Importance of serological cross-reactivity among Toxoplasma gondii, Hammondia spp., Neospora spp., Sarcocystis spp. and Besnoitia besnoiti

SUMMARY Toxoplasma gondii, Neospora spp., Sarcocystis spp., Hammondia spp. and Besnoitia besnoiti are genetically related cyst-forming coccidia. Serology is frequently used for the identification of T. gondii, Neospora spp. and B. besnoiti-exposed individuals. Serologic cross-reactions occur in different tests among animals infected with T. gondii and H. hammondi, as well as among animals infected by T. gondii and N. caninum. Infections caused by N. caninum and N. hughesi are almost indistinguishable by serology. Neospora caninum, B. besnoiti and Sarcocystis spp. infections in cattle show some degree of serologic cross-reactivity. Antibody cross-reactivity between Neospora spp. and H. heydorni-infected animals is suspected, but not proven to occur. We review serologic cross-reactivity among animals and/or humans infected with T. gondii, Neospora spp., Sarcocystis spp., Hammondia spp. and B. besnoiti. Emphasis is laid upon antigens and serological methods for N. caninum diagnosis which were tested for cross-reactivity with related protozoa. Species-specific antigens, as well as stage-specific proteins have been identified in some of these parasites and have promising use for diagnosis and epidemiological surveys.

Toxoplasma gondii, Neospora spp., Sarcocystis spp., Hammondia spp. and Besnoitia besnoiti are closely related tissue cyst-forming parasites that belong to the family Sarcocystidae (Carreno et al. 1998;Mugridge et al. 2000). The complete life cycles of Sarcocystidae organisms are complex and involve several parasite stages in definitive and intermediate hosts. In species with known life cycles, carnivore or omnivore definitive hosts harbour sexual reproduction of the parasites in intestinal epithelium. They may shed large numbers of parasite oocysts or sporocysts in their feces. Intermediate hosts acquire infection upon ingestion of sporulated oocysts or sporocysts in food or water. Sporozoites invade intestinal epithelial cells of the intermediate hosts and spread to other tissues as tachyzoites or merozoites. These latter stages may encyst as slow multiplying forms, called bradyzoites. Ingestion of tissue cysts by carnivorism is the main route of infection for definitive hosts, which culminates with the formation of oocysts in their intestinal epithelium (Levine and Ivens, 1981).
Among the cyst-forming parasites, T. gondii is the most studied. It induces disease in a wide range of warm-blooded animals, including humans , and causes abortion in livestock, especially in sheep and goats (Buxton, 1998). Neospora caninum was originally described in dogs (Bjerkas et al. 1984;Dubey et al. 1988a), but it gradually gained more attention from the scientific community as a major cause of neonatal mortality and abortion in mainly cattle but also other ruminants (O'Toole and Jeffrey, 1987;Parish et al. 1987;Anderson et al. 1991), besides causing neuromuscular disease in dogs (Ruehlmann et al. 1995). Neospora hughesi was proposed as a new species in the genus Neospora and has been associated with myeloencephalitis in horses (Marsh et al. 1998). The genus Sarcocystis possesses more than 100 species, with cattle as intermediate hosts of at least three species (S. cruzi, S. hirsuta and S. hominis). Among these, only S. cruzi is mildly pathogenic for cattle and generally non-pathogenic for its definitive host (dog) (reviewed by Dubey and Lindsay, 2006). Additional Sarcocystis spp. have been observed in bovine tissues, but their nomenclatures are still in debate Gjerde, 2016). Three parasite species compose the genus Hammondia (H. hammondi, H. heydorni and H. triffittae), which have no known association with disease in humans or in naturally infected animals. However, Hammondia spp. are closely related to T. gondii and N. caninum (Mugridge et al. 2000), so diagnostic methods need to discriminate between infections caused by these parasites. Besnoitia besnoiti causes a debilitating disease mainly characterized by both a cutaneous and systemic manifestation (Alvarez-Garcia et al. 2013). Reproductive abnormalities in cattle may also occur, such as infertility in bulls and abortion when cows are infected during pregnancy (Cortes et al. 2014). Bovine besnoitiosis was reported first more than a century ago in Southern France and Portugal, but the parasite has spread to several European countries during the last 10 years and besnoitiosis is now considered as a re-emerging disease in cattle, at least in Europe (Alvarez-Garcia et al. 2013).
Toxoplasma gondii and H. hammondi have cats as definitive hosts, which shed morphologically indistinguishable oocysts in their feces. Dogs and certain canid species serve as definitive hosts for N. caninum (McAllister et al. 1998;Gondim et al. 2004;King et al. 2010;Dubey et al. 2011) and H. heydorni (Blagburn et al. 1988;Slapeta et al. 2002;Soares et al. 2009). Hammondia triffittae has two species of wild canids (red fox and arctic fox) as definitive hosts (Gjerde and Dahlgren, 2011). Three Sarcocystis spp. from cattle, S. cruzi, S. hirsuta and S. hominis, have dogs, cats, and primates as definitive hosts, respectively (reviewed by Gjerde, 2016). Besnoitia besnoiti is suspected to have a carnivore as definitive host, but so far no animal has been identified shedding oocysts of the parasite by natural or experimental infections .
Infections caused by T. gondii, Hammondia spp., Neospora spp., Sarcocystis spp. and B. besnoiti are assessed by a great variety of diagnostic tools, depending on the purpose of the analysis and available biological sample. In clinically affected individuals, detection of parasite-specific antibodies in serum or other body fluids is the most commonly employed diagnostic approach, except for Hammondia spp., and Sarcocystis spp. from cattle. Continuous cultivation of bovine Sarcocystis spp., like S. cruzi, is difficult (Andrews et al. 1990). To date, Hammondia spp. cannot be continuously grown in cell culture, which impedes production of parasite antigens needed to produce serologic tests for these parasites (Riahi et al. 1995;Schares et al. 2003;Gondim et al. 2015). In this review, serologic cross-reactivity is reviewed in detail among infections caused by T. gondii, Hammondia spp., Neospora spp., Sarcocystis spp. and B. besnoiti. Special emphasis is put on serologic cross-reactivity among animals infected with N. caninum and related pathogens. Further consideration is given to the discovery and production of species-specific and stage-specific antigens, which promise to improve diagnostic specificity and may enable discrimination between different modes of parasite acquisition.
During the first decades after T. gondii was discovered, several scientists have attempted to develop serologic tests with high sensitivity and specificity to diagnose T. gondii infection, as well as to understand the antigenic composition of the parasite. The development and improvement of serological tests, such as the Sabin-Feldman dye test (DT) (Sabin and Feldman, 1948;Beverley and Beattie, 1952), direct agglutination test (Fulton and Turk, 1959;Desmonts and Remington, 1980), complement fixation test (CFT) (Sabin, 1949), enzymelinked immunosorbent assay (ELISA) (Walls et al. 1977), immunofluorescence antibody test (IFAT) (Kelen et al. 1962), indirect haemagglutination test (IHA) (Lunde and Jacobs, 1958) and Western blot (WB) (Araujo et al. 1984) favoured a great advance in the study of toxoplasmosis.
Cross-immunity studies between T. gondii and H. hammondi. Mice and hamsters that were experimentally infected with H. hammondi (CR-4 strain) oocysts developed immunity and did not die after challenged with lethal doses of oocysts from a mouse-virulent T. gondii strain (M-7741 strain) (Frenkel and Dubey, 1975). In contrast, cats that were experimentally infected with H. hammondi and shed oocysts were not immunized against excretion of T. gondii oocysts (Frenkel and Dubey, 1975). An additional study approached cross-immunity between T. gondii and H. hammondi, by using six H. hammondi strains in an infection model of mice and hamsters (Christie and Dubey, 1977). The authors observed that 103 of 108 mice that were orally inoculated with H. hammondi oocysts survived a lethal challenge dose (10 5 oocysts) of T. gondii (strain M-7741). The H. hammondi-inoculated hamsters also developed immunity against a lethal dose of T. gondii oocysts, but this immunity was variable depending on the H. hammondi strain. The two most immunogenic H. hammondi strains conferred protection to fatal toxoplasmosis in 100 and 83% of the hamsters, respectively, after challenging with T. gondii oocysts (Christie and Dubey, 1977). Immunization of goats with H. hammondi had a protective effect against abortion induced by T. gondii; however, the immunization did not prevent transplacental transmission of T. gondii in pregnant does (Munday and Dubey, 1988). Partial immunity against toxoplasmosis was also obtained in Tamar wallabies (Macropus eugenii), that were orally infected with 1 × 10 5 oocysts of H. hammondi (Reddacliff et al. 1993).
Serologic cross-reactivity between T. gondii and H. hammondi. When H. hammondi was first described, some rodent species experimentally infected with this parasite developed cross-reacting antibodies against T. gondii antigens in the DT (Frenkel and Dubey, 1975). Further studies were carried out and confirmed that serologic crossreactivity between T. gondii and H. hammondi occurred with sera from other animals, besides rodents. Weiland et al. (1979) investigated crossreactivity between T. gondii and H. hammondi in four animal species (120 mice, six dogs, six rabbits and six pigs) by using five serological tests (DT, CFT, ELISA, IFAT and IHA). Half of the animals were orally inoculated with T. gondii oocysts and the other half received H. hammondi oocysts by the same route. Sera from H. hammondi-infected mice reacted with T. gondii antigens in three tests (DT, ELISA and CFT). Sera from dogs infected with H. hammondi recognized T. gondii antigens by DT and ELISA. Sera from rabbits exhibited cross-reaction between the two parasites by ELISA. The sera of pigs infected with H. hammondi did not cross-react with T. gondii in any of the five serological tests. In this study, the IFAT was considered the most Toxoplasmaspecific method. During the course of infection, the animals infected with T. gondii presented higher titres in the tests when compared with those infected with H. hammondi (Weiland et al. 1979). Munday and Dubey (1986) observed that sheep that were inoculated with H. hammondi oocysts presented cross-reactivity with T. gondii antigen by IFAT. Before infection with H. hammondi, the sheep had no detectable antibodies to T. gondii. After oral inoculation with H. hammondi oocysts, the animals presented antibody titres of 1:16 to T. gondii by IFAT. Goats infected with H. hammondi were shown to produce antibodies against T. gondii tested by DT, with titres up to 1:64 (Dubey, 1981).
The antigenic similarity between T. gondii and H. hammondi was investigated using sera from experimentally infected mice (Araujo et al. 1984). The authors employed T. gondii tachyzoites (RH strain) for two antigen detection procedures: (1) the antigen was labelled with 125 I, immune-precipitated with sera from T. gondii or H. hammondi-infected mice, run by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and autoradiographed; (2) the antigen was lysed, separated by SDS-PAGE and tested by immunoblot. The sera from mice that were orally inoculated with H. hammondi oocysts recognized T. gondii antigens of MW 92·5 kDa (two antigens), 66·2 kDa, between 66·2 and 45 kDa, and between 31 and 21·5 kDa; compared with the T. gondii-positive reaction only a T. gondii antigen of 21·5 kDa was not recognized by the H. hammondi-positive mouse serum (Araujo et al. 1984). This study confirmed that T. gondii and H. hammondi have similar antigenic components in their tachyzoites.
To facilitate scientific communication among research groups and laboratories, Sibley et al. (1991) proposed a nomenclature system for naming mutants, genes and gene products of T. gondii, which was based on the system used for the yeast Saccharomyces cerevisiae. As an example, the surface protein (P30) was designated SAG1, which is the product of the SAG1 gene. However, as it was observed more recently that SAG1 genes belong to a superfamily of related genes, named SRS (SAG1-related sequences), which encode a superfamily of structurally related surface proteins from T. gondii, the name of SAG1 (P30) has been changed to SRS29B . Riahi et al. (1998) studied the antigenic similarity of T. gondii and H. hammondi employing five monoclonal antibodies (MAbs) against T. gondii surface antigens and a polyclonal mouse serum to H. hammondi. In order to produce enough H. hammondi antigen, the authors used an in vitro model that allows the production of H. hammondi cysts up to 3 months in cell culture (Riahi et al. 1995). The cyst formation, confirmed by ultrastructural characteristics of the organism, started from 6 days after sporozoites were inoculated into feline kidney cells (CRFK). At 4 days of infection, the authors assumed that the multiplying H. hammondi zoites were tachyzoites (Riahi et al. 1998).
Tachyzoite antigens from T. gondii and H. hammondi were tested by IFAT and WB. By combining the two serologic techniques (IFAT and WB), five T. gondii antigens (MW of 30, 32, 35, 66 and 90 kDa) were recognized using polyclonal anti-H. hammondi serum. An interesting finding obtained by Riahi et al. (1998) was the recognition of the SAG1 (SRS29B) antigen by the anti-H. hammondi serum, as this protein is a major antigen of T. gondii (Burg et al. 1988) and had been considered to be a specific marker for the parasite (Mineo et al. 1993). Despite the antigenic similarity between T. gondii and H. hammondi, MAbs targeted against H. hammondi antigens were produced, and five of them did not cross-react with T. gondii antigens (Riahi et al. 2000). These findings are promising for the characterization of H. hammondi-specific antigens or epitopes, which could enable development of serologic tests for this parasite that would not cross-react with T. gondii.
The genome of a German strain of H. hammondi was sequenced and the genomic synteny between this parasite and T. gondii was higher than 95% (Walzer et al. 2013). It was found that orthologues of key T. gondii mouse virulence genes are functionally conserved in H. hammondi, but these data were not enough to explain the phenotypic differences observed between both parasites (Walzer et al. 2013). In a recent work, the genomes of 62 strains of T. gondii were compared with those from H. hammondi, N. caninum and Sarcocystis neurona (Lorenzi et al. 2016); these authors demonstrated that T. gondii possesses an expansion of parasite-specific secretory pathogenesis determinants (SPDs) when compared with the three latter parasites. The SPDs encompass genes encoding secretory proteins from micronemes, dense granules, rhoptries and surface antigens, whose expansion and diversity are associated with the patterns of transmission, host range and pathogenicity of T. gondii (Lorenzi et al. 2016). It was reported that T. gondii shares a high number of orthologues with H. hammondi (7095) and N. caninum (6308) (Lorenzi et al. 2016).
The search for more practical and efficient methods to detect specific antibodies against T. gondii pushed the establishment of tests based on antigenic fractions, which are gradually replacing traditional tests based on whole organisms (agglutination, DT and IFAT) or total extracts of parasite antigens. Recognized immunodominant antigens of T. gondii, including the tachyzoite surface antigens SAG1 (p30) (Kasper, 1987;Mineo et al. 1993 -Ferragut et al. 1996) and several other proteins from dense granules, rhoptries and micronemes are being produced as recombinant proteins (reviewed by Holec-Gasior, 2013).
So far, it is not known whether humans may be infected with H. hammondi, and in case it happens, the possibility of serologic cross-reactivity with T. gondii antigens cannot be ruled out. Cats seem to shed T. gondii and H. hammondi in similar proportions as reported in a recent study from Germany , and humans and animals of many species are potentially exposed to H. hammondi. The establishment of H. hammondi infections in most animal species has not been rigorously investigated; however, the life cycle of H. hammondi seems to lack avian hosts (Dubey and Sreekumar, 2003).

Toxoplasma gondii vs N. caninum
The serologic differentiation between T. gondii and H. hammondi infections has clinical and epidemiological relevance. Since H. hammondi is not known to induce disease in animals or humans, the major concern about the serologic differentiation of these parasites seems to be to avoid T. gondii false-positive results in individuals potentially infected with H. hammondi. The identification of N. caninum (Bjerkas et al. 1984;Dubey et al. 1988a) imposed a new challenge for the scientific community, because this non-zoonotic Toxoplasmatinae parasite is able to infect and cause disease in some mammalian animals that are also susceptible to T. gondii infection (Buxton, 1998). The comparison of the wholegenome sequences from T. gondii, H. hammondi and N. caninum showed that these three related tissuecyst-forming coccidian parasites have a similar total genome size of 62-65 Mb, and many similar orthologous groups of proteins involved in key biological functions (Lorenzi et al. 2016).
Neospora caninum can be maintained as tachyzoites in cell culture (Dubey et al. 1988b), which has enabled the development of a range of serological tests and antibodies against the parasite (Björkman and Uggla, 1999). Neosporosis research was accelerated by the prior accumulation of knowledge about and procedures for T. gondii.
The nomenclature system adopted for mutants, genes and gene products of T. gondii (Sibley et al. 1991) was proposed for N. caninum (Howe and Sibley, 1999). In the comparison of T. gondii and N. caninum homologous antigens (e.g. SAG1), the use of a Tg or Nc prefix was recommended to distinguish these gene products (e.g. TgSAG1 and NcSAG1) (Howe and Sibley, 1999). Also, a cluster of cell-surface genes was found in N. caninum, as described for T. gondii, and these gene products were re-named as members of the SRS superfamily .
Cross-immunity studies between T. gondii and N. caninum.
Cross-protection of mice immunized with N. caninum and challenged with T. gondii is probably T. gondii strain-and dose-dependent. In one study, mice were immunized with N. caninum and died after challenging with the highly virulent RH strain of T. gondii (Lindsay et al. 1990). When mice were immunized with N. caninum and challenged with a less-virulent T. gondii strain (PLK, a clone from the ME49 strain), the animals had 100% protection against death; immunization with a higher dose of N. caninum tachyzoites improved protection against T. gondii-induced disease (Kasper and Khan, 1998). Similar levels of protection were observed when mice were immunized with N. caninum tachyzoites and then challenged with oocysts of T. gondii from a moderately virulent strain . In another study, pregnant sheep were immunized with a T. gondii sheep vaccine (Toxovax ® , Intervet, Cambridge, UK) and challenged with a high dose of N. caninum tachyzoites (10 7 tachyzoites per animal) at 90 days of gestation (Innes et al. 2001). No protection against fetal death was observed. The authors speculated that if the sheep were challenged with a lower dose of N. caninum tachyzoites, there perhaps would be some degree of cross-immunity (Innes et al. 2001). The confirmation of cross-immunity between T. gondii and N. caninum in some studies suggests that these parasites share antigens which may also be involved in serologic cross-reactivity.
Serologic cross-reactivity between T. gondii and N. caninum.
In the initial observation of N. caninum in dogs, sera from five animals that were naturally infected with the parasite tested negative for T. gondii by the DT (Bjerkas et al. 1984). In addition, dogs that were naturally or experimentally infected with N. caninum did not cross-react to T. gondii by IFAT, when using 1:50 dilutions as cutoff (Dubey et al. 1988b). Accordingly, the same or higher dilution cutoffs by IFAT have been found to be appropriate to avoid cross-reactivity between N. caninum and T. gondii in serum samples from different hosts (Lobato et al. 2006;Silva et al. 2007;Benetti et al. 2009). Apical reactions, i.e. reactions limited to the apex of the parasite were regarded as non-specific in N. caninum IFAT. Cross-reactivity with apical antigens is potentially caused by the high conservation of antigens in the apical organelles of a variety of Apicomplexan parasites, including T. gondii. In contrast, a complete peripheral fluorescence of the parasite was considered as a positive response (Pare et al. 1995b).
Cross-reactive antigens between N. caninum and T. gondii have been observed by immunohistochemistry using tissues of naturally-or experimentally infected animals. A rabbit anti-N. caninum serum cross-reacted with T. gondii in tissue sections from mice . A bradyzoite antigen from T. gondii, designated as BAG1 (synonymous to BAG5) (Weiss et al. 1992;Parmley et al. 1995), was used to produce hyperimmune serum in rabbit, which cross-reacted with bradyzoites of N. caninum (McAllister et al. 1996). Polyclonal sera against T. gondii tachyzoites induced strong crossreactivity with N. caninum tachyzoites by immunohistochemistry (Sundermann et al. 1997). The use of MAbs specific to T. gondii (Sundermann et al. 1997) or a combination of two MAbs specific to N. caninum (Uzêda et al. 2013), were demonstrated to avoid immunohistological cross-reactivity between these protozoa.
Although species-specific MAbs have been developed, serologic cross-reactions between T. gondii and N. caninum have been shown to occur also by means of MAbs. Sundermann et al. (1997) generated MAbs against T. gondii tachyzoites and observed among 26 MAbs tested by IFAT, five antibodies that cross-reacted with N. caninum tachyzoites. Kobayashi et al. (2013) cloned the NcBAG1 gene and generated MAbs against its recombinant protein, which recognized TgBAG1. Liao et al. (2005a) produced 384 MAbs against N. caninum by immunizing mice with N. caninum tachyzoites; 10 of the 384 MAbs were also reactive against T. gondii tachyzoites. Similarly, Sohn et al. (2011) developed 46 MAbs using a mouse immunized with a mixed fraction of N. caninum organelles and some of the MAbs cross-reacted with T. gondii. MAbs generated to oocyst antigens of T. gondii cross-reacted by immunofluorescence with the sporocyst wall (Dumetre and Darde, 2007) and tissue the cyst wall of N. caninum (Gondim et al. 2016). These findings show that cross-reactive antigens between T. gondii and N. caninum are present in tachyzoites, tissue cysts and oocysts of these parasites.
Antigens of N. caninum tested by WB using monoclonal or polyclonal antibodies, resulted in the identification of a limited number of specific immunodominant bands (often referred to as 'immunodominant antigens') and other less reactive bands (Table 1). The comparison of these antigens based on their molecular weights is difficult, as differences on the SDS-PAGE conditions, especially the use of reducing or non-reducing conditions, cause variation in the estimated molecular weights. Most likely, some of these immunodominant bands observed in WBs may represent more than a single protein.
In WB, non-reduced antigens exhibit much stronger reactivity than reduced antigens (Barta and Dubey, 1992) which is a clear indication that most of the epitopes recognized on these antigens are conformational epitopes. Under non-reduced conditions a large number of researchers observed mainly four areas with specific immunodominant bands in N. caninum tachyzoite antigen: However, other areas with major bands of reactions were also observed with non-reduced antigens, but differed widely between studies (Table 1). Sera from animals immunized with recombinant forms of the major N. caninum antigens NcSAG1 and NcSRS2 provided evidence that these antigens are among those recognized in the range 29-32 kDa and 36-40 kDA bands   (Table 2).
Under reduced conditions there is a dominant band between 17-19 kDa which most likely represent reactions against NcGRA7 (Alvarez- Garcia et al. 2007) and other antigens (Table 2). However, as demonstrated by MAbs (4·7·12; Ncmab-7) in combination with immunoprecipitation, surface biotinylation and immuno-electron microscopy, a surface antigen might also be among those migrating in WB at 17-19 kDa under reduced conditions (Björkman and Hemphill, 1998;Schares et al. 1999a). Further immunodominant banding areas are at 34-36 and 37-46 kDa, eventually also representing reactions to NcSAG1, NcGRA6, NcGRA7 and NcSRS2 ( Table 2). When tested with polyclonal antibodies against T. gondii, only minor reactions were observed with antigen bands regarded as specific for N. caninum in WB ( Barta and Dubey, 1992;Björkman et al. 1994Björkman et al. , 1997. For the diagnosis of T. gondii infection in veterinary investigations, the use of recombinant and synthetic antigens, developed using novel molecular techniques, have expanded diagnostic options as alternatives to native antigens directly isolated from cultivated parasites. For diagnosis of T. gondii infection in cats, sheep and pigs, some species-specific ELISAs are available that have performed well when compared with previous reference serological techniques, as reviewed by Wyrosdick and Schaefer (2015).
ELISAs for N. caninum antibodies in sera from dogs and cattle have been developed in several studies (Table 3). Conventional ELISAs using crude soluble antigen showed higher levels of serologic cross-reactivity to T. gondii when compared with IFAT (Björkman et al. 1994;Silva et al. 2007). Serologic cross-reactivity with T. gondii was also observed when polyclonal mouse and cat sera were tested by a N. caninum ELISA using crude antigen (Nishikawa et al. 2002). In contrast, the same mouse and cat sera did not cross-react by an ELISA based on N. caninum recombinant antigen (NcSRS2) (Nishikawa et al. 2002). ELISAs prepared with N. caninum tachyzoite antigen associated with immunostimulating complexes (ISCOM), and using MAbs as secondary antibodies, presented better specificity than conventional ELISAs (Björkman et al. 1994. The ISCOM particles have affinity for surface proteins, which minimizes interference by internal non-specific antigens (Björkman et al. 1994). Moreover, none of the MAbs developed against N. caninum ISCOM incorporated antigens (including also Ncmab-4, and Ncmab-10 mentioned above) cross-reacted with T. gondii (Björkman and Lunden, 1998). An ELISA based on immuno-affinitypurified native NcSRS2 showed no significant cross-reactions when tested with sera from cattle experimentally infected with a variety of protozoan parasites including also ten cattle infected with T. gondii (Schares et al. 2000). In addition, the TgSAG2A molecule has been demonstrated to be specific to T. gondii, considering that no crossreactivity has been shown with N. caninum when using recombinant protein or even mimotopes derived from this molecular marker, as characterized by A4D12 MAb Carvalho et al. 2008;Cunha-Junior et al. 2010;Santana et al. 2012;Macedo et al. 2013).
Nowadays it is becoming clear that there is a need to characterize new molecular markers that are species-specific for T. gondii and N. caninum for the development of new diagnostic tools (Zhang et al. 2011;Regidor-Cerrillo et al. 2015). In this context, the identification of cross-reactive and species-specific antigens between N. caninum and T. gondii tachyzoites is mandatory and the proteomics approach constitutes an appropriate strategy for this purpose (Zhang et al. 2011). These authors demonstrated the usefulness of proteomics to immuno-screen for cross-reactive or species-specific antigens from both parasites. Moreover, they showed that there was significant homology in the antigenic proteome profiles between the two parasites (Zhang et al. 2011). Taking together, these findings shed light on the process to design new diagnostic tools in order to avoid cross-reactivity between N. caninum and T. gondii diagnostic tests.
Besnoitia besnoiti WBs and indirect ELISAs were developed for detection of antibodies in cattle and were tested for cross-reactivity with sera from animals infected with N. caninum and T. gondii (Cortes et al. 2006;Fernandez-Garcia et al. 2010;Schares et al. 2011). Cross-reactivity with N. caninum was observed in the tested ELISAs and WBs, especially with animals exhibiting high antibody titres for N. caninum. A novel ELISA for B. besnoiti was developed with affinity-purified antigens of confirmed relevance, mostly localized on the surface of tachyzoites. This ELISA has shown a lower degree of cross-reactivity with N. caninum . A recent study demonstrated that cattle exhibiting high antibody levels to N. caninum and/or Sarcocystis spp. presented a higher number of false-positive reactions for B. besnoiti by an in-house ELISA ; thus, infection with Sarcocystis spp. and N. caninum may augment cross-reactions with B. besnoiti.
No serologic cross-reactivity has been observed between N. caninum and H. heydorni. A few serum samples from mice, dog and sheep that were infected with H. heydorni did not cross-react with N. caninum antigens by IFAT, ELISA or WB (Nishikawa et al. 2002;Gondim et al. 2015). However, further studies using larger numbers of sera from H. heydorniinfected animals are necessary to confirm the absence of cross-reactivity between N. caninum and H. heydorni.
Neospora hughesi was proposed as a new species based on seven nucleotide differences in the internal transcribed spacer 1 (ITS1) of the rDNA, as well as on ultrastructural and antigenic differences when compared with N. caninum (Marsh et al. 1998). The acceptance of N. hughesi as a new species was reinforced by differences observed in amino acid sequences of two surface antigens (SAG1 and SRS2) to those from N. caninum (Marsh et al. 1999). However, polyclonal serum from a N. caninum-infected rabbit recognized NcSAG1, NhSAG1, NcSRS2 and NhSRS2 by immunoblot (Marsh et al. 1999). Differences also have been found in gene sequences between the dense granule proteins GRA6 and GRA7 of N. caninum and N. hughesi, although polyclonal serum to N. caninum recognized GRA6 and GRA7 antigens from both N. caninum and N. hughesi by immunoblot (Walsh et al. 2001).
An ELISA for N. hughesi antibodies was developed using a recombinant NhSAG1 as antigen (Hoane et al. 2005); animals infected with N. hughesi presented a higher antibody reactivity to rNhSAG1 than to rNcSAG1, but the test was not able to unambiguously differentiate infections caused by N. hughesi or N. caninum. In another study, pre-and post-infection sera from dog and cattle that were experimentally inoculated with N. caninum were tested simultaneously by IFAT using tachyzoites of N. caninum or N. hughesi (Gondim et al. 2009); all sera that tested positive for N. caninum also reacted with N. hughesi tachyzoites, although the antibody titres for N. hughesi IFAT were slightly lower as compared with the IFAT for N. caninum. To date, infections caused by N. caninum and N. hughesi cannot be serologically discriminated. As horses may be infected by both N. caninum and N. hughesi (Marsh et al. 1999;Pitel et al. 2003;Veronesi et al. 2008), a species-specific serologic test for equines is desired.
Serological tests based on chimeric antigens, synthetic peptides and stage-specific antigens Chimeric antigens.
Almost 30 years ago, the production of recombinant polypeptides derived from genes encoding T. gondii antigens revolutionized the search for more efficient serologic methods (Johnson et al. 1989;Johnson and Illana, 1991). The development of ELISAs based on a mixture of recombinant antigens, rather than the use of a single recombinant protein, has been presumed to increase the sensitivity of the ELISA for human sera, while maintaining the desired specificity Aubert et al. 2000). ELISAs containing a mixture of recombinant antigens were also tested for antibodies against T. gondii in sheep and cats .
A chimeric antigen is the fusion of gene fragments constructed as a single gene and expressed to form a hybrid protein (Yang et al. 2004). The use of chimeric antigens for T. gondii serology is promising. Chimeric proteins are usually larger than single recombinant antigens resulting in a better binding to microtitre plates. In addition, as a chimeric antigen preparation consists of a single antigen, it may be easier to standardize than mixtures of recombinant antigens (Beghetto et al. 2006;Lau et al. 2011;Holec-Gasior et al. 2012a).
Several chimeric antigens have been tested by WB or ELISA for the detection of human antibodies against T. gondii. Among the developed chimeric antigens tested by serology, are preparations including parts of well-characterized immunodominant proteins, such as SAG1, SAG2, MIC1, MIC2, MIC3, MAG1, M2AP, GRA1, GRA2, GRA3 and ROP1 (Beghetto et al. 2006;Lau et al. 2011;Holec-Gasior et al. 2012a, b;Ferra et al. 2015a). The standardization of a chimeric-antigen-based test for T. gondii antibodies depends on various factors, including an optimal selection of antigens and proper expression of all desired epitopes. In addition to humans, a recent paper reports the first trial of chimeric antigens employed for T. gondii serology in farm animals (horses, swine and sheep) (Ferra et al. 2015b); the authors validated their antigen with more than 400 sera and also included 15 sera, which were serologically positive for N. caninum but negative for T. gondii. It is interesting to note that each animal species responded differently to the chimeric-antigen preparations used in the ELISAs, but the SAG2-GRA1-ROP1 L construct reached the best overall specificity and sensitivity for the three tested species (Ferra et al. 2015b). Despite progress in the development of chimeric antigens for T. gondii serology, in particular for humans, inadequate information is available about the serologic cross-reactivity potential of those tests with sera from animals infected with other Toxoplasmatinae parasites.
Synthetic peptides, serotyping and stage-specific antigens.
The combination of molecular engineering and chemical synthesis of antigens has been applied for the development of serological techniques with improved sensitivityand specificity. Synthetic peptides representing several epitopes of numerous antigens have been used in microarray assays for serotyping of viral and bacterial diseases (Neuman de Vegvar et al. 2003;Nahtman et al. 2007).
In T. gondii infections, the humoral response has been demonstrated to be partially strain-specific. In one study, MAbs produced against SAG2A from naturally infected mice recognized the surface antigens encoded by the SAG2 allele of type I and III strains, but not of type II strains (Parmley et al. 1994). Another study identified a MAb showing differences in the recognition of type II and III strains (Bohne et al. 1993). Kong et al. (2003) screened nucleotide sequences from types I, II and III of T. gondii for the identification of polymorphic regions from genes coding selected antigens. Allele-specific peptides were synthetized and screened by ELISA using sera from mice and humans. Synthetic peptides based on SAG2A, GRA3, GRA6 and GRA7 were able to discriminate type II from non-type II infections (Kong et al. 2003). In subsequent studies, serotyping for T. gondii infections using new recombinant polypeptides or new or improved synthetic peptides, as well as target populations from different regions, have been performed by ELISAs (Peyron et al. 2006;Sousa et al. 2008Sousa et al. , 2009. In two studies, peptide-microarrays were validated to discriminate the serological responses against clonal-type T. gondii strains in samples from humans (Maksimov et al. 2012b) and cats . The latter study used sera from experimentally infected cats for validation and showed significant typespecific differences in the IgG response against the tested peptide panel. However, in many peptides, reactions were not clonal type-specific .
A peptide microarray was developed and validated for serotyping T. gondii infections in humans, aiming to differentiate between different manifestations of T. gondii infection (Maksimov et al. 2012a). Thirty eight T. gondii synthetic peptides, consisting of 18 peptides characterized in previous studies, and 20 novel peptides, predicted by bioinformatics approach, were tested to differentiate acute, latent and ocular infections. Some peptides based on dense granule and microneme antigens (GRA2-28,  showed promising results for differentiation between acute and latent infections (Maksimov et al. 2012a).
The use of synthetic peptides for T. gondii serotyping may have a great potential for discriminating between T. gondii infections and infections caused by other parasite species expected to induce some degree of serologic cross-reactivity. Of course, one important prerequisite is that peptides applied in such tests are not specific for particular clonaltypes or genotypes of T. gondii or related parasite species. In addition, potential cross-reactivities need to be addressed; for instance, synthetic peptides considered to react specifically with antibodies generated by T. gondii infections need to be tested with sera from animals exposed to H. hammondi, as this parasite has a very close genetic relationship to the former (Walzer et al. 2013(Walzer et al. , 2014. The use of T. gondii-specific peptides for diagnosis in animals should also be tested with N. caninum, which may cross-react, as shown in Table 3. To our knowledge, there are no published studies using synthetic peptides to differentiate infections in animals cause by T. gondii, H. hammondi or N. caninum. Stage-specific antigens (native or recombinant) from T. gondii have been investigated for decades by several research groups, including oocysts and sporozoites which can be produced in cats (Kasper et al. 1984;Ferguson et al. 2000;Dumetre and Darde, 2007;Possenti et al. 2010Possenti et al. , 2013Bushkin et al. 2012;Fritz et al. 2012). In contrast, antigens from Neospora spp. are mostly obtained from tachyzoites. Only small numbers of studies have produced N. caninum tissue cysts in cell culture or in animals (Weiss et al. 1999;Tunev et al. 2002;Risco-Castillo et al. 2004;Vonlaufen et al. 2004), or have induced oocyst production in canid definitive hosts (McAllister et al. 1998;Schares et al. 2001;Gondim et al. 2002).
The identification of the first bradyzoite-specific gene of N. caninum (NcSAG4), an orthologue to TgSAG4, allowed the production and characterization of the recombinant protein rNcSAG4 (Fernandez-Garcia et al. 2006). This antigen has been applied for the development of a stage-specific ELISA for bovine neosporosis (Aguado-Martinez et al. 2008).
The production of Hammondia spp. antigens is even more difficult, because there is no permanent culture for these parasites (Riahi et al. 1995;Schares et al. 2003;Gondim et al. 2015). Therefore, most antigens from Hammondia spp. are derived from oocysts or from parasite cysts from intermediate hosts (Riahi et al. 1998(Riahi et al. , 1999(Riahi et al. , 2000Walzer et al. 2013). Generation of recombinant antigens from Hammondia spp. is lacking. It would enable studies on exposure of animals and humans to these parasites, as well as on serologic cross-reactivity with related protozoa.
A serologic test for humans based on a T. gondii sporozoite-specific protein has been evaluated and seems to represent a promising method to discriminate between oocyst-induced infections from those transmitted by ingestion of infected meat (Hill et al. 2011), which is an important consideration in epidemiological investigations. This study was the first to use a sporozoite protein, called T. gondii embryogenesis-related protein (TgERP) in a serologic test for T. gondii. In a recent study another sporozoite-derived protein called CCp5A was successfully used in an ELISA to identify oocystinfected animals (Santana et al. 2015); this test was able to identify antibodies in sera from humans, pigs, mice and chickens that were naturally or experimentally infected with T. gondii oocysts and discriminate between animals infected from ingestion of oocysts or by carnivorism. Further investigations are necessary for validation and to confirm whether the sporozoite proteins TgERP (Hill et al. 2011) and CCp5A (Santana et al. 2015) cross-react with sporozoite proteins derived from Hammondia spp. and Neospora spp. It is also important to test whether low infectious dose with oocysts will elicit detectable antibodies to TgERP or CCp5A.
Type or stage-specific antigens have a wide spectrum of applications, including investigations of infection outbreaks and identification of risk factors in selected populations (e.g. ingestion of oocyst or tissue cysts). Those antigens should be used with caution and only after rigorous validation in epidemiological studies.
C O N C L U D I N G R E M A R K S Most currently available serologic tests for T. gondii may show some level of cross-reactivity with related coccidia, in particular with H. hammondi and N. caninum. Serological cross-reactivity between T. gondii and N. caninum has been observed when crude antigen ELISAs are employed. Several ELISAs based on recombinant species-specific antigens did not present cross-reactivity with N. caninum. Therefore, T. gondii shares more surface antigens with H. hammondi than with N. caninum.
When detecting T. gondii antibodies in animals, IFAT seems to be more specific than the ELISAs based on crude antigen. Since N. caninum is not considered to be a human pathogen, the major concern regarding serologic cross-reactivity would be potential exposure to H. hammondi, although it is unknown whether this parasite is able to induce infection in humans.
Antibodies from cattle infected with B. besnoiti cross-react with N. caninum antigens by IFAT in serum dilutions lower than the recommended cutoff (1:200). Novel B. besnoiti ELISAs (e.g. ELISA based on purified surface antigens) showed a lower degree of cross-reactivity with sera from cattle infected with N. caninum or other related parasites. Antibodies against bovine Sarcocystis sp. present negligible crossreactions with N. caninum-immunodominant antigens. The development of ELISAs for N. caninum antibodies based on chimeric peptides specific for the parasite seems to be promising. Chimeric antigens would enable a better standardization of serologic tests, as a single protein is used. Moreover, chimeric antigens would potentially present higher sensitivity than single recombinant antigens. Despite the close phylogenetic relationship between N. caninum and H. heydorni, no evidence of serologic cross-reactivity between these protozoa has been confirmed to date. Toxoplasma gondii infections cause more confusion with N. caninum serology than those infections induced by H. heydorni. Neospora caninum and N. hughesi infections cannot be serologically differentiated, because currently available serologic tests are genus rather than species-specific for these protozoa.