Cryptosporidium

Zahedi, Alireza, Paparini, Andrea, Jian, Fuchun, Robertson, Ian & Ryan, Una, 2016, Public health significance of zoonotic Cryptosporidium species in wildlife: Critical insights into better drinking water management, International Journal for Parasitology: Parasites and Wildlife 5 (1), pp. 88-109 : 95-97

3.1. Cryptosporidium in mammals

Due to the morphological similarity of Cryptosporidium oocysts from different host species, initial findings of Cryptosporidium infections in wild animals were assumed to be due to C. parvum leading to an overestimation of the potential role of wildlife as reservoirs of human disease ( Appelbeea et al., 2005). However, with the assistance of advanced molecular techniques, many of these species were identified as host-adapted genotypes ( Table 2). Both wild terrestrial and marine mammals have been studied as potential reservoirs for human-infectious Cryptosporidium species and genotypes using molecular tools ( Table 2). The prevalence of Cryptosporidium in wild placental mammal hosts has been reported in detail in a recent review ( Feng, 2010) and varies widely between mammalian hosts.

3.1.1. Cryptosporidium hominis

Although humans are the major host species for C. hominis , it has been reported in a number of wildlife hosts including a dugong and non-human primates ( Table 2) ( Xiao et al., 1999; Ye et al., 2012; Karim et al., 2014; Bodager et al., 2015; Parsons et al., 2015). C. hominis / Cryptosporidium parvum -like sequences were identified in red and black-and-white colobus monkeys in Uganda ( Salyer et al., 2012). However, typing was obtained using a short fragment of the Cryptosporidium oocyst wall protein (COWP) gene, which is not reliable for differentiating Cryptosporidium species. In Australia, a number of recent studies have also identified C. hominis / C. parvum - like isolates at the 18S locus in marsupials including bandicoots, brushtail possums, eastern grey kangaroos and brush-tailed rock-wallabies ( Hill et al., 2008; Ng et al., 2011; Dowle et al., 2013; Vermeulen et al., 2015). However, despite efforts, the identification of C. hominis / C. parvum could not be confirmed at other loci. This might be due to low numbers of oocysts and the multi copy nature of the 18S rRNA gene. Another study reported a C. hominis - like sequence at the 18S locus in a wild dingo, but was also unable to confirm this at other loci ( Ng et al., 2011).

Subtyping of C. hominis at the gp60 locus has identified nine subtype families (Ia to Ij) (Ryan et al., 2014). To date, few C. hominis subtypes have been reported in wild mammals but include subtype IbA12G 3 in Rhesus macaques, subtype IiA 17 in Cynomolgus monkeys and Rhesus monkeys and subtype IfA12G 2 in baboons and Mitumba chimpanzees ( Feng et al., 2011b; Karim et al., 2014; Bodager et al., 2015; Parsons et al., 2015).

Recently, C. hominis has been identified and enumerated from eastern grey kangaroos and cattle faecal samples from Sydney catchments and characterised at multiple loci (Zahedi et al., 2015). In that study, C. hominis isolates were typed at three loci (18S, a novel mucin-like glycoprotein that contains a C-type lectin domain and the gp60 gene) (Zahedi et al., 2015). The C. hominis IbA10G2 subtype was identified in the marsupials and cattle (Zahedi et al., 2015), which is the main subtype associated with outbreaks of cryptosporidiosis by C. hominis ( Xiao, 2010) .

3.1.2. C. parvum

C. parvum was first described in mice ( Tyzzer, 1912) and is primarily a parasite of artiodactyls and humans ( Xiao, 2010). C. parvum has however been frequently reported in wildlife, infecting a broad range of wild species including various rodents, bovids, camelids, equids, canids, non-human primates and marine mammals ( Table 2) ( Morgan et al., 1999a; Atwill et al., 2001; Perez and Le Blancq, 2001; Matsubayashi et al., 2004; Ryan et al., 2004; Appelbee et al., 2005; Feng et al., 2007; Meireles et al., 2007; Paziewska et al., 2007; Starkey et al., 2007; Ziegler et al., 2007; Gomez-Couso ́et al., 2012; Ye et al., 2012; Abu Samraa et al., 2013; Liu et al., 2013; García-Presedo et al., 2013b; Reboredo-Ferńandez et al., 2014; Montecino-Latorre et al., 2015; Wells et al., 2015; Matsui et al., 2000).

Few studies have identified C. parvum in captive wild mammals but red deer, fallow deer, addaxes, Arabian oryx, gemsboks, and sable antelopes are among mammals to be infected with C. parvum in captivity ( Perez and Le Blancq, 2001; Ryan et al., 2003; Hajdusek et al., 2004; Abe et al., 2006; Feng et al., 2007; Meireles et al., 2007; Matsubayashi et al., 2004; Bodager et al., 2015; Wang et al., 2015; Zhao et al., 2015a).

Subtyping of C. parvum at the gp60 locus has identified fourteen subtype families (IIa to IIo (Ryan et al., 2014)). Few studies which identified C. parvum in wild mammals have conducted typing at the gp60 locus but a variety of C. parvum subtypes including IIdA15G1, IIdA18G1, IIdA19G1 have been reported from golden takins, lemurs, chipmunks and hamsters, and IIaA15G2R1, IIaA19G2R1, IIaA19G3R1, IIaA19G4R1, IIaA20G3R1, IIaA20G4R1, IIaA20G3R2 and IIaA21G3R1 have been reported from deer and Eastern grey kangaroos (Lv et al., 2009; Bodager et al., 2015; Montecino-Latorre et al., 2015; Zhao et al., 2015a; Zahedi et al., 2015). These are all C. parvum subtypes that have been reported in humans ( Xiao, 2010).

3.1.3. Cryptosporidium cuniculus

C. cuniculus (previously known as rabbit genotype) was first described in rabbits by Inman and Takeuchi (1979), who described the microscopic detection and ultra-structure of endogenous Cryptosporidium parasites in the ileum of an asymptomatic female rabbit. Molecular characterisation of C. cuniculus was first conducted on rabbit faecal samples from the Czech Republic ( Ryan et al., 2003) and C. cuniculus was formally re-described as a species in 2010 ( Robinson et al., 2010). Since then, it has been described from rabbits across a wide geographic area including Australia, China, the UK, the Czech Republic, Poland, France and Nigeria ( Ryan et al., 2003; Nolan et al., 2010; Shi et al., 2010; Chalmers et al., 2011; Zhang et al., 2012; Nolan et al., 2013; Liu et al., 2014; Koehler et al., 2014; Puleston et al., 2014; Zahedi et al., 2015). C. cuniculus has a close genetic relationship with C. hominis and its zoonotic potential became clear in 2008, when it was responsible for a drinking-water associated outbreak of cryptosporidiosis in the UK ( Chalmers et al., 2009; Robinson et al., 2011; Puleston et al., 2014) and has also been identified in many sporadic human cases of cryptosporidiosis ( Robinson and Chalmers, 2011; Chalmers et al., 2011; Elwin et al., 2012; Koehler et al., 2014). It is also the third most commonly identified Cryptosporidium species in patients with diarrhoea in the UK ( Chalmers et al., 2011). Subtyping at the gp60 locus has identified two distinct subtype families, designated Va and Vb ( Chalmers et al., 2009). Most cases described in humans relate to clade Va and the first waterborne outbreak was typed as VaA22 ( Robinson et al., 2008; Chalmers et al., 2009). C. cuniculus has been reported in rabbits and humans (subtypes VaA9 ‾ VaA22 and VbA20 ‾ VbA37 ‾ see Wang et al., 2012) but has recently been identified in marsupials (subtype VbA26) (and a human ‾ subtype VbA25) in Australia ( Nolan et al., 2013; Koehler et al., 2014). The widespread occurrence of C. cuniculus genotypes in rabbits and the fact that it has been now been identified in marsupials in Australia suggests that C. cuniculus might be a species more ubiquitous than previously thought, and might be able to spread to other mammals as well as humans. Therefore, there is a need to diligently monitor for C. cuniculus in the vicinity of drinking water catchments and in drinking water.

3.1.4. Cryptosporidium ubiquitum

C. ubiquitum (previously cervine genotype, cervid, W4 or genotype 3) was first identified by Xiao et al. (2000) in storm water samples in lower New York State (storm water isolate W4, GenBank accession no. AF262328). Subsequently, Perez and Le Blancq (2001) identified this genotype in white-tailed deer-derived isolates from lower New York State and referred to it as genotype 3. Since then it has been described in a wide variety of hosts worldwide including humans and was formally described as a species in 2010 ( Fayer et al., 2010). C. ubiquitum is of public health concern because of its wide geographic distribution and broad host range ( Li et al., 2014). In addition to domestic animals (in particular sheep) and wildlife, C. ubiquitum has been frequently reported from drinking source water, storm water runoff, stream sediment and wastewater in various geographic locations, suggesting potential contamination of water sources with oocysts of C. ubiquitum shed by animals inhabiting water catchments ( Nolan et al., 2013; Li et al., 2014). C. ubiquitum is considered an emerging zoonotic pathogen ( Li et al., 2014), as it has been identified in many human cases of cryptosporidiosis in the United Kingdom, Slovenia, the United States, Canada, Spain, New Zealand, Venezuela and Nigeria (Charlmers et al., 2011; Wong and Ong, 2006; Fayer et al., 2010; Cieloszyk et al., 2012; Elwin et al., 2012; Blanco et al., 2015; Qi et al., 2015a).

In wildlife, C. ubiquitum has been reported sporadically in rodents, wild ruminants, carnivores, marsupials and primates ( Table 2) ( Perez and Le Blancq, 2001; da Silva et al., 2003; Ryan et al., 2003; Feng et al., 2007; Feng, 2010; Karanis et al., 2007; Ziegler et al., 2007; Wang et al., 2008; Fayer et al., 2010; Cinque et al., 2008; Robinson et al., 2011; Feng et al., 2011b; Abu Samraa et al., 2013; Mi et al., 2013; Murakoshi et al., 2013; Li et al., 2014; Ma et al., 2014; Perec-Matysiak et al., 2015; Qi et al., 2015a, 2015b; Stenger et al., 2015b; Vermeulen et al., 2015).

Because C. ubiquitum is genetically distant from C. hominis and C. parvum , until recently, gp60 homologs had not been sequenced. However, the gp60 gene of C. ubiquitum was identified by wholegenome sequencing and six subtype families (XIIa ‾ XIIf) within C. ubiquitum were identified ( Li et al., 2014). Application of this new tool to human, animal, and environmental (water) isolates has suggested that sheep and rodents are a key source of C. ubiquitum transmission to humans, through either direct human contact with infected animals or by contamination of drinking source water ( Li et al., 2014). For example, in the US, all C. ubiquitum specimens from humans characterized belonged to the same subtype families found in wild rodents in the US (XIIb, XIIc and XIId) ( Li et al., 2014). However, as persons in the United States usually have little direct contact with wild rodents, the authots concluded that transmission of C. ubiquitum to humans from rodents was likely to come from drinking untreated water contaminated by wildlife ( Li et al., 2014).

3.1.5. Cryptosporidium muris

C. muris is a gastric parasite and was first identified in the gastric glands of mice in 1907 by Tyzzer (1907). Since then, molecular tools have shown that it has a wide host range, including various mammals (rodents, canids, felids, suids, giraffida, equids, non-human primates and marsupials) and birds ( Tables 1 and 2). C. muris is considered a zoonotic species as there have been numerous reports of C. muris in humans and one report in human sewage ( Guyot et al., 2001; Gatei et al., 2002; Tiangtip and Jongwutiwes, 2002; Gatei et al., 2003; Palmer et al., 2003; Gatei et al., 2006; Leoni et al., 2006; Muthusamy et al., 2006; Azami et al., 2007; Al- Brikan et al., 2008; Neira et al., 2012; Hasajova ́et al., 2014; Petrincováet al., 2015; Spanakos et al., 2015; Hurkova et al., 2003).

In a recent human infectivity study, C. muris was examined in six healthy adults ( Chappell et al., 2015). Volunteers were challenged with 10 5 C. muris oocysts and monitored for 6 weeks for infection and/or illness. All six patients became infected. Two patients experienced a self-limited diarrhoeal illness. C. muris oocysts shed during the study ranged from 6.7 × 10 6 to 4.1 × 10 8, and C. muris - infected subjects shed oocysts longer than occurred with other species studied in healthy volunteers. Three volunteers shed oocysts for 7 months ( Chappell et al., 2015). The authors concluded that healthy adults are susceptible to C. muris , which can cause mild diarrhoea and result in persistent, asymptomatic infection ( Chappell et al., 2015), which confirms the zoonotic status of C. muris and highlights the public health risks of finding C. muris in wildlife in drinking water catchments.

3.1.6. Cryptosporidium andersoni

Like C. muris , C. andersoni is also a gastric parasite and primarily infects the abomasum of cattle and to a lesser extent, sheep and goats (Ryan et al., 2014; Wang et al., 2012). C. andersoni produces oocysts that are morphologically similar to, but slightly smaller than those of C. muris (7.4 ‾ 8.8 × 5.8 ‾ 6.6 M m vs 8.2 ‾ 9.4 × 6.0 ‾ 6.8 M m, respectively) and was originally mistakenly identified in cattle as C. muris based on its oocyst size. In 2000, it was described as a new species based on the location of endogenous stages in the abomasum, its host range, and genetic distinctness at multiple loci ( Lindsay et al., 2000). It has only occasionally being detected in wild animals ( Table 2) ( Ryan et al., 2004; Wang et al., 2008, 2015; Lv et al., 2009; Feng et al., 2010; Zhao et al., 2015a). Several studies have reported that C. andersoni is the dominant species in source and tap water (Feng et al., 2011; Nichols et al., 2010), suggesting that cattle may be the primary source of contamination. Interestingly, in a recent study, it was found at a prevalence of 15.6% (19/122) and 0.5% (1/200) in captive and wild giant pandas, respectively in China ( Wang et al., 2015). It is occasionally detected in humans ( Leoni et al., 2006; Morse et al., 2007; Waldron et al., 2011; Agholi et al., 2013; Jiang et al., 2014; Liu et al., 2014). Two studies in China by the same research group have reported that C. andersoni was the most prevalent Cryptosporidium species detected in humans ( Jiang et al., 2014; Liu et al., 2014). However, further research is required to better understand the zoonotic importance of C. andersoni .

3.1.7. Cryptosporidium canis

C. canis (previously dog genotype 1) was first identified as the dog genotype by Xiao et al. (1999) and described as a species in 2001 ( Fayer et al., 2001), on the basis that C. canis oocysts were infectious for calves but not mice and were genetically distinct from all other species. C. canis and its sub-genotypes ( C. canis fox genotype and C. canis coyote genotype) have been reported in dogs, foxes and coyotes ( Table 2) ( Xiao et al., 2002a; Zhou et al., 2004; Fayer, 2010; Feng, 2010). C. canis has also been reported worldwide in humans ( Lucio-Forster et al., 2010; Fayer, 2010; Elwin et al., 2012; Mahmoudi et al., 2015; Parsons et al., 2015).

3.1.8. Cryptosporidium erinacei

Little is known about epidemiology and pathogenicity of zoonotic C. erinacei in wildlife. C. erinacei (previously known as hedgehog genotype) was first identified morphologically in a captive four-toed hedgehog ( Ateletrix albiventris ) in 1998 ( Graczyk et al., 1998). An isolate from a European hedgehog originating from Denmark was typed in 2002 ( Enemark et al., 2002) and shown to be distinct. Subsequent studies have identified C. erinacei in hedgehogs, horses and humans ( Dyachenko et al., 2010; Laatamna et al., 2013; Kvá̆c et al., 2014a, 2014b; Meredith and Milne, 2009). At the gp60 locus, C. erinacei isolates are identified as subtype family XIII ( Dyachenko et al., 2010; Laatamna et al., 2013; Lv et al., 2009; Kvá̆c et al., 2014b). Previously reported C. erinacei subtypes include XIIIaA20R10 (KF055453), XIIIaA21R10 (GQ214085), XIIIaA22R9 (KC305644), XIIIaA19R12 (GQ214081), and XIIIaA22R11 (GQ259140) Kvá̆c et al., 2014b).

3.1.9. Cryptosporidium fayeri and Cryptosporidium macropodum

The two main species identified in a wide range of marsupials are C. fayeri and C. macropodum (previously marsupial genotype I and II) ( Table 2) ( Morgan et al., 1997; Power et al., 2004, 2005; Power and Ryan, 2008; Ryan et al., 2008; Nolan et al., 2010; Power, 2010; Ng et al., 2011a; Yang et al., 2011; Ryan and Power, 2012; Nolan et al., 2013; Vermeulen et al., 2015; Zahedi et al., 2015). Neither of these species is associated with diarrhoea in their marsupial hosts ( Ryan and Power, 2012). C. macropodum has not been reported in humans but cryptosporidiosis caused by C. fayeri has been reported in a 29-year-old female patient in Australia ( Waldron et al., 2010). The woman was immunocompetent but suffered prolonged gastrointestinal illness. The patient resided in a national forest on the east coast of New South Wales, Australia, an area where marsupials are abundant. She had frequent contact with partially domesticated marsupials ( Waldron et al., 2010). Identification of C. fayeri in a human patient is a concern for water catchment authorities in the Sydney region. The main water supply for Sydney, Warragamba Dam, covers 9050 km 2 and is surrounded by national forest inhabited by diverse and abundant marsupials. At the gp60 locus, the subtype family IV has been identified with 6 subtypes (IVa ‾ IVf) ( Power et al., 2009). Subtyping of the human-derived isolate of C. fayeri identified IVaA9G4T1R1, which has also been identified in eastern grey kangaroos in Warragamba Dam, suggesting possible zoonotic transmission ( Power, 2010; Waldron et al., 2010).

In addition to C. fayeri and C. macropodum , there have been several other host-adapted genotypes identified in Australian marsupials. Possum genotype I has been described in brushtail possums, a host species found in a range of habitats throughout Australia ( Hill et al., 2008) and the novel kangaroo genotype I in western grey kangaroos (Yang et al., 2011). Possum genotype I and kangaroo genotype I have not been reported in humans or other animals and their zoonotic potential is unknown.

3.1.10. Cryptosporidium meleagridis

Although primarily a bird parasite (see section 3.2.1 and Table 3), C. meleagridis has been identified in deermice, mountain gorillas and marsupials ( Feng et al., 2007; Sak et al., 2014; Vermeulen et al., 2015). It is also the third most prevalent species infecting humans ( Morgan et al., 2000; Cama et al., 2003; Gatei et al., 2006; Muthusamy et al., 2006; Leoni et al., 2006; Berrilli et al., 2012; Elwin et al., 2012; Neira et al., 2012; Silverlås et al., 2012; Kurniawan et al., 2013; Sharma et al., 2013; Wang et al., 2014; Adamu et al., 2014; Ghaffari and Kalantari, 2014; Ryan and Xiao, 2014; Ghaffari and Kalantari, 2014; Rahmouni et al., 2014; Wang et al., 2014; Stensvold et al., 2014, 2015). In some studies, C. meleagridis prevalence is similar to that of C. parvum ( Gatei et al., 2002; Cama et al., 2007). The ability of C. meleagridis to infect humans and other mammals, and its close relationship to C. parvum and C. hominis at multiple loci, has led to the suggestion that mammals actually were the original hosts, and that the species has later adapted to birds ( Xiao et al., 2002a,). Subtyping at the gp60 locus has identified seven subtype families (IIIa to IIIg) ( Stensvold et al., 2015). More details on transmission dynamics will be discussed in section 3.2.1.

3.1.11. Other Cryptosporidium species and genotypes reported in wild mammals

A number of other Cryptosporidium species and genotypes have been identified in wildlife ( Table 2). Most are host-adapted genotypes that are not of public health significance, however several have been identified in humans ( Table 2). Of these, the chipmunk genotype I is considered an emerging human pathogen ( Jiang et al., 2005; Feltus et al., 2006; Feng et al., 2007; ANOFEL, 2010; Insulander et al., 2013; Lebbad et al., 2013; Guo et al., 2015). At the gp60 locus, 15 different subtypes have been identified but subtypes differ only in the number of tandem repeats (TCA/TCG/ TCT) and comprise a single subtype family (XIVa). Analysis indicates that subtypes from humans and wildlife are genetically similar and zoonotic transmission might play a potential role in human infections ( Guo et al., 2015). The skunk and mink genotypes have also been reported in a few human cases of cryptosporidiosis ( Robinson et al., 2008; Chalmers et al., 2009; Rengifo-Herrera et al., 2011; Elwin et al., 2012; Ng-Hublin et al., 2013; Ebner et al., 2015).