenic variation21. EMP1 is also responsible for cytoadhesion of infected erythrocytes to vascular endothelial cells, leading to parasite sequestration in the brain, kidneys and placenta, with particular importance in cerebral and pregnancy-associated malaria. High-affinity therapeutic antibodies that bind to functionally important regions of EMP1 could be developed to block cytoadhesion and facilitate phagocytosis of infected red blood cells by antibody binding to FcR on the surface of macrophages. Antibodies from patients' serum have been shown to bind conserved linear epitopes in EMP1 (Ref. 51), suggesting that the development of therapeutic antibodies to target infected erythrocytes is a real possibility.

Novel antibody-based formats

There are many potential advantages to using 'designer antibodies' over conventional mAbs for therapeutic applications. The antibody format available is limited only by the imagination of the investigator, and the choice will depend on the desired outcome (Fig. 1). Reagents can be designed that are optimal for parasite killing, whereas others might be better at presenting antigen for the development of immunity in the context of a vaccine (Fig. 3). Ideally, antibody-based therapies should be designed with both outcomes in mind. For malariologists, particular approaches might be most relevant (Fig. 3). For example, to circumvent the competition from pre-existing antibodies for FcR occupancy, bispecific antibody-based reagents (Figs 1, 3) can be created that bind to the parasite antigen of choice and to FcRs outside their natural antibody-binding site, allowing the recruitment of immune cells already saturated in antibody22 (Fig. 3). This application might be particularly important in the case of malaria, because of the hypergammaglobulinaemia associated with this disease13. This specific targeting also means that smaller quantities of antibody can be given in treatment, therefore providing additional savings in terms of cost.

Figure 3: Bispecific antibodies (BsAbs).
figure 3

BsAbs can be designed in various formats to control the malaria parasite. Injected BsAbs bind to key antigens on the merozoite or infected erythrocyte surface. This can inhibit progression of the disease by inhibiting proteolytic cleavage of antigens required for erythrocyte entry by the parasite (a). BsAbs for optimal killing and phagocytosis of merozoites that bind the Fcα receptor (FcαR) (this is the optimal FcR in terms of bioavailability and killing power) expressed on monocytes, macrophages, polymorphonuclear leukocytes (PMNs), NK (natural killer) cells and some dendritic cells (DCs) are designed to bind the FcR outside the receptor's natural binding site for antibody, to circumvent interference by serum immunoglobulin (b). BsAb can deliver malaria antigen to professional antigen-presenting cells (APCs) including DCs (c) or to NK cells through Toll-like receptors (TLRs) to optimally harness innate immune responses (d). Regulatory cells can also be specifically targeted, thereby modifying the course of disease (e). High-affinity binding BsAb can be designed to prevent pathology (f), for example by blocking the interaction of erythrocyte membrane protein 1 (EMP1) with endothelial receptors such as CD36 or intercellular adhesion molecule-1 (ICAM1). IFN-γ, interferon-γ; IL, interleukin; NO, nitric oxide; TGF, transforming growth factor-β; TNFα, tumour-necrosis factor-α; Treg, regulatory T cell.

Bispecific antibodies (BsAbs) have been developed to redirect or enhance immune effector activity towards particular targets, including tumour cells and pathogens, and they are clinically promising52,53. The prototypic BsAb consists of two covalently linked F(ab′)2 fragments, with one fragment targeting the pathogen or tumour cell and the other fragment targeting a trigger molecule (Fig. 1). We are investigating the BsAb format. Using recombinant methods, we have coupled the variable domains from a mAb recognizing P. yoelii MSP119 to the variable domains from a mAb recognizing human FcγR1, through a flexible peptide linker. This BsAb binds FcγR1 outside the high-affinity IgG-binding site, and simultaneous engagement of MSP119 and FcγRI by this molecule induced a potent neutrophil respiratory burst22. These promising results obtained with FcγRI BsAb in the P. yoelii model system have provided an impetus to generate BsAbs against P. falciparum that target other FcRs.

Other uses of BsAbs can be envisaged. For example, given the importance of innate immune responses in driving adaptive immunity, it might be of interest to link P. falciparum MSP119 to TLRs (Fig. 3). The ability of BsAbs to bind two different epitopes simultaneously indicates that a BsAb could be designed to inhibit the proteolytic cleavage of the MSP1 complex that is necessary for erythrocyte invasion by binding to either side of the cleavage site (Fig. 3). A bispecific reagent for malaria therapy has recently been developed for the treatment of P. falciparum malaria54. This BsAb targeted the CD3 antigen of human T cells and MSP119. In cooperation with T cells, the BsAb induced a significant increase in merozoite phagocytosis and inhibition of P. falciparum growth in vitro. Trispecific antibodies that bind MSP119, AMA1 and FcR could be useful, as this type of reagent is effective against tumours and might offer enhanced protection against malaria55.

The antibody immunotherapy for malaria (AIM) programme is generating intact chimaeric and fully human antibodies against key malaria antigens with which to compare the effectiveness of the BsAbs. These are generated by cloning variable domains from hybridomas or from antibody repertoire phage display libraries (generated from immune Gambian adults) and linking them to human constant domains in suitable expression vectors. Although there is clear evidence that chimaeric antibodies are less immunogenic than murine mAbs, there is little evidence to support claims for further humanization to avoid the antiglobulin response, perceived as an important problem in the clinical development of therapeutic antibodies56.

Although passive antibody therapy can confer immediate protection against malaria, rapid and efficient methods for generating the high-affinity neutralizing mAbs required for use in these novel formats have been slower in arriving. However, considerable progress has now been made in this area, including improved immortalization of human B cells by Epstein–Barr virus57, versatile high-throughput phage, prokaryotic and fungal expression systems for selecting high-affinity human antibody-binding fragments58,59,60, and immunization of transgenic mice carrying human Ig loci, coupled with hybridoma technology61.

Another exciting development has been the discovery of alternative antigen-binding scaffolds to the paired heavy and light chains from mammalian antibody repertoires (Fig. 1). For example, both camelids (camels, llamas and related species) and sharks produce single variable domains capable of high-affinity binding to buried targets62,63. This finding could have a major impact for antibody therapy in malaria, as some target molecules with cavities or clefts might be refractory to conventional antibody development. The camelid repertoire has been used to generate antibodies that can target enzyme active sites and similar targets64,65, indicating that such an approach could be used to inhibit key malaria enzymes, for example those involved in MSP1 and AMA1 processing. The selection and affinity maturation of single-antigen-receptor variable domains from sharks targeting P. falciparum AMA1 has been described and developed for diagnosis66.

Antibodies can also be manipulated for use in innovative vaccine applications, for example targeting liver-stage parasites, where antibody is believed to have a minor role in immunity. 'Troybodies' have been developed that incorporate several T-cell epitopes into antibodies with specificity for the PRR CD14. This approach involves exchanging loops in Ig constant domains with single copies of well-defined T-cell epitopes. Troybodies were found to be 100–100,000 times more efficient than synthetic peptide or native protein at inducing T-cell responses67. These reagents might therefore be usefully incorporated into malaria vaccination strategies that currently use various systems such as viral vectors for epitope delivery.

Antibodies can also be used as vehicles for the targeted delivery of antimalarial drugs. The development of drug resistance through the widespread consumption of antimalarials and adverse drug reactions arising from their toxicity are important global problems. Antibodies can deliver drugs directly to the parasite, thereby optimizing their pharmacokinetics and reducing the risk of adverse drug reactions. A plethora of molecules have been attached to antibodies, including drugs, enzymes, toxins, viruses and radionuclides. Although these approaches have been shown to be effective in cancer therapy, they remain untested in malaria.

Concluding remarks

Antibodies are believed to represent 25% of all proteins in clinical trials68. This interest in antibodies as therapeutic entities stems not only from a detailed understanding of their structure and immunological function but also from technological advances that have allowed their manipulation and expression. As a result, at least 18 antibodies have approval for use from the US Food and Drug Administration, and at least 100 are in clinical trials worldwide, the majority for the treatment of cancer, transplant rejection or autoimmune disease69. Despite these successes, none is in development for any tropical disease, including malaria. Most pharmaceutical companies shy away from investing in diseases they believe will not provide a healthy dividend for the shareholder. The seriousness of the health situation demands that the primary priority should be to find novel effective therapies. Antibody-based reagents and therapies are believed to be expensive and therefore inappropriate for a disease mainly affecting poor and marginalized populations. However, the pace of technological developments, especially in protein expression (for example in plants), should bring costs down substantially. In addition, these novel approaches could be useful as prophylaxis for the increasing number of non-immune individuals travelling to affected countries. Although it is unlikely that antibody-based prophylaxis will be made available to the millions of people at risk from malaria, antibody-based therapies could usefully be incorporated into the hospital setting, for patients non-responsive to conventional drug regimens. Importantly, antibody-based therapies have 'proof of concept' in clinical settings, ensuring that the development time from laboratory to patient would be significantly reduced and the costs constrained, a luxury unavailable to other prospective therapies. Moreover, antibodies are natural adjuvants that can present important epitopes to antigen-presenting cells, providing a 'vaccine effect', and thereby limiting proliferation of the parasite on re-infection. This potential one-off treatment could be cost effective compared to conventional vaccination strategies, which often require complex multiple boosting regimens and harsh adjuvants. The parasite is becoming resistant to conventional antimalarials, and therefore the urgency of the situation leaves little room for complacency.