Molecular Approaches to Vaccinating against Hookworm Disease

Abstract

Anthelminthic drug chemotherapy has failed as an acceptable approach to hookworm control in the less developed countries of the tropics. The development of a genetically engineered vaccine against hookworm infection would be a major advance in our efforts to control this parasitic disease. We have produced several lead recombinant hookworm vaccine antigens. Their development is based on scientific principles that were generated almost 70 years ago when investigators first began to attenuate living infective hookworm larvae. Those early studies on attenuated live vaccines highlighted the importance of secreted larval antigens for eliciting protective immunity in dogs challenged with Ancylostoma caninum. The two major secreted larval antigens have been recently identified as Ancylostoma secreted protein-1 (ASP-1) and ASP-2. The predicted amino acid sequences of the ASP cDNAs together with experimental immunogenicty data using the expressed recombinant protein suggest that the ASPs are promising vaccine antigens. Preliminary hookworm challenge data in mice immunized with recombinant ASP-1 helps to validate this assumption. Alternative vaccines based on either genetic immunization (DNA vaccines) or immunization with recombinant molecules expressed from adult hookworm cDNAs are also under evaluation. Optimization of vaccine route, delivery system, and adjuvant formulations will be required before future planned phase I testing in humans. Vaccine development for a target population living in rural areas of less developed countries will require innovative solutions to financing and manufacture.

HOOKWORMS AS A GLOBAL PUBLIC HEALTH THREAT

The hookworms are major causes of anemia and malnutrition in the developing nations of the tropics. Most global estimates rank hookworms among the most prevalent infectious agents of humans-more than 1 billion individuals are infected with these parasites⁽¹⁾. In China alone, reliable estimates indicate that 194 million individuals harbor hookworms⁽²⁾. Although hookworms cause approximately 55 000 deaths per year^{(3, 4)}, a far greater number of people suffer from the chronic effects of hookworm disease. There is a growing awareness that the major human hookworms, Ancylostoma duodenale andNecator americanus, exert their most detrimental impact on the health of mothers and children^(5–7). The pathogenic effects from hookworms are caused directly by parasite-induced blood loss. Hemorrhage occurs from capillaries torn at the site of hookworm attachment in the mucosa and submucosa of the small intestine. The process is aided by the release of parasite-specific anticoagulants^{(8, 9)}. Chronic hookworm-associated blood loss during moderate and heavy infections causes physical, intellectual, and cognitive retardation in children⁽⁵⁾. Many of these clinical features are directly attributable to the chronic effects of iron deficiency^{(10, 11)}; in some instances these deficits are irreversible⁽¹²⁾. Plasma protein losses also contribute to the pathogenesis of hookworm disease and associated malnutrition. As a consequence of iron deficiency and protein losses, children are rendered susceptible to intercurrent viral and bacterial infections⁽⁵⁾. Chronic hookworm infection prevents children from achieving their full potential to become productive individuals in later life. Hookworms are central to the downward spiral of malnutrition and rural poverty in less developed countries.

RATIONALE FOR A HOOKWORM VACCINE

For reasons that are not well understood, humans have difficulty mounting an adequate natural effector immune response against adult hookworms in the intestine. There is a common wisdom that humans cannot acquire sterile immunity to hookworms, and individuals living in endemic areas harbor worms most of their lives^(13–16). Indeed, in human volunteers, intestinal hookworms do not induce prominent T-cell-and B-cell-dependent immune responses^{(17, 18)}. The inadequate human immune responses to hookworms may be partly explained by the parasite's own unique host immune evasion mechanism⁽¹⁹⁾, whereby the adult hookworm releases immunosuppressive molecules at its site of gut attachment^{(20, 21)}. However, it would also be misleading to say that host immune responses to hookworms are absent entirely. Early investigators noted that human populations that were frequently and regularly exposed to infective hookworm larvae could somehow regulate the numbers of hookworms in their gastrointestinal tract^{(22, 23)}. Except in the case of infantile hookworm infection, resistance to acquiring overwhelming worm burdens appears to develop at some operational level.

Nonetheless, reinfection with hookworms almost always occurs several months after specific anthelminthic administration^{(7, 19, 24)}. This observation has been attributed to both the absence of specific immunity to hookworms as noted above, as well as some circumstantial evidence for genetic regulation of hookworm populations in humans⁽²⁵⁾. By whatever mechanism, high rates of reinfection with hookworms are a major reason why the widespread or targeted administration of anthelminthic agents has failed to control hookworm populations in endemic areas of less developed countries. Reinfection will occur in just a few months after “chemical deworming” in most endemic areas. In China and other countries with large hookworm problems, the disappointing experience with anthelminthic chemotherapy has led many government health officials to the conclusions that this approach is, by itself, futile. Alternative solutions composed of control programs that do not rely predominantly on anthelminthic drugs are therefore urgently needed. With this goal in mind, we have proposed vaccinating large populations with a genetically engineered recombinant hookworm vaccine.

LIVE ATTENUATED LARVAL HOOKWORM VACCINES

By 1940 it had been shown that it is possible to overcome the absence of a vigorous, naturally acquired immune response to hookworms by eliciting a protective response through active immunization. This realization was based on a body of investigative work beginning in 1920 that used a canine animal model and the dog hookworm Ancylostoma caninum^(26–28) (Table 1). Canine infections with A. caninum resemble human infection with A. duodenale because the third-stage infective larvae of both species1) are acquired by their definitive hosts through oral ingestion or skin penetration, 2) can undergo a period of arrested development in host tissues, and 3) may be transmitted vertically to newborns to cause perinatal infection⁽²⁹⁾. It has been proposed thatA. duodenale and A. caninum are closely linked in evolution and that differentiation of these two species occured relatively late in hominid evolution in association with the domestication of wild canines. This hypothesis may account for the observation that cross-infection with A. caninum causes a rare zoonosis in humans^{(30, 31)}. In contrast, the other major hookworm of humans, N. americanus, causes considerably less blood loss thanAncylostoma and does not infect via the oral route or undergo arrested development.

Table 1 Milestones in larval hookworm vaccine development

Work on the vaccinology of hookworm first began in the 1930s when infective larvae of A. caninum were shown to be antigenic⁽³²⁾, and protective immunity was elicited by administering repeated doses of hookworm larvae^{(33, 34)}. This work followed earlier demonstrations of immunity to other animal nematode infective larvae^(35–38). Host immunity was elicited by administering multiple doses of infective larvae over a period of weeks to months. The inoculations precipitated a“crisis,” a term referring to the time of maximum egg production and blood loss caused by adult hookworms. After the crisis, such animals tolerated test doses of 8 000 to 50 000 larvae without anemia. Variability in host immunity among the outbred dogs was partly reduced by controlling for the nutritional status of the canine host^{(39, 40)}. Ultimately, the conditions for reproducing canine immunity against hookworm larval challenges in a reliable manner required 12 y of intense investigative efforts^{(23, 41, 42)}. Even then, specific acquired immunity to canine ancylostomiasis was not absolute, but instead was demonstrable by diminished intensity of infection after larval challenge. Thus it was noted that resistance induced by graded doses of larvae was manifested by 1) reduction in worm burden, 2) delayed development of the worms, 3) reduced size of the worms developing, and 4) lower fecundity⁽⁴³⁾.

Early attempts to understand the mechanisms underlying the effector immune responses to A. caninum infective larvae suggested that both cellular immune responses and humoral antibody responses were in operation⁽⁴⁴⁾. The humoral immunity to A. caninum, as well as other nematode larvae, was determined to be directed against secreted antigens around the oral opening and excretory pore of the infective larvae^{(45, 46)}. The so-called “precipitin reaction” against antigens secreted from larvae in vitro was named after its discoverer, Merritt Sarles, and became known as the“Sarles phenomenon.” Subsequently, sera from children with low hookworm burdens despite repeated larval challenges were shown to exhibit the Sarles phenomenon^{(47, 48)}. The implication of this observation was that antigens secreted by living hookworm larvae are capable of stimulating protective immunity^{(23, 49)}.

Based on the success of eliciting acquired resistance with living infective larvae of A. caninum, efforts were made to attenaute hookworm larvae by ionizing radiation^(50–52). Single and double vaccination schedules with x-irradiated larvae protected vaccinated pups against the establishment of potentially severe challenge infections by reducing worm burdens 37% and 90%, respectively⁽⁵²⁾. Immunity generated by the irradiated vaccine was transferable by a combination of serum and lymphoid cells⁽⁵¹⁾. Laboratory experiments with the irradiated vaccine ultimately led to the industrial development and commercial introduction by 1974 of a gammairradiated attenuated A. caninum vaccine. Despite its success both in the laboratory and in the field at preventing canine ancylostomiasis, the vaccine did not succeed commercially and was taken off the market in 1975⁽⁵³⁾.

GENETICALLY ENGINEERED LARVAL HOOKWORM VACCINES

Immunizing children with live attenuated hookworm larvae is not a practical strategy for hookworm control⁽⁵⁴⁾. However, the obstacles that prevent the development of a live hookworm vaccine could be overcome if an equivalent protective effect could be reproduced by substituting a genetically engineered vaccine. We have used some of the information generated from the development of attenuated infective larval vaccines to design genetically engineered vaccine candidates against the hookworm A. caninum. Because the live larval vaccine highlights the importance of secreted larval antigens in eliciting protective immunity, we have focused our attention on these molecules. To biochemically characterize the secreted larval antigens, we have modeled their release in vitro by maintaining the larvae under simulated host-like conditions for 24 h^{(55, 56)}.

The release of hookworm larval antigens appears to be tightly coupled to the developmental biology of the parasite. Before host entry, the environmental stages of infective hookworm larvae are nonfeeding and developmentally arrested⁽⁵⁷⁾. A. caninum larvae will resume feeding and development upon entry into a permissive host. Antigen release temporally coincides with the resumption of larval development. These events can be modeled in vitro when larvae are stimulated to resume development by exposing them to dog serum or a low-molecular-weight ultrafiltrate of dog serum at 37 °C^{(55, 56)}. Upon receiving the host-specific trigger, we determined that hookworm larvae release into the culture medium two predominant molecules known as Ancylostoma secreted proteins (ASPs). ASP-1 is the major secreted protein, whereas ASP-2 is a less abundant constituent. Our strategy for cloning ASP cDNAs is based on identifying unique internal peptide sequences obtained from the natural products⁽⁵⁸⁾. A degenerate primer derived from a portion of this sequence, together with a primer complementary to flanking vector sequence, amplifies a polymerase chain reaction product from phase λ DNA containing an A. caninum infective third-stage larval cDNA library. This product was used as a probe to isolate cDNA clones encoding ASPs from theA. caninum cDNA library. The 5′ end containing the ATG (Met) start codon is isolated from third-stage infective larval cDNA either by 5′ RACE or by taking advantage of the observation that some nematode mRNAs contain conserved 5′ spliced-leader sequences that are added during trans-splicing⁽⁵⁹⁾. We have previously identified spliced-leader sequences from A. caninum infective third-stage larval mRNAs⁽⁵⁹⁾.

The full-length cDNA for ASP-1 encodes an open reading frame of 424 amino acids, with a predicted molecular weight of 45 735 and a pI of 7.13, and a 34-bp 3′ -untranslated region⁽⁵⁸⁾. The entire sequence of the original peptide purified from larval secretory natural products was present in the open reading frame (aa 255 to 267), confirming that this cDNA encodes the ASP-1 molecule. The amino terminal 18 amino acids are highly hydrophobic and contain a potential eukaryotic signal sequence cleavage site⁽⁵⁸⁾, consistent with the secretory nature of the protein. Database comparison (SwissProt) revealed significant homology between the C-terminal 215 amino acids of the ASP-1 deduced amino acid sequence and the antigen 5/antigen 3 family of molecules fromHymenoptera venoms^{(60, 61)}, and a family of cysteine-rich secretory proteins called CRISPs⁽⁶²⁾(Fig. 1). ASP-2 exhibits similar homologies (data not shown). Because of the apparent homology to Hymenopteran venom antigen 5 (Ag 5), a polyclonal antiserum against Ag 5 from the yellow jacket Vespula squamosa was tested for its ability to recognize secreted and recombinant ASP. Vesq Ag 5 antiserum cross-reacted with both the recombinant and native ASP molecules⁽⁵⁸⁾.

Figure 1

Comparison of ASP-1 deduced amino acid sequence with the amino acid sequences of selected homologous molecules⁽⁵⁸⁾. Shaded residues are identical to those in ASP-1, and position marked with an asterisk (^*) are conserved in all molecules. Numbers in parentheses the represent the percent identity with ASP-1, calculated using pairwise comparisons between ASP-1 and the homolog using the GAP program of the Wisconsin Genetics Program. Solin, antigen 3 from the red imported fire ant, Solenopsis inviata; Vessq, Ag 5 from the yellow jacket V. squamosa; Htpx, human tests-specific protein; Rscg, rat sperm-coating glycoprotein; Helo, helothermine from Mexican beaded lizard (Helodema horridum horridum) salivary gland; Prla, tobacco-pathogenesis-related protein la precursor. The relationship between the insect venom antigens and Htpx, Rscg, and Prla is an interesting one that has been observed and commented on^{(60, 61)}. Helothermine belongs to the CRISP family⁽⁶²⁾.

Because ASP-1 and ASP-2 are released by the parasite in a manner suggesting that they might be the Sarles antigen, and because of their predicted antigenicity (on the basis of homology to Ag 5), we sought to determine whether immunization with the ASPs might protect animals against experimental challenge infections with A. caninum. In mice, rabbits, and dogs, recombinant ASP-1 (rASP-1) was found to be immunogenic as an alum precipitate or in the presence of adjuvants such as complete Freund's adjuvant orCorynebacterium parvum. Pilot canine immunogenicity and safety testing were carried out using Escherichia coli expressed rASP-1. The molecule used for these studies contained a polyhistidine tag and was soluble only in the presence of denaturing agents such as SDS. Thus, there are concerns that the molecule used for these studies does not conformationally resemble the native soluble molecule. For that reason, it was not the major goal of this study to evaluate rASP-1 for its ability to protect against larval challenge. Nevertheless, when the immunize dogs were challenged with infective larvae of A. caninum, several biologic effects were evident. First, no immediate hypersensitivity reactions were observed in the immunized animals. Second, immunization with the denatured rASP-1 showed a trend to reduce the numbers of the hookworms in the intestine compared with the control dogs. Finally, immunization with rASP-1 appeared to reduce the viability of eggs from the female hookworms, so that fewer viable eggs are shed from immunized animals. These data suggest the possibility that transmission of hookworm might also be interrupted.

We found that the canine challenge model is impractical for early stage vaccine development. The number of animals used in our pilot study was too small to adequately evaluate rASP-1 as a protective antigen, so that although there is a promising trend with regard to a vaccine effect, it is premature to make definitive statements about efficacy in this animal model. Unfortunately, the expense of purchasing and housing dogs, as well as space limitations, prevents us from immunizing sufficient numbers of animals. As noted above, there are also concerns about adequate standardization of the canine model. To circumvent the difficulties associated with canine immunizations, we have developed a nonpermissive mouse model of A. caninum infections. Mice infected with A. caninum abort their infection during larval migration through the lungs. Kerr⁽⁴³⁾ reported that although large numbers of larvae migrating through the lungs of mice will produce a fatal hemorrhagic pneumonitis, sublethal inoculations with smaller numbers of larvae will subsequently protect mice against otherwise fatal challenge doses. Therefore, recovery of larvae from the lungs could be useful as a surrogate end point for evaluating hookworm immunity even though it will not permit us to measure hookworm blood loss.

We determined that infective larvae of A. caninum will enter the lungs of 8- to 10-wk-old outbred mice. The maximum number of larvae are recoverd from the lungs 48-60 h after oral infection. These mice develop a hemorrhagic pneumonitis similar to hookworm pneumonia in humans. To determine whether rASP-1 immunization protects mice against challenge infections, the mice were immunized with 20 μg of recombinant antigen, either as an alum precipitate or in solution with dried C. parvum. The mice were then boosted twice at 2-wk intervals. ELISA testing of serum from the mice revealed that both alum-precipitated rASP-1 and rASP-1 with C. parvum were immunogenic and could elicit antibodies. One week after the final boost, the immunized and control mice were challenged orally with 500 infective larvae ofA. caninum. The lungs were harvested at 50 h and the larvae counted. A significant vaccine protective effect was observed: Immunization with rASP-1 and C. parvum resulted in an 63-83% reduction in larvae compared with C. parvum alone. Similarly, immunization with alum-precipitated rASP-1 resulted in an 79-82% reduction in worms recovered compared with PBS alone. These preliminary data help to validate rASP-1 as a promising vaccine antigen. Confirmatory studies are in progress. In addition, studies to exclude the possibility that antibodies to rASP-1 act to only delay larval entry into the lungs are also pending.

GENETIC IMMUNIZATION (DNA VACCINES)

Because E. coli-expressed rASP-1 is not soluble without denaturing agents, we are exploring the possability that expression in a eukaryotic vector might produce rASP-1 that conformationally resembles the native molecule. Alternatively, genetic immunization might obviate the need to express soluble ASP-1. Genetic immunization takes advantage of the fact that the antigens synthesized de novo after DNA injection are both processed within the cell and assembled with MHC molecules. Class I MHC assembly facilitates recognition and stimulation by cytotoxic T lymphocytes, whereas class II MHC assembly stimulates helper T cells and humoral immunity. Under some circumstances, the Th1 subset of CD4+ cells appears to be preferentially induced by naked plasmid DNA in saline immunization. In some cases, plasmid DNA immunization will elicit predominantly Th1 responses even when immunization with the corresponding native protein will not.

Although there is no direct evidence that Th1 responses mediate resistance to hookworm infections, recent studies using attenuated larval schistosome infections suggests that these responses are important for mediating vaccine immunity. Immunization of mice with cobalt-60-irradiated Schistosoma mansoni cercariae will eliminate up to 80% of the worms in a challenge infection. The resistance associated with the irradiated-cercarial vaccine is associated predominantly with the Th1 responses. Indeed, augmentation of the murine Th1 response either through modulation by cytokine-specific antibodies or by coadministration of IL 12 will enhance vaccine-induced immunity toS. mansoni cercariae in mice^{(63, 64)}. Because vigorous Th1 responses might also elicit similar protective responses against attenuated larvae of A. caninum, vaccine efficacy may be improved through Th1-specific stimulation. Th1 responses to ASP-1 could be generated by genetic immunization (DNA vaccines) with ASP-1 cDNA in a plasmid containing viral promoter-enhancer. By so doing, we will obviate the requirement of purified native antigens and instead will induce the host cell to produce antigens de novo. Our assumption is that the generation of Th1 immune responses and cytotoxic T lymphocytes along with de novo antigen production are two of the major features of attenuated worm vaccines, which confer strong protection. Because we hypothesize that rASP-1 immunization will elicit protective responses that are similar to those of the earlier generation attenuated larval vaccines, we propose to optimize vaccine conditions based on this paradigm.

ADULT HOOKWORM ANTIGENS AS VACCINE TARGETS

The concept of using immunotherapy to target the adult stages of hookworms has also been examined. Asa Chandler⁽⁶⁵⁾ was the first to champion the idea that enzymes, such as those released by hookworms at the site of attachment, could elicit specific antienzyme antibodies. The antibodies would then serve to neutralize important survival functions for the parasite⁽⁶⁵⁾. In 1953, R. E. Thorson^{(66, 67)} provided some proof of this concept when he showed that serum from puppies immunized with infective larvae of A. caninum could neutralize a protease from the adult stage. Presumably, at least one of the hookworm proteases contains antigenic determinants that are shared between the larval and adult stages^{(31, 49, 68–73)}. In addition to the protease, other adult hookworm-derived enzymes are under evaluation as potential vaccine candidates including an acetylcholinesterase^(74–76), a superoxide dismustase⁽¹⁴⁾, and a hyaluronidase⁽⁷⁷⁾.

In addition to the enzymes activities mentioned above, there are two recombinant proteins recently cloned from adult hookworm cDNAs that are under evaluation as vaccine candidates. These include 1) hookworm-derived neutrophil inhibitory factor, a 41-kD antinflammatory molecule released byA. caninum that binds to the receptor CD11/CD 18^{(20, 21)} and 2) A. caninum anticoagulant peptide (AcAP). AcAP is a small, approximately 8.7-kD peptide that specifically inhibits the serine protease factor Xa^{(8, 9)}; it is one of the most potent anticoagulant peptides described to date. One strategy currently under investigation is to elicit specific host antibodies to AcAP to block the anticlotting activities of adult hookworms at the site of attachment. This approach would target the hookworm-induced blood loss as an approach to vaccinating against hookworm disease.

Still another approach to vaccinating against adult-stage nematodes is under investigation by E. A. Munn, who identified antigens lining the alimentary canal of the sheep stomach worm Haemonchus contortus⁽⁷⁸⁾. One of these worm gut antigens, called H11, is a membrane glycoprotein that functions as a leucine aminopeptidase. Munn is exploring the concept that these membrane-bound proteins are ordinarily“concealed” from the host humoral immune response because of their location in the parasite. Immunization with these concealed or“hidden” antigens has been shown in preliminary trials to partially protect against sheep Haemonchus infection.

OBSTACLES TO VACCINE DEVELOPMENT FOR THE THIRD WORLD

We have described some preliminary success at identifying viable recombinant vaccine targets against human and animal hookworm infections, including the ASPs and AcAP. Our major goal in this work is to reproduce the immunity elicited by the earlier generation of living attenuated larval vaccines by replacing the larvae with chemically defined recombinant macromolecules. Genetically engineered vaccines, unlike vaccines containing living worms, offer promise as a means to immunize large populations in hookworm endemic areas like China and India^{(7, 79)}. Our goal in immunizing these populations is to reduce hookworm-associated blood loss and anemia either by reducing the number of worms that enter the intestine or by directly interfering with hookworm attachment and anticlotting mechanisms. With regard to the former, we are in the early “proof of concept” stage and are conducting animal immunizations with ASP-1 genetically engineered from infective third-stage larvae. Immunization with AcAP and other recombinant molecules from adult hookworms offers a promising second approach.

If the mouse vaccine model is successful, we will be able to take advantage of the vast amount of information about murine immune responses, as well as a general availability of murinespecific immunoreagents such as recombinant cytokines, cytokine-specific antibodies, and cDNA probes. Because the field of modern immunology has ignored the hookworm, there is a paucity of knowledge about the effector immune responses against hookworms and the cytokine profile of effector cell populations. Without this information, however, it is difficult to knowledgeably choose specific immunization parameters such as antigen dose and route, which might otherwise strongly bias either a Th1 or Th2 host response⁽⁸⁰⁾. The absence of specific information about hookworm immune responses has also made it difficult to design rational approaches for choosing among a wide variety of adjuvants, immunomodulators, and vaccine delivery systems now available⁽⁸¹⁾. Many of these problems could be solved by investigating mouse hookworm infections and by their manipulation through the use of specific immunoreagents and adjuvant formulations. Once vaccine parameters are optimized in the mouse, they could be tested in dogs.

There are ethical, socioeconomic, and political hurdles that must be overcome before vaccines for parasitic diseases such as hookworm are used in less developed countries. Among them will be whether to first carry out phase I testing in the United States or Europe, where it can be overseen by an established government regulatory agency, and whether even the earliest steps of vaccine evaluation should be undertaken in a nation where hookworm is an imporatant public health problem. The high costs of vaccine development and production are also major obstacles. For the most part, the pharmaceutical industries in the United States and Europe have not had the economic incentive to take on vaccine manufacturing for the less developed countries, and international agencies such as the United Nations are not equipped for this task⁽⁸²⁾. In contrast, there is considerable economic interest in the development of nematode vaccines for animal health. Our finding of ASP-related molecules in animal nematode parasites(Fig. 2) suggests that we could develop veterinary vaccines to help underwrite the costs of a human hookworm vaccine. However, Asian countries with rapidly growing economies also recognize the economic impediaments caused by human hookworm infection and related diseases of poverty and indequate sanitation. Therefore, we anticipate that China and other Asian countries will pioneer vaccine technology for parasitic diseases. The Yale Medical Helminthology Laboratory and the Institute of Parasitic Diseases of the Chinese Academy of Preventive Medicine are working together toward that goal.

Figure 2

Western blot detecting ASP-1 in A. caninum and another soil-transmitted parasitic nematode. Lane 1: L₃whole ext., extracts of third-stage infective larvae of A. caninum (10 μg of protein); lane 2: L₃ ES, excretory-secretory products from 5500 host-stimulated third-stage infective larvae of A. caninum; lane 3: rASP (0.8 μg of protein) fromA. caninum; lane 4: L₃ whole ext., extracts of third-stage infective larvae of H. contortus (5 μg of protein).