Surface labelling of nematode antigens

A transformational breakthrough in antigen identification came at the end of the 1970s when Bridget Ogilvie and Michael Parkhouse at NIMR introduced surface labelling as a technique for nematodes. Individual parasite antigens could be visualized for the first time, even if at that time identification at the sequence level was still elusive. It was also startlingly clear that the surface-accessible repertoire was limited to a small number of defined proteins (Philipp et al. Reference Philipp, Parkhouse and Ogilvie1980; Maizels et al. Reference Maizels, Philipp and Ogilvie1982), as illustrated in Fig. 1 A and B. Further, each stage in the life cycle presented a different and unique set of antigens, which only rarely cross-reacted with other stages (Philipp et al. Reference Philipp, Taylor, Parkhouse and Ogilvie1981; Maizels et al. Reference Maizels, Partono, Oemijati, Denham and Ogilvie1983a). In every case, these surface antigens were highly immunogenic in the host, allowing specific serological responses to be mapped, and where tested were also found to stimulate protective immune responses against the parasite (Grencis et al. Reference Grencis, Crawford, Pritchard, Behnke and Wakelin1986).

Fig. 1.

Surface antigens, secretions and surface coat of parasitic nematodes. (A) Brugia adults, surface labelled with iodogen (a non-permeable reagent which iodinates of tyrosine residues) and solubilized in deoxycholate, a non-ionic detergent (DOC) or with SDS and 2-mercaptoethanol (2ME). The major surface glycoprotein, (gp-29, glutathione peroxidase) is released into detergent, while under reducing conditions, the disulphide-linked cuticular collagens are solublilized (Selkirk et al. Reference Selkirk, Nielsen, Kelly, Partono, Sayers and Maizels1989). (B) T. canis larvae, surface labelled as (A), or ES proteins labelled with same reagent. Identities to antigens discussed in the text are indicated. Note that the mucins are the major carriers of the dominant trisaccharide glycan. (C) Cryo-electron micrograph showing Ruthenium Red-stained polyanionic surface coat exterior to the cuticle of T. canis larvae. (D) Cryo-immunoelectron micrograph localizing binding of monoclonal antibody to the trisaccharide glycan binding to the surface coat (Page et al. Reference Page, Rudin, Fluri, Blaxter and Maizels1992b).

With the advent of cDNA cloning, nematode surface antigens were conclusively identified, such as the gp29 glutathione peroxidase from Brugia malayi (Maizels et al. Reference Maizels, Gregory, Kwan-Lim and Selkirk1989; Cookson et al. Reference Cookson, Blaxter and Selkirk1992), also independently discovered by Eileen Devaney in B. pahangi (Devaney, Reference Devaney1988), and the TES-32 C-type lectin of T. canis (Maizels et al. Reference Maizels, de Savigny and Ogilvie1984; Loukas et al. Reference Loukas, Mullin, Tetteh, Moens and Maizels1999), which has homologues in hookworms and other nematodes (Loukas and Maizels, Reference Loukas and Maizels2000).

In 1981, in collaboration with Felix Partono of the University of Indonesia, we had begun to examine the antigens of human filarial parasites, and the immune responses to them. We were able to show that all the lymphatic filariae, including Brugia timori and Wuchereria bancrofti, expressed highly similar sets of surface antigens which were also strongly cross-reactive in terms of human antibody recognition (Maizels et al. Reference Maizels, Partono, Oemijati, Denham and Ogilvie1983a, Reference Maizels, Partono, Oemijati and Ogilvieb, Reference Maizels, Burke, Sutanto, Purnomo and Partono1986; Morgan et al. Reference Morgan, Sutanto, Purnomo, Sukartono, Partono and Maizels1986). These findings sparked curiosity into why there should be antigenic conservation within this set of parasites, rather than the diversity that might be expected under the pressure of the specific immune response. One possible answer is that filarial parasites induced some form of immunogical tolerance, favouring conservation of ‘tolerised’ specificities (Maizels and Lawrence, Reference Maizels and Lawrence1991).

Labelling experiments also investigated the structure of the nematode surface, which is an unconventional structure with a collagen-rich extracellular matrix bounded by a lipidic epicuticle. Interestingly, in Trichinella spiralis the epicuticle is not a conventional lipid bilayer, but is arranged as cylindrical lipid bundles based on a hexagonal pattern (Gounaris et al. Reference Gounaris, Smith and Selkirk1996). In T. canis, the fact that most surface antigens were also secreted (Maizels et al. Reference Maizels, de Savigny and Ogilvie1984) led to the discovery of a labile surface coat, external to the epicuticle (Fig. 1C and D), involved in parasite immune evasion (Page et al. Reference Page, Rudin, Fluri, Blaxter and Maizels1992b).

Secreted ES antigens

In 1975, Don de Savigny published a simple technique for culturing T. canis larvae, in serum-free medium, which allowed the collection of significant quantities of ‘excretory-secretory’ (ES) antigens (de Savigny, Reference de Savigny1975). In collaboration with Don, we identified the ES proteins (Maizels et al. Reference Maizels, de Savigny and Ogilvie1984), as did the group of Bob Grieve (Badley et al. Reference Badley, Grieve, Bowman, Glickman and Rockey1987b), and once cDNA libraries became available each of these antigens were cloned (Maizels et al. Reference Maizels, Tetteh and Loukas2000). In vitro biosynthetic labelling of secreted proteins with radioactive amino acids (Parkhouse et al. Reference Parkhouse, Clarke, Maizels and Denham1985) allowed identification of ES products from species and stages which are available in only small numbers, such as from the infective larvae of W. bancrofti (Maizels et al. Reference Maizels, Burke, Sutanto, Purnomo and Partono1986). While the technologies of the 1980s allowed Gek-Eng Kwan-Lim to visualize the broad array of adult B. malayi ES antigens (Kwan-Lim et al. Reference Kwan-Lim, Gregory, Selkirk, Partono and Maizels1989), nearly 20 years elapsed before these were matched to genome sequence data by proteomics (Hewitson et al. Reference Hewitson, Harcus, Curwen, Dowle, Atmadja, Ashton, Wilson and Maizels2008). The detection of ES products in the serum of infected humans (Des Moutis et al. Reference Des Moutis, Ouaissi, Grzych, Yarzabal, Haque and Capron1983) and animals (Maizels et al. Reference Maizels, Denham and Sutanto1985), also proved important in developing diagnostic circulating antigen tests (Weil et al. Reference Weil, Lammie and Weiss1997).

T. canis provided an unrivalled model system in which to study the relationship between surface and secreted proteins, because larvae could be recovered from eggs in the laboratory, and cultured at 37°C for many months in serum-free medium. With Malcolm Kennedy and Huw Smith, we generated monoclonal antibodies which demonstrated that most ES components were associated with the parasite surface (Maizels et al. Reference Maizels, Kennedy, Meghji, Robertson and Smith1987b). Tony Page then used these antibodies in immuno-electron microscopy to show that while the C-type lectin TES-32 is integrally distributed in the larval cuticle, the mucin-like glycoproteins (TES-120, see below) are synthesized in internal secretory glands (Page et al. Reference Page, Hamilton and Maizels1992a) and on release form a loosely-attached surface coat which envelops the parasite (Page et al. Reference Page, Rudin, Fluri, Blaxter and Maizels1992b). Immune attack, for example by eosinophils, results in shedding of the surface coat and parasite escape (Fattah et al. Reference Fattah, Maizels, McLaren and Spry1986; Badley et al. Reference Badley, Grieve, Bowman and Glickman1987a). Similar surface coats can be observed in a number of other nematode species (Blaxter et al. Reference Blaxter, Page, Rudin and Maizels1992).

Once identified, the challenge has been to determine the functional properties of ES components. ES material is known to have significant immunological properties; for example, Martin Holland showed that N. brasiliensis ES (NES) is an intrinsically potent Th2-inducing adjuvant (Holland et al. Reference Holland, Harcus, Riches and Maizels2000) while Billy Harnett's laboratory have demonstrated that the major component of ES from the rodent filaria Acanthocheilonema viteae (ES-62) directly interferes with host immune cell signalling, offering therapeutic routes to the treatment of immunological disorders such as arthritis (Harnett et al. Reference Harnett, Goodridge and Harnett2005) and allergy (Melendez et al. Reference Melendez, Harnett, Pushparaj, Wong, Tay, McSharry and Harnett2007). Many ES products are proteases, fulfilling key roles in invasion and nutrition (Williamson et al. Reference Williamson, Lustigman, Oksov, Deumic, Plieskatt, Mendez, Zhan, Bottazzi, Hotez and Loukas2006), but their profile as immunomodulators is less certain. For example, although NES contains proteases (Healer et al. Reference Healer, Ashall and Maizels1991), they are not responsible for its Th2-driving properties (Holland et al. Reference Holland, Harcus, Riches and Maizels2000). Similarly, while ES-62 is a leucine aminopeptidase, its immunoregulatory activity is attributable to phosphorylcholine modifications and not to enzymatic activity (Goodridge et al. Reference Goodridge, McGuiness, Houston, Egan, Al-Riyami, Alcocer, Harnett and Harnett2007).

cDNA cloning and the spliced leader

The second transformational change took place from the mid-1980s with the introduction of molecular cloning of cDNA transcripts (Selkirk et al. Reference Selkirk, Rutherford, Denham, Partono and Maizels1987). Screening B. pahangi expression libraries with monospecific antisera, Murray Selkirk's laboratory isolated clones for key surface-associated antigens of microfilariae (MF22, SHP-1) (Selkirk et al. Reference Selkirk, Yazdanbakhsh, Freedman, Blaxter, Cookson, Jenkins and Williams1991) and adults (gp29 and gp15/400) (Cookson et al. Reference Cookson, Blaxter and Selkirk1992; Tweedie et al. Reference Tweedie, Paxton, Ingram, Maizels, McReynolds and Selkirk1993). A new platform for parasite biology and immunology was created with a rapid succession of collagens, muscle and heat shock proteins, enzymes, inhibitors and lipid-binding proteins being identified by investigators around the world. Pleasingly, the Selkirk laboratory also sequenced the acetylcholinesterases of N. brasiliensis first discovered by Bridget Ogilvie (Blackburn and Selkirk, Reference Blackburn and Selkirk1992; Grigg et al. Reference Grigg, Tang, Hussein and Selkirk1997). By differential screening of an Onchocerca volvulus cDNA library with sera from onchocerciasis and lymphatic filariasis patients, Jan Bradley isolated a set of O. volvulus-specific clones, providing both immunodiagnostic probes and potential immunomodulators (Bradley et al. Reference Bradley, Helm, Lahaise and Maizels1991). Most of the individual parasite proteins studied in the field today were first characterized during this era of discovery.

C. elegans molecular biology was developing rapidly at this time, for example with the discovery of a perfectly-conserved 22-nt spliced leader (SL) sequence which is trans-spliced onto the 5′ end of >50% of mRNA transcripts (Blumenthal and Thomas, Reference Blumenthal and Thomas1988). David Gems exploited this feature to amplify mRNA from T. canis by PCR using SL and oligo-dT (complementary to the poly-A tail at the 3′ end of all mRNAs) primers. He found 2 abundant mRNAs, one of which was a phosphatidylethanolamine (PE)-binding protein (Gems et al. Reference Gems, Ferguson, Robertson, Page, Blaxter and Maizels1995), and the other a mucin (Gems and Maizels, Reference Gems and Maizels1996), 1 of 3 associated with the surface-coat (Loukas et al. Reference Loukas, Hintz, Tetteh, Mullin and Maizels2000b). The SL ‘trick’ was then used to great effect on B. malayi by Bill Gregory to discover the abundant larval transcript (ALT) proteins (Gregory et al. Reference Gregory, Blaxter and Maizels1997), subsequently shown to be a prominent larval-specific antigen shared by all filarial parasites, and a promising vaccine candidate (Gregory et al. Reference Gregory, Atmadja, Allen and Maizels2000). Returning the compliment, Mark Blaxter applied parasite surface labelling techniques to C. elegans (Blaxter, Reference Blaxter1993).

Era of transcriptomes, genomes and proteomes

The late 1990s saw the development of systematic sequencing of cDNA libraries, termed Expressed Sequence Tags (ESTs). For B. malayi, this analysis confirmed that the SL-bearing alt transcripts were indeed very highly expressed by L3 larvae (Blaxter et al. Reference Blaxter, Raghavan, Ghosh, Guiliano, Lu, Williams, Slatko and Scott1996). More directed cDNA sequencing projects were then aimed to capture major products of N. brasiliensis (Harcus et al. Reference Harcus, Parkinson, Fernández, Daub, Selkirk, Blaxter and Maizels2004) and T. canis (Tetteh et al. Reference Tetteh, Loukas, Tripp and Maizels1999), as well as – foraying into cestodes – Echinococcus granulosus (Fernández et al. Reference Fernández, Gregory, Loke and Maizels2002). In the case of N. brasiliensis, for example, we were able to show that signal sequence-bearing proteins, presumed to include most of the secreted (ES) products, had a greater proportion of novel sequences with no similarity to C. elegans or other database entries; this was interpreted as showing a greater diversification among proteins interacting with the host (Harcus et al. Reference Harcus, Parkinson, Fernández, Daub, Selkirk, Blaxter and Maizels2004). With large-scale transcriptomics now accomplished across many members of the nematode phylum (Parkinson et al. Reference Parkinson, Mitreva, Whitton, Thomson, Daub, Martin, Schmid, Hall, Barrell, Waterston, McCarter and Blaxter2004), a rich new resource for helminth research had been made available.

The size of nematode genomes, typically 5–30×10⁷ bp, has proved an obstacle to full genome sequences until recent advances in high-throughput sequencing. In addition, there are varying degrees of polymorphism within each parasite species (Maizels and Kurniawan-Atmadja, Reference Maizels and Kurniawan-Atmadja2002; Redman et al. Reference Redman, Packard, Grillo, Smith, Jackson and Gilleard2008), which can frustrate genome assembly. Nevertheless, a real success in the field has been the publication of the draft sequence of B. malayi in 2007 (Ghedin et al. Reference Ghedin, Wang, Spiro, Caler, Zhao, Crabtree, Allen, Delcher, Guiliano, Miranda-Saavedra, Angiuoli, Creasy, Amedeo, Haas, El-Sayed, Wortman, Feldblyum, Tallon, Schatz, Shumway, Koo, Salzberg, Schobel, Pertea, Pop, White, Barton, Carlow, Crawford, Daub, Dimmic, Estes, Foster, Ganatra, Gregory, Johnson, Jin, Komuniecki, Korf, Kumar, Laney, Li, Li, Lindblom, Lustigman, Ma, Maina, Martin, McCarter, McReynolds, Mitreva, Nutman, Parkinson, Peregrin-Alvarez, Poole, Ren, Saunders, Sluder, Smith, Stanke, Unnasch, Ware, Wei, Weil, Williams, Zhang, Williams, Fraser-Liggett, Slatko, Blaxter and Scott2007). This dataset has allowed us to apply proteomics to nematode ES products; for example, recently James Hewitson and Yvonne Harcus identified some 80 proteins in adult B. malayi ES (BES), including a number detailed below (Hewitson et al. Reference Hewitson, Harcus, Curwen, Dowle, Atmadja, Ashton, Wilson and Maizels2008). For our laboratory, the focus has been on identifying which of these many predicted products and/or known secreted proteins may be immunomodulatory (Maizels et al. Reference Maizels, Blaxter and Scott2001a, Reference Maizels, Gomez-Escobar, Gregory, Murray and Zangb). In Fig. 2, and in a later part of this review, I summarize our findings on candidate modulators which have been obtained from a variety of approaches, ranging from in vitro testing of recombinant proteins, to heterologous expression by transfection into Leishmania for assessment of immunological functions in vivo (Maizels et al. Reference Maizels, Gomez-Escobar, Prieto-Lafuente, Murray and Aebischer2008).

Fig. 2.

Immunomodulators secreted by nematode parasites and their host target cells. (A) B. malayi adult worms: CPI-2, MIF-1/2. (B) B. malayi microfilariae: SPN-2. (C) B. malayi infective larvae (L3): ALT-1/2. (D) N. brasiliensis adult worms: NES. (E) H. polygyrus adult worms: HES. (F) T. canis adult worms: glycans and lectins. Images are from our own laboratory, with thanks to Bill Gregory (A), Judith Allen and Andrew MacDonald (B), Bill Gregory and Sinclair Stammers (C), Julie Healer and Sinclair Stammers (D), Constance Finney (E) and Tony Page (F). See text for details of all mediators mentioned.

Cystatins and serpins

Protease inhibitors secreted by nematodes do not simply regulate activity of parasite enzymes, as 2 major ES protease inhibitor families appear to act as immunomodulators by targetting host enzymes. One set, the cystatins (cysteine protease inhibitors) include CPI-2, a 15 kDa surface and secreted protein of B. malayi larvae and adult worms (Gregory et al. Reference Gregory, Blaxter and Maizels1997; Gregory and Maizels, Reference Gregory and Maizels2008), as well as O. volvulus and A. viteae (Hartmann and Lucius, Reference Hartmann and Lucius2003). Although homologues from plants to vertebrates inhibit papain-like cysteine proteinases, only mammals have forms which block asparaginyl endopeptidase (AEP), a key enzyme in antigen processing, through a second inhibitory site. Remarkably, CPI-2 has evolved the same AEP-inhibitory motif, and hence can interfere with antigen processing in human cells (Manoury et al. Reference Manoury, Gregory, Maizels and Watts2001). Consistent with the hypothesis that parasite cystatins have evolved AEP inhibition as an immune interference strategy, Janice Murray showed that C. elegans homologues lacked both the sequence motif and AEP inibition (Murray et al. Reference Murray, Manoury, Balic, Watts and Maizels2005). One unsolved puzzle is how CPI-2, secreted from extracellular Brugia parasites, accesses intracellular antigen-processing compartments of mammalian cells; an untested possibility is that a unique 20-aa N-terminal extension of the filarial CPI-2 protein, absent from other cystatins, is involved in uptake or entry into host APCs.

In the late 1990s, Xingxing Zang started a project to identify T-cell-stimulating antigens from B. malayi. He identified a serine protease inhibitor (serpin) SPN-2, which was also the most abundant transcript from the microfilarial (Mf) stage. In collaboration with a leading serpin laboratory, we found that SPN-2 selectively inhibited 2 neutrophil proteases (Zang et al. Reference Zang, Yazdanbakhsh, Kiang, Kanost and Maizels1999), but not a range of other serine proteases tested. Subsequently, another group noted that the SPN-2 sequence varied from the consensus for inhibitory proteins, and reported lack of inhibitory activity for SPN-2. Irrespective of this controversy, SPN-2 selectively drives Th1 responses (Zang et al. Reference Zang, Atmadja, Gray, Allen, Gray, Lawrence, Yazdanbakhsh and Maizels2000), an interesting finding in view of the fact that B. malayi Mf are one of the few helminth forms which induce Th1 in vivo (Lawrence et al. Reference Lawrence, Allen, Osborne and Maizels1994). SPN-2 is a major target of antibody in infection, and presence of the IgG4 isotype to this antigen can be considered diagnostic of active infection (Zang et al. Reference Zang, Atmadja, Gray, Allen, Gray, Lawrence, Yazdanbakhsh and Maizels2000), as antibody levels decay rapidly following chemotherapeutic clearance of parasites. Notably, no C. elegans homologues encode a signal sequence, indicating that B. malayi has evolved to secrete the serpin for a new function within the host (Zang and Maizels, Reference Zang and Maizels2001).

Glycans, mucins and lectins

Carbohydrate specificities play a major role in the immunology of infections. In the case of T. canis ES, we found that it comprised approximately 40% carbohydrate by weight (Meghji and Maizels, Reference Meghji and Maizels1986); subsequent characterization of the glycan structures by mass spectrometry, the first for a parasitic nematode, identified a dominant O-linked trisaccharide similar to mammalian blood group sugars, but modified with 1–2 additional O-methyl groups (Khoo et al. Reference Khoo, Maizels, Page, Taylor, Rendell and Dell1991). These trisaccharides have more recently been synthesized (Amer et al. Reference Amer, Hofinger and Kosma2003) and confirmed as major antigens in infection (Schabussova et al. Reference Schabussova, Amer, van Die, Kosma and Maizels2007). Monoclonal antibodies reactive to the trisaccharides also bind to the mucin glycoproteins produced by the major larval secretory glands (Page et al. Reference Page, Hamilton and Maizels1992a), and bind the labile surface coat of the parasite (Page et al. Reference Page, Rudin, Fluri, Blaxter and Maizels1992b). Whether the mucins and glycans interact further with the immune system once they are secreted seems likely but has yet to be documented.

Alex Loukas subsequently discovered, by protein sequencing, that T. canis TES-32 was a C-type lectin (CTL), with calcium-dependent carbohydrate specificity (Loukas et al. Reference Loukas, Mullin, Tetteh, Moens and Maizels1999), presaging a much wider presence in helminth secretions of this family of carbohydrate binding proteins (Loukas and Maizels, Reference Loukas and Maizels2000; Loukas et al. 2000 a). Because immune system receptors such as DC-SIGN are members of the same gene family, the possibility arose that secreted parasite CTLs interfere or compete with host lectins for ligands, thereby blocking host immunity. If so, it would constitute a fascinating role reversal, as similar CTLs are employed in C. elegans to mediate immunity against bacterial infection. In a further functional twist, the surface-bound CTL of the marine nematode Laxus captures symbiotic bacteria required for sulphur metabolism (Bulgheresi et al. Reference Bulgheresi, Schabussova, Chen, Mullin, Maizels and Ott2006). It is notable that while CTLs are prominent in the secretions of T. canis and hookworms, they are poorly represented among the filarial parasites; however, adult B. malayi secrete an abundant S-type lectin (galectin), which while bearing no structural similarity to CTLs, might serve a similar physiological role in vivo (Hewitson et al. Reference Hewitson, Harcus, Curwen, Dowle, Atmadja, Ashton, Wilson and Maizels2008).

Cytokine homologues

The immunosuppressive cytokine TGF-β belongs to an evolutionarily ancient extracellular signalling family that also includes bone morphogenetic proteins (BMPs). From observations of filarial parasite hyporesponsiveness, and with identification in C. elegans of the TGF-β homologue DAF-7, we hypothesized that B. malayi may express a functionally active mimic of TGF-β. Natalia Gomez-Escobar used degenerate PCR to isolate a B. malayi homologue (TGH-1), with a 7-cysteine active domain as found in the BMPs, which is primarily expressed in developing larvae (Gomez-Escobar et al. Reference Gomez-Escobar, Lewis and Maizels1998). A second homologue more like DAF-7 and human TGF-β in containing a 9-cysteine active domain emerged from ESTs; moreover, transcription of this gene (TGH-2) was maximal in the developmentally-arrested Mf and mature adult stages, which were found to secrete the protein (Gomez-Escobar et al. Reference Gomez-Escobar, Gregory and Maizels2000). More recently, we have found ES from Heligmosomoides polygyrus contains an active TGF-β mimic able to induce immunosuppressive regulatory T cells (Grainger et al., manuscript in preparation). Currently we are testing whether this biological activity is indeed attributable to a nematode homologue of the mammalian gene so central to immunoregulation.

A second ancient evolutionary family, which encompasses both mammalian cytokines and parasite homologues, is that of the Macrophage Migration Inhibitory Factor, or MIF (Vermeire et al. Reference Vermeire, Cho, Lolis, Bucala and Cappello2008). The first member of this family from B. malayi (MIF-1) was reported by Alan Scott's laboratory (Pastrana et al. Reference Pastrana, Raghavan, FitzGerald, Eisinger, Metz, Bucala, Schleimer, Bickel and Scott1998), and Xingxing Zang isolated a second homologue, MIF-2, which was subsequently crystallized in Edinburgh (Zang et al. Reference Zang, Taylor, Meyer, Wang, Scott, Walkinshaw and Maizels2002). Because of the apparent contradication for a tissue-dwelling parasite to produce a potentially pro-inflammatory mediator, we examined the activity of MIF in more detail. In vivo, MIF was able to induce the production of Ym-1, a key product of alternatively activated macrophages (Falcone et al. Reference Falcone, Loke, Zang, MacDonald, Maizels and Allen2001). Most recently, Lidia Prieto-Lafuente established that in the presence of IL-4, MIF loses ability to induce IL-12 and TNF, and instead promotes the development of these suppressive macrophages (Prieto-Lafuente et al. Reference Prieto-Lafuente, Gregory, Allen and Maizels2009).

Novel proteins – ALTs, VALs and vaccines

Many parasite products involved in establishment in the host will not be recognized or functionally categorized by sequence alone. We were intrigued by novel gene transcripts that are highly expressed in infective larval stages, as judged by their abundance in EST datasets. For B. malayi, 2 such families stood out at levels >1%, the abundant trans-spliced transcripts alt-1 and alt-2 mentioned above, and a Venom Allergen-Like (VAL) homologue similar to Ancylostoma secreted protein (ASP).

ALT-1 and -2 proteins are approximately 80% identical to each other. Homologues exist in all filarial nematodes, for example in O. volvulus in which Ted Bianco's laboratory showed ALT proteins stockpiled in the glandular oesophagus of the vector-borne infective larva (Wu et al. Reference Wu, Egerton, Pappins, Harrison, Wilkinson, Underwood and Bianco2004). ALT proteins are the foremost filarial vaccine candidates, with data from several laboratories showing protection in animal models (Gregory et al. Reference Gregory, Atmadja, Allen and Maizels2000). We wished, however, to test the immunomodulatory properties of ALT proteins, as the dominant parasite products at the first encounter with the immune system. With Natalia Gomez-Escobar, we adopted the strategy of transfection of Leishmania parasites, to allow in vitro effects on macrophages and in vivo outcomes in mice to be evaluated (Maizels et al. Reference Maizels, Gomez-Escobar, Prieto-Lafuente, Murray and Aebischer2008). When either ALT-1 or ALT-2 were expressed as a transgene in L. mexicana, growth in macrophages was enhanced; cellular mRNA analyses showed higher levels of SOCS-1 expression due to ALT transfection. Because SOCS-1 inhibits IFN-γ signalling, the effects of ALT may therefore dampen Th1 effector mechanisms. Consistent with this hypothesis, the virulence of ALT-expressing lines was significantly amplified, with up to 3-fold more parasites in vivo (Gomez-Escobar et al. Reference Gomez-Escobar, Bennett, Prieto-Lafuente, Aebischer, Blackburn and Maizels2005).

Nematode VAL/ASP proteins are similarly elusive in function. Homologues range from pathogenesis-related proteins in plants through to testis-specific proteins in mammals, but have no assigned biological function. It is possible that they represent a conserved structural framework compatible with a variety of functional activities. The first nematode homologue (Ancylostoma secreted protein) was discovered by Peter Hotez' laboratory (Hawdon et al. Reference Hawdon, Jones, Hoffman and Hotez1996), and a closely-related product is now in human vaccine trials. Since then, VAL/ASP proteins have been shown to be dominant ES products of many organisms, such as H. contortus (Yatsuda et al. Reference Yatsuda, Krijgsveld, Cornelissen, Heck and De Vries2003). In our laboratory, Janice Murray isolated the B. malayi VAL protein and showed it to be highly immunogenic although its vaccine efficacy, under the conditions tested, was less potent than that of ALTs (Murray et al. Reference Murray, Gregory, Gomez-Escobar, Atmadja and Maizels2001).

Surprisingly, when we applied the transcript-led approach to T. canis (Tetteh et al. Reference Tetteh, Loukas, Tripp and Maizels1999), we found not only the major secreted lectins and mucin aproprotein, but also a set of 4 abundant novel transcripts (ANTs) with similarity to each other only in their 3′ untranslated regions. No protein corresponding to these transcripts could be identified, but 27 nt in the 3′ UTR bore similarity to the lin-4 suppressive micro-RNA of C. elegans. The possibility of post-transcriptional suppression of protein translation was supported by transfection of C. elegans with GFP reporter constructs containing the ANT 3′ UTR, which resulted in sequence-specific suppression of GFP expression (Callister et al. Reference Callister, Winter, Page and Maizels2008). Hence, empirical proteomic analysis is crucial before interpreting parasite expression profiles based on cDNA alone.

Phosphorylcholine (PC)

The presence of PC in somatic antigens from N. brasiliensis had been discovered in the 1970s (Péry et al. Reference Péry, Luffau, Charley, Petit, Rouze and Bernard1979). Remarkably, many monoclonal antibodies raised to filarial antigens were PC-specific, and reacted with a circulating parasite antigen (Forsyth et al. Reference Forsyth, Spark, Kazura, Broen, Peters, Heywood, Dissanayake and Mitchell1985; Maizels et al. Reference Maizels, Burke and Denham1987a). In human patients, PC is found on a high molecular-weight proteoglycan-like complex, while in the rodent filaria A. viteae, PC occurs on one secreted protein, leucine aminopeptidase (LAP) or ES-62. More recently, James Hewitson identified that in B. malayi, PC is bound not to LAP but to a distinct secreted protein, N-acetylgalactosamine transferase (Hewitson et al. Reference Hewitson, Harcus, Curwen, Dowle, Atmadja, Ashton, Wilson and Maizels2008). Interestingly, in another rodent filarial species (Litomosoides sigmodontis), the PC specificity is replaced by dimethylethanolamine (DMEA), containing one less methyl group (Hintz et al. Reference Hintz, Schares, Taubert, Geyer, Zahner, Stirm and Conraths1998). This microheterogeneity in the presentation of PC-like epitopes may relate to differences in innate recognition systems of the respective host species.

Dendritic cells and macrophages

In the first encounter between pathogens and the innate immune system, dendritic cells (DCs) shape the ensuing response, through recognition of pathogen molecules and selective induction of T-cell differentiation. A central paradigm is that bacterial LPS, ligating TLR4, upregulates surface costimulator molecule expression and induces IL-12 release, driving a Th1 outcome. Adam Balic asked whether DCs exposed to NES would represent the opposite side of the coin, eliciting a Th2 bias, and indeed this was the case (Balic et al. Reference Balic, Harcus, Holland and Maizels2004). Interestingly, the capacity to drive Th2 responsiveness was not accompanied by the upregulation of surface markers observed in LPS stimulation, but by blocking the ability of LPS to induce IL-12 p70, as has also been observed for other helminth products (Cervi et al. Reference Cervi, MacDonald, Kane, Dzierszinski and Pearce2004). Because NES actively promotes Th2 differentiation, even in the presence of Th1-inducing bacterial products, we argued against a ‘default’ hypothesis in which Th2 resulted from an absence of DC stimulation (MacDonald and Maizels, Reference MacDonald and Maizels2008). More recently, with the focus shifting to DC populations in vivo, Katie Smith in our laboratory has shown that nematode infection switches intestinal DCs into a predominantly tolerogenic subset.

An unusual innate immune population was discovered by Judith Allen in the peritoneal cavity of Brugia-infected mice. These were macrophages which exerted a broad and profound anti-proliferative effect on target cells (Allen et al. Reference Allen, Lawrence and Maizels1996), and which required IL-4 but not IL-10 to develop (MacDonald et al. Reference MacDonald, Maizels, Lawrence, Dransfield and Allen1998), While the mechanism of suppression is still unknown, it is cell contact-dependent (Loke et al. Reference Loke, MacDonald, Robb, Maizels and Allen2000). This new Th2-associated macrophage was similar to the ‘alternatively activated’ macrophage identified by Saimon Gordon at Oxford (Gordon, Reference Gordon2003), and was further linked to the synthesis of a chitinase-like molecule Ym1, among other characteristic proteins. From the initial finding in Brugia-infected mice, the study of alternatively activated macrophages has grown into a major field bringing together immunity, wound healing, obesity and innate recognition (Loke et al. Reference Loke, Gallagher, Nair, Zang, Brombacher, Mohrs, Allison and Allen2007; Siracusa et al. Reference Siracusa, Reece, Urban and Scott2008). A functional role for these cells in protective immunity against H. polygyrus, while in its larval phase in the gut wall, has been demonstrated (Anthony et al. Reference Anthony, Urban, Alem, Hamed, Rozo, Boucher, Van Rooijen and Gause2006). A key question is what stimulates the development of this phenotype, and Lidia Prieto-Lafuente has very recently shown that the combination of IL-4 and MIF (whether from host or parasite) can induce suppressive, Ym-1-expressing macrophages in vitro (Prieto-Lafuente et al. Reference Prieto-Lafuente, Gregory, Allen and Maizels2009).

Regulatory T cells and tolerance

The Regulatory T cell (Treg) has represented the latest conceptual tidal wave in immunology, and the concept of a suppressive population inhibiting both Th1 and Th2 was one which resonated immediately in the human filariasis context (Doetze et al. Reference Doetze, Satoguina, Burchard, Rau, Loliger, Fleischer and Hoerauf2000). In 2001 Matthew Taylor started testing whether, in an animal model, Tregs blocked Th2 immunity to L. sigmodontis. Matt showed that ablation of Treg activity, using anti-CD25 and anti-GITR antibody treatment in vivo, significantly reduced worm load (Taylor et al. Reference Taylor, Le Goff, Harris, Malone, Allen and Maizels2005). This was the first immunological intervention able achieve a measure of ‘cure’ in a susceptible infected host. While anti-CD25 alone abated Treg numbers, additional anti-GITR or anti-CTLA-4 (Taylor et al. Reference Taylor, Harris, Babayan, Bain, Culshaw, Allen and Maizels2007) were required to kill parasites. Because infected mice displayed antigen-specific hyporesponsiveness, which was reversed by the combined antibody treatment, we suggested that CD25+ Tregs tolerize the effector T-cell population, which requires re-activation through GITR ligation or CTLA-4 blockade (Taylor et al. Reference Taylor, Harris, Babayan, Bain, Culshaw, Allen and Maizels2007).

Mark Wilson and Constance Finney then studied Tregs in the model system of H. polygyrus, well characterized by Jerzy Behnke and David Pritchard at Nottingham. Mark found a significant rise in Treg numbers during infection, which Constance demonstrated was accompanied by raised expression of CD103 (considered to be a Treg activation marker) and increased suppressive effect per cell (Finney et al. Reference Finney, Taylor, Wilson and Maizels2007). Mark Wilson used H. polygyrus to explore the interaction between airway allergy and helminth infection, because this parasite does not transit the lungs. Infection resulted in a profound suppression of inflammation (measured, for example, by eosinophil infiltration) even in animals rendered fully allergic prior to infection. Most importantly, CD4+CD25+ Tregs from infected, allergen-naive animals can, on transfer to an allergen-sensitized but uninfected recipient, suppress airway allergy (Wilson et al. Reference Wilson, Taylor, Balic, Finney, Lamb and Maizels2005). Thus, helminth-induced Tregs were shown functionally to inhibit bystander reactivities, contributing to our perception of the ‘hygiene hypothesis’ discussed below.

The question remains as to whether helminths stimulate expansion of pre-existing ‘natural’ Tregs, specific for self antigens, or induce the differentiation of new ‘adaptive’ Tregs from naive peripheral cells. We suspect that both may occur (Fig. 3). Matt Taylor showed that L. sigmodontis activates the natural Treg subset early in infection (Taylor et al. Reference Taylor, van der Werf, Harris, Graham, Bain, Allen and Maizels2009), while B. malayi can drive bystander (ovalbumin-specific) T cells into a regulatory phenotype (McSorley et al. Reference McSorley, Harcus, Murray, Taylor and Maizels2008). Most recently, John Grainger has found that H. polygyrus ES contains a TGF-β-like activity which, acting through the TGF-β receptor of mouse T cells, can convert non-regulatory cells into ‘adaptive’ Tregs, confirming that some nematodes at least have evolved to exploit the host regulatory network (Grainger et al., manuscript in preparation).

Fig. 3.

Immunoregulation in nematode infections and the ‘Hygiene Hypothesis’. Infection generates both activating signals (blue arrows), regulatory signals (red arrows) and inhibitory interactions (red bars). Hence, a Th2 response is generated through innate immune system cells, but in chronic infection adopts a modified or muted phenotype in which proliferative responses are poor (due to either or both alternatively activated macrophages and the effects of regulatory T cells). The Th2 response is sufficient, however, to inhibit pro-pathogenic Th1 and Th17 responses, again in combination with Treg activity. IL-10 switches B cells into IgG4 production, reducing IgE, and minimising overall pathology of infection. The high levels of Tregs (induced via DC, or directly by parasite molecules), together with B cells, create a regulatory environment which inhibits pro-allergic and autoimmune T cells from mounting a response. The observed associations between infections and protection from allergies and autoimmunity may be attributed to a regulatory network such as that depicted here.