History

The history of Toxoplasma gondii discovery, identification of hosts, description of its life cycle, diagnostic strategies and treatment have been thoroughly reviewed with ongoing research efforts and clinical experience continuing to advance our understanding of this important pathogen (Dubey, Reference Dubey2008, Reference Dubey2010; Ferguson, Reference Ferguson2009). One investigative team in the early 20th century noted the presence of protozoal organisms in their research gundy (Ctendactylus gundi) in Tunisia and another in rabbits (Oryctolagus cuniculus) in Brazil; both initially attributed the observation to Leishmania sp., before the former group realized it was a previously undescribed species and named it T. gondii (Splendore, Reference Splendore1908; Nicolle and Manceaux, Reference Nicolle and Manceaux1909). One year later, a clinical case of toxoplasmosis was described in the domestic dog (Canis familiaris) in Italy, an early insight into the distribution of this parasite across divergent animal species and geographic areas (Mello, Reference Mello1910). Tissue cyst structures were first noted in a group of rabbits and mice and described cytologically about 20 years after the organism's initial identification (Levaditi et al., Reference Levaditi, Schoen and Sanchis Bayarri1928). Early investigators depended on microscopic work and staining techniques. The advent of serological tests and molecular diagnostic techniques led to a rapidly increased understanding of the parasite and its role in clinical disease.

Life cycle

The demonstration of successful T. gondii transmission by congenital, carnivorous and fecal-oral routes stimulated research interest into the parasite's life strategy that allows survival in widely diverse hosts (Weinman and Chandler, Reference Weinman and Chandler1954; Hutchison, Reference Hutchison1965; Frenkel et al., Reference Frenkel, Dubey and Miller1970; Dubey et al., Reference Dubey, Sundar, Hill, Velmurugan, Bandini, Kwok, Majumdar and Su2008a, Reference Dubey, Velmurugan, Ulrich, Gill, Carstensen, Sundar, Kwok, Thulliez, Majumdar and Sub). T. gondii has a complex life cycle that can only be completed in the feline definitive host (DH), as indicated in Fig. 1. The most prolific and destructive stage is the tachyzoite, formerly called the trophozoite or endozoite. This is the active stage of the parasite and the most likely stage to cause clinical disease. The second life stage is the bradyzoite, which is the slow-replicating chronic stage that is most commonly within a tissue cyst. The oocyst is the life stage that results from sexual reproduction within the intestine and is shed in the feces of the DH, the wild or domestic felid (Dabritz et al., Reference Dabritz, Gardner, Miller, Lappin, Atwill, Packham, Melli and Conrad2007a, Reference Dabritz, Miller, Atwill, Gardner, Leutenegger, Melli and Conradb). In intermediate hosts (IH), sporozoites released from sporocysts penetrate through intestinal epithelial cells and migrate to the lamina propria, where numerous different types of cells are invaded, and two tachyzoites are produced by approximately 12 h post infection. Following invasion of additional tissues with tachyzoites, tissue cysts have been noted to start forming by 6 days post infection and these likely persist for the life of the animal. The life cycle in felids can follow this same course, making felids uniquely both potential intermediate and definitive hosts. The latter, due to their hosting the sexual stages of T. gondii, in which schizogony is noted after intestinal invasion leading to the production of sexual stages (female macrogamonts and male microgamonts) by 2 days after infection.

Fig. 1.

The life cycle of Toxoplasma gondii.

Following the formation of zygotes and oocyst walls, the host cell ruptures, discharging oocysts into the intestine (Dubey, Reference Dubey2010). Both IH and DH life cycles require evasion of the host's immune system to proceed successfully, and the same white blood cells that respond to the presence of the parasite are utilized to spread infection throughout the host (Denkers et al., Reference Denkers, Schneider, Cohen and Butcher2012). Each cat has the potential to shed over 1 million T. gondii oocysts during the initial phase of the infection (Dabritz et al., Reference Dabritz, Gardner, Miller, Lappin, Atwill, Packham, Melli and Conrad2007a, Reference Dabritz, Miller, Atwill, Gardner, Leutenegger, Melli and Conradb). The oocysts become infective when they sporulate in as little as 24 h in the environment (http://www.cdc.gov/parasites/toxoplasmosis/biology). The prepatent period (PPP) between ingestion and shedding of oocysts in domestic cats varies depending on the stage of the parasite initiating infection as noted above (Frenkel et al. Reference Frenkel, Dubey and Miller1970). The PPP after tissue cyst ingestion is considerably shorter (3–10 days) than after sporulated oocyst ingestion (18 days) in cats (Weiss and Kim, Reference Weiss and Kim2007). Considering the abbreviated PPP following tissue cyst ingestion in the cat with the similarly brief time from ingestion to tissue cyst production in the IH, it is easy to surmise how quickly contamination can spread through a felid-intermediate host/prey ecosystem, e.g. felids and rodents.

Human toxoplasmosis

Toxoplasmosis was established as a human disease after observation of a case of congenital transmission in 1939 and experimental transmission to animals from this case was demonstrated (Wolf et al., Reference Wolf, Cohen and Paige1939a; Ferguson, Reference Ferguson2009). The recognition of the potential importance of toxoplasmosis in human health precipitated development of diagnostic tests to allow antemortem detection. Sabin and Feldman recognized that tachyzoites lose their affinity for methylene blue in the presence of anti-T. gondii antibody, leading to the description of the first widely used serological test, the Sabin–Feldman dye test (Sabin and Feldman, Reference Sabin and Feldman1948; Dubey, Reference Dubey2008). Though still considered a gold standard in serological testing, the requirement for live parasites limits the utility of the test both due to the difficulty of acquiring and storing the organism and potential risk to laboratory personnel. Multiple tests have since been developed that utilize the agglutination of killed parasite or particles coated with antigen in the presence of antibody as the basis of serological diagnosis, including the direct agglutination test, the modified agglutination test, indirect hemagglutination and latex agglutination (Desmonts and Remington, Reference Desmonts and Remington1980; Dubey and Desmonts, Reference Dubey and Desmonts1987). These have the advantage of not requiring live organisms, but still being able to test human and animal sera equally well, due to lack of reliance on a species-specific protein conjugate for detection. A disadvantage is their reliance on subjective interpretation by the user as to a positive or negative reaction. The reliance on serological interpretation in the diagnosis of toxoplasmosis leads to difficulties for medical and veterinary clinicians in judging the significance of a positive test, i.e. is the presence of antibody indicative of a clinically significant infection or just a historical artifact of latent infection? The advent of techniques to establish the presence of the parasite in different tissues holds promise in identifying clinically significant infection, with a polymerase chain reaction (PCR) technique targeting T. gondii first being described over two decades ago (Burg et al., Reference Burg, Grover, Pouletty and Boothroyd1989).

Public health

Though felids act as the only definitive host, birds and mammals, including human beings, are known to serve as intermediate hosts (Assadi-Rad et al. Reference Assadi-Rad, New and Patton1995). Transmission to an IH occurs via ingestion of sporulated oocysts from feline fecal contamination, ingestion of tissue cysts from raw or undercooked infected meat, or congenital transmission (transplacentally). In a European study, contact with domestic cats was not identified as a risk factor for seropositivity in people. Rather, eating undercooked meat, contact with soil and travel outside of Europe, the USA and Canada were significant. In all, 30–63% of infections were attributed to consumption of undercooked or cured meat and 6–17% to soil contact (Cook et al., Reference Cook, Gilbert, Buffolano, Zufferey, Petersen, Jenum, Foulon, Semprini and Dunn2000). In the USA, parasite transmission to human adults typically occurs by ingestion of infected meats, ingestion of contaminated vegetables, or aspiration of aerosolized sporulated oocysts from cat litter boxes. In addition, there are an estimated 400–4000 new cases of congenital toxoplasmosis annually in the USA (http://www.cdc.gov/parasites/toxoplasmosis; Lopez et al., Reference Lopez, Dietz, Wilson, Navin and Jones2000). The prevalence of infection in people varies widely, with a seroprevalence of 38% previously noted in the USA and a range of 4–92% noted in other countries. The fecundity of oocyst production (an average of 10,000,000 oocysts shed per infected cat has been estimated) and their environmental durability, with survival times of greater than one year noted in both natural and experimental setting, contribute to the ubiquity of the infection (Dubey, Reference Dubey2010). However, an estimate from the latest National Health and Nutrition Examination Survey (2009–2010) continues to recognize a trend towards decreasing rates of infection in the overall US human population with a seroprevalence of 22.5% noted (Jones et al., Reference Jones, Parise and Fiore2014).

The pathogenicity of toxoplasmosis varies, depending on the age and species of host infected, the mode of infection, and the T. gondii genotype present (Assadi-Rad et al., Reference Assadi-Rad, New and Patton1995). Mortality due to human toxoplasmosis occurs in very few cases, most notably due to reactive encephalitis in immunocompromised patients and abortion in naïve mothers on first exposure (Dabritz and Conrad, Reference Dabritz and Conrad2010). However, morbidity due to toxoplasmosis can take many forms, related to the diverse tissue types that can be invaded in intermediate hosts. Clinical signs and pathology in human adults range from no clinical signs to ocular disease, lymphadenopathy, tissue-specific lesions and localized edema, and death (Weiss and Kim, Reference Weiss and Kim2007). It was linked to potentially fatal central nervous system disease in AIDS patients in the late 1980s (Luft et al., Reference Luft, Conley, Remington, Laverdiere, Wagner, Levine, Craven, Strandberg, File, Rice and Meunier-Carpentier1983; Tuazon, Reference Tuazon1989; Harden and Hair, Reference Harden and Hair1994). Congenital toxoplasmosis occurs when a mother is infected during pregnancy and commonly causes fetal death, though severity of disease and probability of congenital transfer depends on stage of pregnancy at time of maternal infection (Weiss and Kim, Reference Weiss and Kim2007). Though incidence of congenital transmission increases as pregnancy progresses, the pathogenicity of the congenital disease tends to decrease in severity (Remington and Desmonts, Reference Remington, Desmonts, Remington and Klein1990; Lynfield and Guerina, Reference Lynfield and Guerina1997; Freeman et al., Reference Freeman, Oakley, Pollak, Buffolano, Petersen, Semprini, Salt and Gilbert2005; Rorman et al., Reference Rorman, Zamir, Rilkis and Ben-David2006). The most commonly observed clinical and pathological signs in infants range from no clinical disease to chorioretinitis, blindness, severe hydrocephaly, meningoencephalitis, encephalomyelitis, intracerebral calcification, mental retardation, abortion, and postpartum death (Janků, Reference Janků1923; Torres, Reference Torres1927; Wolf et al., Reference Wolf, Cohen and Paige1939a, Reference Wolf, Cowen and Paigeb; Sabin, Reference Sabin1942; Frenkel and Friedlander, Reference Frenkel and Friedlander1951; Hogan, Reference Hogan1951; Remington et al., Reference Remington, McLeod, Thulliez, Remington and Klein2001; Weiss and Kim, Reference Weiss and Kim2007). Methods to reduce human exposure include: proper cooking of meat and vegetables, hand hygiene after handling meat or soil contact, preventing predation by cats, providing litter boxes and controlling feral cat populations (Dabritz and Conrad, Reference Dabritz and Conrad2010). However, the recognition of T. gondii genotypes in wildlife capable of infecting people implies a sylvatic life cycle that can independently perpetuate the parasite's population (Dubey et al., Reference Dubey, Sundar, Hill, Velmurugan, Bandini, Kwok, Majumdar and Su2008a, Reference Dubey, Velmurugan, Ulrich, Gill, Carstensen, Sundar, Kwok, Thulliez, Majumdar and Sub; Wendte et al., Reference Wendte, Gibson and Grigg2011).

Feline toxoplasmosis

Feline toxoplasmosis was first diagnosed in a domestic cat (Felis catus) from Middletown, NY, in 1942 (Olafson and Monlux, Reference Olafson and Monlux1942; Dubey, Reference Dubey2008). The protracted time between initial description (1908) and this first described case in what was later determined to be a definitive host of the organism is a tribute to the covert nature of T. gondii transmission and symptomatology in this host. It was another two decades before an investigator noted that oocysts can be isolated from feline feces (Hutchison, Reference Hutchison1965). He further noted that they can remain infectious in the environment for at least 12 months as demonstrated by experimental transmission to mice. Feline intestinal stages, both sexual and asexual, were described in detail by Frenkel et al. (Reference Frenkel, Dubey and Miller1970). They demonstrated that domestic cats shed oocysts following ingestion of infected mice, and further noted a discrepancy in the PPP for T. gondii in the feline host based on the stage of the parasite that produced the infection. Tissue cyst ingestion, which is thought to be the most common method of feline infection due to predation of intermediate hosts, revealed shedding of oocysts in 3–5 days. In contrast, ingestion of tachyzoites produced oocyst shedding in 8–10 days and ingestion of oocysts produced shedding in 21–24 days. Though oocyst shedding is intensive, with one study noting up to 13 million oocysts per gram of feces, it is temporally abbreviated, at approximately 1 week (Schares et al., Reference Schares, Vrhovec, Pantchev, Herrmann and Conraths2008; Dubey, Reference Dubey2010). This unfortunately makes direct fecal examination a diagnostic test of very low sensitivity in the definitive host, and questionable specificity due to the similar appearance of other protozoal oocysts like those of Hammondia and Besnoitia spp.

T. gondii in intermediate hosts

The search for intermediate hosts of T. gondii has been accomplished by numerous investigators with notable success (Dubey, Reference Dubey2010). Current indication that any warm-blooded animal can serve as an intermediate host for the organism and the identification of marine invertebrates as potential transport hosts significantly broadens the number of animals of potential interest (Lindsay et al., Reference Lindsay, Phelps, Smith, Flick, Sumner and Dubey2001). The earliest animals identified as intermediate hosts of T. gondii were the gundy, rabbit, domestic dog and mouse (Splendore, Reference Splendore1908; Nicolle and Manceaux, Reference Nicolle and Manceaux1909; Mello, Reference Mello1910; Levaditi et al., Reference Levaditi, Schoen and Sanchis Bayarri1928). The first isolation of viable T. gondii was from laboratory mice and led to its characterization as an obligate intracellular parasite (Sabin and Olitsky, Reference Sabin and Olitsky1937). The lack of a requirement for a definitive host in the parasite's propagation was noted by oral infection of pigs with mouse tissue, and of mice with pork, also establishing the potential risk to human beings from meat consumption (Weinman and Chandler, Reference Weinman and Chandler1954). Similarly, minced fetal cotyledons from suspected ovine abortion cases were demonstrated to be infectious to other pregnant ewes and mice by intraperitoneal injection (Hartley and Marshall, Reference Hartley and Marshall1957). More recently, the surprising discovery of toxoplasmosis being a cause of significant mortality in southern sea otters (Enhydra lutris nereis) demonstrated how pervasive environmental contamination with oocysts is in both terrestrial and marine environments (Cole et al., Reference Cole, Lindsay, Howe, Roderick, Dubey, Thomas and Baeten2000). The ability of T. gondii to infect a wide host range and to have a broad geographic distribution (described from all continents) makes it one of the most common protozoal parasites in the world.

Feline health

Toxoplasmosis in domestic cats is a topic of interest primarily due to the threat their infection poses to people and other animals, as feline hosts are typically subclinical. An extensive survey of cats that were clinically ill from any cause in the USA found a seroprevalence of 31.6% overall, with the southwestern USA having the lowest prevalence (16.1%) and the northeastern USA having the highest prevalence (43.5%). Significant factors other than geographic location that predisposed the cats to being positive included: older age, being male and being a domestic short-haired breed (Vollaire et al., Reference Vollaire, Radecki and Lappin2005). A study of Ohio cats brought into a mobile spay and neuter clinic revealed a 48% overall seropositivity, with a 62% prevalence noted in owner-described outdoor cats (Dubey et al., Reference Dubey, Gamble, Hill, Sreekumar, Romand and Thulliez2002a, Reference Dubey, Saville, Stanek and Reedb). A survey of Pennsylvania cats from a peri-urban environment yielded a significantly lower seropositivity of 19.5% (Dubey et al., Reference Dubey, Bhatia, Lappin, Ferreira, Thorn and Kwok2009a, Reference Dubey, Jenkins, Kwok, Zink, Michalski, Ulrich, Gill, Cartensen and Thulliezb). In the USA, the range of seropositivity survey results for cats was a low of 12.1% from Florida and a high of 80% from Iowa (Hill et al., Reference Hill, Zimmerman, Wills, Patton and Clark1998; Luria et al., Reference Luria, Levy, Lappin, Breitschwerdt, Legendre, Hernandez, Gorman and Lee2004). Surveys from other countries produced a wider range of seroprevalence, from a low of 2.1% in a Chinese study to a high of 89.3% in a Colombian study (Shen et al., Reference Shen, Zhichung, Biaucheng and Huayuan1990; Dubey, Reference Dubey2010). A survey of non-domestic felids in the USA found a prevalence of 43.6%, looking at both captive and free-ranging animals (Spencer et al., Reference Spencer, Higginbotham and Blagburn2003). The degree of exposure to T. gondii for both domestic and wild felids as demonstrated by these studies can be significant from diverse geographic and demographic areas. The role of feline predation in the T. gondii life cycle and the lack of an effective vaccine make the reduction of roaming cat populations a critical tool in controlling environmental contamination with oocysts. This remains a challenge in most settings, with the questionable efficacy of many sterilization and release programs potentially complicating population control efforts (Andersen et al., Reference Andersen, Martine and Roemer2004; Jessup, Reference Jessup2004; Nogales et al., Reference Nogales, Martin, Tershy, Donlan, Veitch, Puerta, Wood and Alonso2004; Longcore et al., Reference Longcore, Rich and Sullivan2009).

Canine health

The description of toxoplasmosis in the domestic dog preceded that in the domestic cat by more than three decades (Mello, Reference Mello1910; Olafson and Monlux, Reference Olafson and Monlux1942). While the evaluation of the significance of toxoplasmosis in the dog is complicated by potential infection with Neospora caninum, it is clear that both parasites can cause significant diseases in dogs. An examination of stray dogs and cats in Sao Paulo, Brazil, revealed a T. gondii seroprevalence of 40% in cats and 50.5% in dogs, which the authors attributed to the more selective palate of cats (Meireles et al., Reference Meireles, Galisteo, Pompeu and Andrade2004). A demonstration of the broad appetite of dogs was the identification of viable oocysts in canine feces in a German study, which could have wrongly indicted dogs as an alternative definitive host if the investigators were not familiar with their coprophagic habits (Schares et al., Reference Schares, Pantchev, Barutzki, Heydorn, Bauer and Conraths2005). A study of stray animals in Bangkok, Thailand, revealed significantly lower levels of seroconversion than many studies, 11% in cats and 9.4% in dogs, the authors noted that this population of animals were often fed cooked fish and rice by local monks, foods expected to carry a low risk of serving as routes of infection (Jittapalapong et al., Reference Jittapalapong, Nimsupan, Pinyopanuwat, Chimnoi, Kabeya and Maruyama2007). A study in the USA revealed a seroprevalence of 25% in Kansas dogs, with a range of 5.2–88.5% noted in other areas of the world, the lowest value noted from urban, owned dogs in Brazil and the highest value from dogs in Mato Grosso, the Brazilian state with the lowest human population density (Lindsay et al., Reference Lindsay, Dubey, Upton and Ridley1990; deSouza et al., Reference deSouza, Gennari, Yai, D'Auria, Cardoso, Junior and Dubey2003; Santos et al., Reference Santos, Costa, Toniollo, Luvizotto, Benetti, Santos, Matta, Lopes, Oliveira and Oliveira2009). Domestic dogs, especially free roaming dogs that are more likely to participate in predation, can serve as sentinels for local T. gondii occurrence, as they are exposed through both oocyst and tissue cyst ingestion (Meireles et al., Reference Meireles, Galisteo, Pompeu and Andrade2004).

Large animal health

Toxoplasmosis in food animals is a topic of significant interest, both due to potential public health risk from meat and milk consumption and economic losses due to disease. One of the earliest demonstrations of the importance of toxoplasmosis in sheep production was an investigation of an abortion storm on a farm in New Zealand, in which minced cotyledons from infected ewes transmitted T. gondii to additional ewes and mice (Hartley and Marshall, Reference Hartley and Marshall1957). A similar outbreak in New York revealed that 73.8% of sheep on the farm were seropositive (Dubey and Welcome, Reference Dubey and Welcome1988). The range of seroprevalence in sheep studies in the USA is 27.1–73.8%, with a range of 3–95.7% noted from sheep studies worldwide (Zaki, Reference Zaki1995; Dubey et al., Reference Dubey, Sundar, Hill, Velmurugan, Bandini, Kwok, Majumdar and Su2008a, Reference Dubey, Velmurugan, Ulrich, Gill, Carstensen, Sundar, Kwok, Thulliez, Majumdar and Sub; Dubey, Reference Dubey2010). Similar clinical effects have been noted in caprine toxoplasmosis, with US surveys noting a seroprevalence of 22.1–51.9% in goat herds, and a worldwide seroprevalence of 3.2–90.9% (McSporran et al., Reference McSporran, McCaughan, Currall and Demsteegt1985; Dubey and Adams, Reference Dubey and Adams1990; Patton et al., Reference Patton, Johnson and Puckett1990; Garcia-Vazques et al., Reference Garcia-Vazques, Rosario-Cruz, Diaz-Garcia and Hernandez-Baumgarten1993; Dubey, Reference Dubey2010). Consumption of poorly cooked lamb and goat, raw milk and unpasteurized cheese should be considered potential routes of transmission (Dubey et al., Reference Dubey, Dennis, Verma, Choudhary, Ferreira, Oliveira, Kwok, Butler, Carstensen and Su2014a, Reference Dubey, Verma, Ferreira, Oliveira, Cassinelli, Ying, Kwok, Tuo, Chiesa and Jonesb). Pork may also be a source of T. gondii, as reported in a Massachusetts study in which 51 of 55 pigs destined for human consumption were found to have viable tissue cysts in their heart and tongue tissue (Dubey et al., Reference Dubey, Gamble, Hill, Sreekumar, Romand and Thulliez2002a, Reference Dubey, Saville, Stanek and Reedb).

The role of toxoplasmosis as a significant pathogen of cattle and horses is not well recognized. Both species can experience infection as indicated by seroprevalence and experimental infection studies, but reports of associated pathology are rare (Tassi, Reference Tassi2007; Dubey, Reference Dubey2010). These infections tend to be insignificant in both intensity and duration with a risk of misidentification due to the morphological similarity of Neospora spp. and Sarcocystis spp. The detection of viable tissue cysts and DNA in tissue samples from infected cattle and horses indicates a potential threat to human and animal health from the consumption of undercooked beef or horse meat (Dubey and Desmonts, Reference Dubey and Desmonts1987; Dubey, Reference Dubey2010; Papini et al., Reference Papini, Buzzone, Nardoni, Rocchigiani and Mancianti2015). Dubey et al. (Reference Dubey, Hill, Jones, Hightower, Kirkland, Roberts, Marcet, Lehmann, Vianna, Miska, Sreekumar, Kwok, Shen and Gamble2005) described a large survey of beef, chicken and pork purchased from retail food stores in the USA, utilizing bioassay, enzyme-linked immunosorbent assays (ELISA) and PCR analysis, to find viable T. gondii only in pooled pork samples. The role of unpasteurized cow's milk in transmission of T. gondii has been thought to be minimal, with examination of the subject continuing (Tenter et al., Reference Tenter, Heckeroth and Weiss2000; Dehkordi et al., Reference Dehkordi, Borujeni, Rahimi and Abdizadeh2013; Dubey and Jones, Reference Dubey and Jones2014).

Wildlife health

Many wildlife species are susceptible to T. gondii infection and to clinical toxoplasmosis. One species of interest as a sentinel host for domestic herbivores is the white-tailed deer (Odocoileus virginianus), both due to its tendency to graze alongside livestock and its wide dispersal. A study in Ohio noted a T. gondii seroprevalence of 44% among white-tailed deer, 6% in deer harvested in urban parks (Columbus) and 55% in more rural Hocking County (Crist et al., Reference Crist, Stewart, Rinehart and Needham1999). A more recent study examined seroprevalence in deer in Northeastern Ohio and found evidence of greater exposure among urban rather than suburban populations (Ballash et al., Reference Ballash, Dubey, Kwok, Shoben, Robison, Kraft and Dennis2015). This implies that human (and by extension domestic cat) population density may not be a reliable predictor of environmental contamination by T. gondii. Surveys in other states ranged from 14.8 to 75.7% seroprevalence (Humphreys et al., Reference Humphreys, Stewart and Dubey1995; Vanek et al., Reference Vanek, Dubey, Thulliez, Riggs and Stromberg1996; Dubey et al., Reference Dubey, Bhatia, Lappin, Ferreira, Thorn and Kwok2009a, Reference Dubey, Jenkins, Kwok, Zink, Michalski, Ulrich, Gill, Cartensen and Thulliezb, Reference Dubey, Dennis, Verma, Choudhary, Ferreira, Oliveira, Kwok, Butler, Carstensen and Su2014a, Reference Dubey, Verma, Ferreira, Oliveira, Cassinelli, Ying, Kwok, Tuo, Chiesa and Jonesb; Dubey, Reference Dubey2010; Schaefer et al., Reference Schaefer, Kirchgessner, Whipps, Mohammed, Bunting and Wade2013). The clinical effect of toxoplasmosis on deer is poorly understood, as is the question of bias arising from an interaction between infection and the tendency to be killed by hunters, a requirement for inclusion in most of these studies.

Despite the strictly terrestrial nature of the definitive host, studies over the past 11 years have found T. gondii in marine settings and have attributed marine mammal morbidity and mortality to toxoplasmosis. The first major case in marine mammals occurred in sea otters off the coast of California (Miller et al., Reference Miller, Gardner, Kreuder, Paradies, Worcester, Jessup, Dodd, Harris, Ames, Packham and Conrad2002). These infections were linked to coastal freshwater runoff of feces from feral cat populations (Cole et al., Reference Cole, Lindsay, Howe, Roderick, Dubey, Thomas and Baeten2000; Miller et al., Reference Miller, Gardner, Kreuder, Paradies, Worcester, Jessup, Dodd, Harris, Ames, Packham and Conrad2002), and Shapiro et al. (Reference Shapiro, Krusor, Mazzillo, Conrad, Largier, Mazet and Silver2014) determined that further dispersal of the oocysts into the marine environment was facilitated by aquatic polymers (biofilm) and invertebrates. The diversity of the marine mammal habitats and foraging habits make it challenging to determine how many are infected with T. gondii. Coastal runoff may explain infection in the benthoneritic but does not explain infection in more pelagic marine mammals. A few studies have suggested the dispersal of T. gondii oocysts is possible to more pelagic fish-eating marine mammals via transport host or mechanical vector. Eastern oysters can filter and maintain viable T. gondii oocysts from the seawater within their tissues for up to 85 days (Lindsay et al., Reference Lindsay, Phelps, Smith, Flick, Sumner and Dubey2001, Reference Lindsay, Collins, Jordan, Flick and Dubey2005). Migratory filter feeding fish, including sardines and anchovies, have been implicated in the dispersal of oocysts into the marine environment by serving as transport hosts or biotic vectors. In addition to the oocysts being maintained in the alimentary tract of the fish, 30% of fish retained oocysts infective to mice (Massie et al., Reference Massie, Ware, Villegas and Black2010).

Diagnostic test development

The discovery and subsequent investigation of T. gondii largely depended on bioassays and microscopic examination until the advent of the Sabin–Feldman dye test, which allowed highly accurate antemortem diagnosis of infection through the evaluation of serum samples (Sabin and Feldman, Reference Sabin and Feldman1948). The dye test proved highly sensitive and accurate, but had the distinct disadvantage of requiring live tachyzoites, a potential risk to laboratory staff conducting the test (Reiter-Owona et al., Reference Reiter-Owona, Petersen, Joynson, Aspock, Darde, Disko, Dreazen, Dumon, Grillo, Gross, Hayde, Holliman, Ho-Yen, Janitschke, Jenum, Naser, Olszewski, Thulliez and Seitz1999). In medical diagnostics, a number of immunofluorescence and ELISA tests were developed, but have limited utility in veterinary diagnostics due to their requirement for species-specific protein conjugate. The development of the direct agglutination and later the modified agglutination tests, techniques that utilized formalin-killed rather than live tachyzoites and did not have a conjugate requirement, was revolutionary in the diagnosis and investigation of toxoplasmosis in animals (Desmonts and Remington, Reference Desmonts and Remington1980; Dubey and Desmonts, Reference Dubey and Desmonts1987). These techniques have been further modified to produce additional agglutination tests that utilize the same strategy: the indirect hemagglutination test (IHA) using antigen coated sheep red blood cells and the latex agglutination test (LAT) using antigen coated latex beads. Indirect hemagglutination test results were noted to agree with Sabin–Feldman dye test results in human patients, and similarly remain positive for the life of the patient (Lunde, Reference Lunde1973). In a study of experimental infection using cats that were monitored for up to 950 days after inoculation, the Sabin–Feldman dye test, modified agglutination test (MAT), IHA and LAT were compared and the MAT was found to be the most sensitive of the agglutination methods (Dubey and Thulliez, Reference Dubey and Thulliez1989). The MAT and IHA were found to produce similar results in a study of dairy goats (Patton et al., Reference Patton, Johnson and Puckett1990). The MAT also had good correlation with an ELISA test developed specifically for porcine testing (Gamble et al., Reference Gamble, Dubey and Lambillotte2005). None of the agglutination tests reliably detect IgM on known IgM positive samples, and therefore become positive later in the course of infection than the dye test (Lappin and Powell, Reference Lappin and Powell1991). Agglutination testing remains the most commonly used modality for the diagnosis of toxoplasmosis in animals, despite the loss of availability of some commercial kits. These diagnostic methods and those discussed below are summarized in Table 1.

Table 1.

Summary of veterinary diagnostic test methods for toxoplasmosis

Antibody testing

Diagnosis of human toxoplasmosis is largely accomplished by serological investigation using ELISAs to detect T. gondii-specific IgG and IgM, as the requirement for species-specific conjugate is not as burdensome as in veterinary medicine. There are numerous commercial serological tests available for the detection of infection in human beings; one study compared three ELISAs, two IHAs, one IFA and one LAT to the Center for Disease Control's reference indirect immunofluorescence test and found 100% sensitivity across the tests, with a range of specificities from 77.8 to 100%, and perfect agreement being noted by one representative of each testing method (Wilson et al., Reference Wilson, Ware and Walls1987). The reliance on a single serum test (e.g. an IgM-specific ELISA) to determine the critical question of chronicity of infection, however, complicates the diagnostic picture, as positive results indicating acute infection can have dramatic consequences in the management of patients, especially pregnant women. Several studies have noted high sensitivities but variable specificities for commercial IgM-specific ELISAs, with a positive predictive value of only 43.3% noted in one (Wilson et al., Reference Wilson, Remington, Clavet, Varney, Press and Ware1997; Roberts et al., Reference Roberts, Hedman, Luyasu, Zufferey, Bessieres, Blatz, Candolfi, Decoster, Enders, Gross, Guy, Hayde, Ho-Yen, Johnson, Lecolier, Naessens, Pelloux, Thulliez and Petersen2001; Kodym et al., Reference Kodym, Machala, Rohacova, Sirocka and Maly2006). Confirmation testing is recommended in these cases, two options being paired IgG testing or IgG avidity testing (Liesenfeld et al., Reference Liesenfeld, Press, Montoya, Gill, Isaac-Renton, Hedman and Remington1997). While few IgM ELISAs are utilized for veterinary diagnostics, IgM testing can be accomplished by utilizing modified agglutination methods, but the same cautions must be noted when relying on a single serum sample for the determination of infection chronicity.

In veterinary investigation of T. gondii, the production of recombinant antigen and advance of molecular techniques has expanded diagnostic options, while more established agglutination techniques are becoming less commercially available. Some species-specific ELISAs have been developed and performed well when compared with reference serological techniques, notably for cats, sheep and pigs (Kimbita et al., Reference Kimbita, Xuan, Huang, Miyazawa, Fukumoto, Mishima, Suzuki, Sugimoto, Nagaswaw, Fujisaki, Suzuki, Mikami and Igarashi2001; Hill et al., Reference Hill, Chirukandoth, Dubey, Lunney and Gamble2006; Dabritz et al., Reference Dabritz, Gardner, Miller, Lappin, Atwill, Packham, Melli and Conrad2007a, Reference Dabritz, Miller, Atwill, Gardner, Leutenegger, Melli and Conradb; Klun et al., Reference Klun, Djurkovic-Djakovic and Thulliez2007; Mainar-Jaime and Barberan, Reference Mainar-Jaime and Barberan2007; Shaapan et al., Reference Shaapan, El-Nawawi and Tawfik2008). Some cautions that must be noted in interpretation of these results include the potential of cross-reactivity in the presence of infections with other coccidia and the difficulty of determining true sensitivities and specificities without a reliable gold standard for comparison (Silva et al., Reference Silva, Lourenco, Silva and Mineo2002; Gardner et al., Reference Gardner, Greiner and Dubey2009). Most ELISA procedures require species-specific conjugates, which translates to a large number of individual ELISAs being used to test the range of animal species submitted to veterinary diagnostic laboratories, making them a poor replacement for the established agglutination techniques. Some investigators have examined the use of non-species-specific protein conjugates in veterinary diagnostic testing, e.g. proteins A, G and A/G, with some success (Eliasson et al., Reference Eliasson, Andersson, Olsson, Wigzell and Uhlen1989; Werre et al., Reference Werre, Jacobson, Bowman, Dubey and Mohammed2002; Bhide et al., Reference Bhide, Curlik, Travnicek and Lazar2004; Zhang et al., Reference Zhang, Wang, Fang, Nie, Feng, Zhou and Zhao2010; Schaefer et al., Reference Schaefer, White, Schaaf, Mohammed and Wade2011; Schaefer et al., Reference Schaefer, White, Schaaf, Mohammed and Wade2012; Al-Adhami and Gajadhar, Reference Al-Adhami and Gajadhar2014).

Molecular testing

While serological detection is the primary mode of T. gondii diagnosis in people and other animals, molecular techniques have increasingly been used since the first identification of a gene target for PCR amplification (Burg et al., Reference Burg, Grover, Pouletty and Boothroyd1989). PCR is now commonly used to test amniotic fluid in pregnant women with suggestive serological results, and has been used for environmental detection of oocysts and characterization of T. gondii genotypes in vertebrate hosts (Schwab and McDevitt, Reference Schwab and McDevitt2003; Su et al., Reference Su, Zhang and Dubey2006). The difficulty of acquiring the appropriate tissue sample for antemortem diagnosis makes PCR a modality of questionable sensitivity but potentially excellent specificity when used in the diagnostic setting. Direct detection of the parasite is complicated in the definitive feline host by the transient appearance of oocysts in feces and in intermediate hosts by the difficulty in predicting the precise location of tissue cysts (Dubey et al. Reference Dubey, Christie and Pappas1977, Reference Dabritz, Gardner, Miller, Lappin, Atwill, Packham, Melli and Conrad2007a, Reference Dabritz, Miller, Atwill, Gardner, Leutenegger, Melli and Conradb; Dubey, Reference Dubey1995, Reference Dubey2001; Dabritz and Conrad, Reference Dabritz and Conrad2010). While serology provides evidence of infection, a single sample does not allow the characterization of an acute versus a chronic infection, or the severity of the infection. The direct detection of DNA by PCR is a promising method in that a positive result establishes the presence of the organism in the tested tissue. However, the nature of the organism's complex life cycle poses a challenge in translating a successful PCR assay into a useful clinical test.

PCR

There are two commonly used targets for the detection of genomic DNA by either traditional or real time PCR, the B1 gene and AF146527 (529-bp repeat element). The B1 gene was initially detected as a PCR target with 35 repeats in the T. gondii genome by Burg et al. (Reference Burg, Grover, Pouletty and Boothroyd1989), who determined a limit of detection of one organism in cell lysate and 10 organisms in a human leukocyte preparation by traditional PCR, with no evidence of detection of closely related pathogens. The B1 gene has become a standard target for PCR detection in clinical samples and, significantly for veterinary diagnostics, does not occur in Neospora caninum (Schatzberg et al., Reference Schatzberg, Haley, Barr, deLahunta, Olby, Munana and Sharp2003; Chabbert et al., Reference Chabbert, Lachaud, Crobu and Bastien2004). More recently, a 529-bp target (AF146527) that is repeated 200–300 fold in the T. gondii genome has been described and utilized as a target for PCR detection (Homan et al., Reference Homan, Vercammen, de Braekeleer and Verschueren2000). No statistically significant difference in sensitivity or specificity between the B1 gene and AF146527 targets were noted in some investigations (Filisetti et al., Reference Filisetti, Gorcii, Pernot-Marino, Villard and Candolfi2003; Remington et al., Reference Remington, Thulliez and Montoya2004), whereas other studies noted that assays for 529 were more sensitive for a variety of fluid and tissue samples (Homan et al., Reference Homan, Vercammen, de Braekeleer and Verschueren2000; Reischl et al., Reference Reischl, Bretagne, Kruger, Ernault and Costa2003; Calderaro et al., Reference Calderaro, Piccolo, Gorrini, Peruzzi, Zerbini, Bommezzadri, Dettori and Chezzi2006; Cassaing et al., Reference Cassaing, Bessieres, Berry, Berrebi, Fabre and Magnaval2006; Dubey, Reference Dubey2010; Belaz et al., Reference Belaz, Gangneux, Dupretz, Guiguen and Robert-Gangneux2015). Investigation of sequence variation across strains from human infections noted that the AF146527 element is absent in 4.8% of cases, which may lead to false negative determinations even if the organism is present in the tissue being evaluated (Wahab et al., Reference Wahab, Edvinsson, Palm and Lindh2010).

Both methods are used for detection of T. gondii in diverse tissues in human and veterinary diagnostic investigation. Samples commonly evaluated from suspected human cases include: amniotic fluid, placenta, brain, blood, cerebrospinal fluid, urine, vitreous, aqueous, bronchoalveolar lavage, pleural and peritoneal fluids (Remington et al., Reference Remington, Thulliez and Montoya2004). Veterinary diagnostic laboratories offer evaluation of blood, feces, CSF, brain and other tissues, with fecal examination used to distinguish T. gondii oocysts from morphologically similar Hammondia hammondi in feline feces (Schares et al., Reference Schares, Vrhovec, Pantchev, Herrmann and Conraths2008). Knowledge of the parasite's life cycle is critical to increase the utility of PCR testing, e.g. blood stages are transient and expected early in infection, and detection of the parasite in tissue relies on the presence of a tissue cyst in the aliquot being tested. Even with a good sampling strategy, molecular detection must be considered a method of unproven sensitivity but good specificity, where the predictive value of a positive result is much higher than that of a negative result (Remington et al., Reference Remington, Thulliez and Montoya2004). This makes molecular detection a reasonable adjunct to serological investigation, in which the opposite is true regarding positive and negative predictive values. The establishment of a reliable real-time PCR protocol for detection of T. gondii DNA will allow definitive diagnosis of the presence and stage of infection when serological testing produces equivocal results.