EXPERIMENTAL ANIMAL MODELS TO STUDY VACCINES AND DRUGS AGAINST NEOSPOROSIS

There is a general consensus that small laboratory animals such as mice do not accurately reflect the situation that occurs during N. caninum infection in cattle. Physiological differences can account for discrepancies in drug efficacy, tolerance and/or toxicity and immunological differences between ruminant and mice hamper the use of small laboratory animals as models for a vaccine to be used in cattle (Monney and Hemphill, Reference Monney and Hemphill2014). For instance, while antibodies can pass the placental tissue in mice, this is not the case in cattle. In mice, an increase of IFN-γ production is correlated to increases of IgG2a- and IgG3-synthesis, increased IL-4 production is associated with increased IgG1- and IgE-levels, and elevated TGF-β is associated with higher IgA responses. In cattle, the situation is less polarized and the classical roles of many cytokines in the laboratory mouse do not extrapolate entirely, or at all, to the situation in cattle (Estes and Brown, Reference Estes and Brown2002). Compared with mice, cattle have a higher number of circulating γδ T cells. These cells can respond to antigens and pathogen associated molecular patterns (PAMPs) from different disease agents such as tuberculosis (Telfer and Baldwin, Reference Telfer and Baldwin2015), and may play a role in Neospora infections. In addition, Neospora-specific cytotoxic lymphocytes are CD4(+), while cytotoxic T cells are CD8(+) in the murine model (Staska et al. Reference Staska, Davies, Brown, McGuire, Suarez, Park, Mathison, Abbott and Baszler2005).

Nevertheless, by far the largest numbers of experimental vaccinations and drug treatment studies have been carried out in mice (see Tables 1 and 3 for a summary of selected studies), since experiments in small animals are much more cost-effective, and despite the obvious setbacks can still provide initial information on potentially protective effects for promising vaccine and/or drug formulations. Two types of murine models are available, (i) the non-pregnant mouse model for the assessment of acute disease and/or CNS infection and (ii) the pregnant mouse model that mimics exogenous transplacental infection (Williams and Trees, Reference Williams and Trees2006; Williams et al. Reference Williams, Hartley, Bjorkman and Trees2009). For the latter, mice are made pregnant and then challenged during pregnancy, and the drug- or vaccine mediated protection of dams and offspring is assessed. Despite considerable efforts there is no laboratory animal that allows efficient monitoring of endogenous transplacental infection, namely infection of the fetus following recrudescence of N. caninum infection during pregnancy (Jimenez-Ruiz et al. Reference Jimenez-Ruiz, Alvarez-Garcia, Aguado-Martinez and Ortega-Mora2013a , Reference Jimenez-Ruiz, Alvarez-Garcia, Aguado-Martinez and Ortega-Mora b ).

Table 1.

Overviews on selected vaccine studies on neosporosis carried out in the mouse model

rE, recombinantly expressed in E. coli; DNA, DNA vaccine; s.c., subcutaneous; i. p., intraperitoneal; i. m., intramuscular; i. n., intranasal; p. p., post-partum.

Vaccination studies have been performed using live-attenuated N. caninum tachyzoite isolates, tachyzoite extracts, or specific polypeptides expressed in various systems. In general terms, typical vaccine trials in mice comprise different sequential steps, that include (i) verification that the test animals are not seropositive for N. caninum and/or already infected with a related pathogen; (ii) immunization and boosts with the vaccine formulation to be tested by different routes, and control animals receiving a placebo application; (iii) challenge of animals with a defined inoculum of N. caninum tachyzoites; (iv) monitoring of clinical signs linked to acute or chronic neosporosis. Assessments include survival of dams and offspring, neurological symptoms, parasite burdens in various organs, especially the brain, and humoral and cellular immune responses. In the case of vaccination studies in the pregnant model, mice are usually mated a few weeks following the final vaccination, and experimentally infected with N. caninum tachyzoites of a virulent isolate between days 7 and 9 post-mating, which produces almost 100% vertical transmission and mortality after 1 month post-partum in the offspring from non-vaccinated dams (Lopez-Perez et al. Reference Lopez-Perez, Collantes-Fernandez, Aguado-Martinez, Rodriguez-Bertos and Ortega-Mora2008). Experiments involving the evaluation of drug candidates follow a similar scheme, also ensuring that all animals are sero-negative, but omitting the vaccination steps. Treatments with the compounds of interest are initiated ideally 2–3 days post-infection, in order to give the parasite some time to establish before it encounters the drug treatment. Since it cannot be ruled out that a given compound affects pregnancy and offspring, in some cases, controls with uninfected dams have to be included. Ideally, in both vaccine and drug studies using the pregnant mouse model, the fate of offspring mice is followed-up over a period of 1 month post-partum (Arranz-Solís et al. Reference Arranz-Solís, Aguado-Martínez, Müller, Regidor-Cerrillo, Ortega-Mora and Hemphill2015).

A major setback of many experimental studies in mice carried out to date is the fact that they cannot be accurately compared with each other, since they lack standardization. Different groups have worked with different mouse strains, different parasite isolates; have employed different culture techniques and varying routes of inoculation. With this in mind, defined conditions for the standardization and refinement of the pregnant BALB/c mouse model for N. caninum infection, employing the virulent Nc-Spain7 isolate have recently been suggested (Arranz-Solís et al. Reference Arranz-Solís, Aguado-Martínez, Müller, Regidor-Cerrillo, Ortega-Mora and Hemphill2015). The key finding of this study were: (i) a challenge dose of 10⁵ N. caninum tachyzoites produced identical results as so far most frequently employed dose of 2 × 10⁶ N. caninum tachyzoites; and (ii) inoculation of 100 tachyzoites still resulted in 76% pup mortality at 1 month post-partum. This indicated that most experimental infections in mice carried out so far have used an unrealistically high challenge dose, rendering the identification of immuno-protective antigens or antigen-combinations, or bioactive compounds, a difficult undertaking.

Only few groups have taken on the heroic task to perform experimental vaccinations/challenge infections and even field trials in cattle, some of which are summarized in Table 2. Naturally, such trials require much larger financial resources compared with studies in small laboratory animals, but have the advantage that they are carried out in the natural and economically most important host species. Basically, two experimental vaccine strategies have been followed more recently: (i) vaccination of non-pregnant animals prior to pregnancy and with naturally attenuated low-virulence N. caninum isolates (e.g. NcSpain 1H and NcNowra) followed by challenge during defined time points of gestation; (ii) vaccination of already pregnant cattle with live vaccines, native antigens or recombinant antigens (see Table 2). The only drug that has been evaluated in cattle so far is toltrazuril, which was administered to newborn calves experimentally infected with N. caninum tachyzoites (Häsler et al. Reference Häsler, Regula, Stark, Sager, Gottstein and Reist2006a , Reference Häsler, Stark, Sager, Gottstein and Reist b ). Results on the efficacy of toltrazuril, however, were not conclusive (see Table 3).

Table 2.

Overviews of selected vaccine studies against neosporosis in farm animals

CHV, canine herpes vector; s. c., subcutaneous; i. v., intravenuous; ISCOMs, immune stimulating complexes.

Table 3.

Overviews of selected in vitro and in vivo studies with chemotherapeutics against neosporosis

If not indicated otherwise, in vivo studies were performed with mice. p. o., per os (oral application); i. p., intraperitoneal; HFF, human foreskin fibroblasts.

An attractive and more cost-effective alternative to the bovine model is to perform experimental infection studies in sheep and exploit them as a small ruminant model for neosporosis. As an experimental animal model, sheep exhibit several advantages over cattle, including size, length of gestation and cost. In addition, unlike mice, sheep do not represent an artificial model, as recent evidence suggests that N. caninum is an important abortifacient in small ruminants (Moreno et al. Reference Moreno, Collantes-Fernandez, Villa, Navarro, Regidor-Cerrillo and Ortega-Mora2012), or even the main cause of reproductive losses in some flocks (Gonzalez-Warleta et al. Reference Gonzalez-Warleta, Castro-Hermida, Regidor-Cerrillo, Benavides, Alvarez-Garcia, Fuertes, Ortega-Mora and Mezo2014). This would also account for goats (Dubey, Reference Dubey2003; Moore, Reference Moore2005; Costa et al. Reference Costa, Orlando, Abreu, Nakagaki, Mesquita, Nascimento, Silva, Maiorka, Peconick, Raymundo and Varaschin2014). Experimental infections in pregnant sheep (McAllister et al. Reference McAllister, McGuire, Jolley, Lindsay, Trees and Stobart1996; Buxton, Reference Buxton1998; Weston et al. Reference Weston, Howe, Collett, Pattison, Williamson, West, Pomroy, Syed-Hussain, Morris and Kenyon2009) and pregnant pygmy goats (Lindsay et al. Reference Lindsay, Rippey, Powe, Sartin, Dubey and Blagburn1995) have shown that both are highly susceptible, and disease outcomes display similar characteristics as reported for cattle. As in cattle, it has been suggested that the time point of infection during gestation plays a key role in the pathogenesis of the disease (Dubey and Lindsay, Reference Dubey and Lindsay1990; Buxton et al. Reference Buxton, Maley, Thomson, Trees and Innes1997, Reference Buxton, Wright, Maley, Rae, Lunden and Innes2001; Jenkins et al. Reference Jenkins, Tuo and Dubey2004). However, again comparison between different studies is difficult due to the lack of standardization. Most recently, the outcome of experimental infection by N. caninum in ewes has been investigated under standardized conditions at an early, mid-term and late period of gestation allowing evaluating the effect of the gestation period on the clinical course of disease, lesion development and parasite distribution in different organs (Arranz-Solis et al. Reference Arranz-Solis, Benavides, Regidor-Cerrillo, Fuertes, Ferre, Ferreras Mdel, Collantes-Fernandez, Hemphill, Perez and Ortega-Mora2015) This model for exogenous transplacental transmission for ruminant neosporosis can serve as a model to study the effects of vaccines and drugs.

VACCINES AGAINST NEOSPOROSIS

The costs of management practices used for control of neosporosis indicate that vaccination could be the most efficient intervention strategy (Reichel and Ellis, Reference Reichel and Ellis2009). The only licensed Neospora vaccine, Bovilis Neoguard^®, was composed of a tachyzoite lysate, and was available in selected countries for several years (Barling et al. Reference Barling, Lunt, Graham and Choromanski2003). However, this vaccine exhibited only moderate efficacy in field trials (Romero et al. Reference Romero, Perez and Frankena2004), and one study suggested that vaccination itself could increase the risk of early embryonic death (Weston et al. Reference Weston, Heuer and Williamson2012). As the vaccine has been taken off the market, farmers have been left without an alternative (Reichel et al. Reference Reichel, Moore, Hemphill, Ortega-Mora, Dubey and Ellis2015). Therefore, there is a need for the development of effective vaccines to prevent N. caninum infection (Monney and Hemphill, Reference Monney and Hemphill2014).

More recently, an in silico approach has been proposed for the development of a vaccine against neosporosis (Goodswen et al. Reference Goodswen, Kennedy and Ellis2014). This approach is based on exploiting the current genomic and transcriptomic information on the N. caninum Nc-Liv isolate, and using bioinformatics tools to assess the suitability of expressed proteins as vaccine candidates by identifying those antigens containing T and B cell epitopes by reverse vaccinology. This approach could be helpful in providing a priority list of potential vaccine candidates. However, besides assessing the physicochemical properties of proteins based on their predicted peptide sequence, this approach does not take into account other important factors that are crucial for vaccine development against N. caninum infection, such as the extensive immunomodulation that takes place during pregnancy, routes of vaccine delivery, infection route, genetic background of parasites and hosts, immunization dose, challenge dose and timing, as well as the effects of the adjuvant and the non-protein components of an antigen preparation such as lipids and carbohydrates (Monney and Hemphill, Reference Monney and Hemphill2014). Nevertheless, with the improvement of novel bioinformatics tools in silico vaccinology could become an interesting approach, but will not replace the tedious work in the wet lab in the near future.

Both, live vaccines and subunit vaccines have been studied in experimental settings. So far, the largest numbers of experimental vaccinations has been carried out in mice (see Table 1), while few studies have been performed in cattle and sheep, a selection of which is summarized in Table 2. While live vaccines have clearly shown superior efficacy in experimental trials compared with any subunit vaccine formulation, the pharmaceutical industry has been reluctant to introduce these live vaccines into the market, and have clearly favoured an approach that includes subunit vaccines (Reichel et al. Reference Reichel, Moore, Hemphill, Ortega-Mora, Dubey and Ellis2015). For the closely related T. gondii, considerable efforts have been put into the development of vaccines to reduce oocyst shedding in cats and tissue cyst formation in mammals. However, the only licensed vaccine is a live-attenuated strain (Toxovax^®), which is maintained in cell culture and licensed for veterinary use in sheep only (Zhang et al. Reference Zhang, Chen, Wang, Petersen and Zhu2013). Also for other livestock apicomplexan diseases, including theileriosis, babesiosis and coccidiosis, live vaccines are used as effective tools for the prevention of infection and serious disease (McAllister, Reference McAllister2014). In the case of Theileria parva, the live vaccine is composed of a cocktail of three strains, but it has been shown that there is not always cross-protection against other strains (Morrison et al. Reference Morrison, Connelley, Hemmink and MacHugh2015). For besnoitiosis, a live vaccine has been produced and made available in Israel, but its efficacy and safety have not been reported. A live vaccine for neosporosis prevention ideally would use attenuated parasites, but concerns such as high production costs, short shelf life of the product, maintenance of a cool-chain prior to application of the vaccine and the risk of reversion to virulence have prevented the introduction of several live-Neospora vaccine candidates into the market. In addition, as live vaccines might result in chronic infection of the host, there is a risk that the life cycle could ultimately be completed again, if tissues from vaccinated animals were fed to canid definitive hosts (Reichel et al. Reference Reichel, Moore, Hemphill, Ortega-Mora, Dubey and Ellis2015). Toxovax^® comprises a T. gondii tachyzoite strain that has lost the ability to form tissue cysts in the vaccinated host however, the molecular basis for this failure to persist long-term in the host has not been defined (Monney and Hemphill, Reference Monney and Hemphill2014).

For the development of subunit vaccines, many researchers have exploited the fact that N. caninum has developed distinct adaptations to its intracellular lifestyle. Survival depends on successful invasion of host cells, and on the ability to differentiate into the slowly proliferating and cyst-forming bradyzoite stage under physiological stress conditions (Hemphill et al. Reference Hemphill, Debache, Monney, Schorer, Guionaud, Alaeddine, Müller and Müller2013a ). Knowledge of the molecular basis of these processes is essential for understanding the pathogenic mechanisms underlying infection and for designing strategies to combat these diseases. Thus, the concept of using subunit vaccines has largely relied on identifying defined parasite fractions or proteins that play essential roles in host cell invasion and/or tachyzoite-to-bradyzoite stage differentiation, and targeting these components by generating humoral and/or cellular immunity against them.

Some vaccines based on native antigens of apicomplexan parasites are currently commercially available. For instance, CoxAbic™ is composed of affinity-purified gametocyte antigens from Eimeria maxima and confers protection to hens and their offspring against coccidiosis by transmission of specific antibodies via egg yolk (Sharman et al. Reference Sharman, Smith, Wallach and Katrib2010). Another marketed vaccine, Nobivac Piro™, is composed of soluble antigens from two Babesia species and confers protection against babesiosis in dogs (Schetters et al. Reference Schetters, Moubri and Cooke2009). However, these vaccines are derived from parasite cultures, and to date there has been no vaccine candidate composed of bacterially expressed subunit antigens showing convincing protection against apicomplexan parasites.

There are a few common characteristics in all Neospora vaccines studies carried out so far: a primary characteristic is that only live vaccines have conferred convincing protection against abortion and/or vertical transmission upon experimental or natural challenge in cattle. Vaccination of animals prior to pregnancy with naturally attenuated low-virulence N. caninum isolates (e.g. NcSpain1H and NcNowra) and then challenged during defined time points of gestation demonstrated substantial protection against abortion and fetal loss (Williams et al. Reference Williams, Guy, Smith, Ellis, Bjorkman, Reichel and Trees2007; Rojo-Montejo et al. Reference Rojo-Montejo, Collantes-Fernandez, Perez-Zaballos, Rodriguez-Marcos, Blanco-Murcia, Rodriguez-Bertos, Prenafeta and Ortega-Mora2013; Weber et al. Reference Weber, Jackson, Sobecki, Choromanski, Olsen, Meinert, Frank, Reichel and Ellis2013). Vaccination of pregnant cattle with tachyzoites of a live-attenuated vaccine strain (Hecker et al. Reference Hecker, Moore, Quattrocchi, Regidor-Cerrillo, Verna, Leunda, Morrell, Ortega-Mora, Zamorano, Venturini and Campero2013), iscom preparations of native N. caninum extract (Hecker et al. Reference Hecker, Moore, Quattrocchi, Regidor-Cerrillo, Verna, Leunda, Morrell, Ortega-Mora, Zamorano, Venturini and Campero2013), or recombinant antigens (SAG1, Hsp20, GRA7) incorporated into iscoms (Hecker et al. Reference Hecker, Coceres, Wilkowsky, Jaramillo Ortiz, Morrell, Verna, Ganuza, Cano, Lischinsky, Angel, Zamorano, Odeon, Leunda, Campero, Morein and Moore2014) demonstrated partial protection against transplacental fetal infection with the live vaccine, but not with the native or recombinant subunit vaccines. In addition, a live vaccine isolate (NcIs491) was assessed in a field trial comprised of 520 pregnant and N. caninum seropositive cows, of which 146 were vaccinated at mid-gestation and 374 served as controls (Mazuz et al. Reference Mazuz, Fish, Wolkomirsky, Leibovich, Reznikov, Savitsky, Golenser and Shkap2015). While this field trial showed that the live vaccine reduced the abortion to 16% compared with 25% abortion losses in unvaccinated cows, these losses upon endogenous transplacental transmission are unusually high, much higher than in typical endemic situations. Interestingly, this vaccine reduced abortion, but not the rate of vertical transmission. Similar efficacy of live vaccines was demonstrated in the mouse model (Miller et al. Reference Miller, Quinn, Ryce, Reichel and Ellis2005; Rojo-Montejo et al. Reference Rojo-Montejo, Collantes-Fernandez, Lopez-Perez, Risco-Castillo, Prenafeta and Ortega-Mora2012). A second important point is that subunit vaccines, either tachyzoite-derived or expressed as recombinant antigens exhibited promising protective efficacy in non-pregnant models (mainly mouse models), but in pregnant mice, the same or virtually identical formulations have been found to be largely non-protective (reviewed in Monney and Hemphill, Reference Monney and Hemphill2014; see also Table 1). Thus, in these cases pregnancy has led to the loss of subunit vaccine-induced protective immunity. Protection against infection in non-pregnant cattle was achieved using oligomannose-coated liposome entrapped NcGra7, but there is no further information whether this protection would be retained during pregnancy (Nishimura et al. Reference Nishimura, Kohara, Kuroda, Hiasa, Tanaka, Muroi, Kojima, Furuoka and Nishikawa2013). In addition, in both pregnant and non-pregnant mouse models, combinations of antigens as polyvalent vaccine exhibited a higher efficacy compared with- single antigens applied as monovalent vaccines, indicating that only a combination of recombinant antigens will induce protective immunological responses (Debache et al. Reference Debache, Alaeddine, Guionaud, Monney, Müller, Strohbusch, Leib, Grandgirard and Hemphill2009; Pastor-Fernandez et al. Reference Pastor-Fernandez, Arranz-Solis, Regidor-Cerrillo, Alvarez-Garcia, Hemphill, Garcia-Culebras, Cuevas-Martin and Ortega-Mora2015). However, in some cases, vaccination rendered mice are more susceptible to infection, demonstrating that some antigenic components of the parasite could exhibit immune-modulating properties (Srinivasan et al. Reference Srinivasan, Müller, Suana and Hemphill2007).

Another difficulty in interpreting the results from studies employing recombinant vaccines comes from the fact that many of these antigens are expressed in Escherichia coli and affinity purified from crude extracts. Therefore, the presence of immuno-modulating agents derived from bacterial contaminants such as lipopolysaccharides (LPS) cannot be ruled out, and controls with irrelevant proteins expressed in the same system or with LPS -depleted protein fractions (Basto et al. Reference Basto, Badenes, Almeida, Martins, Duarte, Santos and Leitao2015) should be included. Surprisingly, only a minority of Neospora vaccine studies performed to date have addressed this point. LPS and other bacterial contaminants will greatly influence the way innate immune pathways are activated, and to what extent polarization of the cellular immune response is elicited later during infection (Basto et al. Reference Basto, Piedade, Ramalho, Alves, Soares, Cornelis, Martins and Leitão2012; Basto and Leitão, Reference Basto and Leitão2014). This fact could be exploited, by actually generating LPS-free subunit vaccine candidate formulations, and then mixing them with PAMPs in a controlled way, or even physically link these antigens with PAMPs as described by Basto et al. (Reference Basto, Piedade, Ramalho, Alves, Soares, Cornelis, Martins and Leitão2012, Reference Basto, Badenes, Almeida, Martins, Duarte, Santos and Leitao2015).

DRUGS AGAINST NEOSPOROSIS

In general, chemotherapeutic treatment of Neospora-seropositive animals has not been regarded as an economically viable option, since no effective and safe drugs are available, and depending on the compounds used, milk or meat from drug-treated animals would remain unacceptable for consumption for some time (Dubey et al. Reference Dubey, Schares and Ortega-Mora2007). Nevertheless, experimental studies have revealed potentially interesting effects of several compounds in vitro and in laboratory animal models in vivo (Müller and Hemphill, Reference Müller and Hemphill2011) (see Table 3). Many of these compounds or compound classes were shown previously to be active against other intracellular protozoan parasites, including Trypanosoma cruzi, the causative agent of Chagas Disease, and Leishmania species responsible for cutaneous and visceral leishmaniasis, and some exhibited broad-spectrum anti-parasitic activity against various protozoan and helminth species. However, other approaches identified compounds that inhibited targets that were conserved almost exclusively within the group of apicomplexan parasites, and were shown to be active against different N. caninum isolates, as well as other apicomplexans including Plasmodium, Toxoplasma, Cryptosporidium and others. The most interesting drugs, however, are most notably derived from screenings carried out in the framework of Plasmodium research. Their application against other apicomplexans would thus be a good example of drug repurposing (Andrews et al. Reference Andrews, Fisher and Skinner-Adams2014; Sateriale et al. Reference Sateriale, Bessoff, Sarkar and Huston2014).

The strategies to identify anti-parasitic agents are discussed elsewhere (Müller and Hemphill, Reference Müller and Hemphill2013a , Reference Müller and Hemphill b ). Briefly, drug candidates are identified and initially characterized by in vitro tests, during which suitable host cells (e.g. fibroblasts) are infected with N. caninum tachyzoites in the presence of the test compounds or of a solvent control. After a given time period (i. e. when the controls show a high level of infection) the experiment is stopped and the tachyzoites are quantified by a suitable method. Quantitative real time polymerase chain reaction will work with all N. caninum isolates, but is laborious, time consuming and costly, if many samples need to be processed. For higher throughput screening, transgenic parasites expressing an easily detectable marker are more practical. For instance, a N. caninum NC-1 strain expressing E. coli beta-galactosidase under the control of a GRA1 promotor (Howe and Sibley, Reference Howe and Sibley1997) has been used for such initial drug screenings. At first, these in vitro studies will provide inhibition constants (e.g. IC₅₀) and data concerning host cell toxicity. Furthermore, it can be determined whether a compound is parasitocidal or only parasitostatic, whether it affects intracellular parasites or only parasites prior to infection, and to which extent resistance formation can occur. Combined with morphological studies using scanning and transmission electron microscopy, such in vitro studies have already provided a detailed picture on how a given compound affects the parasite. For recent examples on Neospora see Table 3. Moreover, a detailed study dealing with all these aspects has been performed with T. gondii strains and pentamidine derivatives (Kropf et al. Reference Kropf, Debache, Rampa, Barna, Schorer, Stephens, Ismail, Boykin and Hemphill2012).

Several studies have been performed with toltrazuril, a triazinone derivative effective against various coccidians including Eimeria (Steinfelder et al. Reference Steinfelder, Lucius, Greif and Pogonka2005), and commercialized under the proprietary name Baycox™. The mode of action of toltrazuril and of its main metabolite toltrazuril sulfone (ponazuril) does not consist only in the inhibition of dihydroorotate dehydrogenase and thereby pyrimidine biosynthesis, but also in the inhibition of the respiratory chain of the parasite (Harder and Haberkorn, Reference Harder and Haberkorn1989). Whereas the effects against coccidian infections are well documented in poultry (Mathis et al. Reference Mathis, Froyman, Irion and Kennedy2003) as well as in cattle (Mundt et al. Reference Mundt, Bangoura, Mengel, Keidel and Daugschies2005), it remains unclear whether toltrazuril is a suitable drug against neosporosis because the efficacy results in cattle do not support this conclusion (Haerdi et al. Reference Haerdi, Haessig, Sager, Greif, Staubli and Gottstein2006; see Table 3). Thiazolides including nitazoxanide, the prototype compound of this class (Hemphill et al. Reference Hemphill, Müller and Müller2013b ), have good effects against N. caninum in vitro, but fail in vivo when applied orally or are even toxic when applied intraperitoneally (see Table 3). This is most likely due to induction of host cell apoptosis (Müller et al. Reference Müller, Sidler, Nachbur, Wastling, Brunner and Hemphill2008). The most promising drug candidates against neosporosis come from compounds initially developed against Plasmodium (artemisinin-derivatives) and Leishmania (dicationic pentamidine derivatives (Soeiro et al. Reference Soeiro, Werbovetz, Boykin, Wilson, Wang and Hemphill2013) and spiroindolones, a novel class of antimalarials (Rottmann et al. Reference Rottmann, McNamara, Yeung, Lee, Zou, Russell, Seitz, Plouffe, Dharia, Tan, Cohen, Spencer, Gonzalez-Paez, Lakshminarayana, Goh, Suwanarusk, Jegla, Schmitt, Beck, Brun, Nosten, Renia, Dartois, Keller, Fidock, Winzeler and Diagana2010) inhibiting a Na⁺-efflux pump in Plasmodium (Spillman et al. Reference Spillman, Allen, McNamara, Yeung, Winzeler, Diagana and Kirk2013). In addition, there is a link between anti-cancer chemotherapeutics and anti-parasitic activities, since many anti-cancer drugs, which target mechanisms that lead to increased cellular proliferation, also affect the proliferative stages of parasites (Klinkert and Heussler, Reference Klinkert and Heussler2006). For instance, exoerythorcytic and erythrocytic development of Plasmodium parasites is blocked by the proteasome inhibitor MLN-273 (Lindenthal et al. Reference Lindenthal, Weich, Chia, Heussler and Klinkert2005). Moreover, artemisinin and derivatives, which are active against N. caninum and T. gondii, also impact the proliferation and viability of many cancer cells (Das, Reference Das2015). Another example is given by organometallic ruthenium compounds originally developed for the treatment of cancer, and also active against N. caninum and T. gondii tachyzoites in vitro in the low nanomolar range (Barna et al. Reference Barna, Debache, Vock, Küster and Hemphill2013). Ruthenium drugs were originally thought to bind to DNA, but more recent investigations showed that they also interact strongly with proteins (Ravera et al. Reference Ravera, Baracco, Cassino, Zanello and Osella2004; Scolaro et al. Reference Scolaro, Chaplin, Hartinger, Bergamo, Cocchietto, Keppler, Sava and Dyson2007) and potential targets in cancer cells were postulated, including cathepsin B, P-glycoprotein, and glutathione S-transferase P1 (Casini et al. Reference Casini, Gabbiani, Sorrentino, Rigobello, Bindoli, Geldbach, Marrone, Re, Hartinger, Dyson and Messori2008).

CALCIUM DEPENDENT KINASE 1 (CDPK1) AS A PROMISING DRUG TARGET

Amongst novel compounds, inhibitors of CDPKI, deserve particular interest. CDPK1 is essential for microneme secretion, host cell invasion, and egress of T. gondii (Lourido et al. Reference Lourido, Shuman, Zhang, Shokat, Hui and Sibley2010). A particular class of inhibitors, the bumped kinase inhibitors (BKIs), has bulky C3 aryl moieties entering a hydrophobic pocket in the ATP binding site. BKIs selectively inhibit CDPK1 from apicomplexans in a good structure-activity-relationship (Keyloun et al. Reference Keyloun, Reid, Choi, Song, Fox, Hillesland, Zhang, Vidadala, Merritt, Lau, Maly, Fan, Barrett, WC and Ojo2014; Zhang et al. Reference Zhang, Ojo, Vidadala, Huang, Geiger, Scheele, Choi, Reid, Keyloun, Rivas, Siddaramaiah, Comess, Robinson, Merta, Kifle, Hol, Parsons, Merritt, Maly, Verlinde, Van Voorhis and Fan2014) but do not inhibit mammalian kinases because they have larger amino acid residues adjacent to the hydrophobic pocket, thereby blocking the entry of the bulky C3 aryl group. Some BKIs, especially BKI-1294 (Ojo et al. Reference Ojo, Reid, Kallur Siddaramaiah, Müller, Winzer, Zhang, Keyloun, Vidadala, Merritt, Hol, Maly, Fan, Van Voorhis and Hemphill2014), have a good efficacy against N. caninum in vitro and in vivo (see Table 3). In vitro studies, however, indicate that the BKI 1294 is not directly parasiticidal. Only upon long-term in vitro treatment of infected human foreskin fibroblasts monolayers, a complete clearance of viable tachyzoites can be observed (Ojo et al. Reference Ojo, Reid, Kallur Siddaramaiah, Müller, Winzer, Zhang, Keyloun, Vidadala, Merritt, Hol, Maly, Fan, Van Voorhis and Hemphill2014). For different Neospora isolates, but also for T. gondii strain RH and ME49, clearance of intracellular parasites is preceded by the formation of large, multinucleated complexes with deregulated gene expression as evidenced by the expression of bradyzoite as well as tachyzoite antigens (Winzer et al. Reference Winzer, Müller, Aguado-Martínez, Rahman, Balmer, Ortega-Mora, Ojo, Fan, Maly, Van Voorhis and Hemphill2015). The upregulation of bradyzoite antigen expression, as exemplified by the heat shock protein hsp20 (or BAG1) has also been reported upon treatment of N. caninum infected monolayers with artemisone and derivatives (Müller et al. Reference Müller, Balmer, Winzer, Rahman, Manser, Haynes and Hemphill2015b ). In the pregnant mouse model, BKI-1294 was the only compound tested so far that achieved a good protection against vertical transmission of N. caninum (Winzer et al. Reference Winzer, Müller, Aguado-Martínez, Rahman, Balmer, Ortega-Mora, Ojo, Fan, Maly, Van Voorhis and Hemphill2015).

A COMBINED IMMUNO-CHEMICAL STRATEGY AGAINST NEOSPOROSIS?

During the last decade, numerous in vitro and in vivo trials have yielded several promising vaccine and drug candidates that could be potentially used for the prevention or treatment of neosporosis. None of them achieved complete protection against transplacental transmission of N. caninum, the goal that should ultimately be achieved in cattle. In addition, it is very unlikely that a vaccine can be developed that protects against endogenous transplacental transmission from a chronically infected dam to its offspring (Williams and Trees, Reference Williams and Trees2006). Nevertheless, the promising results of both approaches suggest that they could be combined for immuno-chemotherapy.

One possible strategy could include application of a live-attenuated vaccine in combination with a compound that has exhibited high efficacy in previous in vitro and in vivo studies. Such an approach has been applied for a long time against East coast fever, in which cattle are vaccinated with live sporozoites together with a drug that affects parasite viability such as tetracycline, or alternatively buparvaquone, the only, but highly effective, drug against Theileria available (Irvin and Morrison, Reference Irvin, Morrison and Wright1989; Mutugi et al. Reference Mutugi, Young, Maritim, Linyonyi, Mbogo and Leitch1988; Brown, Reference Brown1990). BKIs (Ojo et al. Reference Ojo, Reid, Kallur Siddaramaiah, Müller, Winzer, Zhang, Keyloun, Vidadala, Merritt, Hol, Maly, Fan, Van Voorhis and Hemphill2014; Winzer et al. Reference Winzer, Müller, Aguado-Martínez, Rahman, Balmer, Ortega-Mora, Ojo, Fan, Maly, Van Voorhis and Hemphill2015) would constitute a suitable class of chemotherapeutics for co-application. They exhibit little or no side-effects and prevent the infection of cells by tachyzoites. The fact that they do not directly kill the parasites, or take an extended period of time until they act parasiticidally, renders them suitable for a combined immunization plus treatment protocol. As mentioned above, intracellular parasites treated with these compounds express a variety of tachyzoite and bradyzoite antigens. By turn-over of the infected cell, these antigens may be presented to the immune system thereby eliciting stable immune responses against tachyzoite as well as bradyzoite stages.

Another approach could include the simultaneous application of suitable chemotherapeutics and polypeptides such as recombinant antigens acting as classical vaccines or immunostimulants. Recombinant proteins produced in E. coli or in any other organism may contain impurities and are very expensive, especially when produced in high purity at a large scale. On the other hand, the chemosynthesis of peptides has become increasingly cheaper. Highly antigenic peptides could thus be produced by chemosynthesis, coupled to a high molecular weight carrier to render them immunogenic and/or to a ligand recognized by a toll-like receptor (Casal et al. Reference Casal, Langeveld, Cortes, Schaaper, van Dijk, Vela, Kamstrup and Meloen1995; Xin et al. Reference Xin, Cartmell, Bailey, Dziadek, Bundle and Cutler2012).

Taken together, the results achieved in the last years suggest that the ultimate aim of a one-shot-therapy against neosporosis in cattle will be difficult to achieve, but still is a realistic goal. Similar to Toxoplasma, Neospora has developed important means of adaptation and protection by differentiating into bradyzoites and forming tissue cysts. These can only be tackled by parasite-specific compounds that cross the blood brain barrier, and they could inactivate reactivated parasites early during recrudescence. A major problem, however, is that we do not know the exact timing of recrudescence during pregnancy. More insight is also needed into the immune mechanisms that are required to combat infection, but also how these can be maintained during pregnancy, without affecting the viability of the fetus. Potentially, a combined immune-chemotherapeutical approach could provide a solution. More in vitro as well as in vivo research is required, and ideally researchers join forces to apply appropriate and standardized models, that allow accurate comparisons of results achieved in different laboratories. In addition, it is important to ensure that any standardized model is truly appropriate otherwise standardized testing methods are not predictive of protection in the field.