Journal of Immunological Methods, 155 (]e~/2) 233-240 © 1992 E l ~ i c r Science Publishers B.V. All righls reserved f1022.1759/9'2/$05.01t

233

JIM 0ff177

Removal of false~positive reactions from plasma in an enzyme immunoassay for bovine interferon-~, Stephen L. Jones ~, John C. Cox a, Jane M. Shepherd ~, .lames S. Rothel b, Paul R. Wood b and Anthony J. Radford b e Researzh and Det,dopraem Division, CSL Limited, Parkcille 3052. Au~wralia, and t, CSIRO Di~,ision ,~fAnimal Health, Animal Health Research Laboratory~ Pack,_:ille3052, Australia (Received 11 October 1991, revised received 9 March I992. accepted 21 May 19921

A monocional antibody-based sandwich enz3,me immunoassay (EIA) for bovine interferon-,/(IFN~/) has been developed and can be used in conjunction with a whole blood culture system to diagnose tuberculosis in cattle. During its development, normal bovine plasma samples were tested to establish background levels of circulatory IFN-~. Cf 191 samples tested, 81 (42.4%) were positive (OD > 0.1) when tested undiluted in intact monoclonal antibody (lgGl).coated wells compared to only 8 (4.2%) in F(ab')2-coated wells, which suggested non-specific interference in the EIA rather than circulatory IFN-y. Reactivity of all remaining samples was removed by diluting plasmas 1/2 with 1% casein-PBS-0.05% Tween 20 supplemented with an optimum amount (5%) of normal mouse serum (NMS). Serum pools derived from BALB/e, DBA/2, C3H/HeJ, CBA/CaH and Swiss, but not C57BL/6J, mice were found to inhibit equally the reactions of five strong false-positive bovine plasma samples but had no effect on the title of IFN-y in the sample. Sera from other species tested were less effective. This suggests that the interfering factors possess a high degree of specificity, since the immunoglobulin heavv chain of IgG1 produced by all these five strains of mice are ailotypically identical and different to lgGl produced by C57BL/6J mice. The use of F(ab')2 antibody fragments to coat plate wells and sample diluent containing 5% NMS has resulted in an EIA for bovine IFN-y that is virtuaily free from false-positive reactions, has a high degree of reproducibility and a ~,ample detection limit equivalent to approximately 80 pg/ml recombinant bovine IFN-~. Key words: F_,nzymeimmunoassay; False-positive reacxion; Interferon-y; Tuberculosis; Plasma; (Bovine)

Introduction

An in vitro cellular assay for b~vin¢ tuberculosis has been developed that can be vsed a: a Correspondence to: S.L Jones, Research and Development Division, CSL Limited, 45 Poplar Road, Par~ille 31252.Victoria, Australia.

convenient and more sensitive alternative to the existing intradermal tuberculin skin test (Wood et al., 1990a, 1991). Animals infected with Mycobacterium bocis, the eausat;.ve agent of b~'Ane tuberculosis, should have circulating T lymphocytes sensitised to mycobacterial antigens. Using a simple whole blood culture system, ,~uch cells normally resound to in vitro stimulation with bovine

234 tuberculin purified protein derivative (PPD) and secrete interferon-~ (IFN-7) which can be detected in the supernatant plasma by bJoassay (Wood et al., 1990a) or sandwich enzyme ira. munoassay (EIA) (RotheI et al., 1990). However, the many reported drawbacks of using bioassays for the measurement of cytukine levels in body fluids (Whieher and lngham, 1990) make the EIA for bovine IFN-9" the assay of choice for routine usage. Nonetheless, a variety of factors, such as rheumatoid facter~ d,RFs) and heterophil antibodies. have been shown to cause non-specific interfemnce in sandwich EIAs similar to that described in this paper (Boscato and Stuart, 1988; Spencer, 1988; Hamilton, 1989; Gosling, 1990). These interfering proteins often have specificities for sites on antibody molecules other than the antigen binding sites, and hence in sandwich ElKs, especially where the two antibodies are of the same species and immunoglobulin isotype, have a propensity to form bridges (between the antibody reagents) that lead to false-positive reactions. Commonly, this type of interference in an EIA can bc eliminated by the use of non-imm u n e serum in diluents and F(ab') 2 antibody fragments instead of intact IgG (Spencer, 1988). The 1FN--r EIA employed two non-competing monoclona[ antibodies (IFN-2, IFN-9) of the same leG1 isotype (Wood et al., 1990b) which could potentially be bridged by bovine RFs that have been shown to cross-react with mouse g a m m a globulin (Waron ctal., 1987). False-positive reactions coutd also arise from other heterophilie antibodies to mouse immunoglobulins that have been described in the sera or plasma of h u m a n s (Boscato and Stuart, 1988) and cats (Lopez and Jacobsen, 1989) and are likely to exist in cattle. Therefore, the assessment and elimination of false-positive reactions in the bovine IFN-y EIA was important for the accurate measurement of IFN-~/ in plasma samples. T h e aims of this study were to determine the levels of IFN- T in the plasma of healthy cattle and to devise methods to eliminate the effects of possible interference factors. This was required to maximise the specificity of the bovine IFN-1, EIA and hence the utility of the assay system for the diagnosis of tuberculosis in ca~tle.

TABLE I ABILITY OF SERUM FROM VARIOUS SPECIES TO BLOCK FALSE-POSITIVE REACTIONS IN A BOVINE [FN-y EIA BLockingserum (5% v/v) Mouse CBA/CaH Swiss CaH/HuI" DBA/2 BALB/c C57BL/6J Rat Goose Monkey Horse Rabbit Sheep Human Cat Fewer Guinea Pig Goal Pig Chicken Camel Dog Diluent

% Reduction in OD False positive" Positive" 97.4 (3.3) 10.6(4.7) 9"1.3 (4.0) 4.t (I.8) 96.7 (4.2~ 12.6 (4.7) 95.9 (3.4) 9.4(4.2) 93.0 (2.5) ~ 1.6{I.OJ 81.4 (33.8) 7.1 (7.1} 47,1 (33.8) 1.7(2.3) 27.7 05.0) 2.5 (6,3) 22.6 (34.7) - 5.6 (4.6) 21.4 (7.O) -0.5 (3.7) 13,0 (5.7) -4.3 (3.3) 12.3 (6.5) ~4,9 (3.6) 10.4 (3,7) -0,6 (Z3) 9.6 (10.9) 2,9 (1.9) 9,1 (5.7) - 8.7 (-1.8) 9.0 (7A} 1.5 (2.7) 8.9 (5.5) 2.6 (3.1) 5.9 (t0.3) 4.1 (I.'/) 4.2 ( 6 . 7 ) -4.5(2.21 0.8 (15.0) 1.0 (0.6) - 24.6 (23.6) - 0.9 (2.3) 0.0 (0.0) 0.0 (0.0)

Mean (SD} of five strongly false-positivebovine plasmas in F(ab')z-eoated EtA wells. b Mean (SD) of two samples containing bovine IF'N.y.

Materials and methods

Animals Mice were purchased from Animal Resources Centre (Murdoch, Western Australia, Australia) and housed in isolation in the Animal Services' facility at CSL Limited (CSL), Parkville, Australia. Local institutional guidelines on animal care were followed throughout.

Serum and plasma Normal mouse serum (NMS) was obtained by pu,~ling sera from 40 outbred Swiss mice and was used throushont this study except where indicated in Table L Serum pools for the other strains were obtained from groups of six 10-14-week-old normal male mice.

235

All other sera used in this sLudywore colloctcd from normal animals and healthy ~.ndividuals. Bleeds from 191 normal cattle were collected in 50 ml tubes containing 500 IU sodium heparin (CSL), in 1 ml phosphate-buffered saline (PBS, pH 7.2), and the plasma separated by eentrifugation+ Positive and negative control samples were prepared according to Rothel et al. (1990). The positive bovine control plasma containing natural IFN-'}, was derived by PPD stimulation of blood cultured from an animal infected with M. bo[~is. A second positive sample contained 1 ng/ml recombinant bovine IFN- 7 in 1% sodium casein. PBS, 0.05% Tween 20 (di[uen0, The negative bovine cent,el plasma was from an uninfected

animal and did not react in the bovine IFN-3, EIA. All scram and plasma samples wore stored in aliquots a: - 20°C.

Preparation of F(ab')~ IFN-9 anll-bot,ine IFN-y Monoclonal antibody IFN-9 (IgGI) against bovine IFN-y was purified from ascitie fluid as

described previously (Wood et al., 1990b). Purified IFN-9 was dialyscd against 0.i M sodium acetate buffer, pH 4.0, adjusted to 10 mg/ml, and I ml was incubated with 2250 U of washed immobilised pepsin (Pierce, Rockford, IL, USA) for 2 h at 37°C with constant rocking. After the addition of 2 ml of L0 M Tris-HCI buffer, pH 7.5, the reaction mixture was centrifuged, and the super-

let 450t~m

{{3/191 133.0%1

49J 191 ( 2 ~ . 7 ~ )

2.C

2.0

t.a 1.1 14

1.4

1,2

1.2

I

%0 O:l

o.a

G,4

0.4

o.=

O3 °:

Or. .

O0

2~1

40

4~, I~ 100 12Q 140 Normal ~¢V}rHi Ptmimo 5 1 m p { #

¢cl OO45enm

I~

100

J,L { L l

.

.iv.

.

20

I.

. 4O

t'd| OD O,SC~m

7t191 1 3 7 % l

20

2o

11

q.e

1,1

Li

1.4

a.4

1.2

L:z

I

h

JL ,t

.{.. ,, . . . . ~2~ 140 1~ NOr m~l gemini P~mmi ~llffl~ ~ l

. ~

.

.

IlO

. ~1,'4

|.

[

~00

Cq191 (G%I

Q.i

O.I oA

0.3 O~

, ..... I r

,

~o

4n

,

.

.,. .

.

.,.,, . .

....

i ,

.

so uo 1® 1~ 14o N~m~,l Bowi~ Pilndn~ $ t ~ 1 4 r 4

lt~

,

, ,

,

,.

%

il

,

,

,I,

~iD

Fig. I. FaLse-positivereactivityof 191 normal bovine plasma samples in bovlnc IFN-y EIA, Smmplem were diluted I,/7.with diluent (a.c) or diluent containing 5% NMS (b, d) in wells teared with either intacl IFN j) monoclonal an|i~odi~s(a+ b) or IFN-Q F(ab')z fragments {c, d).

236 natant dialysed against PBS. Cleavage of IFN-9 to l~redominantly F(ab')2 fragments was confirmed by sodium dodecyl sulphate polyacryiamide gel electrophoresis using the PhastSystem (Pharmacia, Uppsala, Sweden) with 10-15% gradient PhastGel media followed by silver staining. Intact and F(ab') 2 fragments of IFN-9 were stored in aliquots at -4(PC before use.

Bo~,ine IFN-y I~IA The bovine IFN-y EIA has been described elsewhere (Rothcl et al., 1990) end was used with minor modifications. Briefly, EIA plates (MaxiSorp, Nunc, Roskilde, Denmark) were coated with 100 ~l/well of IFN-9 or IFN-9 F(ab') 2 at 5 ~tg/ml in 50 mM carbonate buffer, pH 9.6, for 18 h at 4°C. Unbound antibodies were removed by aspiration, and unreacted binding sites blocked with 150 t~l/well ofa I mg/ml solution of sodium casein in PBS containing 0.05% Tween 20 (PBST) for 1 h at room temperature (RT). Plates were stabilised, dried and stored over silica gel at 4°C until use. Test and control samples (50 #l/well) were added to appropriate wells and incubated either alonc, with 50 /tl/well diluent, or with diluent supplemented with blocking serum at various concentrations, for 1 h at RT. After washing the plates six ~imes with PBST, 100/~l/well of horseradish peroxidase (HRP)-conjugated IFN-2 (prepared according to the method of Wilson and Nakane, 1978) in diluent were added and incubated for I h at RT. Plates wcF¢ washed as above and developed using 100 ~l/well of tetramethylbenzidine (TMB) substrate (Bus et al., 1981) for 30 rain at RT. The reaction was stopped by the addition of 50 ~l/well of 0,5 M H2SO4 and the optical density (OD) at 450 nm of each well was determined using an automated EIA plate reader (Titertek Multiscan MCC). Samples that generated an OD greater than twice that of diluent alone (i.e., > 0.100) were deemed positive. Results

Frequency of false.l~Silive reactions Plasma samples from healthy cattle were tested by EIA to determine whether IFN-3, was detocta~le in their circulation. Of 191 samples, 81

% Illew.tlon In OD i

0O

4o ~7 0

i0

1

2

% Nmmal ~

5

10

i 20

i so

Serum Iv,tv)

Fig. 2~, Determination of optimum blocking concenlration of NMS in 10urine IFN-3, EtA. Mean per cent reduction in OD ~ standard deviation of eight false-I~Ositive nounal bovine p]esma samples diluted 1/2 in wells coated with IFN-9 F~ab') 2 antiM), fragments.

(42,2%) p~oduced OD readings greater than 0.100 when tested undiluted in E1A wells coated with intact IFN-9 monoclonal antibodies, When plasmas were diluted 1/2 in diluent the number of reactive plasmas was reduced to 63 (33,0%) (Fig, la). To determine whether these reactions were specific or due to non-specific factors bridging the two monoclonal antibodies used in the EIA, IFN-9 F(ah') 2 antibody fragments were used to coat wells. This treatment reduced the number of undiluted plasmas giving positive reactions to 8 (4,2%), suggesting that an abso~tion/diluent step wouldbe required, These 8 plasmas were used to determine the optimum concentration of NMS to add to diluent to minimise the level of anti-F(ab') 2 reactivity in normal bovine plasma (Fig. 2). Samples were diluted 1/2 in diIuent alone or in diluent containing various concentrations of NMS and assayed in F(ab')ycoated wells. Concentrations of 5% (v/v) NMS and greater were found to reduce the OD generated by each of the eight plasmas to less than 0.100. Dilution in the absence of NMS reduced the OD of only one of these eight plasmas to below the cut-off (Fig, lc). The OD of all 191 normal bovine plasma samples was similarly reduced to below the cut-off when diluted 1/2 in diluent containing 5% NMS and tested in F(ab')2-coated wells (Fig. ld). When the plasma

237 OlD 4 ! m a n

samples were diluted 1/2 with diluent containing 5% NMS without the use of F(ab') 2 fragments of

iFN-9 to coat the wells 49 (25.7%) of plasma samples were still reactive (Fig. lb),

2.5

2.0

Species speei~city of felse-positive reactions The possibility of substituting mouse serum in the diluent with that from a larger species was investigated using five strong false-I~sitive bovine plasmas and two samples containing bovine IFN-y (Table I). Samples were diluted 1/2 with diluent supplemented with 5% serum and tested in IFN-9 F(ab')2-coated wel|s. Of the t6 species' sera tested, only NMS was able to eliminate nonspecific reactivity although, interestingly, rat and go~e serum removed the reactivity from one sample. The addition of 5% serum to the diluent had little effect on the OD of the samples conraining bovine IFN-?. Serum pools from six strains of mice were also tested to assess more accurately the specificity of the factors generating these false.positive reactions. Sera from BALB/c, DBA/2, C3H/HeJ, CBA/CaH and Swiss mice were found to inhibit false reactions equally, while the C57BL/6J serum pool was effective with only four of the five false.po3itive samples (Table 1). Therefore, all subsequent investigations were performed using IFN-9 F(ab') 2 antibody frag. moots coated onto EIA wells and sample diluent supplemented with 5% NMS derived from Swiss mice. The diluent (50 #l/well) was added to wells prior to the addition of sample (50 p.1/well).

1,6

1.9

O,!

I

Q.O

0

~1~

4]'11

0"32

0"413

%2S

2.S

~

5.0

fez,smeue~ Im~kmm4~edl Fig. 3. Sensit~.vlt~of bovineIFN,-~ EIA for the ~le.~iO~Of recoml~inantbovineIFN.y in diluent, Mean OD±standard deviationof 42determinations.

program (Titersoft !I Version 3.1B, Flow Laboratories, Rickm~sworth, UIO. The mean concentration :1: standard deviation of natural bovine IFN~ in the positive control plasma was 2.42 :t: 0.25 ng/ml. The mean intraassay e,acfficient of variation (CV) was 5.7 :L 2.7% (range 2.5-12.7) for the 19 plates. The iuterassay CV was 7.3% and the overall assay CV for each determination was 10.2%+

N

Optimised bovine IFN-7 EIA Fig. 3 shows the mean OD + standard deviation of 42 separate determinations of various concentrations (up to 5 ng/ml) of recombinant bovine IFN-7 in diluent. The assay was able to detect consistently 0.08 nf~/ml IFN-~+ The reproducibility of the optimised bovine IFN.y EIA was estimated by assaying a positive bovine control plasma 104 times, with 2-8 repli. cares on each of 19 plates tested over a period of ten months (Fig. 4). The concentration (ng/ml) of natural bovine 1FN-y in each replicate p|asm~ was calculated from its OD in relation to that of various levels of recombinant bovine IFN-T (0.08-5.0 ng/ml) using a curve fitting computer

eommml~ mlq+Iml~ Fig+4. Reproducibilityol"l~ov[n¢IN-'y EIA for the deleClkm of natural bovine IF'~+I, in one positS+ conlrM ~imp~ tinted 104 limes in 19 plates over ]0 months. Con~fltralioil c~]¢uluted in relation to slandard curves of recombiqanl ~il1~ IFN.y.

2~

Discussion This paper describes the optimisation of an EIA for the detection of bovine IFN-T in plasma samples. The EIA originally described by Rothel et al. (1990), was developed to replace a bioassay that could be used to diagnose tuberculosis in cattle by measuring specific celI mediated immune responses to PPD stimulation in vitro (Wood eta|., 1990a), Compared with the bipassay, the EIA was less tedious and time consuming to perform, more sensitive and specific, and amenable to routine use (Rothel et al., 1990). In this study we report that a high frequency (42.4%) of unstimulated, undiluted normal bovine plasmas reacted positlvely in the bovine IFN-~, EIA, which dearly indicated the whole test system would be unworkable using the EIA in its original format. Such a level of background IFN-~, was also contrary to other published work (Andersson et al., 1989; Pisa et al., 1990) and suggested that at least some of these reactions were non-specific. ]t was necessary to remove the non-specific reactivity to maxirpise the utility of the IFN-,), assay system for the diagnosis of bovine tuberculosis and this needed to be achieved without sacrificing the very high sensitivity of the bovine IFN-T EIA, obtained using a combination of minimum sampie dilution and the TMB-HRP enzyme chromogen detection system (Bos et al,, 1981; Heytman, 1988; Roberts et at., 1991). Similar n(m-spocific interference in two-site sandwich EIAs has been reported to be a major problem in a numbei" of human (Boscato and Stuart, 1986, t988; Thompson and Hewitt, 1986; Thompson et al., 1986; Andersson et al,, 1989; Bocrman el al., 1990) and veterinary (Lopez and Jacobsen, 1989) diagnostic tests and is believed ~o be due to the presence of heterophU antibodies binding to both the eapt,re and detector antibodies in the assay, mainly through epitopcs in the Fc region of the antibodies. While such heterophil antibody interference can usually be eliminated by the inclusion of large amounts of absorptive, non.immu~e serum in the diluents or the use of antibodies that have he.d their Fc portions en~mically relnoved (reviewed by Boscato and Stuart, 1988; Spencer, 1988; Hamilton, 1989; Gosling, IS00), neither treatment alone was able to eliml-

note totally the background interference in our assay. These findings are in contrast to those of most others who found for immunoassays using murine monoclonal antibodies, that the addition of up to 25% NMS to test diluents or samples removed false-positive reactions without the need to use F(ab') 2 antibody fragments (Thompson and Hewitt, 1986; Thompson et al., 1986; Lopez and Jacobsen, 1989; Boerman et at., 1990). In a similar EIA for human IFN-% Andersson et al. (1989) reported that the use of 1% NMS was necessary to inhibit the false binding of serum IgG. This study and that of Boseato and Stuart (1986) examined considerably larger numbers of samples and showe,d that the addition of NMS dlrcctly to serum or plasma was not always totally effective in removing interference in an immunoassay that used intact monoclonal antibodies. This confirmed the need to use a combination of both treatments in some assays to ensure maximum specificity. The alternate approach of using F(ab') z fragments of monoclonal antibodies has similarly been shown to abrogate totally heterophil interference from human (Sakai et al., 1988) and bovine (AbduI-Ahad and Gosling, 1987) plasma. For example, Ida et al. (1990) found no interference from serum or plasma components in an EIA for interleukin-6 when F(ab') 2 monoclonal antibody fragments were used. However, these authors tested only 12 samples, which may have been an insufficient number to detect such interference given ,-. frequency of around 4%. Nevertheless, coating the solid phase with F(ab') 2 antibody fray,merits was by far the best single treatment in our system but the combination of the two was required for complete removal of interference by plasma-borne anti-mouse antibodies. Our requirement for both treatments may also reflect other methodological differences and minimal sample dilution (Nahm and Hoffman, 1990). When heterophil antibodies are present in serum or plasma they are b e l l i e d to possess a broad specificity, binding to antibodies from a variet~ of species (Tho~npson et al., 1986; Clark and Price, 1987; Boscato and Stuart, 1988; Lopez and Jacobsen, 1989; Boerman et al., 1990), usually to the Fe portion. A few samples may have other specifieities directed against epitopes resid-

2.~ ing on the F(ab') 2 fragment of lgG that ~re also common to several species (Boseato and Stuart, 1988), In this study we have found similarly that most falsely reactive bovine plasmas bind to the Fc portion of mouse lgG (73 of 191). A small number (eight of 191) also reacted with the F(ab') 2 fragment of mouse IgG and of these eight, the reactivity of only one could be removed with rat or goose serum, while that of the others could not be eliminated by absorption with nonimmune sera from any source other than mouse and would therefore seem peculiar to murine IgG. We also showed that one particular bovine plasma contained anti-mouse antibodies of exquisite specificity, heir.: ab'~orbed eu~. en!y by NMS from strains of mice that share the same IgG1 allotypic markers (Green, 1989). Boerman et al. (1990) also noted that hetcroantibody interference was best minimised when serum and reagent antibody were from the same mouse strain. Taken together these data suggest that it is good laboratory practice when using sandwich EIAs for measuring cytokines or other analytes in plasma or serum to include in assay diluents non-immune immunoglobulin which is allotypically as well as isot~pically matched in order to avoid interference from those rare samples that display quite exquisite specificity. Our finding that greater than 99.5% of normal cattle have undetectable levels of IFN-? in their blood is consistent with data obtained from human clinical studies where IFN-.-y was undetectable in serum of normal healthy controls (Andersson et at., 1989; Pisa e t a l , 1~0) and suggests that in the normal state, 1FN-,/ is not present in the circulation at levels detectable by EIA, This is of importance since the diagnostic usefulness of the test system for bovine tuberculosis depends on its ability to detect IFN-y produced by blood lymphocytes after specific stimulation with PPD (Rothel etal., 1990, 1992). In conclusion, the accurate determination of IFN-y in bovine plasma samples requires a combination of IFN-9 F(ab') 2 antibody fragments to coat EIA plate wells and sample diluent supplemented with 5% allotypieally matched NMS. The use of this optimised EIA in conjunction with a whole blood culture system (Ruthel etal., i~'~21

has resnlted in a rapid, sensitive and highly specific in vitro diagnostic test for bovine tuberculosis (Wood ct al., 1991, 1992). The bovine. IFN-'I, EIA (now commercially available from CSL) itself is also a highly sensitive, reproducible and convenient alternative to lymphocyte proliferation for monitoring cell-mediated immune responses in cattle, sheep, goats and buffalo (Rothel etal., 10t)0), and has proven to be of similar value for the diagnosis of other disease states in these animals (Billman-Jacob¢ ¢ta[., 1992). The measurement of IFN-y produced in in vitro cultures has also been useful in the assessment of immunological responses to potential vaccines and adjuvants (Emery ct al., 19thl), as well as in the analysis of T cell antigens/epitopes (BillmanJacobc c t a l . , 1990).

Acknowledgements We gratefully thank Mr, BiLl Smith, Gilberts,3n Wholesale Butchers, Brooklyn, Victoria, and Mr. Jack Porter, Department of Agriculture and Rural Affairs, Victoria, Australia, for the supply of normal bovine blood, Dr. Tony James and the Animal Services staff for their assistance with the collection of sera and Mr, Charles Quinn for the generation of ascitie fluid.

ReFerences Anderson, G., Ekrc, lt.-P.. Aim, G. and Per]mann. P. (1989)

Monuclonal antibody two-siteELISA for human IFN-7. Adaptation for delerminatil~n.sin human,serumor plasmn. J, Immunol. Methods t25, 89. Abdu]-Ahad,W.G.and Gosling,J.P. (1987) An enzyme-linked immunnsorbentassay (ELISA)for bovine Ltt capable of monitoringfluctuationsin baselineconcentrations.J. Reprod. Fort. 81), 653.

Billman-Jacobe. It., Radford.A.J,, Rothcl. J.S. and Wood. P.R. (1990) Mapping of T and B cell cpitnpes of the Mycobacterium bocis protein. MPB70.Immune1.Cell Biol. 68, 359. Billman-Jacobe.H., Carrigan, M.. Cockram.F.. Corner. L.A.. Girt. I.J., Hill. J,F., ,[essup, T., Miln~r, A.R. and Wood, P.R. (1992) A comparison of the interferon gamma assay

with Ihe absorbed ELISA for the diagnosisof Johne's di~ase in cattle. Aust. Vet, J. 69. 25. Bocrman, O.C.. Segcr~ M,F.G., PoeL~, L.G,, Kenemans. P.

and Thomas. C.M.G. (1990) Ileterophilic antibodies in human sera L-ausingfairly increasedresultsin the CA 125 immunolluo~metric ashy. Cite. Chem. 36, t~S,

240 Bus, E.S.. Van tier Doclen, A.A., V~n Rooy, N. and Schuurs, A.H.W.M. (t981) 3,3',5,5'.Tetramethy]heuzidine as an Ames test negative chromogan for horse-radish peroxidase in enzymc-immunoassay. J, Immunoassay 2, 187. Boscato. L.M. and Stuart, M.C. (1986) Incidence and spccifzci~ of interference in two-site immunoassays. Oin. Chum. 32. 1491. Boxcato, LM. and Stuart, M.C. (1988) Heterophilic antibodies: a problem for all immunoassays, Clin. Chem. 34, 2Z Clark. P.M.S. and Price, C.P, (1987) Removal of interference by immrmoglobulins in an enzyme-amplified immanoa.sszy for lhyrotropin in serum. Clio. Chem. 33, 414. Emery, D.L.. Rothcl. J.S. and Wood, P.R. (I¢F)0) Influence of antigens and adjuvants on the production of gamma-interferon and antibody by urine lymphozylcs. Immuno]. Cell Biol. 68, 127. Gosling, J.P. (1990) A decade of development in immunoassay methodology. Clin. Chem. 36. 1408. Green. M.C. (i 989) Catalog of mutant genes and polymorphic loci. in: M.F. Lyon and A.G. Searle (Eds,), Genetic Variants and Strains of the LahoraloqJ Mouse. Oxford Unicef. sily Press, Oxford, p. 12. Hamilton, R.G. (1989) Rheumatoid factor interference in immunological melbca:ls. Manner. Allergy 26, 27, Heytman, MJ. (19118) En~me-coojugates and subslrates. In: G.W. Burgess (Ed.), EUSA Technology in Diagnosis and Rcs.v.arch. James Cook University of Northern Queensland, Townsville, p. 63. Ida, N., Sakurai, S., Hosaka T., Hosoi, K., Kunitomo, T., Malsuura, Y. and Kohase, M. (1990) An enzyme-linked immunosnrbent assay for the measurement of human interlcukin-6. J. immunol. Methods 133, 2W. Lopez, N.A. and Jacobs0n, R.H. (1989) False-positive runelions associated with anti-mouse activity in serolests for feline leukemia virus antigen. J. Am. Vet. Mud. Assoc. 195, 741. Nahm, M.H. and Hoffman, J.W. (1990) tleleroantibody: phantom of the immunoassay. Clio. Chem. 36, 829. Pisa, P., Stenke, L, Bemcll, p., Hansson, M. and Hart, R. (1990) Tumor necrosis factor-alpha and interferon-gamma in serum of multiple m~loma patients, Anticancer Res. 10, 81'7, Roberts, LM.. Jones, S.L., Premier. R,R. and Coo, J.C. (lq9l) A comparison of the sensitivity and specificity of enzyme immunoassays and time-resolved fluoroimmunoassay. J. lmmunol. Methods 143, 49. Rmhe[, J,S., Jones, S.L., Corner. L.A., COx, J,C, and Wood, P.R. (1~)0} A sandwich enzyme immunoassay for bovine intcrferon-~ and its use for the detection of tuberculosis in cattle. Aust. Vet. J. 67, 134. Roth¢l, J.S.. Jones. S.L.. Corner, LA., Cox, .LC, and Wood,

P.R. (1992) The gamma-interferon assay for diagnosis of bovine tuberculosis in cattle: conditions affecting the production of gamma-interferon in whole blood culture. Aust, Vet. J. fig, L SakaL Y., Maeda, M., Takei, F,, Matsumoto, 1"., Nisbi.iimu, Y. and Nakamura, K. (1988) Determination of FDP in human plasma by a novel latex immunoassay. Thrombosis Res. 50, 469. Spencer, T.L (1988) Blockers, diluents and anomalous reactions in ELISA~ In: G.W. Burgess (Ed.), ELISA Technology in Diagnosis and Research. James Cook University of Norlhern Queensland, Townsville, p. 51. Thompson, R.E. and Hewiu, c . g . (1986) Screening for human antibodies thal interfere in monoctonal antibodybased assays. Clio. Chem. 32, 1424. Thompson, R.J,, Jackson, A.P. and Langlois, N. (1986) Circulating antibodies to mouse monoelonal immunnglobulins in normal subjects - incidence, species specificity, and effects on a two-site assay for ematine kinase-MB isocnzyme. Clio. Chem. 32, 476. Waroo, M., Trainin, Z., Gluckmau, A. and Ungnr-Waron, H. (1987) Reactivity of IgM RtJeumatoid factors of human, feline and bovine sera, J. Camp. Pathol. 97, 487. Whither, J. and Inshore, E. (1990) C-'~tokine measurements in body fluids. Eur. Cytokine Netw. 1,239. Wilson, M.B. and Nakane, P,K. (1978) Recent developments in the periodate method of conjugating horseradish peroxidase (HRPO) 1o antibodies. In: W. Knapp, K. Hnlubar and G. Wick (Eds.), Immunafluorescence and Related Staining Techniques, Elsevier, Amsterdam, p. 215. Wood, P.R., Corner, L.A. and Plackett, P. (19t.~a) Development of a simple, rapid in vitro cellular assay for bovine tuberculosis based on the production of 7 iulcrferon, Res. Vet. Sci. 49, 4f~. Wood, P.R., Rott~el, J.S,, McWalers, P.G.D. and Jones, S.L. (1990b) Prod,~¢lion and cbaracterizalion of monoclonal antibodies Sl~cific for bovine gamma-interferon. Vet, lmm,Jnol. Immqlnopalhol. 25, 37. Wood, P.R,, C~:mer, L.A,, Rothe], J.S., Baldock, C., Jones, S.L., Cousin.~q, D.B,, McCormick, B.S., Francis, B.R., Creeper, J. m,d Twaddle, N.E. 0991) Field comparison of the inter:~,,~;l ~ m m a a~ay and the intradetmal tuberculin test for ,,'in diagnosis of bovine tuberculosis. Aust. Vet. J. 68, 286. Wood, P,R., Corner, L.A., Rothel, J.S., Ripper, J.L., Fifis, T., McCormick, B,S., Francis, B., Melville, L., Small. K., De Wittc, K., Tolson, J., Ryan, T J., De Lisle, G.W., Cos, J.C and Jones, S.L. (1992) A field evalualiou of serological and cellular diagnostic tests for bovine tuberculosis. Vel. Micro. 31, 7I.

Removal of false-positive reactions from plasma in an enzyme immunoassay for bovine interferon-gamma.

A monoclonal antibody-based sandwich enzyme immunoassay (EIA) for bovine interferon-gamma (IFN-gamma) has been developed and can be used in conjunctio...
489KB Sizes 0 Downloads 0 Views