al., 1989, 1995; Macmillan et al., 1990) in addition to high sensitivities and specificities and these ..... Med. 8, 283. Nielsen, K., Cherwonogrodzky,. J., Duncan.
JOURNAL OF IMMUNOlOGICAL METHODS Journal of Immunological
A homogeneous
Methods
195 (1996) 161-168
fluorescence polarization assay for detection of antibody to Brucella abortus
K. Nielsen a3*, D. Gall a, M. Jolley b, G. Leishman ‘, S. Balsevicius P. Nicoletti d, F. Thomas a
a, P. Smith a,
aAgriculture and Agri-Food Canada, Animal Diseases Research Institute. Box 11300, Station H, Nepean. Ont. K2H 8P9, Canada b Jolley Consulting and Research Inc., 683 E. Center St., Unit H, Grayslake. IL 60030, USA ’ Packer’s Diagnostics Co., Box 102, Lake Blufi IL 60044-0102, USA * Universiry of Florida, College of Veterinary Medicine, Department of Infectious Diseases, Gainesville, FL 32611-0880, Received 20 February
USA
1996; accepted 22 May 1996
Abstract A homogeneous
fluorescence polarization (FP) assay @PA) was developed for detection of antibody in bovine sera to The assay used 0-polysaccharide prepared from B. abortus lipopolysaccharide in the molecular weight range of 20-30 kDa which was conjugated with fluorescein isothiocyanate and used as a tracer. Fluorescence polarization was measured with a FPM-1 fluorescence polarization analyzer. Sample (20 ~1) was added to 2.0 ml of diluent buffer at ambient temperature. A serum blank reading was taken and tracer (10 ~1) to yield approx. 1.5 nM fluorescein equivalents was added. The FP of the tracer was determined after a period of greater than 2 min. A positive reaction was indicated by a significant elevation of the FP reading over the negative control. In a blind study, 9480 bovine sera were tested in addition to sets of four controls which were included with each lot of 100 samples tested. The controls were a strong positive, a weak positive, a negative and a serum derived from a B. abortus strain 19 vaccinated cow. Test sera included 8669 sera from Canadian cattle which were negative by routine serological tests, 561 sera from cows from which B. abortus had been isolated either from tissues or milk and 250 sera from cattle previously vaccinated with B. abortus strain 19 at various times. One lot of 0-polysaccharide tracer was used for all tests. The initial cut-off for negative samples in the fluorescence polarization assay was set at 107.2 mP. This resulted in a sensitivity estimate of 98.1 f 1.1% and the specificity was 99.8 _t 0.09%. After decoding the samples and retesting false positive and negative reactions, the sensitivity estimate was 98.5 + 1.0% and the specificity was 100%. It became evident that the initial cut-off value was set too high and, using ROC analysis, a cut-off of 90 mP increased the sensitivity to 99.02% while the specificity decreased to 99.96%. Of the 250 sera from vaccinated cattle, 248 were negative giving a point specificity value of 99.2%. Brucella
Keywords:
abortus.
Fluorescence
* Corresponding
polarization;
Brucella abortus; Serology;
author. Tel.: (613) 998-9320;
0022-1759/96/$15.00 Copyright PII SOO22-1759(96)00116-O
Homogeneous
assay
Fax: (613) 95243072
0 1996 Elsevier Science B.V. All rights reserved.
1. Introduction Serological diagnosis of brucellosis was first accomplished using an agglutination test nearly a century ago (Wright and Smith, 1897). The agglutination test is very sensitive; however, its specificity is relatively low because IgM antibody resulting from exposure to microbes containing cross-reacting antigens is capable of causing agglutination as well as residual antibody to Brucella abortus strain 19, a vaccine commonly used in control programs for bovine brucellosis. Since then, many modifications have been made to increase the specificity of agglutination reactions, the most useful employing acidified antigen (Angus and Barton, 1984). In addition, complement fixation tests have been developed and widely used (Hill, 1963). Precipitation tests using subcellular preparations as antigens gave the first indication that the antibody response of vaccinated cattle could be distinguished from field infection (Diaz et al.. 1979) which was not possible using the complement fixation test. Primary binding assays were developed to improve the sensitivity and specificity of serological testing and to replace the more cumbersome conventional tests. The indirect enzyme immunoassay, while very sensitive. lacked relative specificity until the procedure was supplemented with divalent cation chelating agents (Nielsen et al., 1994). However, the indirect ELISA was not able to differentiate vaccinal antibody from that resulting from field infection. Competitive ELISAs had this capability (Nielsen et al., 1989, 1995; Macmillan et al., 1990) in addition to high sensitivities and specificities and these assays are currently replacing existing conventional tests. These tests are fairly rapid, taking about 90 min to complete but require multiple manipulations and extensive equipment for result assessment. In the current format, primary binding assays for serological diagnosis are not suitable for field use. Homogeneous assays have been developed for detection and measurement of small molecular weight molecules (Kabanov et al., 1989; Fernando et al.. 1992) and macromolecules (Hasoda and Yasuda, 1989; Katch et al., 1993). This type of assay requires minimal manipulations and less equipment for data assessment (formats reviewed by Jenkins, 1992). In addition, they can be completed in a few minutes. A
type of homogeneous assay, a fluorescence polarization immunoassay (Jolley, 198 11, has been developed for detection of numerous substances, including illicit drugs (Clark et al.. 1992; Cody and Schwartzhoff, 1993). monitoring of drugs in clinical use (Pesce et al.. 1990) and some macromolecules (Tait and Fujikawa, 1987; Urios and Cittanova, 1990). Because of its characteristics, this type of assay would be an ideal candidate for field and laboratory tests for detection of exposure to pathogens, currently accomplished by much slower and tedious technologies. A homogeneous fluorescence polarization assay for the measurement of bovine antibody to B. abortzrs is reported in this communication. The assay requires a single. one-step serum dilution on which a baseline serum intensity blank is obtained using a fluorescence polarization analyzer. This is followed by addition of a fluorescein labeled B. abortus Opolysaccharide tracer. The fluorescence polarization of the tracer is then measured. The total time required is approx. 2 min.
2. Materials
and methods
2. I. Serological
tests
The buffered plate antigen test (BPAT) was a modification of the method reported by Angus and Barton (1984). Results were obtained as lack or presence of agglutinated cells after 4 min of rotation at 20°C. The complement fixation test (CFT) was a 96-well adaptation incubated overnight at 4°C. The procedure was described by Samagh and Boulanger (1978). Results were observed in 25% hemolysis increments of the highest serum dilution. All sera were tested at I /5. 1/IO and l/20 dilutions and positive sera were titrated to their end point. The indirect enzyme immunoassay (I-ELISA) used B. abortus lipopolysaccharide as the antigen immobilized on 96-well polystyrene plates and a murine monoclonal antibody. specific for an F, epitope of bovine IgG, and conjugated with horseradish peroxidase. Divalent cation chelating agents (EDTA and EGTA) were added to the serum incubation step to minimize non-specific interactions. The assay proce-
K. Nielsen et al./ Journal
of Immunological Methods I95 (1996) 161-168
dure has been described previously (Nielsen et al., 1994). The competitive enzyme immunoassay (C-ELISA) was performed as described by Nielsen et al. (19951. Briefly, B. abortus lipopolysaccharide was used as the antigen. Test serum and monoclonal antibody specific for an 0-polysaccharide epitope were diluted in EDTA/EGTA buffer and added simultaneously. This was followed by addition of a goat anti-mouse IgG-horseradish peroxidase conjugated antibody. The fluorescent polarization assay (FPA) utilized B. abortus 0-polysaccharide, 3.0 mg, molecular weight 20-30 kDa, dissolved in 600 l.~l 0.1 M sodium hydroxide and incubated for 1 h at 37°C. Conjugation was accomplished by adding 300 ~1 of freshly prepared fluorescein isothiocyanate (isomer I, Sigma, St. Louis, MO, USA) dissolved in dimethyl sulfoxide at a concentration of 100 mg/ml. The mixture was incubated at 37°C for 1 h and then applied to a DEAE-Sephadex A25 column with a bed volume of 20 ml and equilibrated with 0.01 M phosphate buffer, pH 7.4. The same buffer was used for eluting the column. The first fraction (7 ml) was buffer and the second, a bright green fluorescent fraction, was contained in the next 7 ml. The buffer was changed to 0.1 M phosphate buffer, pH 7.4 and two further fractions were collected. The initial 10 ml was buffer, followed by 25 ml of bright green fluorescent material. The latter fraction was used as the tracer for the study; however, six other batches of tracer were prepared and when tested gave the same results. The conjugated 0-polysaccharide was diluted in 0.01 M sodium phosphate, pH 7.4, containing 0.15 M NaCl and 0.1% sodium azide (PBSA) until a further l/200 dilution gave the equivalent of approx. 1.5 nM fluorescein equivalents of total fluorescence intensity. This amount of FITC conjugated 0-polysaccharide was used in each test. For testing, sera were diluted l/100 in 2 ml of 0.1 M PBSA containing 0.05% lithium dodecyl sulfate (PBSAL), placed in a 12 X 75 mm glass tube. A serum blank measurement was obtained with the Fluorescence Polarization Analyzer (FPM- 1, Jolley Consulting and Research Inc., Round Lake, IL, USA). The predetermined amount of tracer was added and after mixing and incubation at room temperature for at least 2
163
min, the FP of the tracer was determined (with the blank subtracted). The latter determination indicated the amount of antibody in the serum sample. Data from this assay were expressed as millipolarization units (mP). 2.2. Sera Most sera had previously been tested by the BPAT, CFT, I-ELISA and C-ELISA. These were tested by the FPA. A total of 9480 serum samples were used. Of these, 561 were obtained from cattle from which B. abortus had been cultured either from milk or tissues, whereas 8669 sera were obtained from Canadian cattle, epidemiologically and serologically free from brucellosis and 250 were obtained from B. abortus strain 19 vaccinated cattle at various times post vaccination. All 250 sera reacted in one or more of the BPAT, the CFT or the I-ELISA. All sera were stored at -20°C until tested at which time dilutions in racks containing 100 serum dilutions consisting of approx. 90 negative sera and six positive sera were prepared. The placements were recorded. A second analyst performed the remainder of the test without any knowledge of the distribution of the positive sera among the negative sera. The results and the original placement sheets were collated by a third individual. Any test results that did not conform to the expected values were tabulated and the sera were retested at the end of the study. Included in each rack of 100 serum samples were four controls consisting of a strong positive, a weak positive, a negative and a serum sample from a vaccinated animal. 2.3. Data analysis An initial cut-off value was derived from the data obtained from the negative samples. The data were sorted in ascending order and divided into 100 equal percentiles. The mean of the 100th percentile was calculated and used as the preliminary cut-off value between positive and negative results. The diagnostic sensitivity and specificity could then be optimized by plotting the sorted data for defined negative and positive samples using a frequency histogram. The preliminary cut-off value was
K. Nielserl et al.//ournal
164
of Immunological Methods 195 (1996) 161-168
used for calculating sensitivity and specificity with 95% confidence limits. The data were also analyzed using signal detection analysis (receiver operator characteristics analysis, ROC) according to Stenson (1988) and Metz (1978). This analysis allowed the determination of a more suitable cut-off value estimate by comparing the range of sensitivity and specificity values for a range of cut-off values.
__
50
2
R R
R
2 40 ; $ g 30' z
t
$ 20 ;: 2 z 10 1
0 60
3. Results Assay sensitivity and specificity for the FPA before and after decoding of the serum samples are as follows. Using a preliminary cut-off of 107.2 mP (the 100th percentile of the mean of the negative sera), the point estimates of sensitivity and the specificity of the FPA after initially testing all samples were 98.2 f 1.I% and 99.8 k 0.9%. respectively. After decoding and retesting positive samples. the sensitivity value increased to 98.5 + 1%. Figs. 1 and 2 represent histograms of the distribution of results obtained with the positive and negative sera in the
60 70 80 90
I 100110120130140150160 mP
Fig. I, Frequency distribution of 9230 known positive (open bars) or negative (closed bars) samples tested for their antibody content against B. ubortus. Numbers of samples (Y axis) are plotted against a measurement of fluorescence polarization (mP, X axis). The scale of the Y axis has been adjusted to allow observation of the lower number of observations. The numbers above the bars indicate the number of observations for the point. All data were collected using coded samples and data for retested aberrant samples are not included. Using a preliminary cut-off value of 107.2 mP. there were eight false negative reactions and 24 false positive tests.
70
60
90
1001~pO12013014015016G
Fig. 2. Frequency distribution (same scale as Fig. 1) of 8669 serum samples negative for antibody to B. abortus (closed bars) and 561 serum samples from animals proven to be infected with B. nbortus (open bars). Numbers above bars indicate the number of observations for that point, Data are presented after decoding of samples and retesting false reactions. It is apparent that the cut-off can be lowered to 90 mP with a very small decrease in specificity and a gain in sensitivity
FPA before and after decoding of the samples. While the cut-off from Fig. 1 is difficult to determine because of the overlap between positive and negative samples in the 90-140 mP region, it is clear that the cut-off should be above 100 mP and a 107.2 value as calculated using the 100th percentile of the negative sera is not unreasonable. From Fig. 2, however, after retesting 24 false positive samples, it is apparent that the cut-off should be set at a lower level. If a value of 90 mP is selected, three false negative results and one false positive are evident. These findings were confirmed using ROC analysis (Fig. 3). Thus if a cut-off of 90 mP is selected, the sensitivity and specificity values are 99.02 and 99.96%, respectively. Of the six serum samples from animals with confirmed brucellosis that remained negative in the FPA, one originated from a cow infected with B. abortus strain 19 and was negative in all serological tests. Of the remaining, four animals were infected with field strains of B. abortus and of these, two were serologically negative in all tests while two sera were positive in all tests including the FPA. In addition to these five serum samples, a sample from another strain 19 infected cow was positive in the
K. Nielsen et al. / Journal of Immunological
165
Methods 195 (1996) 161-168
100 99
%#
o*oo~o@woe.%
250
D
~~----______---__----_----~--~---
O~ODg-a e.DOOgOO
0 Pos
98
250
200 i
90919293949596979899100 % Specificity
L-_-----_____-__---_____.-__---____
150
FPA but negative in other serological tests. These data are presented in Table 1. The four control sera, a strong positive, a weak positive, a negative and a serum from a vaccinated cow, included in each 100 samples tested, are plotted in Fig. 4. Each point represents an individual deter-
Table 1 Results obtained in various tests with serum samples preliminary false negative reactions in the FPA
1 2 3 4 5 6
FS FS FS FS s19 s19
+ _ + _
nd nd _ _
0.290 0.750 0.115 0.950 0.310 0.286
d C-ELISA
that gave ’ FPA f
14 36 13 80 6 14
17.3 94.1 81.3 232.8 71.9 93.8
a Results of bacteriological culture of B. abortus strain 19 (S19) or field strain (FS). ’ Buffered plate antigen result, either positive or negative. ’ Complement fixation test titer (reciprocal). d Optical density reading in the indirect enzyme immunoassay. A reading of 0.460 was considered positive. ’ Percent inhibition in the competitive enzyme immunoassay. Inhibition of 30% or greater was considered positive. f Results of fluorescence polarization assay. A reading of 90 mP or higher was considered positive. nd, not done.
DogDIm
Do D .mgno*%~
q0
II
d’nD
----__D_“____“_P~____---_____---___: ‘OOf--.___________ -__--__-____---__ n n .p.. ‘.=.,.m..~....=.m. a,=9 50_t------_____L__!r_L---_~___---JA~
Fig. 3. ROC analysis, plotting % sensitivity (Y axis) against % specificity (X axis) for various cut-off values, after retesting of aberrant samples in the FPA. The various cut-off values are indicated on the graph. From the data, a cut-off of 90 mP gives a specificity value of 99.96% and a sensitivity value of 99.02%.
Serum Culture a BPAT b CFI ’ I-ELISA
!
Do
mm0
q 00
Wk POS
on
rn..rn..,.#
l-r
voc
I --t--_._+
0+-10
20 Number
of
-.-+---
30 observations
40
50
Fig. 4. Plot of the results obtained with four control sera included in each 100 tests. The controls included a strong positive (open circles), a weak positive (open squares), a negative (closed circles) and serum from a B. abortus strain 19 vaccinated cow (closed squares). The level of positivity (mP, Y axis) is plotted against the number of observations for each control. The first 50 observations for each control were selected to eliminate overcrowding of the figure. The broken lines indicate 3 SD on either side of the mean for each entire set of data.
mination. The upper and lower lines for each set of points indicate + 3 SD from the mean. Of the sera obtained from B. abortus strain 19
Table 2 Point specificity values of serological tests used to evaluate the antibody content in sera from 250 cattle vaccinated with B. abortus strain 19 at various times previously Test
Specificity
BPAT CFI I-ELISA C-ELISA FPA
48.6 49.0 56.3 97.1 99.2
(%)
Cut-off used + or Reaction at l/5 46% positivity 30% inhibition 90 mP
a b b
All sera were positive in one or more of the assays (other than the FPA) (Nielsen et al., 1992, 1996). a 87 sera were anticomplementary and no diagnosis could be made. b Compared to a known positive reaction.
166
K. Nielserl et al./Jounml
of immunological A4ethod.s195 (19%) 16/-16X
vaccinated cattle, two gave positive reactions in the FPA using a cut-off of 90 mP, giving a point estimate for specificity of 99.2%. For comparison, the specificity in the BPAT was 48.6%, the CFT was 49.0% (87 of 250 sera were anticomplementary and no diagnosis could be made), the I-ELISA was 56.3% and the C-ELISA was 97.2%. These data are presented in Table 2.
4. Discussion Homogeneous immunoassays are single step assays that require no repetitive washing procedures to remove unbound reactants as with classical primary binding assays (see Jenkins, 1992). This decreases the assay time considerably, since only time for primary immune complex formation is required as well as simplification of the assay, making field application of a primary binding assay a possibility. In the past. attempts to develop homogeneous enzyme immunoassays by either attaching antigenic determinants in the vicinity of the active site of an enzyme (Brontman and Meyerhoff, 1984), of a substrate (Ashihara et al., 1988), by formation of an active enzyme (Engel and Khanna, 1992), by lysis of liposomes (Ligler et al., 1987; Katch et al., 1993) or by measurment of local electrostatic changes of immune complex formation (Huang and Lee, 19921 have found limited applications. Fluorescence polarization immunoassay, initially described by Dandliker and Feigen (1961) and reviewed by Rhys Williams (1988>, lends itself well to use in homogeneous immunoassays. Thus a fluorescent dye can be excited by plane-polarized light of an appropriate wavelength. A stationary molecule, excited with polarized light, will emit light in the same plane. However, molecules are in constant motion, rotating around an axis in any plane. The rotation rate can be assessed by measuring the light intensity in the horizontal and vertical planes. A large molecule rotates at a slower rate than a smaller molecule, thus confining emitted light more to a single plane (more polarized) while small molecules rotate at a faster rate emitting more depolarized light. Thus the time a molecule takes to rotate through a given angle is an indication of its size. This principle was used to develop an assay for detection of bovine antibody to B. ubortus. Thus O-polysaccharide from
B. abortus was hydrolyzed and conjugated with fluorescein isothiocyanate. After blanking on the diluted serum sample, the O-polysaccharide tracer was added, the serum blank subtracted and the FP determined. If antibody was present in the serum, a complex of the tracer-antibody would form, slowing the rotation of the tracer and thereby inreasing the polarization of the emitted light. Of the 8669 sera from B. ubortus negative cattle, 2.5 gave false positive reactions on initial testing. This was reduced to a single serum sample on retesting, resulting in a test specificity of nearly 100%. This compares favorably with the I-ELISA and C-ELISA which were used on the same samples, both resulting in specificity estimates of 99.8%. The complement fixation test had a specificity value of 93.1 or 99.8% if anticomplementary sera were considered positive or negative, respectively, while the buffered plate antigen test correctly identified 97.9% of the negative samples (Nielsen et al., 1995). In the same paper, 654 sera from B. abortus infected cattle were tested by the ELISAs and the sensitivity for each assay was reported at 100%. For comparison, the complement fixation test detected 97.1 or 87.9% of the positive sera (if anticomplementary samples were considered positive or negative, respectively) and the BPAT correctly identified 98.6% of the positive sera. However, due to depletion of samples, only about 500 of the original collection were available for testing by the FPA. An additional 60 sera, also from culture positive animals, were added. All but six of the 561 sera were positive on initial testing in the FPA. These sera were also retested and when the cut-off value was altered in accordance with the data analysis after retesting the false positive samples (set at 90 mP instead of 107.2 mP), three samples remained negative in the FPA. Of the same six sera, four remained negative in the ELISAs (Table 1I. Samples 1 and 3 were obtained from cattle with very low bacterial numbers isolated from a single lymph node. The bacterial load of the animals from which samples 5 and 6 originated is unknown. The serum samples obtained from cattle vaccinated with B. nbortus strain 19 all gave a reaction in one or more of the BPAT. the CFT or the I-ELISA. They were also tested in the C-ELISA. The specificity data are presented in Table 2. Two of these sera gave positive results in the FPA, at 92 and 106
K. Nielsen et al./ Journal
of Immunological
mP, using a cut-off value of 90 mP. This resulted in a specificity of 99.2% for the FPA. It is clear from the ROC analysis presented in Fig. 3 that by adjusting the cut-off value upward, the specificity may be increased to 100% but at a loss of sensitivity. This type of analysis allows decisions to be made regarding the most suitable cut-off value to be used in a given situation, depending on specificity and sensitivity requirements. From Fig. 4, depicting the values obtained with the four control samples included with each 100 tests, the FPA is apparently a highly repeatable test with only minor day to day variations observed. Only one test value for each of the strong and weak positive sera fell outside the bar indication + 3 SD. The FPA has been shown to be a highly accurate assay for detection of antibodies to B. abortus resulting from field infection in bovine sera without detecting vaccinal antibodies. Because of the ease and rapidity of the procedure, it is quite likely that it could be adapted as a field test at a considerable cost reduction over other primary binding assays. It is also quite possible that this procedure is adaptable to testing exposure to other pathogens, the only limitations being the size of the antigen, its ability to be covalently linked with a fluorochrome and its ability to induce an antibody response in exposed animals.
References Angus, R.D. and Barton, C.E. (19841 The production and evaluation of a buffered plate antigen for use in the presumptive test for brucellosis. Dev. Biol. Stand. 56, 349. Ashihara, Y., Nishizono, I., Tanimoto, T., Tsuchiya, H., Yamamoto. K.. Kido Y., Miyagawa, E. and Kasahara, Y. (1988) Enzyme inhibitory homogeneous immunoassay for high molecular weight antigen. J. Clin. Lab. Anal. 2, 138. Brontman, S. and Meyerhoff, M. (19841 Homogeneous enzyme linked assays mediated by enzyme antibodies; a new approach to electrode based immunoassays. Anal. Chim. Acta 162, 363. Clark, G., Rosenzweig, I., Raisys. V., Calahan, C., Grant, T. and Streissguth, A. (1992) The analysis of cocaine and benzoylecgonine in meconium. J. Anal. Toxicol. 16, 261. Cody, J. and Schwartzhoff, R. (19931 Fluorescence polarization immunoassay detection of amphetamine. methamphetamine and illicit amphetamine analogues. J. Anal. Toxicol. 17, 26. Dandliker. W.B. and Feigen, G.A. (19611 Quantification of the antigen-antibody reaction by the polarization of fluorescence. Biochem. Biophys. Res. Commun. 5, 799.
Methods
195 (1996)
161-168
16-I
Diaz. R., Garatea, P., Jones, L. and Moriyon, I. (19791 Radial immunodiffusion test with Brucella polysaccharide antigen for differentiating vaccinated from infected cattle. J. Clin. Microbial. 10, 37. Engel, W. and Khanna, P. (1992)CEDIA in vitro diagnostics with a novel homogeneous immunoassay technique. Current status and future prospects. J. Immunol. Methods 150, 99. Fernando, S., Sportsman, J. and Wilson, G. (1992) Studies of the low does ‘hook’ effect in a competitive homogeneous immunoassay. J. Immunol. Methods 151, 27. Hasoda, K. and Yasuda, T. (1989) Homogeneous immunoassay for alpha, plasmin inhibitor (alpha,PI) and alphazP1-plasmin complex. Application of a sandwich liposome immune lysis assay (LILA) technique. J. Immunol. Methods 121. 121. Hill, W.K. (1963) Standardization of the complement fixation test for brucellosis. Bull. OIE 60. 401, Huang, P. and Lee, C. (1992) Direct and homogeneous immunoassay for IgG analyses. Biotechnol. Bioeng. 40, 913. Jenkins. S.H. (19921 Homogeneous enzyme immunoassay. J. Immunol. Methods 150, 91. Jolley, ME. (1981) Fluorescence polarization immunoassay for the determination of therapeutic drug levels in human plasma. J. Anal. Toxicol. 5, 236. Kabanov, A., Khrutskaya. M., Eremin. S. and Klyachko, N. (19891 A new homogeneous immunoassay: reversed micellar systems as a medium for analysis. Anal. Biochem. 181, 145. Katch. S., Sohma. Y., Mori, Y., Fujita. R., Sade, E.. Kishimura, M. and Fukuda, H. (19931 Homogeneous immunoassay of polyclonal antibodies by the use of antigen coupled liposomes. Biotechnol. Bioeng. 41. 862. Ligler, F.. Bredehorst, R., Talebian. A., Shriver, L., Hammer, C., Sheridan J., Vogel, C. and Gaber. B. (19871 A homogeneous immunoassay for the mycotoxin T-2 utilizing liposomes. Anal. Biochem. 163. 369. Macmillan. A., Greiser-Wilke, I., Moening, V. and Mathias, L. (19901 A competitive enzyme immunoassay for brucellosis diagnosis. Dtsch. Tierarztl. Wochenschr. 97. 83. Metz, C. (1978) Basic principles of ROC analysis. Semin. Nucl. Med. 8, 283. Nielsen, K., Cherwonogrodzky, J., Duncan. J. and Bundle, D. (19891 Enzyme immunoassay for the differentiation of the antibody response of Erucella abortus infected or strain 19 vaccinated cattle. Am. J. Vet. Res. 50, 5. Nielsen, K., Gall. D.. Kelly, W., Henning. D. and Garcia, M. (1992) Enzyme Immunoassay: Application to the Diagnosis of Bovine Brucellosis. Agriculture Canada Monograph. Nielsen, K., Kelly, L., Gail, D., Smith, P., Bosse, J., Nicoletti. P. and Kelly, W. (1994) The use of divalent cation chelating agents (EDTA/EGTAl to reduce non-specific serum protein interaction in enzyme immunoassay. Vet. Res. Commun. 18. 433. Nielsen, K., Kelly, L., Gall. D., Nicoletti, P. and Kelly, W. (19951 Improved competitive enzyme immunoassay for the diagnosis of bovine brucellosis. Vet. Immunol. Immunopathol. 46, 285. Nielsen, K., Kelly, L., Gall, D., Balsevicius, S., Bosse, J., Nicoletti. P. and Kelly, W. (1996) Comparison of enzyme immunoassays for the diagnosis of bovine brucellosis. Prev. Vet. Med. 26, 17.
168
K. Nielsen et al. /Journal
of Immunological Methods 195 (1996) 161-168
Pesce, A., Schroeder, T. and First, M. (1990) An evaluation of cyclosporine monitoring by non-selective fluorescence polarization immunoassay. Transplant. Proc. 22, 1171. Rhys Williams, A.T. (1988) Fluorescence polarization immunoassay. In: W.P. Collins (Ed.), Complementary Immunoassays. John Wiley, Chichester, pp. 135-147. Samagh, B. and Boulanger, P. (1978) Adaptation of the Technicon Autoanalyzer II for an automated complement fixation test for detection of antibody to Brucella abortus in bovine serums. Proc. 21st Annu. Meet. Am. Assoc. Vet. Lab. Diagnost., pp. 347.
Stenson, H. (1988) SIGNAL: a Supplementary Module for SYSTAT and SYGRAPH. SYSTAT Inc., Evanston, IL. Tait, J. and Fujikawa, K. (1987) Primary structure requirements for the binding of human high molecular weight kininogen to plasma kallekrein and factor XI. J. Biol. Chem. 262, 1105 1. Urios, P. and Cittanova, N. (1990) Adaptation of fluorescence polarization immunoassay to the assay for macromolecules. Anal. Biochem. 185, 308. Wright, A.E. and Smith, F. (1897) Application of a serum test to the differential diagnosis of typhoid and Malta fever. Lancet i, 656.
