Immune Responses in Hookworm Infections

Humoral Response

Larval sheath antigens are immunogenic and are stage specific in their expression. Most hookworm L3 cast their outer sheaths on skin penetration (85), but some seem to retain the sheath; antibodies from people infected with N. americanus recognize surface antigens of ensheathed but not exsheathed L3 by immunofluorescence (112). Antibodies have been detected to L3 exsheathing fluid, prompting the suggestion that the cast sheath and associated antigens diverted the immune response away from the potentially vulnerable larval surface (69).

In Papua New Guineans infected with N. americanus, strong antibody responses representing all five human immunoglobulin (Ig) isotypes were found to larval and adult antigens (116), but the responses varied widely in the larval antigens recognized (24). While much of the IgE response in helminth infections is not directed against the parasite, detection of IgE antibodies against Necator L3 proved to be highly specific and sensitive for diagnosing infections, and, furthermore, IgE was the least cross-reactive isotype (41, 115).

Human and canine hookworms do not reach maturity in laboratory hosts such as mice. Infective L3 penetrate the skin and can migrate through the lungs and other extraintestinal tissues similarly to somatic migration in the definitive hosts (12). Mice rapidly become resistant to otherwise fatal doses of L3, and this antilarval resistance is thought to be mediated by significant increases in levels of IgM, IgG1, and IgE (59). Furthermore, antibody-dependent cell-mediated cytotoxicity is thought to play a central role in trapping and killing of L3 in murine infections (59, 136). However, it cannot be overemphasized that these experimental infections do not represent natural exposure in humans: the larval doses used in mice are grossly disproportionate, their routes of administration are often unnatural (e.g., subcutaneous, intragastric, intraperitoneal), and the findings do not reflect the humoral immune response of humans to anthropophilic hookworms, although they could provide valuable information for researchers involved in designing hookworm vaccines.

Cellular Response

Eosinophils predominate in the inflammatory response to hookworm L3 in tissues (11), and, given a sufficient larval dose, this is reflected by peripheral blood eosinophilia. Circulating eosinophils from human volunteers infected with N. americanus were functional and secreted superoxide, while neutrophil numbers and activation remained unaffected (149). Peripheral eosinophil responses in experimental human infections with either N. americanus (86) or A. duodenale (99) were boosted greatly by the arrival and development of worms in the gut, probably reflecting accelerating antigenic output by the feeding L4 and adult stages.

Hookworm larval migration through the lungs is often associated with clinical pulmonary abnormalities, which probably reflect a combination of physical trauma and tissue inflammation (118). Eosinophils have been recovered by bronchoalveolar lavage from mice infected with N. americanus (143, 144), but such findings are less common in humans (118). In vitro, human eosinophil binding to N. americanus L3 becomes maximal in the presence of complement and antibodies (37), although it is not known if this adherence is larvicidal.

The contribution of eosinophils to the in vivo destruction of helminth parasites is even less clear. The evidence suggests that these cells can kill infective larval stages but not the adults of most helminth species investigated (89), and it is supported by studies with genetically modified mice. Interleukin-5 (IL-5) is a critical growth and differentiation factor for eosinophils, and IL-5 transgenic mice have an enhanced capacity to kill larvae of Nippostrongylus brasiliensis but not Toxocara canis (36). However, IL-5 knockout mice infected with A. caninum did not mount blood eosinophilia and yet still had the same numbers of L3 in their tissues as did normal mice after 6 weeks (J. Landmann and P. Prociv, unpublished data), indicating that an absence of circulating eosinophils in mice does not increase susceptibility to hookworm L3 invasion.

Larval Hypobiosis

Invasive larvae of A. duodenale (but not N. americanus) can enter hypobiosis in the definitive host, becoming developmentally arrested in the muscles (and perhaps other tissues) and reactivating when conditions might be more favorable for transmission. Experimental infection of volunteers with A. duodenale triggered a small, initial rise in circulating eosinophil numbers with no further change until week 33 of infection, when they climbed sharply. Eggs were first observed in stools 7 weeks later, indicating a prepatent period five times that expected for this parasite (99). Larval hypobiosis explains why the transmission of A. duodenale coincided with the monsoon season, when environmental conditions were most favorable for larval survival (132). Larval arrest is much more convincingly documented with A. caninum in dogs, where arrested larvae mobilize at the end of pregnancy and infect newborn pups by the transmammary route (133). While L3 have not been detected in human milk, the circumstantial evidence is strong, with documented intense neonatal infections (121, 153) suggesting transmammary (and/or transplacental) transmission of A. duodenale.

Humoral Response

Like most intestinal helminthiases, hookworm infection is characterized by antibody responses dominated by IgG1, IgG4, and IgE, which in turn are under the control of Th2 cytokines, typically mediated by the effects of IL-4. Various methods have been used to identify these antibody responses, ranging from observation of immunoprecipitates around oral openings (102) to more detailed analyses of Ig subclasses against somatic and excretory/secretory (ES) products by enzyme-linked immunosorbent assay and Western blot analysis (79, 80, 112).

Studies of Papua New Guineans infected with N. americanus showed that antibody isotypes responded differently before and 2 years after anthelmintic treatment (116). Egg counts and worm burdens rose to about 50% of pretreatment levels 1 year after the initial chemotherapy. IgG and IgM levels against adult ES antigens had fallen significantly 1 year after treatment, and IgG levels continued to decline over the following year, while IgM levels increased again but did not reach pretreatment levels, consistent with ongoing reinfection. IgE levels fell only slightly, and, in contrast, IgA and IgD levels increased over the 2-year period. The levels of all isotypes raised to larval antigens except for IgD dropped after the first year, but each isotype had recovered to its original level by the following year except for IgG, whose level continued to decline. Interpreting these data proved difficult, given that the patients were reinfected by 12 months posttreatment. The authors suggested that sufficient reinfection might have provided the smaller quantities of antigen required to maintain the local IgA and IgE responses compared with the larger quantities needed to maintain systemic IgG and IgM responses, accounting for the isotype profiles observed. It was concluded that levels of IgG and IgM against adult ES antigens provided the best indicator of both current infection with adult parasites and efficacy of chemotherapy, since these isotypes were the most strongly affected by treatment (116). However, despite this vigorous response, there is little evidence that antibody levels correlate with protection against hookworms (see “Immune protection against hookworms” below).

Like acute allergy, tissue invasion by helminths is associated with elevated levels of IgE in serum. Much of this IgE induced by helminth parasites seems to be directed against heterologous antigens and is tightly regulated by opposing cytokines. Th2 cytokines (typically IL-4, IL-5, and IL-13) predominate, with IL-4 promoting the synthesis of IgE, while Th1 cytokines such as gamma interferon inhibit its production (128). This observation led to speculation that helminth parasites secrete proallergic mediators that induce polyvalent, non-parasite-specific IgE, thus saturating IgE receptors on effector cells (106). A recent study has shown that IgE antibodies were more specifically directed against N. americanus epitopes than were other Ig isotype responses (115). Preadsorption of human N. americanus infection sera with antigens from closely related hookworm species had little effect on anti-Necator IgE levels (1.0 to 2.4% decrease in mean optical density values), while the IgG levels dropped by 19 to 23%. In human intestinal infection with A. caninum, IgE responses proved to be more specific than IgG responses to ES antigens, with selected patients generating detectable levels of IgE but not IgG to a diagnostic antigen (80). Patients infected with A. duodenale produce both systemic and jejunal IgE that specifically binds to larval antigens (41), suggesting that some hookworm allergens are shared by migrating larvae and adults residing in the gut.

Circulating IgG1 and IgG4 levels are frequently elevated in helminthiases, including hookworm infections, and specific IgG4 has been suggested to be a marker of active infection with N. americanus (103). The role of IgG4 is poorly understood, although, like IgE, it is upregulated in atopic conditions and helminthic infections. It is thought to downregulate immune responses by competitively inhibiting IgE-mediated mechanisms, e.g., by blocking mast cell activation (145). In addition, IgG4 does not fix complement and only weakly binds to Fcγ receptors. Therefore, antigen binding by IgG4 is likely to be less harmful to the host than binding by IgE. The potentially harmful effects to the host of an exuberant IgE response are further restricted by the production of IgG autoantibodies to IgE (114). This complex area needs further exploration, especially if vaccines are to be developed against hookworm infections.

Adult hookworms induce the production of secretory IgE (68), IgG, and IgM but not IgA (41), and the levels of these Igs return to normal after anthelmintic treatment. The absence of secretory IgA in these patients was intriguing (11) and might be accounted for by the secretion of hookworm proteases that specifically cleave IgA (107).

Cellular Response

Information on the cellular immune response in human hookworm infections is sparse: essentially, a Th2 response predominates, generating IgE and eosinophilia.

Eosinophils

Eosinophils feature prominently in the leukocytic response to adult hookworms (11, 118, 149), and their circulating numbers reflect the worm burden (112). Infections of volunteers with N. americanus (86, 149) and A. duodenale (99) were characterized by blood eosinophilia, although the dynamics of the response varied with the species. Larval hypobiosis and reactivation in A. duodenale infection was thought to underlie the small initial increase followed by the sudden rise in eosinophil levels at week 33 of infection, shortly preceding the appearance of hookworm eggs in feces (99). With Necator, volunteers exposed to L3 rapidly developed blood eosinophilia, which peaked between days 38 and 64 (86). Systemic eosinophilia is also pronounced in human enteric infections with A. caninum (118–120).

Obviously, eosinophils in the blood are in transit from the bone marrow to their effector sites, the small intestine in this case. In human infections with both anthropophilic and zoonotic hookworms, infiltration of the worm feeding site with eosinophils is an almost universal feature, but it varies according to the stage and intensity of infection as well as to individual host factors. In extreme cases, exemplified by human eosinophilic enteritis associated with A. caninum, all layers of the gut from the mucosa to the serosa can be severely affected by intense eosinophilic inflammation (118, 120).

T Cells and Cytokines

Very little is known about T-cell reactivity in hookworm infections; only two studies have examined human blood lymphocyte responses, both in Necator infections. Peripheral T cells reacted strongly to mitogens (86) and weakly to hookworm antigens (141). Clinical manifestations did not correlate with lymphocyte responses, either in these studies or in animal models.

In general, nematode infections drive a strong Th2 response, promoting IgE synthesis and eosinophil production and migration, as well as production and proliferation of suppressor cell populations (3, 7). This response is mediated by parasite secretions and can be induced by immunizing naive animals with ES products alone (2). While cytokine profiles have not yet been reported from either animal or human hookworm infections, many detailed reports have been published from other helminthiases, including those in cytokine transgenic and gene knockout mice. It is very likely that many of the findings will apply directly to hookworm infection, although this remains speculative in the absence of data.

Mast Cells

Mast cell degranulation in response to IgE-allergen interaction plays a critical role in the local mobilization and activation of eosinophils (40). Mast cell proteases degrade cuticular collagens of adult N. americanus (87) and no doubt are important in the host response to hookworms, but they have attracted sparse research attention. Hamsters infected with host-adapted A. ceylanicum quickly developed resistance to infection with this species, characterized by increased antibody production and mucosal mastocytosis (9, 42).

Mast cells are instrumental in resistance to some other intestinal helminthiases. While the mechanisms are unclear, neutralizing antibodies to IL-9 inhibit the production of mast cell protease I and prevent normally resistant mice from clearing infections with Trichinella spiralis (126). In contrast, abrogation of IL-3 and IL-4 cytokines required for mast cell hyperplasia led to an 85 to 90% decrease in mucosal mast cell numbers in mice infected with Nippostrongylus but had no effect on worm expulsion (82).

Hookworms appear to be more resistant to intestinal inflammation than are most other intestinal nematodes, perhaps reflecting their attachment and feeding strategies. Hamsters coinfected with T. spiralis and A. ceylanicum or N. americanus produced an intense mucosal response that cleared T. spiralis (albeit more slowly than in hamsters with just T. spiralis), while the hookworms remained relatively unaffected (13). Indeed, depressed anti-T. spiralis antibody levels indicate that hookworms might protect other parasites by generally suppressing immune responses (13).

IgE and Protection

Among Papua New Guineans infected with N. americanus, those with higher levels of total IgE had fewer and less fecund parasites, both at initial anthelmintic treatment and 2 years after reinfection (113), prompting the authors to suggest a protective role for polyvalent IgE. However, correlations of hookworm weight and fecundity with levels of specific antibodies of all isotypes to N. americanus ES products were not significant, and levels of anti-ES IgE showed a trend toward a negative correlation, indicating that total IgE rather than anti-parasite antibodies were associated with protection.

It has been suggested that the massive quantities of total IgE induced during helminth infections protect the host against atopic disorders, by saturating mast cell Fcε receptors and suppressing specific IgE production against allergens. The presence of infection with some helminths, such as Ascaris lumbricoides, is often negatively correlated with atopic disorders such as asthma (81). However, this does not seem to apply to hookworms, where no such correlation was found (114).

Age-Dependent Acquired Immunity

In many gut helminthiases, the prevalence and intensity of infection are higher in children (or young animals) and decline with age. In humans, this has been attributed to age immunity and/or to behavioral changes that diminish exposure to reinfection, two influences which have been remarkably difficult to distinguish by epidemiological studies. Certainly, occupational exposure explains the high levels of hookworm infections in some adult groups, such as plantation workers or coal miners, and no convincing evidence of immunity developing with age has been put forward.

While immunity to human (and canine) hookworms is rapidly acquired in experimental animals and acts against invading L3, the development of protective immunity to these same parasites in humans is not obvious. Varying correlations between levels of IgG (anti-larval and anti-adult ES) and worm burdens in different age groups have also been reported but are difficult to interpret, given the repeated exposure and concurrent infections in hookworm-endemic populations (118, 123). Such findings have yet to be reported for natural A. duodenale infections.

On the other hand, dogs do develop age-related immunity to A. caninum (92), and they can be successfully protected by vaccination with irradiated L3 (see “Vaccination against hookworms” below). Dogs also develop resistance to A. ceylanicum (25), which does not undergo larval hypobiosis, indicating that the canine immune system might hold the key to successful protection against human hookworm infections.

Cutaneous

Cutaneous penetration is the exclusive or predominant infection route for all human hookworm infections (even though A. duodenale and A. caninum are also capable of developing to adulthood after oral ingestion of L3). With reexposure to infective hookworm L3, various inflammatory responses can be generated, ranging from asymptomatic to static ground itch lesions to classic, migratory creeping eruption, regardless of the hookworm species. Generally, however, anthropophilic species produce less extensive skin changes, while A. braziliense is the notorious cause of the last manifestation, even with exposure presumably to small numbers of L3 (118). Eosinophilia and elevated serum IgE levels were reported in only a minority of 98 patients who presented with creeping eruption to a travel medicine clinic (63), suggesting that a few L3 are insufficient to trigger a detectable response.

Pulmonary

As larvae of A. duodenale and N. americanus traverse the human lung, they can induce focal hemorrhagic and inflammatory lesions that may manifest clinically, although generally this is not as severe as the symptoms of Loeffler's syndrome associated with migrating Ascaris larvae (8). Invasive L3 of A. caninum rarely trigger respiratory symptoms, and then only at very high doses (118). On the other hand, pulmonary involvement seems to complicate even light exposure to A. braziliense L3 surprisingly often and is associated with blood eosinophilia (118). Given that the L3 of A. braziliense probably do not proceed beyond the human lung, perhaps their disintegration in situ provides an excessive local antigenic stimulus, and inflammation then manifests as respiratory symptoms and radiological abnormalities (whereas L3 of A. caninum simply move on).

The role of pulmonary tissues in immune responsiveness to and protection against hookworm infections has not been studied.

Intestinal

Intestinal hookworm infection is associated with variable levels of local inflammation and clinical illness. The initial symptoms seem to coincide with the arrival in the gut of the L4, with its rapidly developing anterior glands and genital system, and are most probably triggered by secretions injected into the host mucosa (118). Given the involvement of IgE-primed mast cells, eosinophils, characteristic cytokines, and other accoutrements of the Th2 response, it is difficult to distinguish this type of “normal” inflammation from hypersensitivity. In this case, the allergic reaction seems not to remove the parasite, so that the allergenic stimulus persists indefinitely, although a state of tolerance gradually ensues. What is not clear is the timing and location of the sensitizing stimulus: could antigens secreted at the point of skin penetration by L3, or their pulmonary transit, prime the gut for this response, or does it depend on local immune interactions? And does this matter in natural infections, where people are presumably exposed to invasive larvae over prolonged periods?

When A. caninum was incriminated in human eosinophilic enteritis (31, 119, 120), the initial interpretation was that this parasite stimulated excessive inflammatory reactivity because it was in an unsuitable host (29). However, a retrospective analysis of human infection with anthropophilic hookworms indicated that severe enteritis was not so unusual, in extreme cases being comparable with eosinophilic enteritis provoked by the canine hookworm. Furthermore, it seems that the majority of people who develop enteric infection with adult A. caninum remain subclinically infected; i.e., they do not mount a significant inflammatory response (assuming that clinical features reflect the intensity of gut inflammation [118]).

In severe cases of eosinophilic enteritis caused by A. caninum, only solitary worms have ever been recovered from patients, compared with larger numbers in the typical case of anthropophilic worm infection (119). This discrepancy in the numbers of adult parasites might hinge simply on one key difference between the two infections. Most, if not all, canine hookworm L3 that penetrate human skin do not proceed to the gut but settle in tissue repositories from where (only sporadically) individual L3 mobilize to move into the gut. With human hookworms, most if not all L3 go directly to the gut (excluding hypobiotic larvae). In a hypersensitized individual, the quantity of antigens secreted by just one adult worm would be sufficient to trigger severe inflammation, so that the worm burden would be immaterial: either one A. caninum or several or many A. duodenale or N. americanus worms would provoke similar pathological and clinical responses. It also seems that a state of immunotolerance gradually ensues in ongoing intestinal hookworm infection, with a decline in inflammation and symptoms. This does not occur with human A. caninum infections, since the single worm is expelled within weeks (followed by a rapid resolution of inflammation) through mechanisms that might have little to do with immune responses (16, 74, 117). That immunoreactivity might not reflect inflammatory events in the gut is exemplified by one individual who was found harboring a canine hookworm at colonoscopy, had very high levels of specific IgG and IgE antibodies in blood, and yet was asymptomatic (30, 80).

In an experimental study, clinical and immunological responses were monitored in five volunteers who each received a single dose of 50 L3 of N. americanus (86). Their symptoms varied dramatically, with one volunteer experiencing such severe intestinal pain that the infection had to be terminated early. Interestingly, this individual was the only one who failed subsequently to pass hookworm eggs in the feces. For obvious reasons, eosinophil numbers in the gut were not determined; however, the increase in blood eosinophil levels for this patient was unremarkable compared with that for the other four volunteers. It is tempting to assume that the gut inflammatory response successfully removed developing worms from this patient; however, the small sample used in this study means that caution should be used when considering the response to adult worms in the gut, especially since larval rather than adult hookworm antigens were used to measure antiparasite lymphocyte and humoral responses.

It would appear, therefore, that in the majority of infected people, both zoonotic and anthropophilic hookworms develop in the gut without triggering severe inflammation (and ensuing symptoms) whereas a minority of hypersensitive individuals respond intensely, even to just one worm.

Irradiated Larval Vaccine

A highly effective vaccine was developed to control A. caninum infections in canids over 30 years ago. Dogs immunized by exposure to X-irradiated A. caninum L3 were protected against subsequent challenge with normal larvae (93). This immunity could be transferred passively to naive recipients by serum and lymphoid cells (94), but it was not established whether larvae were killed during migration or whether they reached adulthood but were expelled from the gut prior to patency (11). It is also not clear whether vaccination simply drove more larvae into tissue reservoirs rather than allowing tracheal migration and normal intestinal development. The vaccine was developed commercially and yielded 90% protection levels (sufficient to prevent the serious complications of anemia or fatal gut hemorrhage) but was soon removed from the market because pet owners and veterinarians were concerned that sterile immunity was not induced and because manufacturing and transporting irradiated hookworm larvae on a large scale was logistically problematic (91).

Secreted Antigens

Most of the current experimental antihookworm vaccine work utilizes A. caninum in mice, i.e., paratenic hosts, in which larvae migrate somatically. Large inocula (≥1,000) of larvae are lethal, causing extensive pulmonary hemorrhages. However, immunization of mice with smaller doses of L3 protects them against otherwise lethal challenge doses by reducing the numbers of L3 that reach the lungs (152). Given that L3-mediated protection was attainable, larval secretions became the focus of interest, with the subsequent discovery of the Ancylostoma secreted-protein (ASP) family.

Developmentally arrested A. caninum L3 can be stimulated to feed in vitro by exposure to serum (54), with resumption of development being accompanied by secretion of a novel protein, ASP-1 (53), a cysteine-rich protein that has sequence similarity to a family of vespid allergens and to neutrophil inhibitory factor from A. caninum. Vaccination of mice with recombinant ASP-1 led to a 79% reduction in the number of larvae recovered from lungs (45) and stimulated much interest in developing this family of proteins to control helminth infections. Protection is not achieved in mice that do not express the IL-4 gene or in mice that develop a strong Th1 response after DNA immunization, indicating that it is mediated by Th2-dependent mechanisms (59).

ASPs have since been found in other hookworm species (34, 134), as well as other nematode families, including ascarids (142) and filariids (EST sequences deposited in the dbEST database: http://www.ncbi.nlm.nih.gov/irx/dbST/dbest_query.html). Immunization of mice with recombinant ASPs from different hookworm species elicited different degrees of protection, proportional to the extent of amino acid sequence homology between the vaccine ASP and that produced by the challenge A. caninum larvae, ASP-1 (134). The recently published EST data set from adult N. americanus contains nine different cDNA sequences encoding ASP-like proteins (34), which, given their sequence similarity to immunomodulatory proteins, such as neutrophil inhibitory factor (NIF), might be important in regulating inflammation in larval migration and adult attachment sites (see “Neutrophil inhibitory factor” below).

Anticoagulant proteins of hookworms are also of current research interest, both as potential targets for anti-hookworm interventions and as therapeutic agents in clinical medicine. The adult parasites express a range of proteins with potential antihemostatic functions including inhibitors of factor Xa (23, 137), factor VIIa/tissue factor (137), and platelet aggregation and adhesion (26). Other proteins, such as various mechanistic classes of proteases and C-type lectins, that have potential roles in interfering with blood clotting and evasion of immune responses are considered in the next section.

Neutrophil Inhibitory Factor

An almost universal characteristic of tissue helminthic infections is the marked local and systemic eosinophil response. A possible contributor to the small numbers (or perhaps low activity) of neutrophils in hookworm feeding sites is the secretion of NIF, a novel, cysteine-rich, glycoprotein that is secreted by adult A. caninum and potently inhibits CD11b/CD18-dependent neutrophil function in vivo (96). NIF blocks the adhesion of activated human neutrophils to vascular endothelial cells and inhibits their release of H₂O₂. Further studies showed that recombinant NIF bound to the I domain of CD11b/CD18 and, in doing so, blocked the interaction between neutrophils and fibrinogen (97). NIF is being evaluated as a therapeutic anti-inflammatory agent; when administered intravenously as cationic liposomes, it is selectively expressed in the lungs of mice and prevents lung vascular damage after lipopolysaccharide challenge (155). Distant homologues of NIF have recently been reported as ESTs from adult N. americanus (34), although their functional roles have yet to be determined.

C-Type Lectins

C-type lectins (C-TLs) are a family of sugar-binding proteins that act as receptors for glycoprotein ligands. They are found in serum and on the surface of endothelial and immune effector cells (including antigen-presenting cells and T and B cells), where they play pivotal roles in orchestrating the immune system (148). C-TLs with structural and functional similarity to mammalian immune system lectins were recently found in parasitic nematodes, and they represent major surface and secreted products of tissue-dwelling larvae of Toxocara canis (78). These parasite-derived C-TLs might be involved in immune evasion by competing with host lectins for binding to ligands involved in inflammation (77). ESTs encoding C-TLs with sequence similarity to host immune system lectins such as P-selectin and CD23 were recently found in adult N. americanus (34, 77). These proteins might dampen the local inflammatory responses to parasites feeding in the gut. Snake venom contains C-TLs that inhibit blood clotting by binding to coagulation factors (5, 139); hookworms and other hematophagous helminths might also use similar mechanisms. We are presently investigating the possible roles of hookworm C-TLs in immune evasion and other physiological processes.

Anticoagulant Peptides

The anticoagulant properties of hookworms were first reported almost 100 years ago (72), and in 1966 a comprehensive review of hookworm-induced anemia was published (127). At the time, it was known that hookworms secreted molecules that induced anticoagulation to assist feeding and that the inability of the host to form a clot at the site of attachment resulted in anemia if worm numbers were great enough. In more recent years, significant progress has been made in dissecting the molecular mechanisms by which hookworms interfere with the clotting cascade. Metalloproteases of A. caninum that interfered with clot formation were identified and cloned (58, 60; B. F. Jones and P. J. Hotez, Abstr. Am. S Trop. Med. Hyg. Annu. Meet., p. 262 1999).

More recently, numerous families of small, anticlotting peptides have been found in A. caninum. AcAP is an 8.7-kDa peptide secreted by adult A. caninum which has no homology to other anticoagulants but is a potent inhibitor of factor Xa (23). Subsequently, a family of related peptides were identified (137), one of which uniquely inhibited a complex of blood coagulation factor VIIA and tissue factor. In addition to inhibitors of coagulation factors, adult A. caninum hookworms secrete a protein that inhibits the adherence of platelets to fibrinogen and collagen (26). Anticoagulant proteins such as those described here provide novel targets for the development of antihookworm vaccines and specific inhibitors, as well as the development of new antithrombotic therapies for human and veterinary medicine.

Proteases

Proteases are instrumental in clotting (and anticlotting) cascades but also play immunoevasive roles in some parasitic infections. The first protease from hookworms to be described by mechanistic class was a zinc-dependent, elastin-cleaving metalloprotease from A. caninum (see “Anticoagulant peptides” above). Metalloproteases in ES products of adult N. americanus are important in immunomodulation by cleaving eotaxin and inhibiting eotaxin-mediated recruitment of eosinophils to sites of inflammation in vivo (33).

Cysteine proteases are abundant in nematodes, and potent cathepsin L-like activity was detected in ES products of adult A. caninum (38). The cDNAs encoding these proteases were cloned, and their sites of expression were localized to esophageal, amphidial, and excretory glands of adult worms (51). Cysteine proteases have been implicated as allergens in atopic disorders (46, 49), and patients with enteric A. caninum or anthropophilic hookworm infections mount IgE responses to a cysteine protease-like molecule (75, 80). Proteolytic allergens might interfere with antigen processing, and research into house dust mite allergens explains at least one mechanism. Dust mite cysteine proteases, such as Der p I, not only are potent allergens but also can selectively cleave the low-affinity Fcε receptor, CD23, from the surface of B cells (57). Consequent generation of soluble CD23 and the removal of negative-feedback signals for IgE might promote IgE synthesis and enhance type I hypersensitivity. Hookworms secrete cysteine proteases, and patients infected with N. americanus have elevated levels of soluble CD23 (110), suggesting parallel mechanisms.

Cleavage of Igs seems an obvious tactic in immune evasion. Cysteine proteases secreted by the parasitic platyhelminths Fasciola hepatica (14) and Spirometra mansoni (66) specifically cleave IgG molecules at the hinge region. ES products of N. americanus include proteases capable of cleaving IgA (to a greater extent than other Ig isotypes) to yield Fab fragments that can block complement or phagocyte attack mediated by intact IgG or IgM (107). This finding is especially interesting, since secretory IgA seems to play an important role in triggering the degranulation of eosinophils (1).

Aspartic proteases are another class of proteolytic enzyme found in the secretions and intestines of parasitic helminths, and they have been implicated in hemoglobin digestion (73). Extracts of larval and adult N. americanus contain aspartic protease activity (21, 22), and an EST encoding a pepsinogen-like aspartic protease was recently found in adult worms (34). In addition, a cDNA encoding a cathepsin D-like aspartic protease was cloned from A. caninum (50). Computer modeling of this protein (AcASP1) predicted that it would cleave canine hemoglobin more efficiently than it cleaved human hemoglobin (16). Recently, recombinant Schistosoma japonicum aspartic protease was shown to cleave Igs at the hinge region as well as the complement protein C3 in vitro (146). While an immunoevasive role has not been investigated for hookworm aspartic proteases, it is probable that they perform multiple functions in vivo.

Protease Inhibitors

Proteases occur in all life-forms and are involved in virtually all essential physiological processes; salient here are their roles in mammalian nutrition, antigen processing and presentation, and programmed cell death. Specific protease inhibitors can have major physiological effects. Serine protease inhibitors (serpins) of viruses can inhibit the release of proinflammatory cytokines (125), and a serpin from blood-dwelling microfilariae of the lymphatic parasite Brugia malayi specifically inhibits neutrophil cathepsin G in vitro (154). Recently, a Kunitz-type serpin termed AceKI-1 was cloned and expressed from A. ceylanicum (95). AceKI-1 was found in ES products of adult worms and was a potent inhibitor of mammalian serine proteases involved in digestion (trypsin, chymotrypsin) as well as immunity (neutrophil cathepsin G), suggesting multiple potential roles for this inhibitor in vivo (95). A trypsin inhibitor with sequence similarity to potassium channel-blocking toxins was identified among the Necator EST data set (34), and ESTs from A. caninum encoding inhibitors of trypsin and chymotrypsin have also recently been deposited in dbEST. Like serpins, inhibitors of cysteine proteases (cystatins) have immunomodulatory properties, and a cystatin from the filarial parasite Acanthocheilonema vitae downregulates T-cell proliferation and enhances IL-10 production (52). Furthermore, an abundantly expressed cystatin from B. malayi (48) interrupts antigen processing (83). Cystatins have not been found in hookworms but are very likely to be expressed by larvae and/or adults, especially given the recent deposition in GenBank of a cystatin-encoding cDNA from the related blood-feeding trichostrongyle nematode Haemonchus contortus.

Antioxidants

Nematodes secrete antioxidant enzymes that might protect exposed cuticular surfaces by neutralizing host free radicals generated by the oxidative burst of leukocytes. Genes encoding superoxide dismutases (SODs) have been cloned from a range of parasitic nematodes (62, 71, 140, 142). Extracts of N. americanus contain Cu/Zn SOD activity (138), although cDNAs encoding this family of proteins have yet to be identified from hookworms. While their contribution to parasite protection is widely accepted, it has also been suggested that nematode SODs (such as those secreted by hookworms) play an offensive rather than defensive role (19). Another family of antioxidants secreted by parasitic helminths are glutathione S-transferases (GSTs) (17), which are thought to neutralize lipid hydroperoxides (20). GSTs have been purified from extracts and ES products of N. americanus and A. ceylanicum (18), but their gene sequences are not available. Although their biological roles in parasite infections are poorly understood, GSTs are of interest in helminth vaccine research, given the high levels of protection they induce against trematode infections (135, 151).

Acetylcholinesterase

Having cholinergic motor neurons, both free-living and parasitic nematodes are likely to use acetylcholinesterase (AChE) to hydrolyze acetylcholine. The free-living species Caenorhabditis elegans expresses multiple AChE isoforms (47), one of which is secreted by muscle cells and is responsible for coordinated movement (56). However, many parasitic nematodes, including hookworms, secrete AChE from esophageal and amphidial glands (88, 111), and the enzyme is highly immunogenic in necatoriasis (101). Numerous roles have been proposed for nonneuronal AChE in nematode infections, including inhibition of host gut peristalsis, regulation of enteric secretory mechanisms (61), and immunomodulation (107). The binding kinetics of AChE purified from ES products of adult N. americanus have been analyzed in some detail (109). AChE activity has also been detected in Ancylostoma spp. (4, 100) but at significantly lower levels than those in Necator (147). All of these hookworms have comparable effects on host gut motility, suggesting that secretory AChE might not suppress gut motility but serves an as yet unknown function (118). AChE genes have been cloned from the rodent parasite Nippostrongylus brasiliensis (61) and from plant parasitic nematodes (105) but not yet from hookworms.

Six-Cysteine Domains in Hookworm Proteins

Formerly known as NC6, six-cysteine (SXC) domains are evolutionarily mobile cassettes consisting of about 36 amino acids with 6 perfectly conserved cysteine residues. The motif usually occurs as a tandem pair of SXC domains at either the N or C terminus or both termini of the protein. SXC domains were first found in a lipid-binding protein in the infective larval surface coat of Toxocara canis (43), and the only homologs that could be found at the time were predicted proteins from C. elegans. SXC domains have since been identified in other nematode proteins, including mucins from T. canis tissue larvae (44, 76, 84), and in ESTs from B. malayi and over 70 ESTs from C. elegans, where they are found fused to metalloprotease and tyrosinase domains (15). The role of the SXC motif is unknown, but all nematode proteins containing this cassette possess a signal peptide (where the 5′ end of the gene has been obtained), suggesting an extracellular role. This is further supported by expression of SXC-containing proteins on the surface of T. canis infective larvae (44, 84). A recent EST survey of N. americanus cDNAs identified three clusters encoding different SXC-containing proteins (34). These proteins were unique for several reasons: they consisted of a signal peptide followed by a single SXC domain, as opposed to a tandem pair, and they did not accompany other protein domains of known function. The only known nonnematode peptides with a domain structure resembling SXC motifs are from sea anemone toxins (35), and, as in hookworms, the majority of these proteins consist of a signal peptide followed by a single SXC domain. Some of these anemone proteins potently block potassium channels of T cells (65) via a functional Lys-Tyr diad. While this diad is absent from the Necator homologs, it is reasonable to suspect that hookworm SXC-containing peptides play an analogous role in suppressing T cells during chronic hookworm infections. ES products from N. americanus induce apoptosis of T cells in vitro (27), and SXC peptides might be involved in suppressing and/or killing host cells. In addition, given that SXC-containing proteins are expressed at the parasite surface in T. canis at least, proteins containing this domain in hookworms are likely to be exposed to the immune system, and so they deserve further investigation.

Calreticulin

The calcium-binding protein calreticulin, ubiquitously expressed in the cells of higher organisms, has been proposed to have numerous functions, including molecular chaperoning (98), calcium signaling (90), and integrin-mediated signaling and cell adhesion (28). Probably found in the secretions of N. americanus, calreticulin was recently shown to be the major allergen recognized by IgE from patients infected with the parasite (108). Its role as an extracellular secreted product of nematodes remains speculative, but some intriguing immunomodulatory properties can be extrapolated from studies with mammalian calreticulin. First, it can bind to C1q of the complement system (39), and, second, it has lectin-like properties (67), which might be exploited by secreted parasite homologs to bind to and cross-link cell surface glycoproteins involved in the regulation of IgE production (108). Given the lectin-like properties of calreticulin, it might also be involved in skewing the immune response towards the Th2 phenotype (see “C-type lectins” above).

Since investigations in the field of hookworm secretory molecules are still in their infancy, numerous intriguing findings are expected in the near future.