INTRODUCTION
Parasitism of molluscs by nematodes is widespread, but compared to trematode infections is often overlooked. In particular, terrestrial gastropods are commonly infected with a wide range of species, although those of medical and veterinary importance such as Angiostrongylus cantonensis, A. vasorum, and Muellerius capillaris receive the most attention. Nematodes form a diverse range of parasitic associations with molluscs (Grewal et al. Reference Grewal, Grewal, Tan and Adams2003) with increasing evidence to suggest that they may have an important role in regulating host population dynamics (Morand et al. Reference Morand, Wilson, Glen and Barker2004). Although there are many described parasite species only 2 major groups, metastrongyloids and rhabditoids, dominate the fauna. Metastrongyloids use molluscs as intermediate hosts, maturing in a vertebrate; in contrast, rhabditoids utilize only the mollusc to complete their life cycles (Grewal et al. Reference Grewal, Grewal, Tan and Adams2003).
The nature of nematode development in molluscs is dependent on their life cycles, which are particularly varied (Grewal et al. Reference Grewal, Grewal, Tan and Adams2003). In general, metastrongyloids that use molluscs as intermediate hosts enter the host as a first stage (L1) juvenile and develop up to a third stage (L3), further development requiring transmission to the vertebrate definitive host (Grewal et al. Reference Grewal, Grewal, Tan and Adams2003). In contrast, where molluscs act as definitive hosts many different kinds of life cycles occur. It is often the case for rahabditoids that either juvenile stages develop in the host with a free-living adult or the entire nematode life cycle is completed within the mollusc. In some associations, death of the host is required to complete the life cycle with rhabditoid nematodes releasing symbiotic bacteria as a food source that multiply rapidly, and produce an endotoxin that kills the host within a few days allowing the nematode to develop on the cadaver (Grewal et al. Reference Grewal, Grewal, Tan and Adams2003). Some species are capable of adopting different life cycles depending on the environmental conditions. For example, Phasmarhabditis hermaphridita is a facultative parasite that can reproduce in slug faeces and other organic-rich material or infect a mollusc and form a parasitic or necromanic relationship (Rae et al. Reference Rae, Verdun, Grewal, Robertson and Wilson2007).
Many terrestrial nematode species are characterized by the wide range of molluscan species that can act as host (Grewal et al. Reference Grewal, Grewal, Tan and Adams2003). One aspect of this generalist specificity is the ability to use aquatic molluscs as hosts under both natural and experimental conditions. Indeed, because aquatic molluscs demonstrate a high physiological compatibility with terrestrial nematodes and can be relatively easily cultured they have been widely used as laboratory hosts for studies on parasites of medical and veterinary importance such as A. cantonensis (Table 1). Nevertheless, both biological and ecological studies indicate that terrestrial snails and slugs are clearly the principal hosts utilized by these nematodes under natural conditions with aquatic molluscs acting as auxiliary hosts (sensuDogiel, Reference Dogiel1964). Typically, hosts of this kind are classified as having infection levels (prevalence, intensity, and abundance) that are lower than the principal host(s). Nevertheless, occurrence can vary greatly in auxiliary hosts, which may be due to ecological rather than phylogenetic conditions (Poulin, Reference Poulin2005). For example, one of the most important factors influencing specificity is the contact between parasite and host, particularly the behaviour of free-living stages (Dogiel, Reference Dogiel1964). Under the influence of environmental change and the associated degradation of preferred habitats host switching by parasites may increase (Brooks and Hoberg, Reference Brooks and Hoberg2007). Nematode parasites of insects have a well-established capacity to utilize a wide-range of diverse auxiliary host species (Poinar, Reference Poinar1989; Bathon, Reference Bathon1996). However, the significance of aquatic molluscs as auxiliary hosts for terrestrial metastrongyloid and rhabditoid nematodes, with the notable exception of A. cantonensis, has been largely overlooked and the implications for parasite ecology never evaluated. In particular, under climatic extremes such as floods and drought, which can have devastating effects on land snails and slugs, aquatic molluscs may play a more prominent role in parasite transmission. The frequency of these kinds of extreme environmental conditions is predicted to increase under climate change (Morand and Guegan, Reference Morand and Guegan2008). It is therefore the aim of the present article to assess the role of aquatic molluscs in transmitting terrestrial nematode parasites and how climatic conditions may influence their status as hosts in the future.
Table 1. Species of aquatic molluscs determined to be viable hosts for terrestrial nematode parasites under experimental and natural (N) conditions
LIFE CYCLES OF TERRESTRIAL NEMATODES THROUGH AQUATIC MOLLUSCS
Viability of free-living stages in water
The life cycles of many nematode species in aquatic molluscs have been elucidated under laboratory conditions. The first-stage larvae of metastrongyloids leave the definitive host with faeces and remain free living until they locate a suitable molluscan host. They do not feed, relying on energy reserves and consequently their survival is dependent on a range of environmental factors (Boev, Reference Boev1975). Rhabitoides, in contrast, have a free-living adult which produces larvae that develop into an L3 dauer, a non-feeding stage infective to molluscs which, like metastrongloid larvae, may have to survive a long time under varying environmental conditions before encountering a compatible molluscan host (Morand et al. Reference Morand, Wilson, Glen and Barker2004).
Metastrongyloidea L1 have a high survival rate over many months in freshwater (Boev, Reference Boev1975; Kontrimavichus et al. Reference Kontrimavichus, Delyamure and Boev1976; Kontrimavichus and Delyamure, Reference Kontrimavichus and Delyamure1979). Changes to the water's environmental condition can affect L1 viability. For example, increasing temperature gradually reduces survival, whilst infectivity peaks at an optimum temperature before declining (Skorping, Reference Skorping1982; Shostak and Samuel, Reference Shostak and Samuel1984; Lorentzen and Halvorsen, Reference Lorentzen and Halvorsen1986), However, the response to temperature varies between species, probably related to their different geographical distributions and the climatic conditions encountered there (Cabaret et al. Reference Cabaret, Risye Riseani and Baeza1991). Survival in seawater has also been documented but with a reduced activity and, on return to freshwater, infectivity to molluscs remains unaffected (Richards and Merritt, Reference Richards and Merritt1967).
Rhabditoid L3 can also survive for weeks in fresh water although they demonstrate a gradual decline in infectivity (Lewis et al. Reference Lewis, Delvan, Campbell and Gaugler1995; Grewal and Grewal, Reference Grewal and Grewal2003). Changes in temperature, salinity, UV light and hypoxia can all affect viability of larvae in water (Thurston et al. Reference Thurston, Ni and Kaya1994; Grewal et al. Reference Grewal, Wang and Taylor2002; Grewal and Grewal, Reference Grewal and Grewal2003). Nevertheless, ‘fitness’ varies considerably between individual species, possibly related to foraging strategies, which incur different metabolic demands. Survival is related to declining energy reserves; however, metabolic rates increase drastically after a few weeks in water. This may be due to osmotic stress, as the cuticle becomes more permeable with age and consequently osmoregulation is more difficult, ultimately leading to death (Lewis et al. Reference Lewis, Delvan, Campbell and Gaugler1995). The nature of the relationship between nematode and symbiotic bacteria also changes after prolonged periods in water. The number of viable bacteria declines after a few weeks, but varies between individual nematode species possibly because of differences in the way bacteria are stored in the host, which can affect both bacterial viability and their availability as a food source (Lewis et al. Reference Lewis, Delvan, Campbell and Gaugler1995).
Infection of aquatic molluscs
Nematodes are known to infect aquatic molluscs in 2 ways, either being accidentally ingested by the snail and penetrating through the intestinal wall or, less commonly, by direct penetration of the molluscan tegument. Direct penetration is a widespread mechanism of infecting terrestrial molluscs but appears difficult for aquatic species to accomplish (Yousif and Lammler, Reference Yousif and Lammler1977). This is probably because few juvenile nematodes can produce body waves of the amplitude and frequency needed to achieve real swimming (Clark, Reference Clark1994) and consequently they are unable to actively locate, attach, and penetrate a target host. Aquatic molluscs therefore acquire an infection mainly by accidental contact with larvae with little evidence to suggest that terrestrial nematodes can chemotactically locate hosts in water (Banevicius et al. Reference Banevicius, Zanotti-Magalhaes, Magalhaes and Linhares2006). Nevertheless, some species of freshwater pulmonates demonstrate water-leaving behaviour (Green et al. Reference Green, Dussart and Gibson1992) and it is possible that under natural conditions snails could become infected by direct penetration during periods spent on the banks of aquatic habitats.
A range of factors can influence the establishment of nematodes within molluscs. These include size/age of the host, density of larvae that snails are exposed to, age of the larvae, length of the exposure period, and temperature (Boev, Reference Boev1975; Yousif and Lammler, Reference Yousif and Lammler1975a; Li et al. Reference Li, Deng, Zhang and Yang1986; Morley and Moritt, Reference Morley and Morritt2006). However, large variations in the response of individual host-parasite associations occur with few generalizations. For example, smaller/younger hosts have been found to be more susceptible to A. cantonensis and M. capillaris (Boev, Reference Boev1975; Yousif and Lammler, Reference Yousif and Lammler1975a; Solomon et al. Reference Solomon, Paperna and Alkon1996a). In contrast, the size/age of the mollusc had little impact on the infectivity of some protostrongylids (Cabaret, Reference Cabaret1987). For terrestrial molluscs, at least, the relative importance of this factor appears to be associated with the species of host under exposure with some demonstrating higher susceptibility in older/larger snails whilst others demonstrate the opposite (Cabaret, Reference Cabaret1987). Similarly, for the rhabditoid P. hermaphridita infecting L. stagnalis, smaller snails demonstrate a greater susceptibility than larger ones (Morley, unpublished observations).
Development of larvae in molluscs
Having penetrated into the molluscan host, larvae undergo further development. The rate of growth and maturation of nematodes is controlled by temperature. Increases in temperature cause a more rapid parasite development (Boev, Reference Boev1975; Ishii, Reference Ishii1984). Many species, however, have a minimum temperature threshold for development, below which the parasite becomes dormant. Such a threshold varies between species and appears to be related to their geographical distribution. For example, A. cantonensis, a tropical species, has a minimum temperature threshold of approximately 15°C (Ishii, Reference Ishii1984; Lv et al. Reference Lv, Zhou, Zhang, Liu, Zhu, Yin, Steinmann, Wang and Jia2006) whilst M. capillaris, a temperate species, develops down to 5°C in terrestrial molluscs (Rose, Reference Rose1957). In contrast, the subarctic species Elaphostrongylus rangiferi will only develop down to a 10°C threshold. This is an adaptation to minimize larval mortality during winter because developing L2 and L3 have a higher mortality rate than non-developing L1 in the over-wintering snail host (Schjetlein and Skorping, Reference Schjetlein and Skorping1995).
Development rates can vary depending on the species of mollusc infected. For example, A. cantonensis develops more quickly in Lymnaea palustris than Biomphalaria glabrata (Rachford, Reference Rachford1976a) whilst greater numbers of P. hermaphridita developed to L4 stage in L. stagnalis compared to Physa fontinalis after 14 days post-exposure (Morley and Morritt, Reference Morley and Morritt2006). In addition, the site of infection within the host can also influence parasite development due to local fluctuations in the availability of nutrients (Svarc and Zmoray, Reference Svarc and Zmoray1974).
In contrast, aestivation of the molluscan host does not appear to interfere with the development of nematode parasites (Gerichter, Reference Gerichter1948; Richards, Reference Richards1967), although it progresses at a slower rate than in active molluscs (Boev, Reference Boev1975; Solomon et al. Reference Solomon, Paperna and Markovics1996b). This may be because in the aestivating host there is a decline in available nutrients and oxygen uptake which may suppress parasite metabolic activity as well as a build-up of snail excretory by-products which may inhibit larval growth (Solomon et al. Reference Solomon, Paperna and Markovics1996b).
Pathology of infection to molluscs
A number of patho-physiological changes occur in molluscs infected with metastrongyloid nematodes. Larvae cause localized damage to surrounding tissue and are usually encapsulated by the host (Harris and Cheng, Reference Harris and Cheng1975a, Reference Harris and Chengb; Rachford, Reference Rachford1976b; Hourdin et al. Reference Hourdin, Rondelaud and Cabaret1990). Changes in the dimensions of nerve ganglia as well as neural lesions have also been reported in infected molluscs (Szmidt-Adjide et al. Reference Szmidt-Adjide, Rondelaud, Dreyfuss and Cabaret1996) possibly related to parasite-associated tissue necrosis (Hourdin et al. Reference Hourdin, Rondelaud and Cabaret1990). Specific studies on Angiostrongylus spp. show that infected molluscs have higher, but fluctuating, oxygen uptake rates (Rachford, Reference Rachford1976 c) with levels of haemolymph glucose and calcium, but not protein, significantly reduced immediately following nematode infection (Brockelman, Reference Brockelman1978; Brockelman and Sithithavorn, Reference Brockelman and Sithithavorn1980; Stewart et al. Reference Stewart, Ubelaker and Curtis1985). However, after a few weeks of infection, levels of haemolymph glucose and enzymes, as well as digestive gland glucose, are significantly elevated compared to controls (Stewart et al. Reference Stewart, Ubelaker and Curtis1985). These increases may be associated both with glycogen breakdown in the foot muscle, the main site of infection, where tissue glucose levels are significantly reduced and additional tissue destruction caused by the emergence of mature larvae from the snail (Stewart et al. Reference Stewart, Ubelaker and Curtis1985).
Cellular responses to metastrongyloid nematode infections can be classified as either focal or generalized proliferative reactions and appear similar in both terrestrial and aquatic molluscs. The intensity and structure of focal responses depend to some extent on the type of tissue in which the larvae are localized but, regardless, eventually lead to encapsulation of the parasite (Yousif et al. Reference Yousif, Blahser and Lammler1980; Hourdin et al. Reference Hourdin, Rondelaud and Cabaret1990). Generalized responses include an increase in circulating haemocytes immediately following infection, which may be related to the penetration of larvae through the intestinal wall. After a few days levels of these circulating cells decline, probably due to their removal for encapsulation of the parasite. Nevertheless, the haematopoietic organ remains enlarged indicating continued elevated levels of haemocyte production (Noda and Sato, Reference Noda and Sato1990). A further factor determining both focal and generalized responses would appear to correspond to the degree of susceptibility of individual mollusc species. Those species with low susceptibility demonstrate a stronger cellular reaction (Yousif et al. Reference Yousif, Blahser and Lammler1980).
Rhabditoid infections in aquatic molluscs can cause extensive mortalities (Li et al. Reference Li, Deng, Zhang and Yang1986; Azzam and Tawfik, Reference Azzam and Tawfik2003; Azzam and Belal, Reference Azzam and Belal2006; Morley and Morritt, Reference Morley and Morritt2006). Nevertheless, no specific details on pathological changes are known. It is likely that they are similar to the effects on terrestrial molluscs. In these hosts the effects of only a limited number of parasites have been elucidated that indicate, in general, that pathology increases with increasing parasite intensity (Morand et al. Reference Morand, Wilson, Glen and Barker2004). In addition, infections of P. hermaphrodita are known to cause fluid to accumulate in the shell cavity leading to large-scale swelling with both host feeding and activity reduced (Rae et al. Reference Rae, Verdun, Grewal, Robertson and Wilson2007). The nematodes liberate bacteria on which they feed, from their gut into the host haemolymph where the bacteria multiply rapidly and produce endotoxin. Bacterial septicaemia eventually causes death of the host allowing the nematode to feed and reproduce throughout the cadaver (Rae et al. Reference Rae, Verdun, Grewal, Robertson and Wilson2007). In contrast, a protein constituent of the haemolymph plasma of Helix aspersa inhibited maturation and reproduction, but not growth, of the rhabditoid nematodes Rhabditis maupasi and Steinernema glaseri. This inhibitory factor may be part of the snail's defence mechanism or simply be a cue utilized by the larvae to synchronise its life cycle with that of its host (Ratanarat-Brockelman, Reference Ratanarat-Brockelman1975, Reference Ratanarat-Brockelman1977). The presence of such an inhibitor may be an important factor in the low susceptibility of certain mollusc species to infection.
The cellular response to rhabditoid nematodes by molluscs is poorly understood. Encapsulation of larvae has only been reported in terrestrial species that have a low susceptibility to infection (Rae et al. Reference Rae, Robertson and Wilson2008). Interactions with aquatic molluscs remain unknown but are probably similar.
Aquatic molluscs as paratenic hosts
Aquatic molluscs can also act as paratenic hosts for terrestrial nematodes (Table 1) including species that may or may not already utilize snails as intermediate hosts. A paratenic host is one in which the parasite does not undergo any development and may or may not be normally required for transmission to the target host. For nematodes that already utilize molluscs as intermediate hosts the introduction of a further paratenic host can occur because freshwater snails will often feed on the cadavers of dead molluscs (Daldorph and Thomas, Reference Daldorph and Thomas1991). Some nematode larvae can survive and remain viable in decomposing host tissue for many days (Richards and Merritt, Reference Richards and Merritt1967) and if the infected cadaver is consumed by scavenging snails then they will ingest and retain the parasite (Rachford, Reference Rachford1975). In contrast, species such as Syngamus trachea, Pneumonema tiliquae and Rhabdias spp. can complete their life cycles directly without the use of an intermediate host but have been experimentally demonstrated to infect aquatic molluscs (Table 1). Studies on S. trachea indicate that eggs containing fully developed larvae when deposited into water are ingested by the mollusc and hatch in the digestive tract before penetrating the intestinal wall and locating in the surrounding tissue, with little indication of any host immune response and parasite encapsulation. Snails demonstrate reduced activity after infection but often die when exposed to high parasite numbers. Nematodes retain their viability in the snail host and are infective to the definitive bird host. However, after the death of the mollusc, S. trachea larvae did not survive long in the cadaver (Barus, Reference Barus, Ergens and Rysavy1964). The role aquatic molluscs may play as paratenic hosts for these species under natural conditions remains to be determined. Nevertheless, it cannot be ruled out that infections of S. trachea recorded in wild ducks (Lapage, Reference Lapage1961) may be associated with freshwater snails as they form part of the diet for many waterfowl species.
Emergence and further development of nematodes
Once full nematode development in the snail has been achieved a progression to the next stage of the life cycle becomes a priority. For metastrongyloid nematodes this requires transmission to the definitive vertebrate host that, in many reported cases, occurs by ingestion of the infected mollusc (Anderson, Reference Anderson2000). However, emergence of L3 from snails has been documented in a wide-ranging number of species (Heyneman and Lim, Reference Heyneman and Lim1967; Boev, Reference Boev1975; Kontrimavichus et al. Reference Kontrimavichus, Delyamure and Boev1976; Kutz et al. Reference Kutz, Hoberg and Polley2000). Emergence is not without controversy; with conflicting opinions on the relative ability of certain species to not only achieve it (e.g. Boev, Reference Boev1975), but also as to its significance as a route of infection (Kralka and Samuel, Reference Kralka and Samuel1984). Environmental factors, such as rainfall, have been suggested to both trigger and increase the rate of emergence (Boev, Reference Boev1975; Barcante et al. Reference Barcante, Barcante, Dias and Lima2003), and may be associated with the degree of stress that changes in these parameters cause to the host snail (Barcante et al. Reference Barcante, Barcante, Dias and Lima2003). However, in other cases emergence appears to be related to the ongoing development of the L3 stage (Kutz et al. Reference Kutz, Hoberg and Polley2000). Death of the host can also trigger a large-scale emergence of larval nematodes (Crook et al. Reference Crook, Fulton and Supanwong1971). Emerged L3 may survive for many days or even months in water, dependent on the species, (Kutz et al. Reference Kutz, Hoberg and Polley2000; Barcante et al. Reference Barcante, Barcante, Dias and Lima2003) whilst retaining their infectivity to the target host (Barcante et al. Reference Barcante, Barcante, Dias and Lima2003).
Transmission of free-living L3 is likely to be achieved under natural aquatic conditions to target hosts through ingestion either whilst drinking from the edge of water bodies or grazing on bank-side flora that has been contaminated with larvae. Nevertheless, infection can also occur through other routes. Larvae can penetrate through abraided skin (Eckert and Lammler, Reference Eckert and Lammler1972; Ubelaker et al. Reference Ubelaker, Caruso and Pena1981; Wang et al. Reference Wang, Chao and Chen1991) and, in the case of mice exposed to A. cantonensis, successful infections occur through unabraided skin, the footpad, and anal/vaginal/conjunctival mucosa, but not through the thicker skin of the tail (Wang et al. Reference Wang, Chao and Chen1991). However, it remains unknown whether L3 nematodes can penetrate a vertebrate host in a water medium. Infection may also occur through the ingestion of aquatic paratenic hosts such as fish, frogs and crustaceans that acquire and retain L3 whilst feeding on parasitized molluscs (Anderson, Reference Anderson1962; Wallace and Rosen, Reference Wallace and Rosen1966, Reference Wallace and Rosen1967; Ash, Reference Ash1968; Bolt et al. Reference Bolt, Monrad, Frandsen, Henriksen and Dietz1993).
In contrast, the emergence of rhabditoid nematodes from aquatic snails is not understood. It is often the case in terrestrial molluscs that, after developing to the L4, they either migrate out of the host to complete their life cycle in the soil or wait for the death of the host to continue development on the cadaver (Morand et al. Reference Morand, Wilson, Glen and Barker2004). To facilitate this process nematodes may release symbiotic bacteria whose excreted toxins cause the death of the snail. Mortalities of infected aquatic molluscs occur (Li et al. Reference Li, Deng, Zhang and Yang1986; Azzam and Tawfik, Reference Azzam and Tawfik2003; Azzam and Belal, Reference Azzam and Belal2006; Morley and Morritt, Reference Morley and Morritt2006) which could be associated with bacterial toxins. Furthermore, in laboratory studies of aquatic snails infected with L4, many hosts have been found dead, attached to the sides of exposure beakers above the water line (Morley and Morritt, Reference Morley and Morritt2006). Although the reasons for this are unknown, if such water-leaving behaviour was replicated under natural conditions it would facilitate the transfer of larvae to the bank-side soil allowing completion of their life cycle.
COMPARATIVE SUSCEPTIBILITY OF AQUATIC MOLLUSCS TO TERRESTRIAL NEMATODES
The ability of nematodes to infect molluscs varies from one host species to another and is therefore an important consideration for determining the epidemiology of disease occurrence. Davtjan (Reference Davtjan1945) was the first to classify molluscs into categories according to their susceptibility to nematode parasites. (1) The obligatory group with the highest prevalences and intensity of infections in which parasite development proceeds more quickly compared to other molluscs under the same conditions. (2) The subobligatory group that has lower prevalences and intensity infections with longer periods of parasite development. (3) The faculative group in which the parasite occurs rarely and, as a rule, develops only with difficulty. (4) The mortal group in which the nematode larvae die after penetration without achieving further developmental stages. (5) The resistant group in which no nematodes are able to infect the host.
Unsurprisingly, given their phylogenetic closeness to land molluscs, aquatic pulmonates are susceptible to a wide range of terrestrial nematodes (Table 1). Prosobranchs, although studied less often, appear almost equally susceptible; however, their operculum may form an effective barrier to infection when they withdraw into the shell. Nevertheless, in general, species from both groups can be classified as being subobligatory in comparison to the susceptibility of terrestrial molluscs, although large variations in prevalence and intensity can occur between species (Yousif and Lammler, Reference Yousif and Lammler1975b; Azzam and Tawfik, Reference Azzam and Tawfik2003). The notable exception to this generalization being A. cantonensis-B. glabrata where experimental studies have demonstrated a high compatibility between host and parasites comparable with many highly susceptible terrestrial mollusc hosts such as Achatina fulica, and this association is therefore likely to be obligatory. However, under field conditions B. glabrata has proven to be less susceptible than other freshwater snail species (Azzam and Belal, Reference Azzam and Belal2006).
In contrast, infections in bivalve molluscs have only rarely been investigated. It has been established that Crassostrea virginica, C. rizophorne, Mercenaria mercenaria, and Pisidium abditum can act as experimental hosts of Angiostrongylus cantonensis (Cheng and Burton, Reference Cheng and Burton1965; Richards and Merritt, Reference Richards and Merritt1967; Arrinda et al. Reference Arrinda, Perera and Yong1989) although few third-stage larvae develop (Cheng, Reference Cheng1966). However, in contradiction to these studies, Knapp and Alicata (Reference Knapp and Alicata1967) were unable to experimentally infect both C. virginica and Venerupis philippinarum with this nematode parasite. Cheng (Reference Cheng1967) considered that their negative result was not unexpected with an experimental set-up that impeded ‘pumping’, an essential part of the oysters feeding mechanism. Infections of C. virginica only occur if first-stage larvae of A. cantonensis are ingested, and the parasite is able to successfully penetrate the host's intestinal wall (Cheng, Reference Cheng1966). The geographical strain of bivalve used may also be an additional factor in susceptibility (Cheng, Reference Cheng1967). Nevertheless, it is apparent that bivalves are not best suited to be infected by terrestrial nematodes and therefore can only be considered as ‘facultative’ hosts.
Nevertheless, the compatibility of any particular aquatic mollusc as a host may not be fixed. Susceptibility may be fluid and under continuous parasite exposure may increase. The best examples of this adaptability can be found in studies on trematodes parasitizing pulmonates. When these parasites form a new relationship with an unusual snail host the adaptation of the trematode to the mollusc can occur rapidly due to repeated passage between intermediate and definitive hosts (Boray, Reference Boray1973). The speed of such adaptation depends on the relative ecological and biological suitability of the particular snail for larval development and the longevity of the parasite in the definitive host (Boray, Reference Boray1969). In mollusc-nematode associations only limited information of this kind is available. For example, both B. glabrata and Lymnaea palustris are susceptible aquatic hosts for A. cantonensis. However, the percentage of recovered adult parasites from rats was significantly higher following development in B. glabrata than in L. palustris. Nevertheless, repeated serial passages caused a gradual increase in the number of recovered adult nematodes from those cycled through L. palustris until after 4 serial passages the degree of recovered A. cantonensis from rats was the same from development in either snail species (Kocan, Reference Kocan1972). Consequently, aquatic molluscs found to be only ‘facultative’ auxiliary hosts during initial studies may achieve increased compatibility with repeated exposure.
The natural occurrence of parasites within populations may also be profoundly influenced by infra-specific variations in infectivity for either intermediate or definitive host that will shape the ecology of the host-parasite relationship (Webbe, Reference Webbe1971). Nematode parasites of vertebrates demonstrate large geographical variations in infectivity to animal hosts generally regarded as highly susceptible, suggesting that geographical strains are adapted to locally common suitable hosts (Webbe, Reference Webbe1971). Many genetic strains of invertebrate hosts, particularly insects, have also demonstrated variations in susceptibility to nematodes, which may account for many conflicting results of different studies (Webbe, Reference Webbe1971).
Not surprisingly, similar evidence has arisen in studies on mollusc-nematode interactions. For example, an Egyptian and a German strain of Physa acuta demonstrated different degrees of susceptibility to A. cantonensis in the same study (Yousif and Lammler, Reference Yousif and Lammler1975b), whilst discrepancies between different studies undertaken on infectivity of E. rangiferi and M. capillaris to Lymnaea stagnalis (Skorping, Reference Skorping1982; Zdzitowiecki, Reference Zdzitowiecki1976) may also be related to host strain. Studies on terrestrial snail susceptibility to A. cantonensis have demonstrated that when molluscs from geographical areas where the parasite does not occur are introduced into endemic areas they acquire a lower prevalence of infection, at least initially, than native hosts (Noda et al. Reference Noda, Matayoshi, Uchikawa and Sata1985).
It therefore seems likely that the comparative susceptibility of individual aquatic molluscs to specific terrestrial nematode species may not retain either widespread geographical homogeneity or long-term stability and establishment in any particular aquatic habitat will depend on both the nematode's ability to adapt to new situations, and the favourability of ecological conditions for achieving robust host-parasite associations.
NATURAL NEMATODE INFECTIONS OF AQUATIC MOLLUSCS
It is apparent from these extensive experimental studies that aquatic molluscs may act as hosts for a range of terrestrial nematode parasites. Nevertheless, in the laboratory barriers that may regulate a host-parasite relationship (ecological, geographical and seasonal interactions) are largely removed, facilitating infections in potentially more wide-ranging species.
Studies under natural conditions are therefore of paramount importance to assess the role that aquatic molluscs may play. The transfer of free-living terrestrial nematodes present in the soil into aquatic habitats is relatively straightforward and well documented. Larvae deposited on the soil are mainly disseminated over large distances by water, prior to their contact with a host (Croll, Reference Croll1975). Storm runoff water will eventually deposit nematodes into a range of both large and small aquatic habitats but particularly into flowing water of streams, rivers and irrigation/drainage canals (Faulkner and Bolander, Reference Faulkner and Bolander1970; Mott and Harrison, Reference Mott and Harrison1983). However, there is little tendency for nematodes to settle out in flowing water (Faulkner and Bolander, Reference Faulkner and Bolander1966) and it is only in areas where water velocity is very low, such as in static ponds and lakes as well as sheltered bank-side areas of streams and rivers, that nematodes are likely to accumulate and be a source of infection to aquatic molluscs.
Once contamination of a freshwater habitat occurs the risk of long-term establishment will be dependent on a range of ecological and physiological characteristics of the water body. In particular, the density of susceptible aquatic populations and the frequency of visitation by terrestrial hosts.
Although general surveys of nematode parasites in freshwater molluscs remain rare (Bartlett and Anderson, Reference Bartlett and Anderson1985; Azzam and Belal, Reference Azzam and Belal2006), and the occurrence of terrestrial rhabditoids in these snails merely noted (Azzam and Belal, Reference Azzam and Belal2006), a limited number of field studies on specific metastrongyloid species in aquatic hosts have been undertaken (Table 1). In particular, because of the widespread human consumption of raw or partially cooked freshwater snails in endemic areas, the medically important metastrongyloid A. cantonensis has been extensively studied concurrently in both terrestrial and freshwater habitats.
In general, A. cantonensis is a tropical species but has been found in both the colder climate of northern Japan (Ishii, Reference Ishii1984) and at the edge of the Sahara desert in the Egyptian Nile basin (Yousif and Ibrahim, Reference Yousif and Ibrahim1978; El-Shazly et al. Reference El-Shazly, El Hamshary, El Shewy, Rifaat and El Sharkawy2002; Ibrahim, Reference Ibrahim2007). Transmission is favoured in the majority of sites through terrestrial molluscs, particularly the giant African land snail, Achatina fulica (Alicata and Jindrak, Reference Alicata and Jindrak1970). However, in certain habitats where more extreme environmental conditions prevail, such as the flooded rice fields of South-East Asia and the Egyptian waterways, aquatic molluscs take a more prominent role.
Surveys have shown that infected freshwater snails are only recovered on the edges of water bodies (Tesana et al. Reference Tesana, Srisawangwong, Sithithaworn, Laha and Andrews2009) where the ecotone between terrestrial and aquatic hosts occurs. Prevalence of infections can fluctuate from one aquatic habitat to another, probably associated with the density of potential hosts and the ecology of individual water bodies (Tesana et al. Reference Tesana, Srisawangwong, Sithithaworn, Laha and Andrews2009). Indeed, the related species A. malaysiensis has a higher prevalence in freshwater snails sampled from static water sites (rice-fields and ponds) than flowing water sites (streams) (Lim and Ramachaudran, Reference Lim, Ramachandran and Cross1979). Increasing levels of salinity cause a reduction in the occurrence of A. cantonensis infections (Ibrahim, Reference Ibrahim2007). Although L1 have been experimentally demonstrated to tolerate high salinity levels they have a reduced activity (Richards and Merritt, Reference Richards and Merritt1967), this may impair their ability to infect aquatic molluscan hosts under these conditions.
The occurrence of terrestrial nematodes in aquatic molluscs may change temporally as well as spatially. Seasonal variations in A. cantonensis prevalence occur in aquatic molluscs with greater prevalence during the ‘dry’ compared to ‘rainy’ season in the tropics (Yen et al. Reference Yen, Chen and Cheng1990), possibly due to heavy rainfall diluting the number of larvae in the irrigation canals sampled, and the increasing water flow reducing transmission opportunities (Yen et al. Reference Yen, Chen and Cheng1990). Similarly in Egypt, prevalences were highest in spring and summer when aquatic molluscs were most active (Ibrahim, Reference Ibrahim2007). Periodic disturbance of the habitat can also affect infection levels in freshwater snails. Angiostrongylus malaysiensis has higher prevalences in aquatic molluscs sampled from rice-fields used continuously than in those fields allowed to dry out for several months after crop harvesting (Liat et al. Reference Liat, Fong and Krishnansamy1977). The drying out of temporary water bodies such as rice-fields may interfere with the interactions between aquatic and terrestrial host-parasite systems at a local level. Nevertheless, definitive rodent hosts can simply switch to feeding on aquatic molluscs at nearby fields that remain flooded, ensuring that transmission dynamics are unaffected at a regional scale.
It is likely that the risk of aquatic mollusc acquiring infections will vary from one terrestrial nematode species to another. For rhabditoids infecting terrestrial molluscs the chance of transfer into aquatic habitats is high. Many species of both land snails and slugs have a high affinity for moist soil conditions and can often be found in relatively high densities close to freshwater habitats (Boycott, Reference Boycott1934; Horsak and Cernohorsky, 2008). Such a close association provides a relatively straightforward route for the transfer of nematode larvae, whether in the soil or within terrestrial mollusc cadavers, into the water medium during periods of rainfall.
In contrast, for metastrongyloids the relationship the definitive vertebrate host has with open bodies of water will be a key factor in determining infection levels in freshwater snails. For example, infections of M. capillaris from aquatic sources to sheep and goats are considered to take place mainly when these animals are drinking from shallow water bodies. However, because they drink only a few times during the day and the act of drinking is itself relatively brief they have only a transient relationship with water, and consequently the transmission window for these nematodes to both infect and emerge from freshwater snails is limited (Egorov, Reference Egorov1960). Such a narrow window of opportunity may be a factor, in conjunction with geographical variations in susceptibility, for the conflicting results of field studies on M. capillaris infections in freshwater molluscs. Trushin (Reference Trushin1976) examined almost 14 000 aquatic snails but found none infected in Tver province, northwest of Moscow. In contrast, Egorov (Reference Egorov1960) found 0·3% of freshwater molluscs infected, which compared favourably with the 1·07% of terrestrial snails and 5·2% of slugs infected in the same endemic M. capillaris area near St Petersburg, Russia.
However, other definitive hosts, such as rodents infected with A. cantonensis, are often found associated with aquatic habitats feeding on freshwater molluscs and other potential paratenic animal hosts (Yousif and Ibrahim, Reference Yousif and Ibrahim1978), facilitating a smoother cycling between terrestrial and aquatic environments. This is reflected in the widespread occurrence of A. cantonensis in freshwater molluscs at these localities (Ibrahim, Reference Ibrahim2007; Tesana et al. Reference Tesana, Srisawangwong, Sithithaworn, Laha and Andrews2009). Indeed, in the related species, A. malaysiensis, variations in the worm-burden of four species of rat sampled from the vicinity of aquatic habitats were not due to differences in their susceptibility but rather were associated with local habitat preference and fluctuations in prevalence and intensity within molluscan hosts found there (Liat et al. Reference Liat, Fong and Krishnansamy1977). Mustelids, such as mink, also have strong affinities with water bodies, readily feeding on fish and frogs that may act as paratenic hosts for species such as Aelurostrongylus pridhami and Filaroides martis, increasing the likelihood of aquatic molluscs acting as intermediate hosts (Anderson, Reference Anderson1962).
However, natural transmission of metastrongyloids to vertebrate hosts via the aquatic medium is difficult to prove conclusively. In particular, infections from drinking water contaminated with L3 appear almost impossible to determine in any individual case or large outbreak (Rosen et al. Reference Rosen, Loison, Laigret and Wallace1967). Nevertheless, sources of human infections of A. cantonensis can be sufficiently narrowed down through patient interviews to a few probable routes of exposure. These findings suggest that in a number of cases transmission occurred through ingestion of infected aquatic molluscs, crustaceans, frogs, fish, or bank-side fruit contaminated with L3 (Witoonpanich et al. Reference Witoonpanich, Chuahirun, Soranastaporn and Rojanasunan1991; Brown et al. Reference Brown, Mohareb, Yousif, Sultan and Girgis1996; Thobois et al. Reference Thobois, Broussolle, Aimard and Chazot1996; Lai et al. Reference Lai, Yen, Chin, Chung, Kuo and Lin2007; Malvy et al. Reference Malvy, Ezzedine, Receveur, Pistone, Crevon, Lemardeley and Josse2008). It seems logical to assume that wildlife infections may also be acquired in a similar manner.
AQUATIC MOLLUSCS AS HOSTS IN A CHANGING CLIMATE
Global climate change resulting from anthropogenic activities is now a widely accepted phenomenon. Alterations in temperature, rainfall patterns and soil humidity are likely to occur on a global, regional, and local scale causing changes in average values and an increase in the frequency of extreme events such as flooding and droughts (Morand and Guegan, Reference Morand and Guegan2008). A large range of important factors that structure complex host-parasite systems may alter under the pressure of climate change (Hoberg et al. Reference Hoberg, Polley, Jenkins and Kutz2008). These alterations may be numerical (density, prevalence, and abundance changes in both hosts and parasites), functional (altered ecological structure, geographical distribution, phenology or host associations), or micro-evolutionary through local adaptation (Hoberg et al. Reference Hoberg, Polley, Jenkins and Kutz2008). Although changes in host-parasite dynamics will likely fluctuate between geographical localities and the individual requirements of each species interactions, it is probable that climate change will influence helminths most strongly in temperate and colder latitudes where modifications of climatic variables are more pronounced (Mas-Coma et al. Reference Mas-Coma, Valero and Bargues2008). Evidence for this is already emerging in a nematode-mollusc-ungulate system within the Arctic (Kutz et al. Reference Kutz, Hoberg, Polley and Jenkins2005; Jenkins et al. Reference Jenkins, Veitch, Kutz, Hoberg and Polley2006) because the transmission of parasites through terrestrial molluscs is particularly susceptible to fluctuating climatic conditions (Georgiev et al. Reference Georgiev, Kostadinova and Georgiev2003; Jenkins et al. Reference Jenkins, Veitch, Kutz, Hoberg and Polley2006; Morley and Lewis, Reference Morley and Lewis2008). Extreme changes in climate can therefore have profound effects on these host-parasite relationships. The resulting ecological perturbations will alter the dynamics of parasite transmission, increasing the potential for host switching and facilitating new host-parasite associations (Brooks and Hoberg, Reference Brooks and Hoberg2007). It therefore seems logical to assume that mollusc-nematode interactions will be particularly susceptible to both large and small scale shifts in the spectrum of hosts utilized and the pressure of extreme climatic events may lead to a greater role for aquatic snails.
Changes in temperature, rainfall and vegetation could modify transmission through the molluscan host causing changes in geographical distribution, density and survival of both parasite and snail (Mas-Coma et al. Reference Mas-Coma, Valero and Bargues2008). For example, an increase in average temperatures has the obvious effect of accelerating development of nematodes within the mollusc. This may be particularly important in high latitudes where cycling of the parasite may be accomplished in a single summer rather than over multiple years (Kutz et al. Reference Kutz, Hoberg, Polley and Jenkins2005). At temperate latitudes it may facilitate cycling to occur numerous times over the course of a year whilst the geographical range of certain species may expand e.g. A. vasorum (Morgan et al. Reference Morgan, Jefferies, Krajewski, Ward and Shaw2009).
Rainfall is a particular key factor in defining the scope of aquatic molluscs' involvement in terrestrial nematode transmission. Increased occurrence of heavy, persistent rainfall provides more and improved aquatic habitats for snail development, whilst at the same time escalates the risk of widespread flooding. This can be particularly devastating to terrestrial molluscs causing ‘over-hydration’ in exposed communities resulting in high mortalities from drowning (Peake, Reference Peake, Fretter and Peake1978), and in protractedly wet environments population densities can often be reduced (Peake, Reference Peake, Fretter and Peake1978; Plum, Reference Plum2005) whilst new species of aquatic molluscs may appear (Plum, Reference Plum2005). Storm and flood runoff water will deposit cadavers of drowned infected terrestrial molluscs into both still and flowing water environments (Wallace and Rosen, Reference Wallace and Rosen1967) where they can form a source of infection for a range of paratenic aquatic hosts, including molluscs, crustaceans, frogs and fish, that scavenge on the carcasses. Mature L3 may also emerge from drowned terrestrial molluscs in the freshly contaminated body of water increasing the risk of infecting wildlife that may frequent it. Flooding will also cause an increased frequency in the transfer of free-living nematode stages into aquatic habitats. Under such circumstances aquatic molluscs will naturally acquire infections and play a greater role in the transmission of many parasites. This scenario has been elegantly demonstrated by field studies on the occurrence of Angiostrongylus spp. within the flooded rice-fields and waterways of South-East Asia (Liat et al. Reference Liat, Fong and Krishnansamy1977; Lim and Ramachandran, Reference Lim, Ramachandran and Cross1979; Tesana et al. Reference Tesana, Srisawangwong, Sithithaworn, Laha and Andrews2009).
At the other end of climatic extremes drought, especially during the summer when it may be accompanied by a heat wave, can also have devastating effects on parasite transmission. Terrestrial molluscs, under such conditions, spend prolonged periods aestivating, reducing the likelihood of contact with free-living parasite stages (Morley and Lewis, Reference Morley and Lewis2008). Under such constraints nematodes may have to spend extensive periods in the environment exposed to adverse climatic conditions. Survival in direct sunlight may be impaired by the heat generated in or on the soil, by the increased risk of desiccation, and by the pathological effects of elevated ultraviolet light levels (Kates, Reference Kates1965; Gaugler and Boush, Reference Gaugler and Boush1978; van Dijk et al. Reference Van Dijk, Louw, Kalis and Morgan2009). Consequently under arid or semi-arid conditions only ‘minimal transmission’ will normally occur (Kates, Reference Kates1965), infections being maintained by small areas where soil moisture and some shade are provided (Kates, Reference Kates1965; Glazer et al. Reference Glazer, Liran and Steinberger1991). Bank-side areas with overhanging vegetation of both small and large water bodies provide the conditions necessary to maintain ‘minimal transmission’ under such adverse climate and are likely to prove attractive to many potential terrestrial hosts. With increased contact with water comes a more likely chance of aquatic molluscs becoming infected as these hosts do not need to aestivate to the same extent as their terrestrial counterparts, and elevated numbers of nematodes will be deposited in the soil within the wash zone of the aquatic habitat. Again the occurrence of A. cantonensis in the desert environment of the Egyptian Nile basin (El-Shazley et al. Reference El-Shazly, El Hamshary, El Shewy, Rifaat and El Sharkawy2002; Ibrahim, Reference Ibrahim2007) provides an elegant model for nematode transmission under these conditions.
Nevertheless, climate does not act gradually or entirely predictably on ecosystems. Instead, in combination with other influences, such as geographical variables, it produces threshold effects on populations inducing changes in infectious disease dynamics (IOM, 2008). The nature of host utilization by nematodes is therefore likely to be influenced either directly or indirectly by local, regional, or global variations in a range of factors, that are nevertheless closely interrelated and influenced by climate, but may produce variable pressure on nematode transmission dynamics and consequently result in fluctuating utilization of aquatic molluscs as hosts.
CONCLUDING REMARKS
It is apparent from this study that much of what is known on terrestrial nematodes parasitizing aquatic molluscs arises from species that are of veterinary, medical, or commercial interest. Such a selection will tend to limit the extent of our understanding and therefore more basic research is necessary, particularly for rhabditoids of which almost nothing is known about their natural occurrences in aquatic molluscs.
Nevertheless, it seems likely that, geographically, infections of aquatic molluscs most likely occur adjacent to those terrestrial habitats associated with a relatively high nematode density and where the ecotone between land and water is particularly broad and continuous.
That aquatic molluscs are the dominant hosts for A. cantonensis in both arid and flooded environments is particularly striking and suggests that the environmental conditions in a habitat play a key role in structuring the intermediate host fauna utilized by nematodes. As the impact of climate change becomes increasingly apparent, the deterioration of standard host-parasite associations is increasingly likely as traditional habitats decline in transmission viability. Under such extreme environments the role of aquatic molluscs in transmitting terrestrial nematodes may therefore become increasingly prominent.
