A sensitive ELISA technique for the diagnosis of Anaplasma marginale infections

Veterinary Parasitology, 29 (1988) 1-7 Elsevier Science Publishers B.V., Amsterdam - - Printed in The Netherlands

1

A S e n s i t i v e E L I S A T e c h n i q u e for the D i a g n o s i s of Anaplasma marginale Infections A. DUZGUN 1, C.A. SCHUNTNER 2, I.G. WRIGHT 2, G. LEATCH 2 and D.J. WALTISBUHL2

lLalahan Nuclear Research Centre in Animal Health, Lalahan-Ankara (Turkey) 2CSIRO, Division of Tropical Animal Science, Long Pocket Laboratories, Private Bag No. 3, P.O., Indooroopilly, Qld. 4068 (Australia) (Accepted for publication 27 August 1987)

ABSTRACT Duzgun, A., Schuntner, C.A., Wright, I.G., Leatch, G. and Waltisbuhl, D.J., 1988. A sensitive ELISA technique for the diagnosis ofAnaplasma marginale infections Vet. Parasitol., 29: 1-7. A sensitive enzyme-linked immunosorbent assay (ELISA) technique using a horse radish peroxidase conjugate is described for measuring Anaplasma marginale antibodies in bovine serum. This technique utilizes two antigen preparations, a 'negative' antigen derived from an animal prior to infection and a 'positive' antigen derived from A. marginale-infected red cells from the same animal following infection. This markedly reduces cross-reactions which are a result of isoantigens. Absorbance values obtained using the 'negative' antigen are subtracted from those obtained using the 'positive' antigen to give a net figure. Of 100 A. marginale-positive sera tested no false negative results were obtained. All 11 animals maintained tick-free after initial diagnosis of naturally transmitted anaplasmosis were positive 3 years later, 15 A. marginale-infected animals maintained with ticks were positive 27 months after initial infection and a further 26 animals infected with A. marginale by blood inoculation were positive 3 months later. Three per cent of negative sera, 2% of B. bovis and 4% of B. bigeminapositive sera gave positive reactions.

INTRODUCTION

The card test (CT) and the complement fixation (CF) tests have been developed for the diagnosis of anaplasmosis, but both suffer from lack of sensitivity and/or specificity and do not diagnose long-term infections (Kuttler, 1975). Although both tests are relatively specific in the U.S.A. where A. marginale is the only haemotropic disease of cattle (Kuttler, 1975) such is not the case in countries where other protozoa, and in particular Babesia spp., are prevalent ( Callow, 1984). In these countries lack of test specificity is a real problem. Three papers describing an enzyme-linked immunosorbent assay (ELISA) to detect A. marginale antibodies have been published. In the first test (Thoen 0304-4017/88/$03.50

© 1988 Elsevier Science Publishers B.V.

et al., 1980) crude antigen at a dilution of 1 in 40 was used, with a horse radish peroxidase ( H R P ) -Protein A conjugate (prepared by the authors) as a detection system. The major drawback of this test was that at neutral pH, Protein A does not react with either bovine IgM or IgG1 (Goudswaard et al., 1978), which are the immunoglobulin isotypes that have been shown to be sero-reactive in A. marginale infections (Murphy et al., 1966). This test lacked sensitivity because it could not differentiate between positive and negative sera at dilutions greater than 1 in 80. In the second and third tests either a rabbit antibovine IgG-HRP or rabbit anti-bovine IgG alkaline phosphatase conjugate was used; they were also very insensitive, being used routinely at serum dilutions of 1 in 100 (Long and Wagner, 1981; Barry et al., 1986). This communication reports the development of an ELISA test which is based on a radioimmunoassay (RIA) technique developed in this laboratory for the detection of A. marginale antibodies (Schuntner and Leatch, 1987 ). MATERIALSAND METHODS

Test sera A. marginale positive One hundred animals known to be infected with A. marginale were tested. Of these, 15 were naturally infected and were maintained in a Boophilus microplus-infested paddock from birth (B. microplus is the tick vector of A. marginale in Australia). Infection of these animals had been previously demonstrated by monthly CA and indirect fluorescent antibody (IFA) tests and by microscopic examination of thin blood films stained with Giemsa. Sera were collected 6 and 27 months post-infection. Infection was further confirmed at 27 months by subinoculation. Serum samples were also obtained from an additional 11 naturally infected animals immediately after diagnosis of Ariaplasma infection (by subinoculation into susceptible splenectomized calves), and 1, 2 and 3 years later. These animals were maintained under tick-free conditions after initial diagnosis. Twenty-six animals infected with A. marginale by blood inoculation were also tested between 9 days and 3 months postinfection.

Negative One hundred sera were collected from cattle which were at least second generation in an area over 1000 km south of the anaplasmosis, babesiosis and theileriosis endemic area, a well-delineated area in Australia (Rogers et al., 1978). These sera had been shown to be Babesia-free by use of ELISA techniques ( Waltisbuhl et al., 1987; I.G. Wright, unpublished results ).

B. bovis positive One hundred sera were collected from cattle experimentally infected by blood inoculation with a number of B. bovis strains. Severe clinical reactions were controlled by a non-sterilizing dose of the babesiacide Diampron (May & Baker, U K ) and the animals were maintained under tick-free conditions for 4 years. Blood samples were collected monthly. These animals were described by Mahoney et al. (1979).

B. bigemina positive Fifty sera were collected from animals experimentally infected with B. bigemina by blood inoculation and maintained under tick-free conditions for a further 6 months. All sera were stored at - 2 0 ° C

after the addition of thiomersal to 0.1%

(w/v). Antigen 'Negative' antigen Five hundred millilitres of blood was collected into Na2EDTA from a splenectomized 3-month-old calf, purchased from an Anaplasma-free area and shown by E L I S A and IFA to be negative for A. marginale antibodies. The cells were washed three times with phosphate buffered saline ( P B S ) and centrifuged at 5000 g for 10 min at 5 ° C. After the final centrifugation, white blood cells were removed by diluting cells ten times in P B S and passing them through a W h a t m a n C F l l cellulose column, using a technique described previously ( Richards and Williams, 1973 ). Following further centrifugation, one volume of cells was diluted in two volumes of P B S and stored as 2.5-ml aliquots in the vapour phase of liquid nitrogen. As required, an aliquot was thawed at room temperature then sonicated for 30 s at 100 W using a Labsonic 1510 (B. Braun) with a small probe at 0 ° C. The whole sonicate was then diluted as required for use in the ELISA.

'Positive' antigen The calf previously used to produce 'negative' antigen was infected with A. marginale by blood inoculation. W h e n the parasitaemia had reached 80-85%, 500-1000 ml of blood was collected, processed and stored as described above. W h e n required, aliquots were sonicated as before and diluted appropriately.

Normal red cell sonicate (Nor S) Pooled blood from at least six negative cattle was processed as above and stored in 10-ml aliquots at - 70 ° C. As required this material was sonicated as described above.

4

Second antibody Commercially available goat anti-bovine IgG ( H & L ) conjugated with horse radish peroxidase ( H R P ) (Kirkegaard and Perry, Maryland, U.S.A. ) was used.

Absorbing medium This consisted of 100 vol. of 0.05 M sodium phosphate buffer (pH 9.5 ) ( PB ) containing 0.1% Tween 20 to which was added 4 vol. Nor S.

Test procedure Flat-bottomed 96-well microELISA plates (Dynatech M129B Denkerdark, West G e r m a n y ) were used exclusively. Standard antigen checkerboard titrations were performed with each new batch of antigen, both 'positive' and 'negative', to determine the working range which was usually 1 in 400 (150-200 #g protein m l - 1 estimated by the method of Bradford, 1976). Two hundred microlitres of diluted antigen were added to each well of a microELISA plate, one plate for 'positive' antigen, a second for 'negative' antigen. The plates were sealed with plate sealers (Dynatech, West Germany) and incubated at 4 °C overnight. After incubation the antigen solution was removed and the plates washed three times in PB containing 0.1% Tween 20 ( P B T ) , and then blocked with a 1% solution of normal rabbit serum in PB for 60 min at 22°C. After blocking, the plates were washed three times in P B T prior to addition of 200 pl of absorbed test sera. Test sera were absorbed prior to use by diluting to 1 in 800 in the absorbing medium and incubated for 30 min at 37 ° C or overnight at 5 ° C. After addition of absorbed test sera to both plates, the plates were sealed with plate sealers and incubated for 5 h at 22 ° C. The wells were then washed three times with P B T and 200 ]A of conjugated second antibody, at a dilution of 1 in 400 in PBS containing 1% horse serum, was added. The plates were t h e n sealed and incubated overnight at 4 ° C. After incubation the conjugate was removed and the wells washed three times in PBS prior to addition of 200 #1 of freshly prepared substrate. Substrate was prepared as follows: recrystallized 5-amino salicylic acid ( 5-AS ) (Sigma) was dissolved at 1 mg m l - 1 in phosphate buffer pH 6.8 at 37°C; 2 ~l of 30% hydrogen peroxide was then added to each millilitre of solution. After addition of substrate, the plates were sealed and gently agitated for 30 min at 22 ° C. After 30 min the reaction was stopped by the addition of 50 ttl 1 M sodium carbonate per well. The plate was t h e n immediately read at 492 n m with a Titertek Multiskan (Flow Laboratories). Column one of each plate was always used as a blank.

Analysis of data Net readings were calculated for each sample by subtracting the 'negative' antigen result from the 'positive' antigen result. A group of 20 negative sera and a standard positive A. marginale serum were routinely tested with each batch of assays. A 'threshold' was calculated as the mean net absorbance reading plus two standard deviations of the 20 reference negative sera. Ratios of all other net absorbances were determined against the 'threshold' as unity. The positive serum was used to determine t h a t each assay batch was performing normally. RESULTS Sera from 100 animals infected with A. marginale were all positive. The mean ratio was 2.43 (s.d. + 1.16) with a range of 1.15-6.86. Of 100 known negative sera three gave positive results. The m e a n ratio of the whole group was 0.52 ( s.d. _+0.25 ). The three false positive sera had ratios of 1.2, 1.33 and 1.48. Two sera from B. bovis-infected animals gave positive results. The mean ratio of the B. bovis sera was 0.39 ( s.d. _+0.40). The two false positives had ratios of 2.2 and 3.23, respectively. Two sera from B. bigemina-infected animals gave positive results. The mean ratio of all B. bigemina-positive sera was 0.46 (s.d. +0.27); the two false positive sera had ratios of 1.22 and 1.35, respectively. All 11 A. marginale sera tested 3 years after a single infection were positive, whilst all 15 animals maintained in the B. microplus-infected paddock had positive ratios 27 months after initial infection. Sera from 26 animals infected with A. marginale by blood inoculation were all positive 3 months postinoculation. DISCUSSION The results obtained using the two antigens with either the ELISA or its antecedent RIA method ( S c h u n t n e r and Leatch, 1987) show a higher degree of both specificity and sensitivity t h a n all extant tests previously described. Furthermore, these appear to be the only reports of the detection of long-term single infections. This lack of specificity is especially evident from the work of Barry et al. (1986) who reported that sera from 14.6% of cattle inoculated twice with B. bovis-infected erythrocytes were positive in their A. marginale ELISA test. These cross-reactions were considered to be the result of antibodies produced against red-cell antigens in the inoculum. A similar result would have been demonstrated in the test described in this paper if the two antigen system had not been adopted; only 2% of animals inoculated with B. bovisinfected erythrocytes and 4% of animals inoculated with B. bigemina-infected erythrocytes cross-reacted, albeit weakly in this current test. Approximately

37% would have given false positive results if a 'negative' antigen had not been used (data not shown). The reason for this high degree of specificity is due to the use of two antigen preparations and to the pre-absorption of sera with a range of normal red cell antigens. This largely overcomes the red cell isoantigen cross-reactions which have been noted in a number of bovine haemotropic diseases. In particular, Goodger and co-workers have observed such phenomena in B. boris infections and have also reported non-specific reactions with sera from animals undergoing non-specific inflammation (Goodger et al., 1985). It is also of interest t h a t both laboratory and field infections were detected effectively by this test, for it has been noted t h a t under Australian and Colombian conditions at least, whilst existing tests (complement-fixation, capillary tube agglutination, card agglutination and indirect fluorescent antibody) were accurate for diagnosing laboratory infections, there was a significant disagreem e n t between t h e m when field sera were tested (Rogers, 1971; Gonzalez et al., 1978). In both of these countries concurrent infection with other haemoprotozoa is usual and these may have contributed to the discrepancies observed. Whilst our assay is more time-consuming t h a n existing tests, its advantages of high sensitivity and specificity, and its ability to detect long-term chronic infections far outweigh t h a t disadvantage. Until such time as defined antigens can be readily obtained for use in highly specific tests, this concurrent usage of 'negative' and 'positive' antigens cannot be avoided. This test should now be utilized in countries such as Colombia, in concert with existing tests to see whether current problems of incorrect diagnosis of field sera can be overcome. ACKNOWLEDGEMENTS The authors wish to t h a n k the International Atomic Energy Agency, Vienna for financial support for this work and for providing a training fellowship for Mr. A. Duzgun.

REFERENCES Barry, D.N., Parker, R.J., De Vos,A.J., Dunster, P. and Rodwell,B.J., 1986. A microplateenzymelinked immunosorbentassay for measuringantibody to Anaplasma marginale in cattle serum. Aust. Vet. J., 63: 76-79. Bradford, M.M., 1976. A rapid and sensitive method for the quantitation of microgramquantities of protein utilizing the principle of protein dye binding. Anal. Biochem.,72: 248-254. Callow,L.L., 1984.AnimalHealth in Australia.Vol.5. AustralianGovernmentPublishingService, Canberra, pp. 181-199. Gonzalez, F.F., Long, R.F. and Todorovic, R.A., 1978. Comparisonsof the complement-fixation, indirect fluorescentantibody, and card agglutinationtests for the diagnosisof bovine anaplasmosis. Am.J. Vet. Res., 39: 1538-1541. Goodger, B.V., Wright, I.G. and Waltisbuhi, D.J., 1985. Babesia bovis: the effect of acute inflam-

mation and isoantibody production in the detection of babesial antigens. Experientia, 41: 15771579. Goudswaard, J., Van der Donk, J.A., Noordzij, A., Van Dam, R.H. and Vaerman, J.-P., 1978. Protein A reactivity of various mammalian immunoglobulins. Scand. J. Immunol., 8: 21-28. Kuttler, K.L., 1975. Diagnosis of anaplasmosis: a review. In: E.A. Wells (Editor), Workshop on Hemoparasites (Anaplasmosis and Babesiosis), 17-22 March 1975, Cali. Colombia. Centro Internacional de Agricultura Tropical, 1977, 93-103. Long, R.F. and Wagner, G.G., 1981. Antigen binding buffers in the enzyme-linked immunospecific assay (ELISA) for Anaplasma marginale antibody activity in bovine serum. Proc. 7th Natl. Anaplasmosis Conf., Mississippi State University, Starkville, pp. 321-331. Mahoney, D.F., Wright, I.G. and Goodger, B.V., 1979. Immunity in cattle to Babesia boris after single infections with parasites of various origin. Aust. Vet. J., 55: 10-12. Murphy, F.A., Osebold, J.W. and Aalund, 0., 1966. Kinetics of the antibody response to Anaplasma marginale infection. J. Infect. Dis., 116: 99-111. Richards, W.H.G. and Williams, S.G., 1973. The removal of leucocytes from malaria infected blood. Ann. Trop. Med. Parasitol., 67: 249-250. Rogers, R.J., 1971. Bovine anaplasmosis: an evaluation of the complement-fixation and capillary tube-agglutination tests and the incidence of antibodies in northern Queensland cattle herds. Aust. Vet. J., 47: 364-369. Rogers, R.J., Blight, G.W. and Knott, S.G., 1978. A study of the epidemiology of Anaplasma marginale infections of cattle in southern Queensland: clinical disease and the prevalence of complement fixing antibodies. Aust. Vet. J., 54: 115-120. Schuntner, C.A. and Leatch, G., 1987. A radioimmunoassay for Anaplasma marginale antibodies in cattle. Am. J. Vet. Res., in press. Thoen, C.O., Blackburn, B., Mills, K., Lomme, J. and Hopkins, M.P., 1980. Enzyme-linked immunosorbent assay for detecting antibodies in cattle in a herd in which anaplasmosis was diagnosed. J. Clin. Microbiol., 11: 499-502. Waltisbuhl, D.J., Goodger, B.V., Wright, I.G., Commins, M.A. and Mahoney, D.F., 1987. An enzyme linked immunosorbent assay to diagnose Babesia boris infection in cattle. Parasitol. Res., 73: 126-131.