INTRODUCTION
In his book ‘Parasitism the Ecology and Evolution of Intimate Associations’ Combes (Reference Combes2000) has eloquently described ‘the inequality of host individuals in the face of parasitism’. This inequality, the epidemiological patterns that it generates and the possible mechanisms behind it, has intrigued parasitologists for many decades. More specifically, the number of macroparasites a host carries is fundamental to our understanding of helminth parasite epidemiology (Anderson, Reference Anderson1986). Crofton (Reference Crofton1971), in his seminal paper on a quantitative approach to parasitism, characterized the frequency distribution of parasites in a host as clumped or over-dispersed and best described by the negative binomial. Essentially, most hosts carry few or no parasites and a small proportion of hosts carry heavy infections; these humans or other animals are sometimes described as ‘wormy’ (Croll and Ghadirian, Reference Croll and Ghadirian1981). The ubiquity of this distribution was demonstrated by Shaw and Dobson (Reference Shaw and Dobson1995), amongst a survey of over 270 host-macroparasite systems, with only a single exception to the rule of aggregation.
This inequitable distribution has important implications for parasite transmission, strategies for control and host morbidity. Heavily infected hosts are likely to suffer greater morbidity and mortality, to make a greater contribution to the number of infective stages in the environment than hosts carrying lower intensity infections and to contribute to host-parasite regulatory processes. Consequently, developing an understanding of the mechanisms that generate susceptibility and resistance to helminth parasites, and how these generate the observed patterns of distribution is central to our understanding of parasite epidemiology and the development of appropriate strategies for control.
ASCARIS AND ASCARIASIS
In this retrospective, I have chosen to focus upon the human roundworm Ascaris lumbricoides for two reasons. Firstly, this geohelminth has been and continues to be a focus of my own research; but, perhaps more importantly, I regard it as becoming increasingly understudied relative to its counterparts Trichuris trichiura and the two hookworm species, Ancylostoma duodenale and Necator americanus.
A. lumbricoides is a remarkably infectious and persistent parasite (Crompton, Reference Crompton2001). An estimated 1472 million people (a quarter of the world's population) are infected with it and the level of morbidity, assessed as disability-adjusted life years, is about 10·5 million (Holland, Reference Holland2005). Chronic ascariasis is known to contribute to insidious morbidity, including growth retardation and effects on cognitive development, particularly in growing children (O'Lorcain and Holland, Reference O'Lorcain and Holland2000). Additionally, both A. lumbricoides and Ascaris suum, its counterpart in pigs, are known to cause liver and lung pathology as a consequence of their migration through these organs (Stephenson, Reference Stephenson and Stephenson1987). Pulmonary symptoms can be pronounced and are described as Loeffler's syndrome (Loeffler, Reference Loeffler1932, Reference Loeffler1956). Acute ascariasis, associated with very heavy worm burdens, can result in intestinal obstruction, other serious complications and, in extreme cases, death (O'Lorcain and Holland, Reference O'Lorcain and Holland2000).
Furthermore, the growing perception that infection with Ascaris and other geohelminths have a broad effect on the host's immune system, with consequences for concurrent important infectious diseases, such as malaria (Mwangi et al. Reference Mwangi, Bethony and Brooker2007), renewed interest in the consequences of early infection with worms in the context of the hygiene hypothesis (Yazdanbakhsh et al. Reference Yazdanbakhsh, Kremsner and van Ree2002) and the modulatory consequences for development of allergies (Maizels and Yazdanbakhsh, Reference Maizels and Yazdanbakhsh2003), all greatly enhance the public health significance of geohelminths such as Ascaris.
In this paper, I will first provide a brief historical perspective on the phenomenon of aggregation and predisposition with respect to ascariasis. Secondly, I will evaluate the studies that have explored these epidemiological patterns and the possible mechanisms that generate them in both humans and animal model systems. Thirdly, I will outline how these patterns influenced the development of strategies for control and the broader implications of co-infections between macroparasites such as Ascaris and microparasites.
HISTORICAL PERSPECTIVE
Aggregation
The first paper to describe the frequency distribution of a number of human helminth infections was that of Croll and Ghadirian (Reference Croll and Ghadirian1981). This work, based in 3 Iranian villages, described the over-dispersed distribution of Ascaris lumbricoides, Trichuris trichiura and two species of hookworm, Ancylostoma duodenale and Necator americanus, after anthelmintic treatment and worm collection. Over-dispersed frequency distributions such as these are best described by fitting the data to the negative binomial distribution and estimating the parameter k (Bliss and Fisher, Reference Bliss and Fisher1953; Anderson, Reference Anderson1986). Values of k provide an inverse measure of the degree of over-dispersion or aggregation, with smaller values denoting greater aggregation. Anderson and May (Reference Anderson and May1991) described 4 key consequences of aggregation for a parasite such as Ascaris: influence on mate-finding choices and hence egg production and transmission, density-dependent regulation of parasite populations, host morbidity and mortality in heavily infected hosts and implications for the design of effective control strategies.
Predisposition
In addition to the observations on human helminth distributions, a further important epidemiological parameter was identified during the late 1980s. Longitudinal field-based studies measured the patterns of helminth re-infection in individual patients after the provision of anthelmintic treatment and an assessment was made of the degree to which individuals re-acquired both light and heavy infection (Anderson, Reference Anderson1986). Worm burdens or faecal egg counts of individuals were measured and compared statistically before and after treatment, and a significant positive correlation between these was identified indicating consistency in the extent of re-infection. This phenomenon is known as predisposition. Elkins et al. (Reference Elkins, Haswell-Elkins and Anderson1986) were the first authors to describe predisposition to A. lumbricoides among both child and adult inhabitants of Indian villages. An important age-related pattern was also observed, with Ascaris infection declining with age as did the strength of observed predisposition.
Predisposition to ascariasis was subsequently observed in Burma (Thein-Hlaing et al. Reference Thein-Hlaing, Than-Saw and Myint-Lwin1987), Nigeria (Holland et al. Reference Holland, Asaolu, Crompton, Stoddart, MacDonald and Torimiro1989), Mexico (Forrester et al. Reference Forrester, Scott, Bundy and Golden1990), Malaysia (Chan et al. Reference Chan, Kan and Bundy1992) and Bangladesh (Hall et al. Reference Hall, Anwar and Tomkins1992), confirming it as a widespread phenomenon in endemic communities. In contrast, Croll and Ghadirian (Reference Croll and Ghadirian1981) had failed to detect a relationship between pre- and post-treatment worm burdens in their Iranian populations. Another significant contribution to our understanding of predisposition was our demonstration, that predisposition could be maintained over multiple rounds of treatment among Nigerian children aged 5–16 years (Holland et al. Reference Holland, Asaolu, Crompton, Stoddart, MacDonald and Torimiro1989). In an impressive long-term study, Quinnell et al. (Reference Quinnell, Griffin, Nowell, Raiko and Pritchard2001) revealed that predisposition to N. americanus could be detected 6–8 years after a single round of chemotherapy, but was not detectable after repeated chemotherapy. Utilizing a different approach, Peng et al. (Reference Peng, Zhou, Cui, Crompton, Whitehead, Xiong, Wu, Yang, Wu, Xu and Yan1998) explored natural re-infection with A. lumbricoides, in a rural Chinese community, over a one-year period in the absence of chemotherapy. These authors provided evidence of predisposition to light and medium intensity infections, in the absence of repeated expulsion chemotherapy.
An important review of the phenomenon of predisposition was published in 1990 and much of its content remains pertinent (and understudied) 18 years later (Keymer and Pagal, Reference Keymer, Pagal, Schad and Warren1990). After reviewing the existing field-based data, the authors concluded that the observed predisposition is relatively weak and factors such as sample size, method of worm intensity measurement and the duration of re-infection were important contributory factors to the sensitivity of detection. Echoing this, McCallum (Reference McCallum1990) used probability theory to model predisposition and concluded that this epidemiological pattern is subject to the influence of both short and long-term transitory factors.
To summarize, the patterns of aggregation and predisposition for ascariasis had been well established by the late 1980s, and this extended to the other important geohelminths, T. trichiura (Bundy et al. Reference Bundy, Cooper, Thompson, Didier, Anderson and Simmons1987) and hookworm species (Schad and Anderson, Reference Schad and Anderson1985), in addition to multiple-species predisposition (Haswell-Elkins et al. Reference Haswell-Elkins, Elkins and Anderson1987). However, the mechanisms that contribute to the observed epidemiological patterns have proved to be much more refractory to investigation. These are likely to be multi-factorial and include the relative contributions of ‘exposure’ versus ‘susceptibility’ or, as described by Combes (Reference Combes2000), ‘the encounter filter’ versus ‘the compatibility filter’. Exposure can include attributes of the host, such as their behaviour, environment, socio-economic status and culture. Susceptibility can encompass host physiology (including age, sex and hormonal status), immunity and genetics, although immunity may be influenced by both genotypic and phenotypic factors such as nutrition, reproductive status or the presence of co-infections. Clearly, these relative contributions are complex, difficult to disentangle and likely to vary in both space and time.
STUDIES ON THE MECHANISMS THAT EXPLAIN AGGREGATION AND PREDISPOSITION
Human subjects
The number of studies that have attempted to unravel the factors that contribute to susceptibility and resistance to Ascaris infections in humans remains comparatively small. This is not surprising, given the logistical and particularly the ethical constraints that largely restrict human studies to correlational approaches between parameters, suspected to be involved in influencing susceptibility/resistance to infection. Furthermore, despite the pressing need for longitudinal studies of parasitology in human subjects, funding for such research remains an ongoing challenge.
In seeking explanatory mechanisms, the majority of studies have focused upon differences in susceptibility (host humoral and cellular immune response, genetics) rather than exposure, given the difficulties in measuring the later. Factors that are likely to influence the successful execution and outcome of such studies include sample size, study design, measurement of current intensity of parasite infection, length of follow-up and types of host factors investigated (Anderson, Reference Anderson1986; Keymer and Pagal, Reference Keymer, Pagal, Schad and Warren1990; Holland and Boes, Reference Holland, Boes, Holland and Kennedy2002). If a predisposition study is designed to identify individuals that demonstrate consistency in parasite intensity and assign them to a particular group, the initial sample size needs to be very large indeed in order to compensate for the relatively weak correlations that are observed between worm burdens or egg counts (Holland and Boes, Reference Holland, Boes, Holland and Kennedy2002). This is borne out by our own experience working with Nigerian children in the late 1980s and early 1990s. Our original predisposition study commenced with 808 children at the beginning of Phase 1, fell to 580 by the third round of treatment and, after selection for consistency to predisposition status, narrowed to 120 (Holland et al. Reference Holland, Asaolu, Crompton, Stoddart, MacDonald and Torimiro1989, Reference Holland, Asaolu, Crompton, Crichton, Torimiro and Walters1992; McSharry et al. Reference McSharry, Xia, Holland and Kennedy1999).
Exposure
Within the context of geohelminth infection, the measurement of exposure has proved to be consistently problematical. Wong et al. (Reference Wong, Bundy and Golden1988) developed a method for measuring soil-derived silica in faeces as a measure of geophagia and hence a proxy for exposure to geohelminths. However, this relatively time-consuming and complex method is not appropriate for routine use. In an interesting and novel approach exploiting an invertebrate host-parasite system, Kuris et al. (Reference Kuris, Goddard, Torchin, Murphy, Gurney and Lafferty2007) demonstrated how a rhizocephalan/crab host-parasite relationship could be used to explore the encounter and compatability filters. Specifically, when rhizocephalans fail to establish they leave a tell-tale record in the thoracic ganglion of the crabs, thereby providing a means of quantifying the encounter filter. Unfortunately, such tools are not easily available to us in the human geohelminth field. Henry (Reference Henry1988) explored the relationship between re-infection with Ascaris and risk factors for exposure in pre-school children in St Lucia. Crowding and sanitation were reported to be the most significant factors that influenced re-infection.
The humoral immune response
Several studies have focused upon the relationship between humoral immune responses and re-infection or predisposition to A. lumbricoides. The earliest study, undertaken in Venezuela, did not establish predisposition in 5 to 10-year-old children, but rather compared groups of children who did and did not exhibit re-infection (Hagel et al. Reference Hagel, Lynch, Di Prisco, Rojas, Perez and Alvarez1993). A significant association between re-infection and high pre-treatment total IgE levels, but low parasite-specific IgE levels, was recorded. This led the authors to conclude that parasite-specific IgE may have a protective role against Ascaris infection.
In 1995, Palmer et al. undertook a case-control study among Bangladeshi children, with an average age of 9 years. A comparison of consistently lightly-infected versus heavily-infected subjects revealed higher concentrations of a range of antibody isotypes including parasite-specific IgG1, IgG4 and IgE. Therefore, in contrast to the findings of Hagel and colleagues, parasite-specific IgE levels mirrored the infection intensity of the children. The authors did not rule out an effector role for these antibodies and suggested that the utilization of more specific antigens might sharpen the detection of epitope-specific protective responses.
This suggestion was borne out by our observations on the Nigerian children deemed to be predisposed to Ascaris infection. Children aged 5–16 years were identified as predisposed to remain uninfected, lightly infected and heavily infected. This selection was made by means of a statistically based cut-off and children were classified as follows: predisposed to remain uninfected (no worms on either of the two occasions), those with consistently light infections (1–24 worms in Phase 1 and 1–8 worms in Phase 2), and those with consistently heavy infections (>25 worms at Phase 1 and >9 worms at Phase 2) (Holland et al. Reference Holland, Asaolu, Crompton, Stoddart, MacDonald and Torimiro1989). The 3 groups of children were compared using a range of serum factors and 3 different sources of Ascaris antigen were utilized.
Only the defined allergen, Ascaris ABA-1 as a bacterial recombinant protein, provided evidence for a significant relationship between predisposition status and parasite-specific IgE. The putatively immune group tended to have higher levels of rABA-1 specific IgE, compared to the susceptible group that had low levels. In addition, higher levels of inflammatory indicators were detected in the putatively immune group (serum ferritin, eosinophil cationic protein and C-reactive protein) (McSharry et al. Reference McSharry, Xia, Holland and Kennedy1999). This finding is of particular interest with respect to our recent findings in the mouse model (see below).
Cytokine responses
The first data on cytokine responses in humans infected with Ascaris were obtained by Cooper and colleagues (Reference Cooper, Chico, Sandoval, Espinel, Guevara, Kennedy, Urban, Griffin and Nutman2000), who described a sample of young Ecuadorian adults with moderate Ascaris infections compared to uninfected control subjects. Infected individuals manifested a highly polarized Th2 response with a predominance of IL-4 and IL-5. Two important recent studies have followed concerning the influence of cytokines upon infection intensity and re-infection in Cameroonian children and adults (Turner et al. Reference Turner, Faulkner, Kamgno, Cormont, Van Snick, Else, Grencis, Behnke, Boussinesq and Bradley2003; Jackson et al. Reference Jackson, Turner, Rentoul, Faulkner, Behnke, Hoyle, Grencis, Else, Kamgno, Boussinesq and Bradley2004). In contrast to data from laboratory models (Urban et al. Reference Urban, Katona, Paul and Finkelman1991; Richard et al. Reference Richard, Grencis, Humphreys, Renauld and Van Snick2000; Schopf et al. Reference Schopf, Hoffman, Cheever, Urban and Wynn2002), no evidence for a role for Th2 cytokines in protective immunity in humans had been previously provided.
In the first study, the relationship between cytokine production and intensity of A. lumbricoides was investigated in Cameroonian children and adults aged 2–36 years (Turner et al. Reference Turner, Faulkner, Kamgno, Cormont, Van Snick, Else, Grencis, Behnke, Boussinesq and Bradley2003). Intensity was measured by means of egg counts and subjects were divided into child and adult age classes. Evidence was provided that Th2 cytokines, and in particular IL-9, IL-10 and IL-13, are associated with reduced parasite intensity in older children and adults. The authors concluded that their data were consistent with the hypothesis that Th2 cytokines have a role in mediating Ascaris intensity. Evidence from murine models suggests that this control may be via changes in intestinal environment that may impinge upon both larval and adult survival (Turner et al. Reference Turner, Faulkner, Kamgno, Cormont, Van Snick, Else, Grencis, Behnke, Boussinesq and Bradley2003).
In a second study, the egg counts of A. lumbricoides (and T. trichiura) were recorded before and after treatment amongst children and adults aged 4–57 years (Jackson et al. Reference Jackson, Turner, Rentoul, Faulkner, Behnke, Hoyle, Grencis, Else, Kamgno, Boussinesq and Bradley2004). Egg counts for A. lumbricoides were recorded pre- (n=191) and post-treatment (n=135), 8–9 months later. Because of the level of complexity of the data, a principal components analysis was required to explore the multivariate variation in both parasitological and cytokine data sets. The results from this paper added weight to their earlier observations and confirmed the importance of Th2 cytokines in immunity to ascariasis.
After controlling for age and gender, Th2 cytokines were significantly negatively associated with re-infection and this association was particularly pronounced for IL-5 and IL-13. The authors believe that their observations suggest that persistent susceptibility is, at least in part, due to variation in Th2-driven responses. In contrast to the earlier paper by Turner, the effects were more pronounced in younger children. This led the authors to suggest that heterogeneity in cytokine responses may operate differently depending upon the geographical location of the study. This may be due to differences in transmission patterns or even historical differences in parasite dynamics. The authors conclude that these age and location-related differences may have implications for the differential impact of de-worming programmes on immune responses.
Genetic studies of susceptibility to ascariasis
Williams-Blangero and colleagues have been at the forefront of the investigation of the genetic epidemiology of infectious disease in general and ascariasis in particular. Their studies of the Jirel population in Nepal represent a unique context within which the conduct of detailed pedigree studies provides an opportunity to disentangle environmental from genetic effects on the phenotype of Ascaris intensity.
Prior to the Jirel investigations, relatively little work had been undertaken on the relationship between susceptibility to ascariasis and host genetics. Familial predisposition had been established by Forrester et al. (Reference Forrester, Scott, Bundy and Golden1990) in Mexico and Chan and colleagues in Malaysia (1994a). However when Chan et al. (Reference Chan, Bundy and Kan1994b) explored the correlation between parasite intensities within families they failed to detect a significant role for host genetic factors. We undertook an associational study on the distribution of the major histocompatability complex (MHC) alleles in HLA-A, HLA-B and HLA-C among the same groups of Nigerian children that had been selected for their predisposition status (Holland et al. Reference Holland, Asaolu, Crompton, Crichton, Torimiro and Walters1992). None of the children who were predisposed to remain uninfected possessed the HLA-A30/31 alleles, and the frequency of occurrence of this haplotype was significantly higher in the children observed to be consistently re-infected.
Arguing that association studies of candidate genes can lead to false-positive results, Williams-Blangero et al. (Reference Williams-Blangero, Subedi, Upadhayay, Manral, Rai, Jha, Robinson and Blangero1999) reported the results of their first study of 1261 Jiral individuals belonging to a single pedigree. Individuals were assessed for Ascaris intensity (both egg counts and worm burden) at 2 time-points, post-treatment follow-up occurred 1 year after treatment. This analysis demonstrated that between 30 and 50% of the variation in worm burden of Ascaris was due to genetic factors.
These authors have recently published a whole-genome scan of 1258 members of the Jirel population (Williams-Blangero et al. Reference Williams-Blangero, VandeBerg, Subedi, Jha, Correa-Oiveira and Blangero2008). This represents a refinement of an earlier study of 444 individuals (Williams-Blangero et al. Reference Williams-Blangero, VandeBerg, Subedi, Aivaliotis, Rai, Upadhayay, Jha and Blangero2002), and, provides greater statistical power. These 2 later studies, measured susceptibility to Ascaris infection by means of faecal egg counts at the same time-point (i.e. not with respect to re-infection as in their first study). Prevalence values for Ascaris between the 3 studies range from 21·3% to 27·2%. A variance component linkage analysis was used to measure the covariance between infection and genotype, using microsatellite markers at approximately 10 cM intervals among individuals in a family, and hence to identify quantitative trait loci (QTL). The genome scan revealed 6 chromosomal regions with evidence of QTLs influencing intensity of infection. Three significant QTLs and 3 suggestive QTLs were identified on chromosomes 13, 11, 8 and 1 respectively. Future work is likely to focus upon the identification of specific candidate genes underlying the 3 significant QTLs, 2 of which were identified in the previous scan (13q and 8q). The peak on chromosome 13 is close to the known locus of a major candidate gene, TNFSF13B, involved in the regulation of B cell activation and immunoglobulin secretion (Yan et al. Reference Yan, Marsters, Grewal, Wang, Ashkenazi and Dixit2000).
Animal models
Keymer and Pagal had stated in 1990 that sufficient data on predisposition were available for the erection and testing of specific hypotheses concerning causal mechanisms. Given the logistical and ethical constraints operating on human studies, they advocated the use of laboratory models, where variables such as nutritional status, genetic background, immunocompetence and behaviour can be manipulated experimentally.
Given the difficulties in the conduct of human studies, outlined above, animal models of aggregation and predisposition clearly have many obvious advantages in terms of the degree of control that we can employ. The availability of defined inbred rodents (e.g. congenic and recombinant, knockout and transgenic strains) and an array of immunological reagents, make rodent models of disease an attractive option for the investigator seeking to unravel cause and effect. However, with respect to ascariasis, we lack a natural rodent model system in which the parasite completes its life cycle in a manner similar to that of humans. This is in marked contrast to Trichuris muris, the murine species is closely related and has a similar host-parasite relationship to that of its human counterpart, T. trichiura. The dissection of the immune responses to T. muris, in inbred and knockout mice, has been highly informative (Cliffe and Grencis, Reference Cliffe and Grencis2004). Heligimosomoides bakeri has proved to be an excellent model organism for nematodes of importance in livestock (see Behnke et al. Reference Behnke, Iraq, Menge, Baker, Gibson and Wakelin2003; and current issue Behnke et al. 2009). An enduring example of the use of an animal model to explore the generative mechanisms of aggregation and predisposition was employed by Tanguay and Scott (Reference Tanguay and Scott1992) utilizing both inbred and outbred mice infected with H. bakeri. The investigators were able to simultaneously explore host behaviour, genetics and acquired resistance using a controlled exposure regime.
The mouse model
However, comparative studies on larval migration in pigs and mice have led to a revised view of the mouse as a suitable model for exploring susceptibility to the early phase of Ascaris infection (Murrell et al. Reference Murrell, Eriksen, Nansen, Slotved and Rasmussen1997; Slotved et al. Reference Slotved, Eriksen, Murrell and Nansen1998). As discussed previously, we do not fully understand the mechanisms of predisposition, but it is likely that events early in infection determine the subsequent status of the host in terms of susceptibility/resistance. It has been suggested that pulmonary migration by Ascaris larvae could create a highly polarized Th2 immune environment in the lung mucosa, and genetic variation in the host response at this level may influence the course of infection (Cooper et al. Reference Cooper, Chico, Sandoval, Espinel, Guevara, Kennedy, Urban, Griffin and Nutman2000). Early work by Mitchell et al. (Reference Mitchell, Hogarth-Scott, Edwards, Lewers, Cousins and Moore1976) had demonstrated variability in the susceptibility of different strains of inbred mice to ascariasis.
After optimization of the methodologies for infection and recovery, we have identified a mouse model for susceptibility and resistance to early Ascaris infection (Lewis et al. Reference Lewis, Stafford, Behnke and Holland2006). C57BL/6j mice are consistently highly susceptible to Ascaris larvae in the lungs, peaking at day 7 post-infection, in contrast to CBA/Ca mice that remain consistently resistant to lung infection. Subsequently, we used this model to assess the significance of inflammatory processes within the murine lung (leucocyte population in bronchalveolar lavage (BAL) fluid and lung histopathology) (Lewis et al. Reference Lewis, Behnke, Cassidy, Stafford, Murray and Holland2007). We concluded that the responses mirrored larval intensity and that the pulmonary inflammatory immune response was not prominently involved in primary protection of mice to Ascaris infection in the latter days of infection in the lungs.
The lack of support for a pulmonary mechanism led us to suggest that a hepatic/post-hepatic factor, that varies between C57BL/6j and CBA/Ca mice, may play a critical role in the successful migration through host tissues. Our most recent data, concerning the comparative histo-pathological hepatic inflammatory response in C57BL/6j and CBA/Ca mice, have revealed an important difference in the timing of the response between the two strains (Dold et al. Reference Dold, Cassidy, Behnke and Holland2008). Specifically, in resistant CBA mice the most pronounced response occurred on day 4 of infection, this coincides with the migration of lower larval numbers from the liver to the lungs. In contrast, the severe inflammatory response in C57BL/6j mice took place on day 6 post-infection, a time by which most larvae have already, successfully migrated to the lungs. Therefore, these data suggest a possible inflammatory mechanism of resistance to larval Ascaris migration, acting rapidly and focused in the liver.
To conclude, we are now confident that we have a model of differential hepatic inflammation that we can use to tease out the potential mechanisms that influence susceptibility and resistance to larval ascariasis. This has both public health and veterinary significance, given that we know that liver inflammation occurs in both humans (Javid et al. Reference Javid, Wani, Gulzar, Khan, Shah, Shah and Khan1999) and pigs (Perez et al. Reference Perez, Garcia, Mozos, Bautista and Carrasco2001; Frontera et al. Reference Frontera, Roepstorff, Gazquez, Reina, Serrano and Navarrete2003). Furthermore, for ethical reasons we know virtually nothing about the dynamics and impact of larval migration in humans; and in pigs, we still lack the versatility provided by inbred and knockout mice.
The pig model
Nevertheless, aggregation and predisposition have been modelled successfully in outbred pigs and epidemiological patterns in pigs have been described that were analogous to those of A. lumbricoides in humans (Boes et al. Reference Boes, Medley, Eriksen, Roepstorff and Nansen1998). Furthermore, as argued by Coates (Reference Coates2000) the sample sizes utilized to establish predisposition were much smaller than those required for human studies. Roepstorff et al. (Reference Roepstorff, Eriksen, Slotved and Nansen1997) infected pigs with 3 doses of Ascaris eggs and were able to demonstrate that although each pig became infected, an expulsion event occurred between day 14 and day 21 post-infection that resulted in an overdispersed distribution of adult Ascaris worms.
These observations indicate the considerable potential of the Ascaris-pig model to explore the mechanistic basis of the observed aggregation and predisposition particularly as exposure can be controlled or explored using natural and trickle infection protocols. Furthermore, in exciting and novel work, Nejsum et al. (Reference Nejsum, Roepstorff, Anderson, Jorgensen, Fredholm and Thamsborg2009a) have recently reported a study that explored the dynamics of genetically marked A. suum infections in pigs. The authors described significant differences in the distribution, abundance and size of a variety of worm genotypes among individual pigs. This confirms the idea that genetic heterogeneity among parasites as well as among hosts, plays an important role in determining the success of infection. Other work by Nejsum and colleagues (Reference Nejsum, Roepstorff, Jorgensen, Fredholm, Goring, Anderson and Thamsborg2009b) on host genetics, confirmed that heritability estimates in the Ascaris-pig model were similar to those estimated in humans by Williams-Blangero (Reference Williams-Blangero, Subedi, Upadhayay, Manral, Rai, Jha, Robinson and Blangero1999, Reference Williams-Blangero, VandeBerg, Subedi, Aivaliotis, Rai, Upadhayay, Jha and Blangero2002, Reference Williams-Blangero, VandeBerg, Subedi, Jha, Correa-Oiveira and Blangero2008). These observations demonstrate the potential of identifying the underlying QTL using the Ascaris-pig model system.
THE SIGNIFICANCE OF GENETIC VARIATION IN HELMINTH PARASITES
Genetic variation in helminth parasites has received comparatively less attention than heterogeneity in host responses. Such polymorphism is likely to have profound epidemiological effects and is essential for host-parasite co-evolution (Read and Viney, Reference Read and Viney1996). Furthermore, there are important public health implications including the influence of parasite variation upon the development of anti-helminth vaccines and anthelmintic resistance (Maizels and Kurniawan-Atmadja, Reference Maizels and Kurniawan-Atmadja2002). Certainly, helminth genetics are likely to be an important contributor to host responsiveness and this has been demonstrated for some helminth species including T. muris (Bellaby et al. Reference Bellaby, Robinson and Wakelin1996).
Considerable attention has been focused upon the investigation of human and pig Ascaris using genetic markers to explore cross-infection and hybridization and microspatial genetic variation within Ascaris populations that may shed light upon modes of transmission (reviewed by Anderson et al. Reference Anderson, Blouin and Beech1998). Anderson et al. (Reference Anderson, Romero-Abal and Jaenike1995) identified considerable genetic diversity within populations of worms infecting humans – 41 distinct mitochondrial genotypes in over 200 worms. Furthermore, certain parasite genotypes were found more frequently than expected by chance in certain hosts. However, there is a paucity of data on genetic variation in Ascaris and its relationship with host responsiveness.
Fraser and Kennedy (Reference Fraser and Kennedy1991) examined heterogeneity in the expression of surface antigens of A. lumbricoides infective larvae by assessing variation in antibody binding taken from the same population from whom the infective eggs were collected. Considerable heterogeneity was observed between larvae in their capacity to bind to antibody from a given donor. The authors suggested that this might reflect quantitative differences arising from surface antigen polymorphism or qualitative effects from differential gene expression. It has been suggested that the marked differences in the fecundity of A. lumbricoides observed between Nigeria and Bangladesh (Hall and Holland, Reference Hall and Holland2000) may reflect parasite differences in parasite immunogenicity (Wakelin and Bradley, Reference Wakelin, Bradley, Holland and Kennedy2002). What is certain is that the contribution of parasite genetics remains an understudied component of the mechanisms that explain host susceptibility and resistance to ascariasis and other helminth infections
IMPLICATIONS OF THESE EPIDEMIOLOGICAL PATTERNS FOR PARASITE CONTROL
The practical implications of the observed frequency distribution of, and predisposition to, human helminth infections are significant and not only for host morbidity but also for the development of the most effective strategies for control (Anderson and Medley, Reference Anderson and Medley1985; Anderson and May, Reference Anderson and May1991). Specifically, the hypothesis to be tested was whether the treatment of heavily infected individuals within a community would not only reduce morbidity and mortality in the treated group, but also modify the transmission dynamics of the community as a whole. This so-called ‘selective approach’ was viewed as a potentially attractive option because the use of relatively few doses of chemotherapeutic drugs, might reduce costs and the likelihood of the development of anthelmintic resistance (Crompton, Reference Crompton, Scott and Smith1994).
We compared universal, targeted and selective treatment for ascariasis in Nigerian villages using levamisole (Asaolu et al. Reference Asaolu, Holland and Crompton1991) and subsequently evaluated the cost-effectiveness of these approaches (Holland et al. Reference Holland, O'Shea, Asaolu, Turley and Crompton1996). The universal approach involved the treatment of all willing members of the community with the exception of pregnant women and children under the age of 2 years. The targeted approach consisted of children aged 5–15 years receiving treatment and the selective approach provided chemotherapy to 20% of the population with the heaviest Ascaris infections (Asaolu et al. Reference Asaolu, Holland and Crompton1991). Targeted treatment of school-age children proved to be an attractive option, in terms of impact of treatment upon a high-risk group, reduction in intensity among untreated adults and cost-effectiveness in comparison to the universal approach. Furthermore, the selective option proved to be difficult to administer, costly and unpopular with the community.
Hall and colleagues (Reference Hall, Anwar and Tomkins1992) observed that the disadvantages of the selective approach included the necessity to identify heavily infected people before they are treated and the likelihood of re-infection, particularly heavy infection, among such subjects was likely to occur over a relatively short period. Other studies also reported reductions in the prevalence and intensity of ascariasis in untreated adults within communities where children had been targeted for treatment (Thein Hlaing et al. Reference Thein-Hlaing, Than Saw and Myat Lay Kyin1991; Bundy et al. Reference Bundy, Wong, Lewis and Horton1990).
Subsequently, the World Health Organization have placed the treatment of school-age children at the heart of its commitment to the control of soil-transmitted helminths, with a target of regular treatment of at least 75% of school-age children by 2010 (Albonico et al. Reference Albonico, Montresor, Crompton and Savioli2006). Helminth control within schools has proved to be highly effective in developing countries, with trained teachers distributing the drugs in combination with the provision of health education (Montresor et al. Reference Montresor, Crompton, Gyorkos and Savioli2002).
More recently, the epidemiology and impact of soil-transmitted helminths in pre-school children has come into focus and a WHO informal consultation (2002) recommended that pre-school children over 1 year old should be included in regular de-worming programmes. Furthermore, the consultation and other commentators (Mwangi et al. Reference Mwangi, Bethony and Brooker2007; Albonico et al. Reference Albonico, Allen, Chitsulo, Engels, Gabrielli and Savioli2008), have specifically highlighted the need to define the interactions between geohelminths and important co-infections, such as malaria, in these young children, particularly if they are to be included in large-scale de-worming programmes. We have recently undertaken a randomized placebo-controlled intervention study on the impact of 4-monthly anthelmintic treatment (with albendazole) on ascariasis in Nigerian children aged 1–5 years as part of a wider study to assess the impact of anthelmintic treatment upon the prevalence and acquisition of malaria. A. lumbricoides proved to be the most prevalent geohelminth among this age group (48%) and repeated 4-monthly treatment with albendazole was effective in reducing prevalence and intensity of infection (Kirwan et al. Reference Kirwan, Asaolu, Molloy, Abiona, Jackson and Holland2009). Clearly we require good quality data (from a range of geographical locations) on the epidemiology and impact of helminth infections in these younger children and appropriate strategies for their inclusion in control programmes analogous to what we have obtained for school-age children (WHO, 2002; Albonico et al. Reference Albonico, Allen, Chitsulo, Engels, Gabrielli and Savioli2008).
CONCLUDING REMARKS
Ascariasis, despite its global distribution and the sheer numbers of individuals it infects, remains a classic neglected disease (Hotez et al. Reference Hotez, Molyneux, Fenwick, Kumaresan, Ehrlich Sachs, Sachs and Savioli2007). Furthermore, it is not an infection that exists alone but is found in the company of other geohelminths, particularly Trichuris (Holland et al. Reference Holland, Asaolu, Crompton, Stoddart, MacDonald and Torimiro1989; Tchuente et al. Reference Tchuente, Behnke, Gilbert, Southgate and Vercruysse2003) and important microparasitic diseases such as malaria and HIV. This polyparasitism leads to important interactions both in terms of pathogenesis and immunity. The field of Ascaris epidemiology and, specifically, that of understanding the mechanisms responsible for the observed aggregation and predisposition, is undoubtedly a challenging one. On the one hand, we are faced with the difficulties inherent in working with human subjects who live in challenging environments and asking them to participate in longitudinal studies with potentially multiple interventions and measurements. On the other hand, we look to the more refined animal model systems for greater control and ease of manipulation, but these vary in what we can derive from them.
Undoubtedly, significant progress has been made and parallel observations in humans and other animals have provided us with important lines of enquiry to pursue. This is exemplified by the correspondence between our observations of elevated inflammatory indicators in putatively immune Nigerian subjects and the differential hepatic inflammation in a susceptible and resistant strain of inbred mouse. This emphasizes how selected phases of infection, specifically the role of the liver migratory phase, are tractable within a murine model for ascariasis. In addition, observations on cytokine responses in Cameroonian subjects infected with Ascaris support observations made in other murine models of helminth susceptibility. This underlines the importance of undertaking both field-based studies in human subjects (involving both epidemiologists and immunologists) in tandem with laboratory-based animal model investigations.
Furthermore, the demonstration of genetic control of the intensity of ascariasis raises a number of questions. Work in humans has now established the QTL for intensity of infection within a unique human population and the identification of specific genes is likely to follow. However, it is of interest to speculate as to how widespread this heritability is likely to be in different geographical locations and under different conditions of transmission and regulation. Quinnell (Reference Quinnell2003) suggested that heritability of resistance to hookworms might be higher in environments where exposure is relatively homogenous. Clearly, studies of what Petri et al. (Reference Petri, Kirkpatrick, Haque and Duggal2008) refer to as human genetic epidemiology, in the context of infectious disease, require investigation in a variety of environments that vary in helminth transmission and epidemiology. The development of the pig-Ascaris model now provides a superb opportunity to explore the interaction between host and parasite genetics utilizing a high degree of experimental manipulation and control.
The observation by Turner and colleagues (Reference Turner, Faulkner, Kamgno, Cormont, Van Snick, Else, Grencis, Behnke, Boussinesq and Bradley2003) whereby apparent plasticity of age-specific cytokine responses is observed between different populations reveals another layer of complexity that requires excavation. This suggests the need to be aware of the impact of de-worming upon the immune responses of different human populations living under different regimes of helminth intensity and regulation. Furthermore, exploration of the interplay between helminth infections, immune responses to their presence and that of other important bystander infections, and their control by chemotherapy, remains one of the most crucial challenges facing parasite epidemiologists and immunologists.
I am most grateful to Jerzy Behnke, Chris Dold, Patrick Kirwan, Malcolm Kennedy and Rory O'Donnell for their helpful comments on early drafts of this paper.
This paper is dedicated to Lani Stephenson, a pioneer in field-based helminth studies.
