key: cord-0008477-rz3ftx18
authors: Gottstein, B.; Hentrich, B.; Wyss, R.; Thür, B.; Busato, A.; Stärk, K.D.C.; Müller, N.
title: Molecular and immunodiagnostic investigations on bovine neosporosis in Switzerland
date: 1998-07-26
journal: Int J Parasitol
DOI: 10.1016/s0020-7519(98)00006-x
sha: 6b2f17cad0d24aaaa28ca6399127a8faeb002b51
doc_id: 8477
cord_uid: rz3ftx18

Neospora caninum has gained considerable attention through its role in the aetiology of bovine abortion. Due to its close phylogenetic relationship with Toxoplasma gondii, respective unequivocal differential diagnosis deserves special consideration. In order to evaluate the diagnostic performance of molecular and immunodiagnostic techniques and to provide insights into the epidemiological significance of bovine neosporosis in Switzerland, we conducted a study on 83 cases of bovine abortion: of these, 24 (29%) foetal brains were positive by Neospora-PCR, six of these foetuses were simultaneously seropositive in Neospora-IFAT and/or somatic antigen-ELISA. Conversely, four (5%) foetal brains were considered positive by Toxoplasma-PCR, two of which were also seropositive in the Toxoplasma-P30-ELISA and/or direct agglutination test. The seroprevalence in 1689 cattle sera obtained from 113 dairy farms was 11.5% (95% confidence interval: 9.2–13.8) by Neospora-somatic antigen-ELISA and 10.7% (95% confidence interval: 8.3–12.6) by Toxoplasma-P30-ELISA. From the same samples, 1.1%, less than statistically expected, were positive in both ELISA. Within selected groups of cow-calf farms, the seroprevalence determined using the Neospora-somatic antigen-ELISA was 14% (95% confidence interval: 5.0–23.0) for dams and 15% (95% confidence interval: 3.0–28.0) for offspring calves. Seroprevalences determined by Toxoplasma-P30-ELISA were 8% (95% confidence interval: 4.0–12.0) for dams and 3% (95% confidence interval: 0.3–6.0) for calves. None of the sera gave a positive reaction in both ELISA. Our data indicated that prenatal neosporosis appears as an important cause of bovine abortion in Switzerland.

mals, and is a significant causative awnit of abortion and stillbirth in cattle [l] and, occa&onaUy, in other Neospora caninum is an apicomplexan parasite mammals of veterinary importance (reviewed by causing neuromuscular diseases in various host ani- [2] ). The parasite is morphologkaff y S&.&W to Toxoplasma gondii, but distinct in its ultra&rWure, immunogenicity and host-related p nicity f3]. [4, 51. In contrast to N. caninum, T. gondii causes prenatal infection and consecutive abortion or foetal malformations in non-immune sheep and humans [6] . Both diseases occur world-wide [l] ; however, there is a paucity of information available on the epidemiological situation of neosporosis in Switzerland. Neosporosis was described as an important cause of economic and reproductive loss to the livestock industry [7, 81 . Outbreaks of confirmed Neospora-related abortions as well as the prevalence of neosporosis-associated abortions reported from several countries (reviewed by [2] ) justify an epidemiological pilot study in Switzerland. For an accurate determination of the prevalence of infection and disease, reliable techniques for the direct or indirect demonstration of the two parasites, of respective specific host immune reactions and of the damage induced by infection, are required.

Conventional diagnosis of neosporosis in cattle has been based largely on the post mortem demonstration of the organisms in clinical material, e.g., by histology or immunohistology performed on tissue sections of the brain and other organs. Histopathological examination may reveal minimal histologic lesions, such as focal gliosis in the CNS, [9] as well as severe neural lesions, such as nonsuppurative encephalomyelitis characterised by multifocal non-suppurative infiltration, with or without multifocal necrosis and multifocal to diffuse non-suppurative leukocytic infiltration of the meninges [2] , and occasionally calcifications. Diagnosis may also be achieved by the cultivation of the parasites in experimental animals or cell cultures. Moreover, demonstration of parasite antigens in tissue sections, e.g., by means of immunohistochemistry [lo] , is a valuable adjunct diagnostic approach. An indirect method refers to the assessment of the host's humoral immune response by the demonstration of parasite-specific antibodies in serum or other body fluids, such as by means of the IFAT [9,1 l] or the ELISA [12, 13] . These tests deserve special attention since they are required to discriminate between infection with different apicomplexan parasites [14] . Immunodiagnostic tools have been used to determine epidemic parameters associated with Neospora-induced abortion in cows and to study prenatal and postnatal transmission [I 51. With respect to the molecular identification of Neospora, a substantial improvement in the diagnostic sensitivity was achieved by designing oligonucleotide primers suitable for the PCR amplification of N. caninum-specific genomic sequences isolated from clinical samples [16-l 91 . Subsequently, the PCR established in our laboratory [18, 191 was developed further to achieve routine practicability by the introduction of the uracil DNA glycosidase (UDG) system [20] and a rapid and simple amplification-detection system known as DNA-Hybridisation Immunoassay (DIA) [21] .

In the present study, we have compared the use of diagnostic methods for the demonstration of infection and disease occurrence of N. caninum and T. gondii in a series of abortions in Swiss dairy cattle. Methodologically, we employed two highly sensitive PCR assays and indirect serological tools. Based upon the evaluation of these primary diagnostic parameters, the relative seroprevalence was determined in Swiss dairy cattle and in animals of cow-calf farms.

Experimental design. (1) Different diagnostic tools were established to demonstrate infection with Neospora and/or Toxoplasma in tissues obtained from clinical cases of bovine abortion. These tools were (a) in-vitro isolation of parasites from tissues, (b) histological demonstration of parasites or parasiteinduced lesions in tissues, (c) molecular identification of parasite DNA in tissues by PCR, and (d) detection of host immune responses to infection by IFAT, ELISA and direct agglutination (DA).

(2) Serological tools suitable for mass screening were used to determine the seroprevalence of Neospora and of Toxoplasma by ELISA in Swiss dairy cattle and in animals of cow-calf farms.

Parasites. An original isolate of N. caninum (NC-1 strain [3] ) was obtained from Dr J.P. Dubey, USDA, Beltsville, MD, U.S.A. Parasites were grown in Vero cells as described previously [22, 191. All reagents for cultivation were obtained from Gibco-BRL. Tachyzoites were separated from host cells by passing the trypsinised cells through a 23gauge needle, followed by purification on Sephadex G-25M (PD-IO@ columns, Pharmacia). Purified parasites were stored as pellets at -80°C until required.

Diagnostic in-vitro isolation of parasites. For invitro isolation of parasites from host tissue (foetal brain, heart, maternal placenta), 20ml of homogenised tissue of each organ was incubated with 50ml trypsin at 37°C for 1 h. The samples were subsequently washed several times with sterile PBS (pH 7.2) by centrifugation at 1200g for 10min at 4°C until the supematant was translucent. After discarding the supematant, 10 ml of each pellet was incubated on a 24h-old Vero cell monolayer in a 25cm3 tissue culture flask in 10 ml RPM1 1640 containing 10% FCS and I % amphotericin B. The cultures were incubated at 37°C and the medium replaced 12-16 h after inoculation. The cell cultures were monitored microscopically, and the cultivation medium was exchanged daily. Fourteen days after inoculation, the cultures were considered to be "negative" for Neospora growth unless parasites were detected microscopically [23].

PCR. Further processing of diagnostic tissue samples from brain, musculature and maternal placenta for PCR were carried out as described previously 1211, with the exception that a High Pure PCR Template Preparation Kit (Boehringer, Mannheim) was used according to the manufacturer's recommendations for the preparation of DNA. Molecular diagnosis by Neospora-or Toxoplasma-specific PCR and identification of amplification products were also according to our previous description [2 1 J. Putative false-negative results, which may have been caused by inhibitory compounds in the PCR tests, were excluded by performing a parallel inhibition control reaction (data not shown) on the samples in the presence of N, caninum or T. gondii DNA equivalents from about five parasites per reaction (exact performance data in [2 11 ).

Histological und immuno-histological examinations,

One portion of each tissue sample obtained at nec-ropsy was fixed in 4% formalin ti pm for routine paraffin embedding. P~~~~ sections of 3-4 pm thickness were rnouu$ed on poly-L-lysin precoated glass shdes, depar&Wed in xylene and stained with H gi E or for immunohistology, slides were rehydrated in descending concentrations of methanol, rinsed in H@, and non-specific binding sites were biooked with PBS containing 0.03% Tween and 3% BSA (Fluka AG) (designated PBS-Tween-BSA) for 1 h. Slides were then incubated with the respective rabbit hyperimmune sera [22] for 1 h at 37°C The anti-N. cuninum and the anti-T. go~i~~~~ rabbit sera, and the respective pre-sera (as controls) were diluted 1:200 in PBS-Tween-BSA, The slides were washed in P3S to reaulve unbound antibodies and subsequently incubated for 1 h at 37°C with FITC-conjugated goat+anti-rabbit IgG (The Binding Site, Birmingham, U.K.) at a dilution of 1:200 in PBS-Tween-BSA.

The preparations were washed three times for Smin in PBS, mounted in Fluoprep (BioMerieux Suisse S.A.) and finally examined using a Leitz Laborlux S fluoreseonce microscope. The specificity of the two hyperimmune sera was initially confirmed by the absence of immunoreactivity on heterologous T. gondii or N. caninum tachyzoites, respectively 1221.

Serological assays. For the NeosprPra-IFAT, Neospora tachyzoites were obtained from cell cultures as described above and subsequently coated onto 12-spot IFAT glass shdes and air dried. Zmmediately after drying, the slides were fix& in acetone and stored at -20X For the purposes of the investigation, diagnostic bovine sera were dih&d 1:&l in sterile PBS with subsequent two-fold titration. Positive and negative control sera were eaeb diluted at 1:40 and 1:80. The positive control serum was obtained from a pregnant cow e~~~ntally infected with live N. caninum taehyzo&es (Her&rich et al., unpublished) according to the procedure described by others [9}. InfeGtion in this cow was proven by PCR-based detection of N~w+w-DNA in the foetal brain using the same technique as described previously for mice 1211. The mgative control serum was obtained from the same cow (first gestation) prior to experimental iztfation. The absence of antibodies against Neospora and Toxo-pZusma in the negative control sera was confirmed using the same specific tests as employed subsequent to infection. The diluted test and control sera were incubated on the slides for 30 min at 37°C in a moist chamber. The slides were subsequently washed three times with PBS, and the last washing step was followed by a rinse in distilled water. The slides were then incubated for 30min at 37°C in a moist chamber with a FITC-conjugated monoclonal mouse-anti-bovine IgG antibody (Sigma Immunochemicals, Clone BG-18) at a dilution of 1:300 in PBS. After washing the slides twice in PBS for 5min, they were mounted in FluoPrep (Bio-Merieux) and examined on a Leitz Laborlux S fluorescence microscope. In the current study, infection with N. caninum was identified when foetal antibody titres at a dilution of >, l/80 were detected. For cows, the antibody-binding activity against Neospora was determined to be "positive" at a serum dilution of 2 1: 160.

For ELISA, the following procedures were used. Neospora tachyzoites obtained and purified from in-vitro cultures as described above were washed in sterile PBS. The pellet was resuspended in PBS containing 0.01% NaN, and homogenised. Subsequent treatment by three freezing-thawing cycles (-50"/+ 37°C) and final processing by ultrasonication (3 x 20 s at 65 W) at 1°C yielded the extract solution. The solution was sedimented at 10 OOOg at 4°C for 30 min. The soluble supernatant was used as somatic ELISA-antigen (hereinafter referred to as Neospora-SA-ELISA) or stored at -80°C until use. Coating Dynatech polystyrene plates was done at a concentration of 5 pg protein per ml carbonate buffer (pH 9.6) at 4°C for 12 h. For the detection of antibodies against T. gondii, ELISA plates were coated in a similar manner with 1 ,ug protein per ml of an affinity purified P30 surface antigen (SR2B, Arville, France, cat. No. TXP30B), hereinafter referred to as Toxoplusma-P30-ELISA.

All protein concentrations were assessed by the Bio-Rad protein assay using bovine albumin as the standard. Neospora-and Toxoplasma-ELISA plates were subsequently processed as described for other ELISAs performed in our laboratories [24] . In brief, the washing, blocking and serum dilution solution was PBS containing 0.3% Tween 20 (PBS-Tween). The test and control sera were diluted 1: 100 in PBS-Tween. Serum incubations were for 2 h at 37°C. The second antibody was an alkaline phosphatase-conjugated rabbitanti-bovine antibody (Sigma Immmunochemicals, cat. No. A 0705) diluted at 1:500 in PBS-Tween. The conjugate was incubated for 1 h at 37°C. Antibody reactivity was detected by adding 4-nitrophenylphosphate at a concentration of 1 mgml-' 0.1 M-diethanolamine, pH 9.8. The reaction was stopped after 15 min with 3 N-NaOH, and absorbance values were read at A405nm using a Dynatech MR7000 reader coupled to a Macintosh Performa 6300 computer with Biocalc@ software (Dynatech).

Positive and negative control sera for the Neospora-SA-ELISA were the same as used for the IFAT. For the Toxoplasma-P30-ELISA, the same negative control serum was used as for the Neospora serology. The Toxoplasma-positive control serum was a gift from Dr David Buxton (Moredun Research Institute, Edinburgh, U.K.). The basic test parameters for the Neospora and Toxoplasma antigens were established by investigating 50 sera of animals (no history of abortion; anamnestic lack of clinical or epidemiological signs indicative for neosporosis or toxoplasmosis) negative in Neospora-IFAT and Toxoplasma-DA (see below) using the respective ELISA mean value+4 SD. as the cut-off value. Any value higher than this cut-off value was considered to be "positive"; lower values were considered as "negative". Threshold values discriminating between diagnostically positive and negative reactions were thus determined for both antigens listed above. Reproducibility of ELISA results was monitored by including a low reactive bovine control serum in triplicate, this in addition to the negative and positive standard sera, both also tested in triplicate.

In addition to the tests listed above, antibody concentration against T. gondii was determined in parallel by the commercially available agglutination test Toxo-Screen DA (BioMCrieux). Test performance was determined according to the manufacturer's instructions, which corresponds to the United States Department of Agriculture (USDA) procedure described elsewhere [14] . Tissue collection. Organs and blood samples from 83 naturally aborted foetuses (all originating from different farms in Switzerland) were collected during 1993-1994 and were examined pathologically at the Institute of Veterinary Pathology of the University of Zurich, using conventional macroscopic and histopathologic examination techniques. At the same institute, investigation for the presence of bovine viral diarrhoea virus (BVDV) by immunohistochemistry (LSAB: labelled Streptavidin-Biotin) was performed. Other conventional microbiological investigations representing the spectrum of micro-organisms recommended in the survey programmes of Swiss veterinary health offices (Rotavirus, IBR/IPV, Listeria, Coxiella, Brucella, Actinomyces) were done in routine diagnostic laboratories.

At necropsy, the following samples were ob++ned from the foetuses and processed separately for parasitological analyses: (i) a foetalheart blood or body-cavity fluid sample; (ii) the skull was opened under aseptic conditions and a sample of the cerebrum was transferred into a sterile 50-ml Falcon tube prefilled with 20ml sterile PBS containing 200 U ml-' of penicillin, 200 pg mll ' streptomycin and 0.5 (ug ml-' fungizone; (iii) heart tissue was processed as described above for brain tissue; (iv) placental tissue was obtained in 32 cases;

(v) blood samples were collected from 12 cows after abortion. Classical histopathological examinations of the organs, in particular the foetal brain and the maternal placenta, were performed at the Institute of Veterinary Pathology, University of Zurich. Precolostral-IgG determination was carried out by using a radioimmunodiffusion IgG-Determinationkit (RID-Kit;

The Binding Site), results being expressed in mg IgG per litre.

Tissue samples not processed for histological, immunological or molecular diagnosis were stored at -80°C.

Bovine sera. A total of 1689 bovine sera (adult cows) from 113 dairy farms collected for a different project [25] were subsequently investigated in the present study. The animals and farms had been selected as representative of the Swiss dairy cattle population in 1994. Baseline information on health and fertility problems was available for all farms. The samples were investigated for the presence of antibodies against Neospora sp. and T, gondii using the respective ELISA (see above). Sera from 85 dams and from 102 correspomiing o&p&g c&es were obtained from 36 different cow-&f farms. The age of the calves at the time of Wood-sample colle&ion was between 8 and 10 months [26] .

The association between serological Neospora and Toxoplasma results in dairy and cow-calf serum samples, and between seropositivity of dams and offspring calves, was assessed by the Fisher's Exact test (two-tailed) using EpiInfo v. 6.04 (CDC, Atlanta). Confidence intervals (CI) for clustered data were set at 95% and calculated as described elsewhere [27].

From the overall number of 83 abortion cases, brain material from 27 foe&sees and five corresponding placentas could be prosed for invitro isolation of Neospora sp. (and putatively T. gondii). For 56 cases, the in-vitro isolation could not be performed because either the biological samples were of too poor condition, the size of the sample was too small, or cultivation led to premature microbial contamination of cultures. The 27 samples investigated included two groups: (i) I5 cases negative for any diagnostic Neaswra or Toxoplasma test (serology and PCR) as listed below; (ii) 12 cases positive in Neospora tests as follows: four cases Neospora-seropositive and positive in Neospora-PCR (case Nos 3-6, see Table l) , and eight cases Neospora-seronegative but positive in Meospora-PCR (case Nos 7-14, see Table 1 ). Isolation was also attempted from the placental material obtained from two cases (Nos 2,4) sero-and PCRpositive in Neospora assays and from three cases (Nos 8, 10, 25) only Neospora-XR-ps&ke.

The diagnostic in-vitro isolation of the parasite f&led in all cases, including especially also those cases where the mother-cow had been serologi&!p "Neosporapositive" or where the Neospora-IYX had been positive with foetal brain material. 

and histopathological examinations All brain samples obtained from the 83 abortion cases were investigated by PCR for the detection of parasite-specific DNA. Neospora was detected in 24 (29%) cases and Toxoplasma in four (5%) cases. In one of these cases (No. 24), the simultaneous presence of both parasites was demonstrated ( Table  1) . Additional results of pathological findings are also included in Table 1 . Histological examination of the 24 Neospora-PCR-positive samples revealed that 18 samples exhibited pathologic findings compatible with neosporosis. For samples obtained from either the brain and/or the placenta, seven cases included mukifocal necrotising encephalitis, two cases hypopksia of the cerebellum, one case hydrocephalus internus and five cases multifocal necrotising placentitis or calcification within the placenta, among other features (for details see Table l ), which were not indicative of a protozoa1 infection in most cases. In most of the 24 Neospora-PCR-positive cases, no other infectious agents were found. Exceptions were two foetal BVDV cases and one case of Streptococcus detection in the placenta. From the three foetal brain samples which were only positive by Toxoplasma-PCR (but Neospora-PCR-negative), the histological examinations, performed prior to PCR investigations, had already provided pathologic findings indicative of a protozoal infection (random multifocal necrotic lesions in the brain, calcified microlesions in the placenta and conjunctivitis). The single case positive for both Neospora-and Toxophma-PCR showed also multifocal calcifications in the placenta, but no specific pathology in the foetal tissues could be found. The mean age of the Neospora-PCR-positive aborted foetuses was 6.8 months (95%CI: 6.1-7.5), with 18 cases (75%) being 4-7 months of age. Another parameter which significantly (P= 0.95) discriminated between the group of Neospora-PCR-positive cases and respective negative cases concerned the precolostral IgG. The mean concentration of pre-colostral IgG in 21 Neospora-PCR-positive cases suitable for respective analysis was 1195 mg ml-' versus 5 12 mg ml-' in 45 Neospora-PCR-negative cases. For the 56 cases which were negative by both Neospora-and Toxoplasma-PCRs, characteristic pathologic findings indicating neosporosis were considerably less frequent for cerebral involvement (subsequent data refkr to case Nos S83 in Table  1 ve cases, two were serologically positive for Tox them was also positive by Ne these two cases, the seroposi&e res&&r could not be associated with any charaote* pa@ologic finding (for details, see Table 1 ). The moan age of aborted foetuses negative by all Neospora and Toxoplasma tests (PCR and serology) was 8.1 months (95%CI: 7. 5-8.8) . From these 54 cases, 20 (37%) were 4-7 months of'age.

In 32 cases, a plaeer&al tissue able in parallel to the foe&s and subsequent PCR analysis. The comp&r&ve results between PCR perfcrrmed with foe@ .brain and placenta were the following 23 cases were nq@ive by both Neospora-and ToxotpW~KR in t&e two tissue samples; seven cases were &k positive only in the foetal b&n Toxoplasma-PCR) and negative i cental PCR investigations; one case was positive only by Neospora-PCR fo placenta and one case was tive only in the foetal bram corresponding placental PCR resentative PCR results with a abortion associated with Neozpzwa or Toxoplasma are shown in Fig. 1 Immunohistological examination of fixed foetal brain samples using anti-Neospora or anti-Toxoplasma hyper-immune sera provided results which did not allow a specific and unambiguous identification of parasites for the first 10 cases tested. Consequently, and due to the time-consuming aspects of the technological procedure, we decided that this approach was not practicable for largescale rapid routine diagnosis.

Serum from foetal heart blood or body cavity fluid samples from all 83 abortion cases was tested for the presence of antibodies against Neospora and Toxoplasma using IFAT, ELISA or DA. The results are shown in Table 1 . For Neospora, four out of seven IFAT-positive sera reacted positively by SA-ELISA (case Nos 4, 5, 6, 28) . One serum (case No. 28) proved double-positive by both Neospora and Toxoplasma serology, but it was negative by either of the two PCR tests. For Toxoplasma, one out of two DA-positive sera reacted positively by P30-ELISA. Both Toxoplasma-seropositive foetuses were PCR-negative (for both parasites). In 12 abortion cases, it was possible to determine the serological parameters in association with a maternal serum sample drawn on the day of abortion. With respect to Neospora, seven out of 12 cases were negative for both dam and foetus using all immunodiagnostic and molecular tools. There was one single case with a seropositive dam and a foetus being also positive by Neospora serology (SA-ELISA and IFAT) and Neospora-PCR (case No. 4, Table 1 ). There were three cases where the dams were Neospora-seropositive (P30-ELISA and IFAT) and the foetuses Neosporu-PCR-positive, but dmdtaneously these offspring were negative by Neospora serology. Finally, there was one seropositive @A-ELISA and IFAT) dam with a seronegative foetus which had no Neospora DNA detectable by PCR. With respect to Toxoplasma, there were seven cases where all serological and molecular tests were negative. Then, there were five cases with the dams being Toxoplasma-seropositive (five by P30-ELISA, three by DA) but with the foetuses being negative by both Toxoplasma-PCR and Toxoplasma serology.

From 1689 dairy cattle sera investigated, 194 sera (prevalence= 11.5%; 95%CI 9.2-l 3.8) were positive by Neospora-SA-ELISA and 180 (preva-lence= 10.7%; were positive by ToxopEasnna-P30-ELISA (Table 2) . From those 194 Neospora-positive sera, 19 also contained antibodies reactive in the Toxoplasma-P30-ELISA; thus, the overall rate of double-positivity was 1.1%. The double-positivity rate with respect to reactive sera only (19 double-positives per 374 positives by Neospora and/or Toxoplasma serology) was 5.1%. There was no statistically significant association between seropositivity for Neospora and seropositivity for Toxoplasma (P=O.80). The seroprevalence concerning anti-Neospora antibodies in Swiss dairy cattle was plotted against the incidence Table 2 Serodiagnostic association between seropositivity in the Neospora-SA-ELBA and the Toxoplasma-P30-ELBA with regard to animals kept in conventional dairy cattle farms of (a) fertility problems, (b) return to oestrus problems, (G) cumulative incidence of abortions (results not shown) and (d) herd/farm size. Agab, no statistically significant correlation was observed.

Serological results of this group of GQWS and calves were as follows. For Neaspora: 10 out of 73 sera of dams and 12 out of 78 correfpoading offspring calves tested positive by Neospura-ELISA, resulting in a seroprevaknce of 14% (95%CI: 5.0-23.0) and of 15% (9S%CI: 3.0-28.0), respectively. For Toxoplasma: seven out of 85 dams and three out of 95 corresponding calves tested positive by Toxoplasma-PB-ELfSG resulting in a seroprevalence of 8% (95%CI: 4.0-12.0) and of 3% (95%CI: 3.0--&O), respectively. None of the Neospora-positive sera were potitive by Toxoplasma-P30-ELISA and vice versa; thus, w&in this animal population, specificity was 100%. The association between serological Neospora or Toxoplasma results was assessed by the Fisher's exact two-tail test. Hence, there was no statti&cally significant correlation between seropositivity against Neospora versus seropositivity against Toxctphma, neither at herd level, nor for dams or calves. A significant association, however, was found between dam and offspring calf seropositivity at the individually paired level.

The present study was designed to assess the performance of diagnostic tools for the identification of Neospora infection and, comparatively, Toxoplasma infection in 83 aborted foetuses, and to determine preliminarily the potential occurrence and thus the importance of Neospora in bovines of Switzerland.

By using PCR, we observed an unexpectedly high percentage (29%) of foetuses in which Measpora DNA was detected in the dam@ brain tissue. This percentage appeared to be high ins&ar as the other infectious organisms searched for within these foetuses occurred at only a very low rate. Unexpectedly, Toxoplasma DNA was also detected, but in only a few brains (5%). According to conventional knowledge, T. gondii is principally not considered to be a cause of abortion in bovines [28] . Therefore, our respective results need to be interpreted with caution. Our findings showed the presence of Toxoplasma DNA in the damaged foetal brain, which does not prove a causative correlation between infection and abortion by itself. Nevertheless, histopathologic findings unambiguously supported a protozoa1 infection and induced tissue damage in all three cases.

As the seroprevalence was similar for both parasite species in adult cattle, but the PCR findings were significantly higher for Neospora in foetuses, the comparison between seroprevalence and PCR findings consequently indicates a higher prenatal transmission rate for Neospora to the foetus than for Toxoplasma. The diaplacental transmission seems to be epidemiologically more important for Neospora, whereas postnatal infection seems to be more important for Toxoplasma. Beside the pathologic findings depicted in prenatal neosporosis, another parameter, that of elevated pre-colostral IgG, indirectly pointed at the pathologic significance of a Neospora infection in the foetus.

In contrast to PCR-based diagnosis, demonstration of the parasites by in-vitro culturing, or demonstration of parasite antigens by immunohistochemistry, were unsuitable as diagnostic tools. Although the two latter techniques may be considered as gold-standard methods to diagnose neosporosis, they both lack methodical and diagnostic sensitivity, as already discussed by several authors [2, 91. This lack of technical sensitivity may also be one explanation why, in previous attempts, demonstration of T. gondii in bovine foetuses has failed in many cases [28] . The unavailability of sensitive techniques may have considerably hampered previous studies on the reliability of prevalence assessment [ 151. In these studies, conventional histopathology criteria, beside immunohistochemistry, were also included to select positive cases of bovine neosporosis related to abortion. In a preliminary study (not included in the present study) based on experimental infection of pregnant cows, we observed a considerable difference between the technical sensitivity of DNA detection by PCR and aetiological diagnosis by in-vitro cultivation or immunohistochemistry.

However, more data are needed to significantly evaluate and elucidate the differences between PCR and immunohistochemistry. With regard to diagnostic in-vitro cultivation, we selected for our study an incubation period of 2 weeks to assess in-vitro parasite growth or growth absence from diagnostic brain samples. This period was selected based upon another report [23] and on local practicabilities in our routine laboratory process. Retrospectively, diagnostic experiences made by others [29] have shown that a minimal cultivation time of 8 weeks would have provided a much higher potential to obtain in-vitro growth of parasites, a fact which will have to be considered in future studies.

The application of different threshold dilutions to discriminate between positive and negative serological reactions in IFAT for foetuses, calves and cows has been discussed by several authors [30, 31] . Although applying two different cut-off titres in our IFAT for foetuses and calves/cows, the results obtained indicated that this would not have been necessary in that antibody titres of all animals were either at dilutions < l/80 or 2 l/l 60. For ELISA, it had been shown previously that the use of different cut-off absorbance values for foetal or adult cattle sera may lead to false-positive results [31] . Due to the lack of appropriate experiences, we had decided not to apply different threshold ELISA values indicative for N. caninum infection in foetuses. The full characterisation of operating characteristics of our ELISA still requires further parametric investigations with special emphasis on the problem in question, and we will have to consider the determination of different threshold values for different groups of animals in future investigations. Furthermore, new developments in the purification or biotechnologic synthesis of Neospora antigens with improved immunodiagnostic potential will have to be considered in future studies [12, 13, 321 . Our thus relatively preliminary serological results have shown that only 25% of the foetuses with a positive cerebral Neospora-PCR had detectable antibodies against Neospora at the time of abortion. It has been shown in previous work that the seroprevalence in foetuses increased with gestational age [30] . The same authors had found anti-N. caninum antibodies in 37 of 74 foetuses with Parasitology 28 (1998) 679491 689 suspected or confirmed neosporosis-associated abortion; 31 of these 37 seropositive foetuses were 6months of gestational age or older. In our study, all except one Neospora-seropositive foetus were 7-9 months of gestational age or older, whereas the larger part of seronegative foetuses was significantly younger. Thus the failure to find anti-N. caninum antibodies indicated that most of our foetuses may not yet have reached the appropriate humoral immunocompetence during the acute phase of infection until abortion. In addition, we have studied in 12 cases the association between the serological results of the foetus and the corresponding dam in parallel. Only one of the foetuses was seropositive, whereas five of the dams were seropositive. From the seronegative foetuses, four were positive by Neospora-PCR performed on brain tissue. Although the data are based upon only a few cases, we could thus obtain an indication that prenatal neosporosis seems to be related predominantly to seropositive dams, but that most of the foetuses were not able to seroconvert during infection prior to abortion. This information indicates that serological screening in foetuses is of relatively little diagnostic value, whereas serological pre-screening in dams may provide some indication for a potential prenatal transmission of the parasite. However, more field data are needed in order to provide clear statistical evidence for these observations.

With regard to the time of abortion with respect to the duration of gestation, our findings fitted a presumed expectation based upon other publications (see [2] ). The mean age (Mmonths) of aborted foetuses positive by Neospora-PCR was significantly lower than the mean age (8.1 months) of aborted foetuses negative by any Neospora and Toxoplasma test. Abortion occurred in 18 (75%) of 24 cases positive by Neospora-PCR between the 4th and 7th month of gestation, whereas for the cases negative for any Neospora and Toxoplasma test, only 20 (37%) of 54 cases occurred in the same period. As already discussed by other authors [2, 151, Neospora-induced abortion predisposes for the approximate period of gestation as mentioned above. Later on, infection may not necessarily lead to abortion, but may result in still-birth or weak calves. Neospora-associated abortion does not imply that the parasite has a&m&y ca damage primarily responsible for Nevertheless, in those cases where cental material was tested, the showed that there was no was detectable only in the foetus. However, seven cases were positive in foetal brain only one case was sim positive for both the foeta Among the 83 foetuses tested, BVDV was demonstrated by immunohistology in two of 24 Neospora-PCR-positive cases and in four of 59 Necrsgora-PCR-negative cases. Thus, 80 s&&is&al association became evident betweenJ!$VD~ and any diagnostic findings relating to N~OS~IXCX or Toxoplasma. With regard to the ser vestigaticms, there was an interesting co of iYeq.wra results when comparing dairy c&tle and cow-calf populations. In both groups, the Newyw-a seroprevalence of adult cows was very similar (12% and 14%, respectively) and mately equivalent level (15%) in cow-calf farms. For Tox prevalence in adult cattle was not' different between dairy cattle and co lations (11% and 8%, respectively), but clearly lower (3%) for offspring calves. It may be that diaplacental transmission the epidemiology of neosporosis, thd: akwst identical seroprevalence in dams .ti cakves. Ilk3 the other hand, peroral infection with T, Oocys~ may be the predominant mode of ion for cattle. This hypothesis would be ~~~ by the lower seroprevalence in young anbals.

Technically, there is still some need for further elaboration of the exact o~r~ti~.~~~e~~~s of the ELBA systems compared with the PAT for neosporosis and DA for For foe&l Neospora serology, the I mare sensitive (six IFAT-positive foetuseq of which only three were positive by SA-ELISA). Ad%& cow sera positive by IFAT were always positive by SA-ELISA, although the number of animals tested did not allow an accurate statistical analysis. Therefore, the seroprevalence data obtained from the two cattle populations should be interpreted with caution. By using a somatic Neospora-tachyzoite antigen similar to that employed in our study, an ELISA sensitivity of 88.6% and specificity of 96.5% had been estimated earlier [13] . When compared with the IFAT, the ELISA appeared more sensitive and specific for serodiagnosis of Neospora infection [ 131. We believe that new technical approaches are needed to improve the diagnostic parameters and especially standardisation of serological tools, such as proposed by others [32] . It has been shown that recombinant antigen ELISAs are capable of distinguishing between sera from Neospora-infected cows and sera from non-infected control cows. As the recombinant antigens in question show no evidence of cross-reactivity with sera from animals inoculated with T. gondii or some Sarcocystis species, this kind of approach may be suitable for routine application.

Our results indicate that PCR may be a valuable diagnostic alternative with a superior performance to other tests such as in vitro isolation or immunohistology. Furthermore, we showed that PCR combined with serology does accurately discriminate between Neospora and Toxoplasma infection. Finally, the association between specific pathologic findings in the foetal brain, the demonstration of Neospora DNA at the same site, and the absence of other aetiological abortive microorganisms in most cases not only indicate foetal infection, but also support the hypothesis of a causative role of the parasitic infection for abortion in Switzerland. The high rate of Neospora-PCRpositivity within abortion events in Swiss cattle justifies, therefore, further studies on the molecular epidemiology and pathogenesis of neosporosis in Switzerland.

A review of Neospora caninum and neosporosis

Newly recognized fatal protozoan disease of dogs

Neospora-like protozoan infection as a major cause of abortion in California dairy cattle

Congenital Neospora caninum infection in dairy cattle and associated calfhood mortality

Hehninths, arthropods and protozoa of domesticated animals

The phylogeny of Neospora caninum

Investigation of pregnancy losses in beef cattle herds associated with Neospora sp. infection

Detection of serum antibody responses in cattle with natural or experimental Neospora infections

Detection of Neospora caninum in tissue sections using a murine monoclonal antibody

Prevalence of antibodies to Neospora caninum in a population of urban dogs in England

Neospora caninum in dogs: detecton of antibodies by ELISA using an iscom antigen

An enzyme-linked immunosorbent assay (ELISA) for serological diagnosis of Neospora sp. infection in cattle

Serologic responses of cattle and other animals infected with Neospora caninum

Herd-based diagnosis of Neospora caninum-induced endemic and epidemic abortion in cows and evidence for congenital and postnatal transmission

Neospora parasites by PCR amplification and specific smallsubunit rRNA sequence probe hybridization

Detection of Neospora caninum DNA by the polymerase chain reaction

Discrimiaatiom of Neospora caninum from Toxoplasma gondii and other apicomplexan parasites by hybridisation and PCR

Neospora caninum: specific oligonucleotide primers for the detection of brain "cyst" DNA of experimentally-infected nude mice by the polymerase chain reaction (PCR)

The use of uracil DNA glycosylase to control carry-over contamination in polymerase chain reaction

Diagnosis of Neospora sp. and Toxoplasma gondii by PCR and DNA-Hybridization Immuno-Assay (DIA)

Adhesion and invasion of bovine eadothelial cells by Neospora caninum

In vitro isolation and characterization of a Neospora sp. from aborted bovine foetuses

II& t und Kosten voa Gesundheitsprob bei Schweizer Milchktihea und derea Kaber (1993-I 994)

Htiufigkeiten uad Ursachea von Kllberverlusten tad K~lberkrankheitea in Mutterkuhbetrieben. III. Seropr&valenz ausgewilhlter Krankheiten und pr'dvaenz voa Endoparasiten heim Absetzen

Sampling of populations, methods and applications

Characterization of a Swedish bovine isolate of Neospora eaninum

Diagnosis of bovine fetal Neospora infection with an indirect fluorescent antibody test

Diagnostic interpretation of Neospora serology

Evaluation of two Neospora caninum recombinant antigens for use in an enzymelinked immunosorbent assay for the diagnosis of bovine neosporosis

Acknowledgements-We thank Dr H. Kaufmann and 0. Flechtner for their technical and logistic support as well as for fruitful discussions, and are indebted to Dr A. Hemphill for stimulating discussions and critical comments on the manuscript. Many thanks to Dr J.P. Dubey (United States Department of Agriculture, Beltsville, U.S.A.), and Dr D. Buxton (Moredun Research Institute, Edinburgh, U.K.) for providing parasites and reagents and to Dr Bernadette Connolly for critical reading of the manuscript. This work has been supported by a research grant from the "Bundesamt fur Veterinarwesen" (grant No. 012.4.93.4) and by the "Bundesamt fur Bildung und Wissenschaft" (COST-820 project BBW C96.0068).