J. Vet. Med. B 37, 753-759 (1990) 0 1990 Paul Parey Scientific Publishers, Berlin and Hamburg ISSN 093 1 - 1793

Department of Bacteriology, Central Veterina y Institute, Lelystad

Enzyme Immunoassay Using Mouse Monoclonal Anti-bovine Antibodies for the Detection of Brucella abortus Antibodies

in Cow Milk

2. BERCOVICH and R. TAAIJKE

Address of authors: Central Veterinary Institute, Department of Bacteriology, P. 0. Box 65, 8200 AB Lelystad, The Netherlands

With 3 tables

(Received for publication April 18, 1990)

Summary Individual milk samples and artificially constructed tank milk samples from cows with naturally

occurring brucellosis were examined by the enzyme-linked immunosorbent assay (ELISA) using sonicated B. ubortus S-99 antigen, and mouse monoclonal anti-bovine IgM, IgA, and IgG, conjugates. ELISA results were compared with the results of the milk ring test using either 1 ml milk (MRT-1) o r 8 ml milk (MRT-X).

The ELISA using mouse monoclonal anti-bovine IgG, conjugate was sensitive and specific. In testing individual milk samples and constructed tank milk samples containing milk with low titers in the MRT-I the ELISA was superior to the MRT-1, and MRT-S. In testing other milk samples, the ELISA was as sensitive or slightly less sensitive than the MRT-B. From a total of 5,910 milk samples collected from cows free from brucellosis, only 24 (0.4 %) samples tested positive in the ELISA. All 500 tank milk samples collected from farms negative for brucellosis tested negative in the ELISA.

We concluded that the ELISA is a good substitute for the MRT-I to detect antibodies against Brucellu in milk from individual cows. When tank milk is tested for antibodies against Brucellu, however. both the MRT-8 and the ELISA should be used.

Introduction Although Brucella abortus has long been eradicated in cattle in the Netherlands,

infected cattle occasionally have been detected on farms that purchased imported cattle. The milk ring test, which uses 1 ml milk (MRT-I), is the principal test for surveillance

of brucellosis on dairy farms in many countries. Certain kinds of milk however, may react non-specifically in the MRT-I. Milk containing colostrum, milk examined at the end of the lactation period, milk from cows with hormonal disorders, and milk which fat globules d o not cluster (KEER et al., 1959; MCCAUGNY, 1972; BERCOVICH and MOERMAN, 1979; CORBEL et al., 1984; PATTERSON et al., 1976). Moreover, the MRT-1 is inadequate to detect antibodies against Brucellu in tank milk (HILL and CREMERS, 1967). Although the sensitivity of the MRT-1 was considerably improved after the introduction of the MRT that uses 8 ml milk (BERCOVICH and LAGENDIJK, 1978), 3 of the reactions are still non- specific.

U.S. C y y r i g h r C l c x ~ n c e Centcr Code Statement: 0931 - 1793/90/3710-0753$02.50/0

754 BERCOVICH and TAAIJKE

T h e enzyme-linked immunosorbent assay (ELISA) used to detect antibodies against Brucella in milk f rom individual cows is sensitive and specific, and may also be suitable for the detection of antibodies against Brucella in tank milk (THOEN et al., 1979; HECK et al., 1980; BORAKER, 1981; OLIVER and COOPER, 1981; FORSCHNER et al., 1989). FORSCHNER and BUNCER (1986) improved the sensitivity of the ELISA by concentrating milk serum gammaglobulins with ammonium sulfate and were able t o detect antibodies against Brucella in tank milk comprising milk samples f rom 50 cows or more. Routine use of this procedure o n a large scale is impractical.

In the assay reported here w e used mouse monoclonal anti-bovine IgM, IgA, and IgG, conjugates t o detect Brucella antibodies in milk collected f rom individual cows and in artificially constructed tank milk.

Material and Methods Milk and serum samples

Milk was collected from two cows with naturally occurring brucellosis. Both cows tested positive to serologic tests, the delayed-type hypersensitivity test performed according to BERCOVICH et al. (1989), and bacteriologic examination performed according to BERCOVICH and TER LAAK (1990). Milk from cow 4473 tested negative and milk from cow 4474 tested positive in the MRT-1.

To determine the sensitivity of the ELISA, we prepared whey samples from 39 milk samples that were positive or negative in the MRT-1 and the ELISA. The whey samples were examined by the serum agglutination test, complement fixation test, and Coombs test.

Tank milk samples were constructed as follows. Milk samples from 14 cows positive for Brucella in at least one serologic test were examined with the MRT-1 and the ELISA to establish the titers. The milk samples were then diluted to mimic tank milk samples. Milk from cows 4473 and 4474 was diluted to construct samples “representing” cow’s milk with MRT-1 titers of 2, 4, 8, or 16. Tank milk samples were constructed by further diluting these samples. All dilutions were made with milk that tested negative in the MRT-1. The constructed tank milk samples were examined by the MRT-8 and the ELISA.

To determine the specificity of the ELISA, we examined the following milk samples in an ELISA using mouse monoclonal anti-bovine IgG,: 4,072 milk samples from healthy cows, 943 quarter milk samples from cows with mastitis, 35 milk samples from cows treated with prostaglandins for hormonal disorders, and 500 tank milk samples. An additional 860 milk samples were examined with the MRT-1 and the ELISA. All samples were collected from farms that were officially recognized as free of brucellosis.

Milk ring test Milk samples from individual cows were examined with the MRT-1 according to HILL (1966).

Constructed tank milk samples were examined with the MRT that uses 8 ml milk MRT-8 according to BERCOVICH and LACENDIJK (1978). The antigen was produced and standardized at the Central Veterinary Institute, Lelystad.

Serologic examination of whey and serum Whey was prepared from milk as follows: 5 ml of each milk sample was centrifuged for 20 min at

1,300 g. After the fat was removed, 2 drops of rennet and 2 drops of saturated CaClz were added to each sample. The sample was then mixed and incubated in a 37°C waterbath for 30 min. The whey was collected by centrifugation. -

Whey and serum samples were examined serologically as described earlier by BERCOVICH et al. (1989).

ELISA Antigen

Brucella abortus strain 99 was used to produce the antigen. The bacteria were grown on tryptose agar in Roux flasks for 4 days at 37°C. The bacteria were harvested with sterile saline, autoclaved for 25 min at 121 ’C, cooled to room temperature, and filtered through a sterile gauze to remove agar particles. The bacteria were then disrupted by sonification for 20 min at 20Kc/s on a 500W disintegrator. The resulting suspension was stored in stock solution at 4-6 “C. To prevent bacterial contamination, 0.1 % formaldehyde (final concentration) was added.

Enzyme Immunoassay Using Mouse Monoclonal Anti-bovine Antibodies

Preparation of antigen-coated plates

755

The wells of the ELISA plates were filled with 0.1 ml antigen diluted in 0.01 M phosphate- buffered saline (PBS) pH6.3, sealed, and incubated overnight at 4°C. The coated plates were stored at - 20 "C until used. The optimal antigen concentration for coating the ELISA plates was determined by a block titration against the second international anti-Brucella standard serum. Antigen and conjugate concentrations were adjusted to detect 0.0025 -0.005 Brucella antibody unit per well.

Conjugates and substrate Mouse monoclonal anti-bovine IgM (CVI 17.4.lab), IgA (CVI 15.8.la), and IgGl (CVI 15.8.1)

antibodies labelled with horseradish peroxidase were used as conjugates. The monoclonal antibodies were produced by Dr. D. VAN ZAANE and J. IJZERMAN (1984). The conjugated monoclonal antibodies were kindly supplied by Dr. F. M. VAN ZIJDERVELD from the Bacteriology Department of the Central Veterinary Institute, Lelystad. The conjugates were diluted in 0.5M PBS (pH 7.2), consisting of 0.12 g KH,PO,; 16 g Na2HP04.2 H,O; 29.2 g NaCI; and 0.5 ml Tween 20 dissolved in 1 liter distilled water. The working dilutions were determined by block titration.

The substrate prepared by dissolving 1 g 5-amino-2-hydroxybenzoic acid (Merck, art. 819019) in 1 liter 0.01 M PBS containing 0.04 g Na2EDTA had a pH of 5.9-6.0. The substrate solution was divided into 9ml amounts and kept at -20 "C until used. Just before use, 1 ml of 0.1 YO H,02 solution was added to 9 ml defrosted substrate solution.

ELISA procedure The antigen in the coated ELISA plates was defrosted in a 37°C waterbath and discarded. The

plates were washed five times with a total amount of 500ml washing solution, consisting of demineralized water and 0.05 % (v/v) Tween 20. Appropriate controls were examined on each plate in columns 1 to 3. Milk samples were examined in columns 4 to 12. Two-hundred microliters of each sample was dispensed in a well in row A. The other wells were filled with 0.1 mlO.01 M PBS (pH 7.2) containing 0.05 % (v/v) Tween 20, and twofold dilutions of control and test samples were made. Samples used to evaluate the specificity of the ELISA were undiluted and were examined in duplicate.

After 1 h incubation at 3 7 T , the plates were washed as described above, and 0.1 ml of appropriate mouse monoclonal anti-bovine conjugate, diluted in 0.5 M PBS (pH 7.2), was added to each well. After 1 h incubation at 37"C, the plates were washed again as described above, and 0.1 ml substrate solution containing 0.1 H202 was added to each well. After overnight incubation at room temperature, the plates were shaken on a micro-shaker, and the results were recorded on an Easy reader EAR 400 at 450 nm (SLT-Labinstruments, Austria). When testing milk from individual con \. we chose as a cut off point an extinction value equal to the extinction value of 0.005 Brucella antibody unit per well. To increase the sensitivity of the ELISA in detecting antibodies against Brucella in tank milk, we chose as a cut off point an extinction value equal to the extinction value of 0.0025 Brucella antibody unit per well.

Results When milk samples f rom cows 4473 and 4474 were tested, mouse monoclonal anti-

bovine IgM and IgA conjugates yield titers lower than the anti-bovine IgGl conjugate, T h e IgGl titer in milk f rom c o w 4473 was 1,200 whereas the IgGl titer in milk f rom cow 4474 was 300. A mixture of the three conjugates did not yield titers higher than IgGl conjugate alone.

Sensitivity of the ELISA. Whey samples made from milk that tested negative in the ELISA tested also negative in serologic tests. Whey samples made from milk with low titers in the ELISA tested positive in at least one of the serologic tests. Whey samples made f rom milk that tested positive in the ELISA were also positive in serologic tests (Table 1).

When constructed tank milk samples f rom cows wi th naturally occurring brucellosis were tested, the ELISA detected antibodies against Brucella when both the MRT-1 and MRT-8 were negative (Table 2). When constructed tank milk samples containing cow's milk with a high concentration of IgGl were tested, ELISA titers were higher than MRT-8 titers (Table 3, cow 4473). When constructed tank milk samples containing cow's milk with a low concentration of IgGl were tested, ELISA titers were slightly lower than MRT-8 titers (Table 3, c o w 4474).

756 BERCOVICH and TAAIJKE

Table 1. Comparison of serologic test titers and IgG,-ELISA titers of whey and milk collected from cows with naturally occurring brucellosis

Whey titers Milk titers Cow No. SAT;> CFT Coombs ELISA

~

1-2 neg. < 10 neg. neg. 3-6 neg. < 10 7.5 2 7 neg. < 10 7.5 4 8-9 7.5 < 10 7.5 8

10 7.5 < 10 15 16 11-13 7.5 10 30 64 14 15 10 60 32 15 neg. 20 N D 128 16 15 50 N D 128 17 15 100 N D 128 18 30 20 N D > 128 19-22 30 50 N D > 128 23-29 30 10 N D > 128 30 60 10 N I) > 128 31 -32 60 50 N D > 128 33-36 60 100 N D > 128 37 60 200 N D > 128 38 120 100 240 32 39 120 10 960 32

.' SAT = serum agglutination test; CFT = complement fixation test; N D = not done.

Specif;czty of the ELZSA. Only eight (0.2 'Yo) of the 4,932 milk samples collected from cows from brucellosis-free farms, reacted positively in the ELISA. Only 16 (1.5 Yo) of the 943 quarter milk samples from cows with mastitis were positive in the ELISA. Of the milk samples from cows treated with prostaglandins, tank milk samples, and the milk samples examined by the MRT-1 none were positive in the ELISA.

Discussion Testing milk for brucellosis is economical, because milk is readily available on dairy

farms. The MRT-1 depends on lacteal IgA and IgM immunoglobulins bound to fat globules and on factors that stimulate the clustering of fat globules (COLLIN, 1976; SUTRA et al., 1986; PATTERSON et al., 1976). Individual milk samples lacking these factors or containing a low concentration of IgA or IgM, however, may test false negative in the MKT-1. Tank milk may contain a batch of milk that is positive for brucellosis, but this batch becomes diluted when added to other milk in the tank. Thus, tank milk can test false negative when tested with the MRT-1. It may also test negative because of the rapid decline in MKT titers after abortion or parturition (HILL, 1966).

In comparison to the MRT-1, the ELISA which does not depend on clustering of fat globules, is considered sensitive and specific (THOEN et al., 1979; BORAKER et al., 1981). We tested whether the ability of the ELISA to detect Brucellu antibodies in tank milk depends on the type of anti-bovine conjugate used. We found that when IgA and IgM conjugates were used in the ELISA, antibody titers were no higher than when IgG, was used, although IgA is produced, at least partially, on site (BUTLER et al., 1986). These results indicate that the ELISA may not be able to detect low concentrations of lacteal IgA and IgM in tank milk. Mixing the three conjugates did not improve the sensitivity of the ELISA either. In addition, bovine milk IgM can react non-specifically or actually inhibit the detection of low concentrations of anti-Brucellu IgG (BORAKER et al., 1981; LAMB et

Enzyme Immunoassay Using Mouse Monoclonal Anti-bovine Antibodies 75 7

Table 2. Comparison of serologic tests in serum, and MRT-1, MRT-8 and the ELISA in milk collected from individual cows and constructed tank milk

~ _ _ _ _ _

Milk titers Serum titers Individual Tank

COW NO. MRT-1" ELISA MRT-8 ELISA SAT Coombs CFT

1. 2 < 7.5 15 10 2. 4 15 15 10 3. 4 10 16 < 7.5 15 < 10 4. 4 15 30 < 10 5. 8 16 15 15 10 6. 2 - 7.5 15 < 10 7. 2 4 10 8 15 30 20 8. 2 8 ND 16 15 60 50 9. 16 10 32 15 480 100

10. 8 64 50 128 30 480 100 11. 32 250 250 600 240 1,920 500 12. 64 600 500 1,200 240 3,840 500 13. 8 500 20 1,000 480 960 500 14. 32 500 250 1,600 480 3,840 1,000

'' MRT-1 = milk ring test using 1 ml milk; MRT-8 = milk ring test using 8 ml milk; SAT = serum agglutination test; CFT = complement fixation test; + = positive in the MRT-8; ND = not done.

- - -

- - -

-

- - -

- - - -

-

al., 1979). Because IgG,, however, is the only immunoglobulin freely present in the milk (SUTRA et al., 1986), we used it exclusively as conjugate in our subsequent experiments.

Because we were unable to examine bacteriologically the infected cows, we serologi- cally tested whey and serum, and compared the results with ELISA results. Milk samples and whey made from the same milk samples tested comparably in the ELISA and serologic tests (Table 1).

Since tank milk samples that were positive for brucellosis were not available, we artificially constructed them. Although not enough samples were examined for a statistical analysis, ELISA results, however, were confirmed by at least two serologic tests (Table 2).

Table 3. Comparison of MRT-8 titers and IgG, ELISA titers of constructed tank milk samples containing milk with high or low IgGl concentrations (cows 4473 and 4474, respectively)

- ~ _ _ _ _ _ _ _ ~

Tank milk titer Cow No. Constructed MRT-8 ELISA

milk samples with MRT-1" titer

4473 2 10 320 4 30 640 8 60 1,280

16 120 2,560

4474 2 4 8

16

40 80

120 320

20 40 80

160

'6 See table 2.

758 BERCOVICH and TAAIJKE

The ELISA detected Brucellu antibodies in individual milk samples when the MRT-1 detected none. Likewise, in constructed tank milk samples that had low titers in the MRT-8 the ELISA yield titers higher than the MRT-8 (Tables 2 and 3 ) . In other constructed tank milk samples the ELISA was as sensitive as the MRT-8 (Table 3).

Although cows with a low concentration of lacteal immunoglobulins in milk are rare, they may occur. Moreover, because lacteal immunoglobulin concentrations can vary during lactation, IgG, can escape detection in an ELISA. In our study, milk from cow 4474 with a relatively low IgG, concentration was used to construct tank milk samples. The MRT-8 was slightly more sensitive in detecting antibodies against Brucellu in these samples than the ELISA was (Table 3 ) .

THOEN et al. (1979) suggested that non-specific reactions in the ELISA can be eliminated by diluting the milk before testing. We used undiluted milk, but were able to prevent non-specific reactions by diluting the conjugate in a buffer containing 5 O h NaCI, and a carefully chosen antigen en conjugate concentrations. The ELISA correctly identified as negative 99.8 % of the individual samples collected from brucellosis-free cows, and 98.5 Yo of the quarter milk samples suspected of mastitis. All other milk samples examined tested negative.

The ELISA titers of constructed tank milk (Table 3 ) were higher than titers of individual milk samples because the cut off points for milk from individual cows and tank milk differed. An extinction value equal to the extinction value of 0.005 Brucellu antibody unit per well was used as cut off point to prevent possible non-specific reactions caused by mastitis. Since only 0.4 % milk samples of the total 5,410 individual milk samples tested positive in the ELISA, the test was considered to be highly specific. Because of this high specificity, a lower cut off point was used in examination of tank milk samples. An extinction value equal to the extinction value of 0.0025 Brucellu antibody unit per well was considered positive.

O u r results show that although the ELISA is sensitive and specific in detecting antibodies against Brucellu in tank milk, the results can be affected by concentration of immunoglobulins in the milk. Although the ELISA may fail some times to detect antibodies against Brucellu in tank milk containing milk from a single infected cow (with a low concentration of IgG,), it is nevertheless highly specific. When different batches of antigen or conjugate are being used in the ELISA, the international anti-Brucellu standard serum should be used to achieve reproducable results.

We conclude that the ELISA described in this report may be used to verify ambiguous MRT-I results in milk from individual cows or to entirely replace this test. Both the MRT-8 and the ELISA should be used in testing tank milk for antibodies against Brucellu.

Acknowledgements We thank the following persons for providing milk samples: Dr. J. P. W. M. AKKERMANS, Dr.

A . C . A . VAN EXSEL, Dr. G.M. LAMRERS, and Dr. J.H.H. VAN LIPZIG from the provincial animal health services; Dr. R. WIJFFELS from the INRV Brussel; and Dr. G. BERTELS from the animal disease control center in Alken, Belgium. We also thank Dr. J.HAAGSMA, Dr. E.A. TER LAAK, Mrs. V. W~NSVOORT-THATCHER for their comments on the manuscript.

References BERCOVICH, Z., and W.LACENDIJK, 1978: A modified milk ring test for detecting Brucellu

agglutinins in bulk tank coolers. Tijdschr. Diergeneesk. 103, 407-416. BERCOVICH, Z., and A.MOERMAN, 1979: Non-specific positive milk ring test(s) in tank milk and

Estrumate@ in the treatment of cattle. Tijdschr. Diergeneesk. 104, 713 -716. BERCOVICH, Z., W. LAGENDIJK, and B. A. BOKHOUT, 1989: Evaluation of the delayed type hyper-

sensitivity test for the diagnosis of B.abortus infection in cattle. Vet. Immun. Immuno. 21,

BERCOVICH, %., and E. A. TER LAAK, 1990: An evaluation of the delayed type hypersensitivity test for diagnosing brucellosis in cattle: A field study. The article has been published in Vet. Microhiol.

2 13 -21 8.

22, 241 -248.

Enzyme Immunoassay Using Mouse Monoclonal Anti-bovine Antibodies 75 9

BORAKER, D. K., W. E. STINEBRING, and J. R. KUNKEL, 1981: BrucELISA: An enzyme-antibody immunoassay for detection of Brucellu abortus antibodies in milk: Correlation with the Brucellu ring test and with shedding of viable organisms. J. Clin. Microbiol. 14, 396-403.

BUTLER, J. E., V. L. FERNYO, S. C. WHIPP, R. A. WILSON, and T. E. KOERTGE, 1986: The metabolism and transport of bovine serum SIgA. Comp. Immun. Microbiol. Infec. Dis. 9, 303-315.

COLLIN, J. C., 1976: The role of the different immunoglobulins in the milk ring test used for the detection of brucellosis. Ann. Microbiol. (Inst. Pasteur) 127 B, 117-187.

CORBEL, M. J., F. A. STUART, and R. A. BREWER, 1984: Observations on serological cross-reactions between smooth Brucellu species and organisms of other genera. Develop. Biol. Standard. 56, 314-348.

FORSCHNER VON, E., und I. BUNGER, 1984: Nachweis von IBR/IPV, Leukose und Brucellose- Antikorpern in Bestandmilchproben mit ELISA nach einer einfachen Konzentrierungsmethode. Dtsch. Tierarztl. Wsch. 93, 112- 115.

HECK, F. C., J. D. WILLIAMS, J. PRUETT, R. SANDERS, and D. L. ZINK, 1980: Enzyme-linked immunosorbent assay for detecting antibodies to Brucellu abortus in bovine milk and serum. Am. J. Vet. Res. 41, 2082-2084.

HILL, W. K. W., 1966: Een onderzoek over de waarde van de Abortus Bang-Ringtest bij melk- monsters uit melktanks. Tijdschr. Diergeneesk. 14, 885-895.

HILL, W. K. W., and F. X. M. M. CREMERS, 1967: Voortgezet onderzoek naar de waarde van Abortus Bang Ringtest bij melkmonsters uit melktanks. Tijdschr. Diergeneesk. 92, 1467-1473.

KEER, W. R., J. K. L. PEARSON, and J. E. F. RANKIN, 1959: The bovine udder and its agglutinins. Brit. Vet. J. 115, 105-119.

LAMB, V. L., L. M. JONES, G. G. SCHURIG, and D.T. BERMAN, 1979: Enzyme-linked immunosorbent assay for bovine immunoglobulin subclass-specific response to B . ubortus lipopolysaccharide. Infect. Immun. 26, 240-247.

SUTRA, L., J. P. CAFFIN, and G. DUBRAY, 1986: Role of milk immunoglobulins in the Brucellu milk ring test. Vet. Microbiol. 12, 359-366.

OLIVER, D. G., and R. S. COOPER, 1981: Comparative study of enzyme-linked immuno assay (IgG and IgM) with standard serological tests for diagnosis of brucellosis in cattle. Proc. Annu. Meet. US Anim. Health Ass. 85, 90-140.

PATTERSON, J.M., and B.L. DEYOE, 1976: Effect of physical properties of milk fat globules on Brucellu ring test sensitivity. J. Dairy Sci. 60, 851-856.

THOEN, C. O., D. E. PIETZ, A. L. ARMBRUST, and R. HARRINGTON, 1979: Enzyme immunoassay for detecting Brucellu antibodies in cow’s milk. J. Clin. Microbiol. 10, 222-225.

ZAANE VAN, D., and J. IJZERMAN, 1984: Monoclonal antibodies against bovine immunoglobulins and their use in isotype-specific ELISAs for rotavirus antibody. J. Immun. Methods 72, 427-441.