Committee of ethics on animal experimentation

This study is in accordance with the Ethical Principles of Animal Experimentation and was approved by the Commission of Ethics in the Use of Animals of the Universidade Federal do Paraná-Palotina Sector, with protocol number 50/2014.

Area of study and sampling calculation

The western region is the main dairy basin in the state of Paraná, with the cities of Marechal Cândido Rondon, Toledo, and Cascavel being the largest producers in the region (^{IBGE, 2016}), which is the reason this region was chosen for the seroprevalence study of N. caninum.

Sampling was performed using the EpiInfo program (version 7.2.0.1), where an expected prevalence of 35% was used based on other studies in the state of Paraná, such as the study by ^{Locatelli-Dittrich et al. (2008)}. An expected maximum error of 5% and 95% confidence interval and design effect (DEFF) of 1.5, which should be used when there is no random sample collection to reduce error, were used. This resulted in a minimum of 525 samples to be collected. However, 600 animals, which were distributed in four cities of the western region of Paraná (Cascavel, Marechal Cândido Rondon, Palotina, and Toledo), were included.

The number of rural cattle farms in each city was provided by the Agricultural Defense Agency of Paraná (ADAPAR), and the number of properties where samples will be collected was evaluated, with proportional distribution shown in Table 1.

The properties in each city were randomly selected, ranging from small- to large-scale production. It was stipulated that 10 animals would be sampled from each of the 60 properties.

Sample collection and application of the epidemiological questionnaire

Cow blood samples from caudal veins were collected using 40 X 1.2 mm needles and 10 mL syringes. After collection, blood samples were stored in properly labelled test tubes without anticoagulant and kept refrigerated until processing. Data on the age, race, and abortion history of each animal were also collected.

During the collection period, 17 bovine abortions occurred in nine properties: five in Palotina, two in Cascavel, one in Toledo and one in Marechal Cândido Rondon. Fragments of the heart, brain, and placenta of these fetuses were collected, packed in Microtubes, and stored at -20 ºC. After the collection of the fetuses, another visit was made to the farms to collect blood samples from the cows that had abortions, if they were not part of the 10 selected cows from the property. Thus, in addition to the 600 samples, eight blood samples were also collected.

Furthermore, samples from 163 dogs residing in 52 properties were collected (Table 1). Some properties did not allow collection of blood samples from dogs. Samples from the dogs were collected through the puncture of the cephalic vein or jugular vein using 20 X 0.55 mm needles and 5 mL syringes. After collection, blood was stored in labelled test tubes and kept under refrigeration until processing. During collection of blood samples, data from animals such as age, sex, and race were also collected.

In the laboratory, samples were centrifuged at 1500rpm for 10min for serum separation, after which they were stored in microtubes at -20°C until processing.

Data were collected using an epidemiological questionnaire for the 60 properties to evaluate the main risk factors associated with neosporosis. The data collected included presence of dogs on the properties, contact of the dogs with the cattle, size of the herd, reproductive and herd management, milk production, breed, age and history of cattle abortion, as well as data related to dogs, such as feeding, race, age, sex, whether they were released or trapped, use of vermifuge and vaccines.

Analysis of samples

For the serological analysis of the samples, the indirect fluorescent antibody test (IFAT) was performed for the bovine and canine samples (^{PARÉ et al., 1995}).

The IFAT slides were adsorbed with N. caninum strain Nc-1, from the Laboratory of Parasitology of the University of São Paulo (USP), fixed with methanol, and stored at -20°C.

Each serum sample was diluted with the dilution solution (0.0084 M Na2HPO4, 0.0018 M NaH2PO4, 0.146 M NaCl, and 1% BSA). Then, 20μL was added to each well of the immunofluorescence slide, and they were incubated for 30min at 37°C. After that, three washes were performed with a wash solution (0.0268M Na2CO3, 0.0975M NaHCO3, and 0.036M NaCl). Then, the slides were oven dried at 37°C and then 20μL of the anti-bovine and anti-canine IgG conjugate Sigma, incubated for 30 min at 37 °C, then three washes were performed using the stock solution. Then, the slide was oven dried at 37°C. Afterwards, a coverslip was placed with two drops of buffered glycerin (50%) and viewed under an epifluorescence microscope at a magnification of 400X (^{PARÉ et al., 1995}).

Samples with total fluorescence of tachyzoites from the 1:100 dilution for cattle and 1:50 for dogs were considered positive, while samples that only presented with apical reaction were considered negative (^{ACOSTA et al., 2016}; ^{CAMILLO et al., 2010}). Positive sera were then diluted in ratio two series to the final titre.

For the analysis of the fetuses, polymerase chain reaction (PCR) was performed. About 5 g of the collected organs (heart, brain, and placenta) were individually macerated; then, DNA extraction was performed using the commercial kit DNeasy Blood and Tissue (Qiagen). For PCR, the Nc5 region was selected as the target sequence for DNA amplification. The primers Np21/Np6 (5'-CCCAGTGCGTCCAATCCTGTA-3')/(5'-CTCGCCAGTCAACCTACGTCTTCT-3') (^{MÜLLER et al., 1996}) were used. The reaction was carried out with a final volume of 50μL, containing 10X buffer, 200μM DNTP, 1.5mM MgCl₂, 20μM of each primer, 1.25U Taq DNA Polymerase, with initial denaturation at 95ºC for 5 min, 40 cycles at 94ºC for1 min/63ºC for 1 min/74ºC for 3,5 min and final extension at 74ºC for 10 min. Amplified samples were subject electrophoresis in a 1.5% agarose gel for visualization. The positive controls used were tachyzoites of the Nc-Bahia strain, and negative controls were autoclaved ultrapure water.

For the sequencing, a positive sample from each property was selected and sent to ACTGene Análise Moleculares Ltda. (Biotechnology Center, UFRGS, Porto Alegre, RS). For this, samples were sequenced by ACTGene Análises Moleculares Ltda. (Centro de Biotecnologia, UFRGS, Porto Alegre, RS, Brazil) using the automatic sequencer AB 3500 Genetic Analyzer armed with 50 cm capillaries and POP7 polymer (Applied Biosystems). DNA templates were labeled with 2.5 pmol of the primer Np21/Np6 and 0.5μL of BigDye Terminator v3.1 Cycle Sequencing Standart (Applied Biosystems) in a final volume of 10μL. Labeling reactions were performed in a LGC XP Cycler termocycler with a initial denaturing step of 96ºC for 3 min followed by 25 cycles of 96ºC for 10 sec, 55 ºC for 5 sec and 60ºC for 4 min. Labeled samples were purified by isopropanol precipitation 75% followed by and 60% ethanol rinsing. Precipitated products were suspended in 10μL formamide Hi-Fi (Applied Biosystems), denatured at 95ºC for 5 min, ice-cooled for 5 min and electroinjected in the automatic sequencer. Sequencing data were collected using the software Data Collection 2 (Applied Biosystems) programmed with the following parameters: Dye Set “Z”; Mobility File “KB_3500_POP7_BDTv3.mob”; BioLIMS Project “3500_Project1”; Run Module 1 “FastSeq50_POP7_50cm_cfv_100”; e Analysis Module 1 “BC- 3500SR_Seq_FASTA.saz”.

Research of risk factors

The seroprevalence and variables related to epidemiological questionnaire data were analyzed in the Epi Info program (version 7.2.0.1) using the chi-square test and Odds Ratio (OR), with a 5% as a significant level.

Seroprevalence and risk factors

The IFAT-based seroprevalence of N. caninum in dairy cattle in the western region of the State of Paraná was 23.67% (142/600). Of the 60 properties included in this study, 80% (48) had at least one bovine seropositive to N. caninum. All the properties surveyed in the city of Palotina were positive, whereas the cities of Cascavel, Toledo, and Marechal Cândido Rondon presented positive results in 77% of properties (Table 2).

In the studies carried out in recent years in Brazil, the occurrence of N. caninum in cattle ranged from 10.9 to 50.74% depending on the state, method of diagnosis, cut-off point used, main risk factors, and number of samples analyzed can also influence the results of the research (^{TEIXEIRA et al., 2010}; ^{AGUIAR et al., 2011}; ^{AMARAL et al., 2012}; ^{BRUHN et al., 2013}).

The seroprevalence of N. caninum is rare in dairy cattle in Paraná. One of the reasons is the number of samples collected, which most of the time only allows the detection of occurrence and not of prevalence. The last study was by Locatelli-Dittrich et al. in 2008, who reported a seroprevalence of 33% upon analysis of 1263 sera samples from the state of Paraná by the ELISA method. ^{Rocha et al. (2015)} evaluated 367 animals in the region of Francisco Beltrão (Southwest of Paraná) and found 35.1% of the samples to be positive, a higher percentage than the one found in the west region, which can be justified by the cutoff point of 1:50.

^{Camillo et al. (2010)} analyzed samples from the southwest region of Paraná using the IFAT with 1:100 cutoff, and found a prevalence of 24.2%. The similar seroprevalence in the southwest and west regions can be explained by geographical proximity, as well as the constant purchase or sale of cattle between the two regions.

When the properties were analyzed, ^{Locatelli-Dittrich et al. (2008)} found that 77% of the properties had at least one seropositive animal in the state of Paraná, which is similar to the results of this study in the western region of Paraná. In the state of Rio Grande do Sul, ^{Corbellini et al. (2005)} found that 93.3% of 60 properties had at least one positive animal.

Cattle titers ranged from 100 to 12800, with most animals having titers of 800, followed by 1600 and 400. Only one animal had a titer of 6400 and another one had 12800. ^{Koiwai et al. (2006)} analyzed 2420 bovine sera using the IFAT for N. caninum and found more animals with a titer of 800, followed by 1600 and 400. However, other authors found a higher frequency of titers below 800 (^{KASHIWAZAKI et al., 2004}). Titers of IgG antibodies equal to or greater than 800 are generally associated with clinical signs of the disease in cattle (^{DUBEY, 1999}; ^{LINDSAY et al, 1999}). However, as verified by ^{Teixeira et al. (2010)}, not all animals with high titers had a history of abortion, although they might have given birth to congenitally infected normal calves.

High antibody levels are associated with the reactivation and reconversion of bradyzoite into a tachyzoite, which will reach the placenta and the fetus through the circulatory system. This phase usually occurs from the fourth to the fifth month of gestation but may occur throughout the pregnancy period. This results in miscarriages or birth of congenitally infected calves. After abortion, antibody levels tend to decrease (^{CONRAD et al., 1993}; ^{MCALLISTER, 2016}).

Of the 163 samples from dogs, 19 (11.66%) were positive for anti-N. caninum antibodies. When the properties were analyzed, 28.84% (15/52) had at least one seropositive dog. Palotina had a higher percentage than the other three cities (Table 2).

In some experiments, dogs that ingested animal tissue with N. caninum cysts did not seroconvert after eliminating oocysts in the feces (^{LINDSAY et al., 1999}; ^{DIJKSTRA et al., 2001}; ^{SCHARES et al., 2005}). In another study, which was conducted by ^{Barber & Trees (1998)}, dogs with low titers of antibodies against the protozoa became seronegative over time. Due to these factors, the actual protozoan seroprevalence in dogs may be higher than that found in this study.

Several authors have compared the seroprevalence of N. caninum in dogs from urban areas with those from rural areas, and this one is generally greater in the wandering dogs or domiciled in farms, fact that is related to the cycle of the parasite. As transmission to carnivores occurs through the consumption of raw meat or fetal remains with protozoan cysts, the chances of dogs becoming infected in rural areas are greater than that of dogs in urban areas (^{FERNANDES et al., 2004}; ^{LASRI et al., 2004}; ^{KING et al., 2012}; ^{NOGUEIRA et al., 2013}).

^{Locatelli-Dittrich et al. (2008)}, analyzed 129 canine samples using IFAT from 35 properties from the entire state of Paraná and obtained 25% positive results. This percentage is considered high compared to that found in higher tham the present study.

^{Acosta et al. (2016)} analyzed samples of 187 dogs from 30 properties in the state of Espirito Santo and found an occurrence of 11.76%, which is similar to the results of this study in the west of Paraná. ^{Nogueira et al. (2013)} evaluated the prevalence of protozoa in dogs from urban and rural areas in the State of Minas Gerais and found that 11.4% of the samples had positive results: 13.78% in dogs from rural areas and 8.24% in dogs from urban areas.

The titer of the dogs measured using IFAT ranged from 50 to 800, with a higher number of animals showing titers of 100 (47.36%), followed by 50 (21.05%) and 200 (15.79%). Only two animals showed titers of 400 (10.52%) and one animal had a titer of 800 (5.26%). However, ^{Nogueira et al. (2013)} found anti-N. caninum antibodies in dogs ranging from 50 to 6400, with a higher number of animals showing titers of 50. ^{Acosta et al. (2016)} found a greater number of dogs with titers between 800 and 1600.

The presence of N. caninum seropositive dogs was not considered a risk factor for seropositivity in cattle, as it was not statistically significant (Figure 1). Other studies have also evaluated the seropositivity of dogs as a risk factor for bovine neosporosis and did not find a significant difference (^{GUIMARÃES et al., 2004}; ^{AGUIAR et al., 2006}; ^{LOCATELLI-DITTRICH et al., 2008}; ^{BENETTI et al., 2009}). However, some authors reported a higher seroprevalence of N. caninum in cattle when they had contact with dogs, especially if the dogs were seropositive. Thus, the dogs could be involved in the transmission of neosporosis (^{PARÉ et al., 1998}; ^{WOUDA et al., 1999}; ^{SANCHEZ et al., 2003}).

Figure 1 Correlation of the results of the properties with seropositive dogs and cattle, properties with seropositive dogs, properties with seropositive cattle, and the total data collected from properties in the west of the state of Paraná, Brazil. 

The insignificant difference may be an indication that the protozoan is kept in the herd by vertical transmission, not necessarily needing a definitive host (^{PARÉ et al., 1997}; ^{MCALLISTER, 2016}). Another factor that must be considered is the possibility of the disease being transmitted by stray dogs or wild canids that can feed on fetal remains with the N. caninum bradyzoites and then disseminate the oocysts excreted in the feces through grass or water (^{BENETTI et al., 2009}, ^{MCALLISTER, 2016}). It can be related to seroconversion, since some dogs may not seroconvert or may not be seronegative, which would influence the results of occurrence of the protozoan in dogs and its correlation with seroprevalence in bovines (^{BARBER & TREES, 1998}; ^{LINDSAY et al., 1999}; ^{DIJKSTRA et al., 2001}; ^{SCHARES et al., 2005}).

The risk factors related to N. caninum infection in dairy cattle in the western region of the state of Paraná included the following. Milk production: properties that produce 100 to 200 liters are 1.96 times more likely to have N. caninum than other properties. Reproductive management: properties that use artificial insemination and bull are 2.02 times more likely to have the protozoan than properties that use other reproductive management, such as bull only, or the association of bull and artificial insemination at fixed times. The type of production: cattle raised in extensive production are 1.76 times more likely to have neosporosis. A history of abortion: animals with a history of abortion have about 3.81 times more chances of having N. caninum. The breed Jersey are 2.18 times more likely to have N. caninum than other breeds (Table 3).

The main risk factors are associated with the production e reproduction profile, usually family farms with low milk production, and extensive farms, where animals raised mainly in pasture have a higher chance of contact with the protozoan.

Although the Jersey breed in this study was identified as a risk factor for N. caninum infection, this result reflects the sampling, since blood samples were collected from a small number of animals of this breed. Only two properties had Jersey and most of those were seropositive.

Semi-intensive production and Dutch breed were found to be protective factors. These results may have been influenced by the number of samples collected that did not fit these conditions, such as Jersey breed and extensive production that are risk factors.

Among the seropositive cattle (59/142), 41% had a history of abortion, which is an important risk factor (p <0.001) and is the main clinical sign of neosporosis in cattle. No significant difference was found when this factor was correlated with the titers, although the majority of the animals with a history of abortion presented titers from 400 to 1600 (Table 4).

Other factors analyzed through the epidemiological questionnaire, such as presence of dogs on the property, contact with dogs, number of dogs on the farm, age of cattle, and replacement of cattle, among others, did not have significant differences.

The major risk factors for N. caninum infection in dogs found by other authors are as follows: age, older dogs are generally more likely to have the protozoan; feeding, dogs that consume food and/or raw meat are more likely to acquire the parasite and have contact with cattle, as they are the main intermediate hosts of N. caninum (^{NOGUEIRA et al., 2013}).

However, in the present study, dogs did not present risk factors associated with N. caninum. The main factors investigated were age of the animal, contact with cattle, feeding, sex, if the dog is arrested or released, vaccination, race, use of vermifuge, and history of diseases.

Aborted bovine fetus

Seventeen dairy cattle fetuses were collected from nine properties in the region studied: four from Cascavel, two from Toledo, one from Marechal Cândido Rondon, and ten from Palotina. Of these, 52.94% (9/17) were positive for N. caninum using the PCR technique and all of the samples showed a high similarity (99%) to the protozoan by DNA sequencing.

The percentage found is considered high compared to that reported by other studies. ^{Antoniassi et al. (2013)} performed histopathological analyses of 490 fetuses and found 32.6% to be positive for N. caninum. ^{Corbellini et al. (2000)} found 10% of the fetuses to be positive when they analyzed tissues of 30 fetuses using the immunohistochemical technique (IHC). ^{Orlando et al. (2013)}, after analyzing tissues of 60 fetuses through PCR and IHC, found 23.33% of the samples to be positive for the protozoan. Meanwhile, analyzing the brain, lung, liver, heart and intrathoracic fluid of 30 fetuses of dairy cattle from the State of Santa Catarina, ^{Macedo et al. (2017)} found 38.8% positives through the PCR technique, 25% through immunohistochemistry and 26.7% through of the ELISA.

The high positivity found in the present study may be due to the technique used, since PCR is a more sensitive and more specific technique than histopathology and IHC, especially in autolysis fetuses. Another factor that may have influenced the result is the number of fetuses collected. Additionally, two properties located in Palotina had many symptomatic animals that were seropositive for N. caninum. The animals had abortion frames, and these fetuses were collected.

We analyzed the brains and hearts of the 17 fetuses and the placenta of only three fetuses because of the difficulty in collecting placenta. The 17 fetuses were tested using PCR: nine were positive in the brain, five in the heart, and three in the placenta. All fetuses that showed positive results in the heart and placenta also presented positive results in the brain (Table 5).

The most affected animal tissue by N. caninum is the central nervous system (^{MCALLISTER, 2016}). In Brazil, there are four N. caninum isolates from bovine brain: BCN/PR3 (^{LOCATELLI-DITTRICH et al., 2003}), BCN/PR1 (^{LOCATELLI-DITTRICH et al., 2004}), Nc-Goias 1b (^{GARCÍA-MELO et al., 2009}) and Nc-SP1 (^{OLIVEIRA et al., 2017}). Thus, it is where the parasite is most commonly found through molecular techniques (^{SANTOS et al., 2011}; ^{ORLANDO et al. 2013}, ^{BROM et al., 2014}). The protozoan can also be found in the heart, as it is one of the muscular tissues of predilection of N. caninum. ^{Amaral et al. (2012)} found, using PCR, a higher positivity in the heart than in the brain in fetuses coming from slaughterhouses. Therefore, it is important to evaluate and to associate the results found in both organs to decrease the number of false negatives in the diagnosis.

The placenta is an important organ that should be evaluated when possible to study the causes of abortion. It is essential to emphasize the collection of adequate material, since it is in cotyledons that most of the pathogens responsible for abortion can be found (^{MCALLISTER, 2016}).

The gestational age with the highest occurrence (35.29%) of the abortions was eight months; however, when the age with the highest number of fetuses positive for N. caninum was analyzed, it was five months (75%), followed by four months (66.66%), seven months (50%), and eight months (33.33%). These results confirm those found by other authors. in addition to having a correlation with the gestation time in which there is a greater chance of reactivation and reconversion of the bradyzoite into a tachyzoite in the maternal organism, reaching the placenta and then the fetus (^{ORLANDO et al., 2013}; ^{MCALLISTER, 2016}).

When the fetuses were analyzed based on sex, most were females (9/14), but only 33.33% of them were positive for the protozoan studied. Of the males, 60% (3/5) were positive for N. caninum. It was not possible to identify the sex of three fetuses since many structures were degraded.

One animal had negative serology results for N. caninum, but the aborted fetus was found to be positive; this can be explained by the false negatives that the serology technique can generate. Another animal had positive serology results, but its aborted fetus was negative for the protozoan. This can be justified by the mummified condition of the fetus, which may have degraded the N. caninum DNA and generated a false negative result; it is also possible that there was a different cause for the abortion.