DOES PROTECTIVE IMMUNITY TO E. HISTOLYTICA ARISE AFTER NATURAL INFECTION?

Many of the most successful vaccines have been developed for diseases, such as smallpox or measles, where natural infection leads to lifelong protective immunity. Where natural infection leads to little or no long-term protection (e.g. HIV infection), vaccine development may be much more difficult, the concept of ‘trying to improve upon nature’. For this reason, and for the purposes of identifying correlates for protective immunity, it has been a high priority to try and characterize the nature of immunity to amoebiasis after infection. Clinical studies in this area have given mixed results. Studies in Bangladesh suggest that children with mucosal IgA antibodies to the surface Gal/GalNAc lectin of E. histolytica show a reduced risk of acquiring a new infection with E. histolytica over the next year (Haque et al. 2001, 2002). However, it is worth noting that immunity was short lived and 20% of children in the cohort had a second episode of E. histolytica infection during the study period (Haque et al. 2001; Stanley, 2001). Independently, a second group, studying patients from S. Africa, reported that individuals with a prior history of amoebic liver abscess and mucosal antibodies to the E. histolytica lectin, were less likely to become infected with the related Entamoeba, E. dispar, over the follow-up period (Ravdin et al. 2003). Taken together, these studies suggest that the presence of mucosal antibodies to the amoebic lectin is associated with a reduced risk for reinfection with E. histolytica or E. dispar, but raise important questions about the efficacy and sustainability of immunity. A very different scenario has been reported where individuals in Vietnam with a prior episode of amoebic liver abscess showed the same or slightly higher risk for a second case of amoebic liver abscess (Blessman et al. 2002). This is particularly interesting, because individuals with amoebic liver abscesses produce high quantities of E. histolytica-specific serum antibodies, and develop T cell proliferative responses that recognize amoebic antigens (Jackson, Gathiram and Simjee, 1984; Velazquez et al. 1995). In addition, while work in a gerbil model of amoebic liver abscess indicates that Th1-mediated immunity plays a role in protection against amoebiasis (Campbell and Chadee, 1997; Gaucher and Chadee, 2003), there is little data in humans to suggest this is the case. If T cell mediated immunity, and particularly Th1 responses, were important in host defence against amoebiasis, one would predict that individuals with AIDS would be at increased risk for either acquiring disease or developing more severe disease than normal hosts (as is the case with Leishmania infections where Th1 is critical in host defence) (Locksley & Louis, 1992). However, despite some anecdotal reports, individuals with AIDS appear to be at no increased risk for amoebic disease, even though they have a higher rate of colonization with E. histolytica (Moran et al. 2005). To summarize, while there is evidence indicating that reinfection rates may be lower in individuals with mucosal antibodies against a surface antigen of amoeba, it has not been established that any long-term natural immunity to disease, either amoebic colitis or amoebic liver abscess develops after infection. In addition, most of the available data indicates that individuals with even the most profound immune defects do not appear to be at increased risk for amoebic colitis or amoebic liver abscess. Thus, the role of adaptive immunity in protection against amoebiasis remains uncertain, and a vaccine to prevent amoebiasis may have the daunting task of trying to better the immune response that develops after natural infection (Stanley, 2001).

The uncertainty regarding the role of adaptive immunity in providing protection against E. histolytica infection or amoebic colitis and amoebic liver abscess, is pertinent to vaccine development in another way. It raises the question of what immunological responses should be followed to assess the immunogenicity of an amoebiasis vaccine? In animal models of amoebic liver abscess, protection has been correlated with the development of serum antibody levels and cell mediated immunity (Zhang et al. 1994a,b; Zhang and Stanley, 1996; Ivory, Keller and Chadee, 2006). However, in individuals with amoebic liver abscess high antibody titers or cell mediated immune responses appear to be short lived, and have not been associated with clearance of disease, or protection against subsequent infection (Velazquez et al. 1995; Blessmann et al. 2002; Vohra et al. 2003). The data from Bangladesh suggests that mucosal immune responses to the target antigen may be the best parameter for assessing vaccine efficacy (at least in terms of preventing infection), and in a recent study looking at vaccination of mice with a portion of the Gal/Gal NAc lectin, there was an association, which did not reach statistical significance, between developing stool IgA antibodies to the target antigen and protection against colonic infection (Houpt et al. 2004). While significant progress has been made, clearly there is a need for additional prospective studies to better define whether effective adaptive immunity to amoebiasis occurs and to characterize how it is mediated and how long it lasts.

WHAT ARE THE BARRIERS TO USING NATURALLY OCCURRING AVIRULENT OR ENGINEERED AVIRULENT E. HISTOLYTICA TROPHOZOITES AS VACCINES TO PREVENT AMOEBIASIS?

A basic concept in vaccinology is to use a live attenuated strain of a virus or bacteria to stimulate immunity that will prevent or ameliorate subsequent infection with the parent, virulent strain. This approach requires that the vaccine strain be both immunogenic and stably attenuated, with little risk of reversion. In guinea pigs, prior infection with a non-invasive strain of E. histolytica resulted in some protection against subsequent intracaecal challenge with a virulent strain of E. histolytica (Jain, Sawhney and Vinayak, 1980). Thus, there is a precedent in animal models for protective immunity against amoebiasis arising from infection with avirulent Entamoeba species. It is now possible to genetically engineer E. histolytica parasites to create attenuated E. histolytica strains. One approach to attenuation is to target known E. histolytica virulence factors. Amoebapores are pore-forming proteins that mediate target cell lysis and may play an important role in liver abscess formation (Leippe, 1997; Bracha et al. 1999; Zhang et al. 2004). Mirelman and colleagues were able to silence expression of the amoebapore A gene in E. histolytica trophozoites and demonstrate that intraperitoneal inoculation of hamsters with these trophozoites induced IgG anti-amoebic antibody production (Bujanover et al. 2003). Unfortunately, there were no challenge experiments to determine whether this form of vaccination could prevent amoebic liver abscess, but the results indicate the safety and immunogenicity of the avirulent strain administered in this manner. A cautionary note is sounded by the finding that the amoebapore-deficient strain was capable causing amoebic colitis in human intestinal xenografts in severe combined immunodeficient mice (SCID-HU-INT model), indicating that further targeted attenuation may be necessary to render this strain suitable for vaccine use (Zhang et al. 2004).

An alternative approach is to use naturally occurring E. histolytica strains that show markedly reduced virulence compared to the prototype E. histolytica HM1:IMSS strain. With the completion of the genome project we now have the ability to define genetic differences between the virulent E. histolytica HM1:IMSS strain and reduced virulence strains like E. histolytica Rahman or the avirulent species Entamoeba dispar. Knowing the nature of the genetic differences between these strains could greatly facilitate the use of a strain like Rahman as a potential vaccine or could allow one to more precisely attenuate a given E. histolytica strain. However, preliminary studies from our laboratory using an E. histolytica microarray suggest there are multiple differences in gene expression between the two strains, indicating that it may be difficult to attribute the virulence phenotype to the action of one, or a few, genes. (Davis and Stanley, unpublished).

If a strain like Rahman or a genetically engineered avirulent E. histolytica strain proved safe in animal models, multiple barriers to this form of vaccine would still be present. First, how would infection be established in the vaccine recipient? Trophozoites do not survive the acidity of the stomach, making oral delivery difficult. Feeding of cysts could be effective, but to date it has been impossible to get E. histolytica trophozoites to encyst in culture, providing a barrier to using a laboratory-derived strain. Thus, large-scale production of such a vaccine would be nearly impossible with current methods. Perhaps more importantly, there is little clinical evidence to indicate that infection with a non-invasive strain would stimulate an effective immune response that would protect against E. histolytica colonization or disease. Clearly, additional prospective studies that look at the impact of colonization with avirulent Entamoeba on subsequent acquisition of E. histolytica colonization or disease will be necessary before we can adequately assess the potential value of this approach.

HOW WOULD AN AMOEBIASIS VACCINE BE TESTED, HOW WOULD IT BE USED AND HOW WILL IT BE FINANCED?

While the tone of the previous sections may seem pessimistic, there has been important progress in amoebiasis vaccine development that makes consideration of how such a vaccine might be brought to widespread use more than an abstraction. A number of viable vaccine candidates have been identified, including the 170 kDa heavy chain and a 150 kDa ‘intermediate chain’ of the Gal/GalNac binding lectin, the serine-rich Entamoeba histolytica protein (SREHP), the peroxiredoxin molecule, and the EhCP112 molecule. (See Stanley, 2003 for a recent review). For some of these antigens, novel means of delivery, including fusion of recombinant amoebic antigens with cholera toxin subunits to increase mucosal immunogenicity, expression in attenuated Salmonella strains and introduction as DNA vaccines has been used (Zhang and Stanley, 1995, 1996, 1997, 1999; Ryan et al. 1997; Sultan et al. 1998; Vohra et al. 2003). Most importantly, several different antigen preparations have provided protection against amoebic liver abscess in rodent models of disease and a vaccine formulation containing a recombinantly prepared region of the heavy chain of the lectin provided some protection against disease in a murine model of amoebic colitis (Zhang et al. 1994a; Zhang and Stanley, 1996, 1999; Soong et al. 1995a,b; Ivory et al. 2006).

How is one or more of these promising candidates going to move forward? First there is a clear need for a non-human primate model for amoebic colitis and/or amoebic liver abscess. It is evident from a variety of settings that vaccine studies in rodent models, while absolutely essential, do not always predict vaccine efficacy in humans. Non-human primate models, especially in a disease where human challenge testing for efficacy may be difficult to justify, are critical to moving any vaccine forward. While there were early studies suggesting that experimental E. histolytica infections can be established in old- and new-world monkey species, there is little recent work and, in limited trials, we were unable to establish amoebic liver abscesses in African Green Monkeys (Hegner, Johnson and Sabler, 1932; Miller, 1952; Loomis et al. 1983; Marquez-Monter et al. 1991; Beaver, Blanchard and Seibold, 1988; Stanley et al. 1995). Baboons can be naturally infected with E. histolytica and there has been preliminary work to establish a baboon model for colonization and disease (Jackson et al. 1989). Continuing efforts in this area are a prerequisite to any serious effort towards developing and testing an amoebiasis vaccine in humans.

Assuming a viable amoebiasis vaccine candidate emerges from the appropriate animal studies, and demonstrates safety and immunogenicity in phase 1 and 2 studies, how would one assess efficacy in humans? It is an open question whether challenge experiments in humans would be ethical. Walker and Sellards (1913) successfully infected human volunteers with E. histolytica cysts derived from the stools of individuals with amoebic colitis and some, but not all, recipients developed amoebic dysentery. While these experiments clearly demonstrate the feasibility of this approach there are many areas for question and concern. The current state of the art would require the challenge cyst dose be derived from human stool, and making sure such a sample was free of other potential pathogens yet maintained potency for infection could be challenging. There are additional questions about the composition of the challenge dose – would cysts be clonal, or would there potentially be multiple trophozoite genotypes in a challenge dose, creating an issue about the reproducibility of the challenge across groups? As alluded to above, Walker and Sellards (1913) found significant variability among volunteers in their response to cyst challenge and this could make comparisons between groups difficult, as one might not know whether a failure to establish an infection represented individual variability or a protective effect from a vaccine. Finally, amoebic colitis and amoebic liver abscess can be fatal diseases, so careful monitoring and drug treatment of all volunteers in the study, whether or not they had manifested any evidence for colonization or disease, would be mandatory.

There will also have to be careful consideration given to the target populations and potential sites for field testing an amoebiasis vaccine. Recent epidemiological studies in Bangladesh focused on a cohort of children in a region where sanitary conditions were poor and the prevalence of E. histolytica infection was high (Haque, Ali and Petri, 1999; Haque et al. 2001, 2002). This would appear to be an excellent population for studying any vaccine designed to prevent colonization with E. histolytica, but may be less desirable for studying vaccines designed to prevent invasive disease, since the rates of amoebic colitis and amoebic liver abscess were quite low in this cohort. In contrast, the high incidence of amoebic liver abscess in the male population in Hue, Vietnam would present an opportunity to study a vaccine designed to prevent invasive disease (Velazquez et al. 1995). The epidemiological studies in each of these areas, as well as studies in South Africa, have laid important groundwork for future clinical trials, but additional careful epidemiological studies in other areas with high endemicity for E. histolytica would be helpful (Gathiram and Jackson, 1987; Ravdin et al. 2003).

Finally, perhaps the most critical barrier to developing a vaccine to prevent amoebiasis is financial, the probable lack of any monetary return on an amoebiasis vaccine. There is little to no market for an amoebiasis vaccine in developed countries, hence it is highly unlikely that any pharmaceutical company will initiate or significantly support product development in this field. Innovative approaches, such as appealing to public/private partnerships, large charitable organizations or directly involving governments in areas where amoebiasis is highly endemic, may overcome this barrier. If this type of funding is to be obtained, a powerful argument must be made for the global health benefit of an amoebiasis vaccine. There is extraordinary competition for the very limited health dollars available to developing countries and no one can question that diseases like HIV, tuberculosis and malaria are high priority needs. In addition, there are still on-going programmes to eradicate polio and measles that merit continued support. In this landscape, how does one obtain support for diseases that are important, but do not have the global impact of malaria, TB or AIDS? A key is to try and make the vaccine as cost-effective as possible and this is the driving force behind the concept of combination vaccines, designed to prevent multiple infections. There is no question that the cost-benefit ratio of an effective anti-diarrhoeal vaccine that delivers protection against amoebiasis, cryptosporidiosis, shigellosis, pathogenic E. coli and other critical enteric pathogens in a single vaccine will be much more favourable than a ‘stand-alone’ vaccine for amoebiasis. Such multivalent vaccines are feasible now, but there are still some technical barriers. An anti-diarrhoeal vaccine may be the vehicle that can make an amoebiasis vaccine, and the eradication of E. histolytica, a reality and should be a major consideration in the next phase of research in amoebiasis vaccines.