ELSEVIER Veterinary Parasitology 68 (! 997) 251-260

veterinary parasitology

An indirect enzyme-linked immunoassay (ELISA) for demonstration of antibodies to Neospora

caninum in serum and milk of cattle

Camilla BjSrkman a, *, O. Joakim M. Holmdahl b, Arvid Uggla b a Dept. Cattle and Sheep Diseases, Swedish University of Agricultural Sciences, Box 7019, S-750 07, Uppsala,

Sweden b Dept. Parasitology, National Veterinary Institute and Swedish University of Agricultural Sciences Box 7073,

S-750 07, Uppsala, Sweden

Received 23 April 1996; accepted 2 July 1996

Abstract

An indirect enzyme-linked immunosorbent assay (ELISA) for detection of antibodies to Neospora caninum in serum from cattle is described. Extracted tachyzoite proteins incorporated into immunostimulating complexes (iscoms) were used as coating antigen and a mouse mono- clonal antibody to bovine immunoglobulin G1 as conjugate. Western blot analysis of the iscom preparation revealed a restricted number of antigens compared with whole parasite homogenates. When probed with a serum from an experimentally infected calf, heavily stained antigens with apparent molecular masses of 28, 35, 45 and 78 kDa were seen. The sensitivity and specificity of the ELISA was 100% and 96%, respectively, against an indirect fluorescent antibody test as indicator of true status. The applicability of the ELISA for demonstration of antibodies in milk was evaluated and the agreement between serum and milk ELISA was 95%.

Keywords: Neospora caninum; Cattle-Protozoa; Antibodies; ELISA; lscom

I. Introduct ion

Neospora caninum is a coccidian parasite which causes abortion in cattle and feeble calves at birth (Dubey and Lindsay, 1993). The parasite has been found in various parts of the world and is considered to be one of the major causes of bovine abortion in

" Corresponding author. Tel: +46-18 671778; fax +46-18 673545; e-mail: Camilla.Bjorkman@idmed.slu.se

0304-4017/97/$17.00 Copyright © 1997 Elsevier Science B.V. All rights reserved. PH S0304-4017(96)01076-X

252 C. Bji~rkman et aL / Veterinary Parasitology 68 (1997) 251-260

several countries (Anderson et al., 1991; Thornton et al., 1991). The post mortem diagnosis of Neospora infection is established by demonstration of the parasite in tissue samples by specific immunohistological staining or by isolation in mice or cell culture (Dubey and Lindsay, 1993). In the live animal, presence of antibodies to the parasite can be used as an indication of Neospora infection. The indirect fluorescent antibody test (IFAT) was first introduced for detection of antibodies to Neospora in cattle serum (Conrad et al., 1993b). An enzyme-linked immunosorbent assay (ELISA) for detection of antibodies in serum from cattle using a crude parasite preparation has been described by Pare et al. (1995a). In an ELISA developed for demonstration of antibodies to N. caninum in serum from dogs (Bjbrkman et ai., 1994), N. caninum antigens were incorporated into immunostimulating complexes (iscoms) and a monoclonai antibody to dog IgG was employed as conjugate. This concept was used to avoid the problems with high background absorbances and cross-reactivity with antibodies to related parasites, which was experienced with a conventional ELISA employing a crude antigen prepara- tion and a polyclonal antiserum.

We describe here an iscom ELISA for demonstration of antibodies to N. caninum in serum and milk from cattle.

2. Materials and methods

2.1. Outline

The Neospora iscom antigen used in the ELISA was characterized by Western blot analysis. A cut-off absorbance for the ELISA was obtained by analysis of 200 IFAT negative cattle sera. The ELISA was evaluated by analysis of sera from 245 cattle with both the iscom ELISA and the IFAT, and comparison of the results from the two tests. Since cross-reaction with closely related parasites, especially Toxoplasma gondii, was considered as a potential problem, sera from animals experimentally infected with such parasites were analysed in the ELISA. All sera were also analysed for antibodies to T. gondii by a commercially available direct agglutination test (Toxoscreen, BioM~rieux, Charbonni~res-les-Bains, France).

To investigate the feasibility of the iscom ELISA for use with milk, samples from 30 seronegative cows were used to calculate a cut-off absorbance. Paired milk and serum samples from a further 85 cows were used to calculate the agreement between the results of the serum and milk ELISA.

2.2. Serum samples

Bovine blood was collected by jugular vein puncture into Vacutainer tubes without additives (Becton and Dickinson, Meylan-Cedex, France). The blood was centrifuged at 1000 X g for 10 min and the serum separated and stored at - 20°C until analysed.

The IFAT negative sera used to calculate a cut-off value for the serum ELISA were from 200 cattle submitted to the Department of Cattle and Sheep Diseases, Swedish University of Agricultural Sciences, Uppsala, Sweden. The cattle were of different age,

C. Bji~rkman et al. / Veterinary Parasitology 68 (1997) 251-260 253

breed and clinical condition, the majority suffering from various disorders or diseases. They were from dairy farms situated within 150 km of Uppsala. It was not known if the farms or the individual cows had any history of abortions.

The following sera were used to evaluate the ELISA: (1) 182 serum samples collected in four dairy herds which, during the last year, had experienced an increased rate of abortions (in two of these herds Neospora organisms had been demonstrated in aborted foetuses); (2) 28 sera from cows which had aborted in different dairy herds; (3) 27 sera from cows submitted to the Department of Cattle and Sheep Diseases; (4) sera from five cows which had aborted foetuses in which Neospora infection had been diagnosed by immunohistochemistry; (5) sequential sera collected once a week during 8 weeks from three calves inoculated with N. caninum Nc-1 isolate (Dubey et al., 1988) at 12-14 weeks of age. One of the calves (No. 1) was inoculated intravenously with 1 x 105 live tachyzoites. The two other calves (Nos. 2 and 3) were injected subcuta- neously with 1 X 108 killed tachyzoites disintegrated by ultrasound and mixed with Freund's complete adjuvant. Nineteen and 29 days later they were given booster doses with 1 × 108 killed tachyzoites with Freund's incomplete adjuvant; (6) sera from cattle experimentally infected with T. gondii (n = 3), Sarcocystis cruzi ( n - - 3 ) , Eimeria alabamensis (n = 3) and Babesia divergens (n = 3).

2.3. Milk samples

Milk was collected in plastic tubes without additives. Skim milk was obtained by centrifugation of the whole milk at 1000 × g for 20 min, and subsequent collection of the milk below the fat layer. The skim milk was stored at - 2 0 ° C until analysed by the ELISA.

Milk samples were collected together with blood samples from 115 dairy cows. Thirty milk samples from cows with no demonstrable antibodies in serum (as measured by iscom ELISA and IFAT) were used to calculate a cut-off level. Milk samples from the remaining 85 cows were used to calculate the agreement between the results obtained by the serum and milk ELISAs, respectively.

2.4. Antigen

Antigen for the IFAT and iscom ELISA, respectively, was prepared as previously described (Bj6rkman et al., 1994) but during the in vitro cultivation of the Nc-1 isolate, 5% normal horse serum instead of foetal calf serum was included in the medium.

2.5. Western blot analysis

Western blot analysis was performed under non-reducing conditions essentially as described by Bj~Srkman et al. (1994). Two batches of iscoms and a crude N. caninum lysate were probed with serum from calf No. 1, experimentally infected with live N. caninum tachyzoites, to evaluate the content of antigens. Iscoms were probed with sera that were positive (n = 3) or negative (n = 3) in both the ELISA and the IFAT, and sera which were positive in the ELISA and negative in the IFAT (n = 3). One of the sera that

254 C. Bji3rkman et a l . / Veterinary Parasitology 68 (1997) 251-260

were negative in the ELISA and IFAT was from a calf experimentally infected with T. gondii. Iscoms were also probed with milk from three cows which were positive in both the serum ELISA and the IFAT. All sera were diluted 1:10 in phosphate buffered saline, pH 7.2 with 0.05% Tween-80 (PBS-T) before analysis, and the milk samples were used undiluted. A mixture of mouse monoclonal antibodies to cattle IgG1 (SVANOVA, Uppsala, Sweden) and alkaline phosphatase conjugated rabbit antibodies to mouse Ig (Dakopatts, Copenhagen, Denmark) was used as conjugate.

2.6. IFAT

The IFAT was performed essentially as described by Trees et al. (1994). The sera were diluted in two-fold serial dilutions from 1:320 to 1:5120 in phosphate-buffered saline, pH 7.4 (PBS). Fluorescein isothiocyanate labelled rabbit antibodies to cow IgG (Sigma Immunochemicals, St. Louis, MO, USA) were used as secondary antibodies at a dilution of 1:200 in PBS. Sera from non-infected and infected animals were included as negative and positive controls in each series of analyses. A serum was regarded as negative if no fluorescence was detected at a dilution 1:640.

2.7. Iscom ELISA

The ELISA procedure was modified from Bj6rkman et al. (1994). The N. caninum iscom antigen was diluted to a final protein concentration of 3 /zg ml - i and 1 /xg ml - in coating buffer (0.05 M sodium carbonate buffer, pH 9.6, containing 0.1% sodium nitrate and 0.05% sodium azide) for the serum and milk ELISA, respectively. The sera were diluted 1:100 and the skim milk 1:2 in PBS-T. One hundred microlitres of diluted sample were added to each microplate well and the plate was incubated at 37°C. The incubation time was 60 min for sera and 120 min for milk. After three washes with PBS-T, 100 /xl horseradish peroxidase conjugated mouse monoclonal antibodies to cow IgG1 (SVANOVA, Uppsala, Sweden) diluted in PBS-T with 1% horse serum, were added to each well. After a further 1 h of incubation and subsequent washings, 200 /zl of 0.10 mg ml-~ 3,5,3',5'-tetramethylbenzidine and 0.006% H202 in 0.1 M acetate buffer, pH 5.0, were added, and the plate was subsequently incubated for 10 min at 20°C after which the reaction was stopped with 50 /zl 10% sulphuric acid. The absorbance was read at 450 nm in a Labsystem Multiscan Plus microplate reader (Labsystems AB, Stockholm, Sweden). Three IFAT negative and positive sera were included on each plate as controls. Before the assay was established it was optimised by standard checkerboard analysis using serial dilution of antigen, conjugate and IFAT negative and positive sera.

3. Results

3.1. Western blot analysis

When two batches of iscoms and a crude N. caninum lysate were probed with serum from an experimentally infected calf, several protein bands showed up both in the

C. Bjbrkman et al. / Veterinary Parasitology 68 (1997) 251-260 255

9 9 ~ 66 - -

4 5 - - 3 1 - -

2 2 - -

14 - -

I 2 3

Fig. I. Western blot analysis of N. caninum (3 /zg protein per lane) probed with serum from Calf I inoculated with live N. caninum tachyzoites. Lanes 1 and 2: two different batches of iscoms containing proteins from N. caninum tachyzoites. Lane 3: lysed N. caninum extract.

iscoms and in the crude preparation. Three groups of bands with molecular masses

28 -35 kDa, 45 -52 kDa and 64-78 kDa, respectively, were found in all three prepara-

tions. In the iscoms, no band with molecular mass above 78 kDa was demonstrated, and

the bands with apparent molecular masses of about 28 kDa, 35 kDa and 45 -52 kDa

were more heavily stained than in the crude N. caninum extract (Fig. 1).

Sera which were positive in both the ELISA and the IFAT recognized a major band

with apparent molecular mass of 35 kDa in the iscoms. Protein bands with apparent

molecular masses between 40 and 51 kDa, and 69 to 79 kDa were also seen. The sera

that were positive in the ELISA but negative in the IFAT reacted with two heavily

stained bands of about 35 and 45 kDa. The serum from a T. gondi i - in fected calf did not

recognize any antigen in the iscoms. The other sera that were negative in both the

ELISA and the IFAT did not react with any antigen or only weakly with one band of

about 40 kDa (Fig. 2).

When the iscoms were probed with samples from cows that were positive in both the

serum and milk ELISA, fewer and weaker bands were seen with milk than with serum

9 9 - - 6 6 - -

4 5 - -

3 1 - -

2 2 - -

1 4 ~

1 7 3 4 5 6 7 8 9 10 11

Fig. 2. Western blot analysis of proteins from N. caninum tachyzoites incorporated into iscoms (2.7 tzg protein per lane), probed with bovine sera diluted 1:10 and undiluted milk. Lanes 1-3: sera positive in both the ELISA and IFAT; the serum in Lane 1 was from a cow which had aborted a Neospora-infected foetus. Lanes 4-6: sera positive in the ELISA and negative in the IFAT. Lanes 7-9: sera negative in both the ELISA and IFAT; the serum in Lane 9 was from a calf experimentally infected with T. gondii. Lanes 10 and 11: milk from cows positive in the serum ELISA and IFAT, as well as in the milk ELISA.

256 C. Bj5rkman et a l . / Veterinary Parasitology 68 (1997) 251-260

ELISA absorbance

1.2

0.8

0,4

|

1 negative positive n=194 n=48

Fig. 3. lscom EL ISA absorbances in IFAT negative and ]FAT positive bovine sera. - - -

IFAT

• , CUt-off.

collected at the same time. However, bands of about 35 and 45 kDa were seen with all the positive milk samples, of which two are shown in Fig. 2.

3.2. Evaluation of the serum ELISA

The cut-off level, calculated as mean absorbance + 3 standard deviations for 200 sera which were negative in the IFAT, was 0.20 absorbance units. When 242 serum samples

ELISA a b s o r b a n c e

1.2-

; i o . 8 - //

// //

;7 h

o . . . / / ¢ . . . . ¢ . . . . . . . . . . . . .

o 1 o zo 30 40 5o 6'o A A A days a f te r inoculation

Fig. 4. i scom E L I S A absorbances for sera o f throe calves inoculated with N. caninum tachyzoites, Cal f l , - • - was inocula ted with 1 × l05 live tachyzoi tes on Day 0. Calves 2 ( - - - ) and 3 ( - o - ) were inoculated

with I × 108 killed tachyzoi tes on Days 0, 19 and 29. - - - , cut-off .

milk

1.2-

C, Bfiirkman et al. / Veterinary Parasitology 68 (1997) 251-260

ELISA absorbanee

257

0.8"

0 .4-

I B

serum ELISA result

negative positive

n=58 n=27

Fig. 5. Iscom ELISA absorbances in mi lk from 85 cows, as compared with serum collected at the same time.

- - - , cut-off.

were analysed in both the IFAT and the ELISA the agreement between the two tests was 96%. The sensitivity and specificity of the ELISA were 1130% and 96%, respectively, when compared with the IFAT (Fig. 3). The five cows which had aborted Neospora-in- fected foetuses were all positive in both the ELISA (0.53-1.10 absorbance units) and the: IFAT (titres 1:640-1:1280). Sera from seven cattle had antibodies to T. gondii and three of these were also positive in the iscom ELISA and Neospora IFAT.

The ELISA results from the three calves inoculated with live or killed N. caninum tachyzoites are shown in Fig. 4. The IFAT gave the same pattern as the ELISA, i.e. the IFAT titres were below 1:640 the first week after inoculation in calf No. 1, and during the first 3 weeks in calves Nos. 2 and 3. Thereafter, the IFAT titres increased to 1:20 480.

All sera from cattle experimentally infected with other protozoan parasites were negative in ELISA and IFAT.

3.3. Milk ELISA compared to serum ELISA

The cut-off level calculated on 30 milk samples collected from animals that were seronegative in both the IFAT and the ELISA, was 0.15 absorbance units. When milk and serum samples from another 85 cows were analysed in the serum and milk ELISAs the agreement between the two tests was 95% (Fig. 5).

4. Discussion

The ELISA is a technique that enables rapid determination of antibody levels in body fluids like serum and milk. When trying to develop an ELISA for detection of antibodies

258 C. BjiJrkman et al. / Veterinary Parasitology 68 (1997) 251-260

to Neospora in cattle serum using a crude tachyzoite antigen, we faced problems with high background absorbances and poor discrepancy between sera from animals experi- mentally infected with N. caninum and animals expected to be free of the parasite.

Iscoms are cage-like structures of about 40 nm composed of Quillaja saponin, cholesterol, phospholipids and antigen. Amphipathic proteins, i.e. membrane proteins, are readily incorporated into iscoms (Morein et al., 1987). In the present study iscoms were used as a tool to select membrane proteins from the parasites, thus decreasing the number of intracellular proteins that might cause problems with non-specific binding and cross-reactivity in an ELISA. Iscoms have earlier been used as antigen in ELISAs for detection of antibodies to T. gondii in sheep and cattle sera (LSvgren et al., 1987) and to N. caninum in sera from dogs (BjSrkman et al., 1994). It was then found that the specificity of the ELISAs increased when the antigen used consisted of parasite proteins incorporated into iscoms rather than crude tachyzoite extracts.

In the present study, the canine N. caninum isolate Nc-I (Dubey et al., 1988) was used to prepare the iscoms for the ELISA, as this was the only isolate available to us when this work started. It could be argued that a Neospora isolate of bovine origin would be preferable for analysis of bovine samples. There is, however, no convincing evidence that the Neospora parasites occurring in cattle and dogs should not be of the same species (Conrad et al., 1993a; Marsh et al., 1995). Furthermore, Park et al. (1995a) reported no major discrepancy in results when Neospora isolates of bovine and canine origin were used as antigen in an ELISA for bovine sera.

Western blot analysis revealed that the Neospora iscoms contained several antigens, but considerably fewer than the crude Neospora extract. Furthermore, the antigens with molecular masses of 31-36 kDa and 45-52 kDa appeared to be more concentrated in the iscoms. The reactivity with the different proteins varied between the sera, but a protein with an apparent molecular mass of 35 kDa was detected with the sera which were positive in both the ELISA and the IFAT. Earlier Western blot studies on N. caninum iscoms probed with sera from naturally infected dogs showed heavily stained bands with apparent molecular masses between 30 and 42 kDa (BjSrkman et al., 1994). Bjerk~s et al. (1994) found that N. caninum proteins with molecular masses between 17 and 37 kDa were recognised by sera from infected animals of several species. Therefore, we suggest that these antigen proteins should be candidates in the further development of serological diagnostic systems for N. caninum infections.

To evaluate the sensitivity and specificity of a serologic test, sera should be analysed from known infected and non-infected individuals. In the case of Neospora infection, this is not easily accomplished. The five cows that had aborted confirmed Neospora-in- fected foetuses were all positive in the iscom ELISA, and the three calves inoculated with N. caninum gave increasing absorbance values in the iscom ELISA. These results indicate that the sensitivity of the method was satisfactory. The problem was to define non-infected animals since cattle can harbour a chronic Neospora infection without showing any clinical signs (Bj~Srkman et al., 1996). Therefore, the further evaluation of the iscom ELISA was done by comparison with the IFAT, the most widely used test for detection of antibodies to N. caninum in serum. Using a conservative 1:640 cut-off in the IFAT and an absorbance of 0.20 as cut-off in the ELISA the specificity of the iscom ELISA compared with the IFAT was 96%, i.e. some sera that were negative in the IFAT

C. Bjfirkman et al. / Veterinary Parasitology 68 (1997) 251-260 259

were positive in the ELISA. Similar results were previously found with dog sera (BjSrkman et al., 1994). This means either that the iscom ELISA may give false positive results, that it recognises antibodies raised to other antigenic proteins than does the IFAT, or that the IFAT, with the chosen cut-off level, has low sensitivity. In the IFAT, whole tachyzoites are used as antigen. The technique of preparing iscoms selects for membrane proteins, but whether these originate from the surface membrane or from intracellular organelles depends on the preceding detergent extraction of the parasite. No studies were made to investigate the cellular localization of the proteins incorporated into the Nespora iscoms. However, studies by both Barta and Dubey (1992) and Bjerk~s et al. (1994) suggest that at least some antigenic proteins with molecular masses between 16 and 46 kDa present in N. caninum are of intracellular origin. Antibodies raised against these intracellular antigens should be detected in the iscom ELISA but presum- ably not with the IFAT. Pare et al. (1995b) highlighted the difficulties in interpretation of Neospora IFAT results. When analysing sera from cows and precolostral sera from their calves, they found that 24% of the cows that had given birth to seropositive calves, were themselves seronegative. This might be due to a lack of antibodies in the cow at the sampling occasion, but more probably it is the result of low sensitivity of the IFAT.

Since cross-reaction with antibodies to related parasites is a potential problem of serological assays, sera from animals experimentally infected with T. gondii, S. cruzi, E. alabamensis and B. divergens were tested in the iscom ELISA. There was no cross-reactivity with any of the sera, indicating satisfactory specificity in this respect.

Recently, Par6 et al. (1995a) described an ELISA based on whole sonicated Neospora tachyzoites as antigen. Their ELISA had high sensitivity (89%) and specificity (96.5%) to Neospora infection, which was comparable with results they obtained using an IFAT with a cut-off titre >_ 1:640. They did not encounter any problems with poor specificity, as we did in our preliminary studies using a crude N. caninum antigen.

When evaluating the feasibility of the ELISA for demonstration of antibodies to Neospora in milk, using samples from seronegative and seropositive cows, the agree- ment between the serum and milk ELISA was found to be 95%. This is very promising, since milk samples are easier to collect than blood samples and sampling can be done by the animal owner or attendant. Before using the milk ELISA for epidemiological surveys and clinical samples, however, further studies have to be performed to evaluate factors that might affect the ELISA results, such as stage of lactation and presence of subclinical mastitis.

Acknowledgements

We thank Dr. J.P. Dubey (Livestock and Poultry Sciences Institute, Beltsville, MD, USA) for kindly supplying the N. caninum Nc-1 isolate and Britt-Louise LjungstrSm for technical assistance. Financial support was obtained by Agria Insurances, and in part by the Swedish Council for Forestry and Agricultural Research and the Swedish Farmers! Foundation for Agricultural Research. The study was part of the EU research collabora- tion COST 820.

260 C. Bjbrkman et aL / Veterinary Parasitology 68 (1997) 251-260

References

Anderson, M.L., Blanchard, P.C., Barr, B.C., Dubey, J.P., Hoffman, R.L. and Conrad, P.A., 1991. Neospora- like protozoan infection as a major cause of abortion in California dairy cattle. J. Am. Vet. Med. Assoc., 198: 241-244.

Barta, J.R. and Dubey, J.P., 1992. Characterization of anti-Neospora caninum hyperimmune rabbit serum by Western blot analysis and immunoelectron microscopy. Parasitol. Res., 78: 689-694.

Bjerk~is, I., Jenkins, M.C. and Dubey, J.P., 1994. Identification and characterization of Neo~pora caninum tachyzoite antigens useful for diagnosis of neosporosis. Clin. Diagn. Lab. Immunol., 1: 214-221.

Bji~rkman, C., Lund6n, A., Holmdahl, J., Barber, J., Trees, A.J. and Uggla, A., 1994. Neo~pora caninum in dogs: detection of antibodies by ELISA using an iscom antigen. Parasite Immunol., 16: 643-648.

BjSrkman, C., Johansson, O., Stenlund, S., Holmdahl, J. and Uggla, A., 1996. Neo.wora species infection in a herd of dairy cattle. J. Am. Vet. Med. Assoc., 208: 1441-1444.

Conrad, P.A., Barr, B.C., Sverlow, K.W., Anderson, M., Daft, B., Kinde, H., Dubey, J.P., Munson, L. and Ardans, A., 1993a. In vitro isolation and characterization of a Neo.wora sp. from aborted bovine foetuses. Parasitology, 106: 239-249.

Conrad, P.A., Sverlow, K., Anderson, M., Rowe, J., BonDurant, R., Tuter, G., Breitmeyer, R., Palmer, C., Thurmond, M., Ardans, A., Dubey, J.P., Duhamel, G. and Barr, B., 1993b. Detection of serum antibody responses in cattle with natural or experimental Neospora infections. J. Vet. Diagn. Invest., 5: 572-578.

Dubey, J.P. and Lindsay, D.S., 1993. Neosporosis. Parasitol. Today, 9: 452-458. Dubey, J.P., Hattel, A.L., Lindsay, D.S. and Topper, M.J., 1988. Neonatal Neo,~pora caninum infection in

dogs: Isolation of the causative agent and experimental transmission. J. Am. Vet. Med. Assoc., 193: 1259-1263.

LSvgren, K., Uggla, A. and Morein, B., 1987. A new approach to the preparation of a Toxoplasma gondii membrane antigen for use in ELISA. J. Vet. Med. B., 34: 274-282.

Marsh, A.E., Barr, B.C., Sverlow, K., Ho, M., Dubey, J.P. and Conrad, P.A., 1995. Sequence analysis and comparison of ribosomal DNA from bovine Neospora to similar coecidial parasites. J. Parasitol., 81: 530-535.

Morein, B., LSvgren, K., H~Sglund, S. and Sundquist, B., 1987. The ISCOM: an immunostimulating complex. lmmunol. Today, 8: 333-338.

Par6, J., Hietala, S.K. and Thurmond, M.C., 1995a. An enzyme-linked immunosorbent assay (ELISA) for serological diagnosis of Neospora sp. infection in cattle. J. Vet. Diagn. Invest., 7: 352-359.

Par6, J., Hietala, S.K. and Thurmond, M.C., 1995b. Interpretation of an indirect fluorescent antibody test for diagnosis of Neospora sp. infection in cattle. J. Vet. Diagn. Invest., 7: 273-275.

Thornton, R.N., Thompson, E.J. and Dubey, J.P., 1991. Neospora abortion in New Zealand cattle. N.Z. Vet. J., 39: 129-133.

Trees, A.J., Guy, F., Low, J.C., Roberts, L., Buxton, D. and Dubey, J.P., 1994. Serological evidence implicating Neospora species as a cause of abortion in British cattle. Vet. Rec., 134: 405-407.