- Email: [email protected]
Testing of bovine sera by ELISA for IgG, IgM and IgA rheumatoid factors
Veterinary Immunology and Immunopathology 61 Ž1998. 239–250
Testing of bovine sera by ELISA for IgG, IgM and IgA rheumatoid factors D.A. Graham
a,)
, K.A. Mawhinney a , B.M. Adair a , M. Merza
b
a
Veterinary Sciences DiÕision, Department of Agriculture for Northern Ireland, Stoney Road, Stormont, Belfast BT4 3SD, UK b National Veterinary Institute, SVANOVA Biotech, Glunten, Uppsala Science Park, Uppsala S-751 83, Sweden Accepted 6 October 1997
Abstract Sera from 19 colostrum-deprived calves less than 1 week old, 24 colostrum-supplemented calves less than 1 week old, 36 3–5-month-old calves and 200 females greater than 9 months of age were tested by ELISA for the presence of IgM, IgG and IgA rheumatoid factors ŽRF.. An increasing level of IgM- and IgG-RF with age was found. IgG-RF levels in the colostrum-supplemented calves were significantly higher than in the non-supplemented calves Ž p - 0.001.. Individual IgG-RF values correlated with serum IgG levels, as determined by zinc sulphate turbidity testing Ž r s 0.59, p - 0.01.. No IgA-RF was detected. The cross-reactivity of IgM-RF with heterologous IgG was found to be greatest with rabbit IgG, followed by mouse and chicken IgG. The significance of rheumatoid factors in relation to diagnostic testing is discussed. q 1998 Elsevier Science B.V. Keywords: Bovine; ELISA; Rheumatoid factors
1. Introduction Rheumatoid factors ŽRFs. were initially described in human sera by Waaler in 1939 and subsequently by Rose in 1948 Žreviewed by Kahn, 1995., and they have been Abbreviations: BABG: Biotinylated anti-bovine IgG monoclonal antibody; BABM: Biotinylated anti-bovine IgM monoclonal antibody; COD: Corrected optical density; DMSO: Di-methyl sulphoxide; EPO: Extravidin peroxidase; GAMIgGŽFc.PO: Peroxidase conjugated goat anti-mouse IgGŽFc.; GARIgGPO: Peroxidase conjugated goat anti-rabbit IgG; 2-ME: 2-Mercaptoethanol; MAb: Monoclonal antibody; OD: Optical density; PBS: Phosphate-buffered saline; RFŽs.: Rheumatoid factorŽs.; RABIgGPO: Peroxidase-conjugated rabbit anti-bovine IgG antiserum; sd: Standard deviation; ZST: Zinc sulphate turbidity ) Corresponding author. Tel.: q44 1232 525749; fax: q44 1232 525787; e-mail: [email protected] 0165-2427r98r$19.00 q 1998 Elsevier Science B.V. All rights reserved. PII S 0 1 6 5 - 2 4 2 7 Ž 9 7 . 0 0 1 5 1 - 7
240
D.A. Graham et al.r Veterinary Immunology and Immunopathology 61 (1998) 239–250
extensively studied and characterised since. They are auto-antibodies of IgM, IgG and IgA isotypes associated with, but not exclusive to, patients with rheumatoid arthritis ŽTeitsson and Valdimarsson, 1984; Tuokko, 1984.. Serum RFs have a specificity for the Fc region of IgG1 , IgG2 and IgG4 , Žbut not IgG3 . ŽNormansell and Young, 1975., and exhibit a variable degree of cross-reactivity with IgG of other species ŽButler and Vaughan, 1965.. It has been shown that RFs share a common binding site on IgG with staphylococcal protein A ŽOppliger et al., 1987., and further mapping has identified the cg 2–cg 3 interface as containing the major antigenic determinants of polyclonal IgM-RF ŽSasso et al., 1988; Peterson et al., 1995.. It has been suggested that RFs represent anti-idiotype antibodies to bacterial and viral Fc receptors ŽMouritsen, 1986; Oppliger et al., 1987; Raeder and Boyle, 1991.. This is consistent with the Jerne network hypothesis, which proposes the existence of a series of immuno-regulatory idiotype–anti-idiotype networks ŽRoitt, 1991.. In comparison with the published literature on human RFs, relatively little information is available relating to species of veterinary interest. Rheumatoid factors have been reported in feline ŽUngar-Waron et al., 1985., canine ŽHalliwell et al., 1989; Neilsen, 1992. chicken ŽRaeder and Boyle, 1991; Sugii, 1994. and equine serum ŽCarter et al., 1995.. IgM-RF has been detected in cattle serum ŽUngar-Waron et al., 1985. and its reactivity with heterologous IgG reported ŽWaron et al., 1987.. IgM-RF levels in cattle have been shown to increase with age ŽUngar-Waron et al., 1991a., but be unaffected by infection with bovine leukaemia virus or use of a Babesia boÕis vaccine. Bovine sera have not previously been evaluated for the presence of IgG- or IgA-RF. In this paper, we describe the testing of a large number of bovine serum samples for IgG-, IgA- and IgM-RF, and examine the influence of age and maternally derived immunoglobulins on the levels detected. The implications of bovine RF for serological testing are discussed.
2. Materials and methods 2.1. BoÕine sera Four panels of bovine sera were examined. These consisted of sera from 19 colostrum-deprived calves ŽCD panel. and 24 colostrum-supplemented calves Žnon-CD panel. less than 1 week old, 36 3–5- month-old calves Ž3–5-month panel. and 200 heifers and cows, aged 9 months and older Ž) 9-month panel.. The latter panel was assembled by selecting 10 sera from each of 20 herds submitted for Brucella serology testing. The immunoglobulin levels in the sera of the CD and non-CD panels were measured by the zinc sulphate turbidity ŽZST. test, as previously described ŽMcEwan et al., 1970.. Briefly, 50 m l volumes of each serum were incubated at room temperature with 3 ml of zinc sulphate solution Ž208 mgrl in distilled water. and 3 ml distilled water, and the absorbance at 623 nm read after 1 h. Using a standard curve, the corrected absorbance of each serum was then expressed in ZST units Ž1 unit s 1 mgrml..
D.A. Graham et al.r Veterinary Immunology and Immunopathology 61 (1998) 239–250
241
2.2. Reagents 3X ,3X ,5X ,5X-tetramethylbenzidine ŽTMB., bovine, mouse, rabbit and chicken IgG, biotinylated antibovine-IgM ŽBABM. and -IgG ŽBABG. monoclonal antibodies, extravidin peroxidase ŽEPO., and peroxidase-conjugated goat anti-rabbit IgG ŽGARIgGPO. and anti-mouse IgGŽFc. ŽGAMIgGŽFc.PO. were purchased from Sigma-Aldrich, Poole, UK. 2.3. ELISA testing All ELISAs were performed on Immulon 1 flat-well microtitre plates ŽDynatech Laboratories, Chantilly, VA.. With the exception of stop solution Ž50 m lrwell., all volumes were 100 m lrwell. Plates were coated overnight at 48C, using 0.05 M carbonaterbicarbonate ŽpH 9.6. as coating buffer. The concentration of all immunological reagents, including coating immunoglobulins, was optimized prior to use. All dilutions were made in PTN Ž0.01 M PBS containing 2% wrv NaCl and 0.05% vrv Tween 80., and wells were washed with PBS-Tween Ž0.01 M PBS containing 0.05% wrw Tween 20. between each incubation. Substraterchromogen solution was prepared by adding 9 ml phosphate–citrate buffer ŽpH 5.0. and 2 m l 30% H 2 0 2 to 1 mg TMB dissolved in 1 ml DMSO. All incubations were done for 1 h at 378C, with the exception of the chromogenrsubstrate solution, which was incubated for 10 min before the reaction was stopped with 2.5 M H 2 SO4 . After stopping, plates were read at 450 nm on a Titertek Multiskan MCC plate reader ŽLabsystems, Helsinki, Finland.. 2.4. IgM-RF ELISA Serum samples were tested in duplicate at dilutions of 1r100 and 1r800 on plates coated with bovine IgG at 2 m grml. Bound IgM-RF was detected by successive incubations with BABM Ž1r35 000. and EPO Ž1r4000.. Controls on each plate consisted of duplicate wells to which PTN was added in place of serum ŽPTN control., or in which the serum and BABM liquid phases were replaced with PTN, and peroxidase-conjugated rabbit anti-bovine IgG ŽRABIgGPO. ŽNordic Immunology, Tilburg, Netherlands. Ž1r4000. was substituted for EPO Žcoating control.. The mean optical density ŽOD. of each serum at 1r100 and 1r800 was calculated and the corrected OD ŽCOD. obtained by subtracting the mean OD of the PTN control wells. 2.5. Pre-treatment of serum samples with 2-mercaptoethanol (2-ME) Pre-treatment with 2-ME was performed by incubating equal volumes of undiluted serum and 0.2 M 2-ME for 1 h at 378C. Twenty sera from the ) 9-month group were pre-treated in this way and re-tested by IgM-RF ELISA, as described above. 2.6. IgA-RF ELISA This followed the IgM-RF protocol, except that sera were tested at dilutions of 1r20 and 1r100, and the detection phases consisted of antibovine IgA MAb ŽVMRD, Pullman, WA. Ž1r200., followed by GAMIgGŽFc.PO Ž1r8000..
242
D.A. Graham et al.r Veterinary Immunology and Immunopathology 61 (1998) 239–250
2.7. Selection of heterologous coating IgG for IgG-RF ELISA Microtitre plates were coated in four blocks with bovine, rabbit, mouse and chicken IgG at 2 m grml, and 17 sera from the ) 9-month panel that gave a range of COD values when tested by IgM-RF ELISA were selected. These were then tested in duplicate at a dilution of 1r100 against each of these coating immunoglobulins, following the IgM-RF ELISA protocol. The heterologous IgG giving the highest degree of cross-reactivity, relative to bovine IgG, was selected and used as coating antibody for the IgG-RF ELISA Žrabbit IgG, see Section 3.. 2.8. IgG-RF ELISA Serum samples were tested in duplicate at dilutions of 1r20 and 1r100. Bound IgG-RF was detected using BABG Ž1r50 000.. GARIgGPO Ž1r5000. was substituted for RABIgGPO in the coating control wells. Otherwise, the protocol was as described for IgM-RF.
3. Results 3.1. ZST testing The ZST values of the CD panel ranged from 0–0.9 units Žmean s 0.2, standard deviation Žsd. s 0.3., confirming the colostrum-deprived status of this panel. In contrast, the ZST values of the non-CD panel ranged from 2–32 units Žmean s 14.5, sd s 8.7.. A ZST value of 20 units is considered to represent an adequate intake of colostrum ŽDr. D. Mackie, personal communication.. 3.2. IgM-RF ELISA The results of testing the serum panels at serum dilutions of 1r100 and 1r800 by IgM-RF ELISA are summarised in Table 1. An increase in the mean and range of IgM-RF COD values with age was detected when sera were tested at a dilution of 1r100 ŽFig. 1.. At this dilution, the differences between panels was significant
Table 1 Mean"one standard deviation of IgM-RF and IgG-RF ELISA COD values of 4 serum panels Panel
na
IgM-RF 1r100
CD Non-CD 3–5 months )9 months a b
19 24 36 200
b
0.043"0.025 0.013"0.011 0.173"0.120 0.592"0.287
Number of sera in each panel. Dilutions at which sera were tested.
IgG-RF 1r800
b
0.031"0.016 0.003"0.008 0.024"0.021 0.124"0.145
1r20
1r100
0.008"0.015 0.323"0.322 0.123"0.077 0.398"0.316
0.006"0.011 0.109"0.099 0.046"0.032 0.108"0.127
D.A. Graham et al.r Veterinary Immunology and Immunopathology 61 (1998) 239–250
243
Fig. 1. IgM-RF ELISA COD values for individual sera from 4 serum panels tested at a serum dilution of 1r100 Ž ns19 ŽCD., 24 Žnon-CD., 36 Ž3–5 months. and 200 Ž )9 months... Horizontal dashed lines indicate mean COD of each panel. Horizontal dotted lines indicate mean COD of each panel q1 sd Žnot shown for CD and non-CD panels..
Ž p - 0.001. in all cases. A reduced mean COD value for each panel was observed at a serum dilution of 1r800 ŽTable 1.. This reduction was significant Ž p - 0.001. for all except the CD panel Ž p s 0.81.. The mean signal in the coating and PTN control wells were 1.556 Ž"0.347. and 0.064 Ž"0.017., respectively. 3.3. Pre-treatment with 2-ME The mean IgM-RF COD values Ž"1 sd. of the 20 sera when tested at serum dilutions of 1r100 and 1r800 were 0.806 Ž"0.329. and 0.267 Ž"0.153., respectively. Pre-treatment with 2-ME significantly reduced the mean COD at both serum dilutions Ž p - 0.001. to 0.081 Ž"0.066. and 0.025 Ž"0.012., respectively.
244
D.A. Graham et al.r Veterinary Immunology and Immunopathology 61 (1998) 239–250
3.4. Selection of heterologous coating IgG for IgG-RF ELISA The mean IgM-RF COD of the 19 sera from the ) 9-month panel was greatest on wells coated with bovine IgG, followed by rabbit, mouse and chicken IgG ŽFig. 2.. The mean COD with bovine IgG was significantly greater Ž p - 0.001. than the mean value observed with any of the heterologous IgGs. The mean COD detected on rabbit IgG-coated wells was significantly greater than on chicken IgG-coated wells Ž p - 0.01.. The differences in the mean CODs between rabbit and mouse IgG and between mouse and chicken IgG were not significant Ž p s 0.081 and p s 0.336, respectively.. On the basis of these results, rabbit IgG at 2 m grml was chosen for coating IgG-RF ELISA plates.
Fig. 2. Mean IgM-RF corrected optical density ŽCOD. values of 19 sera from )9-month panel, tested in duplicate on wells coated at 2 m grml bovine, rabbit, mouse and chicken IgG. Horizontal dashed lines indicate mean COD of each panel. Horizontal dotted lines indicate mean COD of each panel q1 sd. ) Coating immunoglobulin.
D.A. Graham et al.r Veterinary Immunology and Immunopathology 61 (1998) 239–250
245
3.5. IgG-RF ELISA The results of testing the serum panels at serum dilutions of 1r20 and 1r100 by IgG-RF ELISA are summarised in Table 1. Fig. 3 shows the IgG-RF COD values of the panels when tested at a serum dilution of 1r20. At this serum dilution, an increase in the mean and range of IgG-RF COD values was detected with age. The means of the CD
Fig. 3. IgG-RF ELISA COD values for individual sera from 4 serum panels tested at a serum dilution of 1r20 Ž ns19 ŽCD., 24 Žnon-CD., 36 Ž3–5 months. and 200 Ž )9 months... Horizontal dashed lines indicate mean COD of each panel. Horizontal dotted lines indicate mean COD of each panel q1 sd Žnot shown for CD panel..
246
D.A. Graham et al.r Veterinary Immunology and Immunopathology 61 (1998) 239–250
Fig. 4. Correlation between IgG-RF ELISA COD values at a serum dilution of 1r20 and ZST units of the 24 sera of the non-CD panel. Ž r s 0.59, p- 0.01..
and ) 9-month panels were lowest and highest, respectively. However, the mean of the non-CD exceeded that of the 3–5-month panel. No significant difference was detected between the non-CD and ) 9-month panels Ž p s 0.28.. All other inter-panel comparisons indicated significant differences Ž p - 0.001.. The results of testing at a serum dilution of 1r100 yielded similar results, although the mean values were significantly reduced in all cases Ž p - 0.01., with the exception of the CD panel Ž p s 0.61.. At this dilution, the means of the non-CD and ) 9-month panels were almost identical Ž0.109 and 0.108, respectively.. Otherwise, all differences between panels were again significant Ž p F 0.01.. The mean signal in the coating and PTN control wells was 2.200 Ž"0.362. and 0.094 Ž"0.060., respectively. Regression analysis, by the least squares method, showed that IgG-RF values Ž1r20. of the non-CD panel were correlated with ZST levels ŽFig. 4. Ž r s 0.59, p - 0.01.. No correlation was found by this method between IgG-RF Ž1r20. and IgM-RF Ž1r100. values of the ) 9-month panel Ž r s 0.05, p s 0.47.. 3.6. IgA-RF ELISA The mean COD values at a serum dilution of 1r20 were extremely low Ž- 0.006.. Similar results were obtained when 50 sera from the ) 9-month panel were re-tested
D.A. Graham et al.r Veterinary Immunology and Immunopathology 61 (1998) 239–250
247
using a different anti-bovine IgA MAb ŽSerotec, Oxford, UK.. The mean signal in the coating and PTN control wells was 1.780 Ž"0.256. and 0.043 Ž"0.004., respectively.
4. Discussion Bovine IgM-RF has been reported previously ŽUngar-Waron et al., 1985., but we believe that this is the first time that bovine sera have been assayed for IgG- and IgA-RF. The strength of a specific signal should decrease as the serum test dilution increases; therefore, sera were tested at two different dilutions on each ELISA. The serum dilutions used in each ELISA were based on those used successfully by others to detect human RFs by indirect ELISA ŽTeitsson and Valdimarsson, 1984; Jonsson et al., 1995.. The IgG-RF and IgA-RF ELISAs were performed at lower serum dilutions than the IgM-RF ELISA to compensate for the use of rabbit IgG as coating immunoglobulin in the former, and the anticipated effect of inter-isotype competition in the latter. Ungar-Waron et al. Ž1991a. examined a total of 41 sera ranging from 2–4 week- to 9–10 year-old bovines and reported increasing levels of IgM-RF with age. Our findings, based on a much larger number of samples, and including neonatal calves of defined colostral status, confirm these results. IgM and to a lesser extent, IgG, are susceptible to reduction by 2-ME ŽCapel et al., 1980.. Pre-treatment of sera with 2-ME produced a reduction in IgM-RF COD value of approximately tenfold at both serum dilutions, confirming the specificity of the signal detected by IgM-RF ELISA. Low levels of IgG- and IgM-RF were detected in the CD panel. Idiotype networks, including low-affinity IgM with IgG-specificity ŽRoitt, 1991., have been described in foetal mice. The existence of similar idiotypic networks in the bovine foetus may be postulated as the source of these antibodies. The levels of IgG-RF detected in the non-CD panel were significantly higher than those in the CD panel. This is expected, given the higher levels of serum IgG present in these calves, as indicated by the ZST test. The correlation of IgG-RF values with the ZST reading of each sample indicates that IgG-RF is transferred to the neonate as part of the overall colostral IgG load. If IgM-RF is found in the IgM component of colostrum in the same way as IgG-RF is found in the IgG component, it would be expected that IgM-RF would be detected in the serum of colostrum-supplemented calves, given that IgM and IgG are absorbed from colostrum with equal efficiency ŽKlaus et al., 1969.. The finding that the IgM-RF levels of the non-CD panel were lower than those of the CD panel by a small, but significant, amount was therefore unexpected. It may be that the much higher levels of serum IgG in the calves of the non-CD panel results in a greater percentage of IgM-RF being complexed with serum IgG ŽMoore et al., 1995., and therefore unavailable for reaction with solid phase IgG. The increasing levels of IgM-RF and IgG-RF with age are in line with expectations. If RF antibodies are indeed auto-anti-idiotype antibodies, as predicted by the Jerne network hypothesis, then levels would be expected to increase as the immune system matures, and is exposed to an increasing range of antigens. Similarly, if RF antibodies are anti-idiotype antibodies produced in response to infection with organisms expressing
248
D.A. Graham et al.r Veterinary Immunology and Immunopathology 61 (1998) 239–250
Fc-binding proteins, an increased exposure to such organisms Žand therefore RF levels. with age is again to be expected. The wide range of IgM- and IgG-RF COD values seen in the - 9-month panel most probably reflects the wide age range of the cattle represented. The decrease in IgG-RF levels between the non-CD and 3–5-month groups reflects the lower level of maternally derived IgG-RF in the latter. The design of the IgG-RF ELISA required a system that allowed bound IgG-RF to be differentiated from solid phase IgG. The use of rabbit IgG as coating antibody, and an anti-bovine IgG-specific MAb as detector, made this possible. The presence of IgM-RF in sera being assayed for IgG-RF represented a possible source of error. Due to its pentavalent nature, IgM-RF can potentially bind to solid phase IgG and subsequently also capture the detecting antibody, generating a false signal ŽTeitsson and Valdimarsson, 1984.. The absence of a signal in the IgA-RF ELISA, which was performed at the same serum dilution, and which also used an IgG1-isotype detector MAb, and the lack of correlation between IgG- and IgM-RF results, indicate that the IgG-RF signal was specific. The greater cross-reactivity of bovine IgM-RF with rabbit than with mouse IgG has been reported previously ŽUngar-Waron et al., 1985; Waron et al., 1987., and our findings are consistent with these reports. We believe that this is the first time that chicken IgG has been included in this type of evaluation. Ž1988. have shown that human RF does not react with chicken Larsson and Sjoquist ¨ IgG, and based on this finding, the use of chicken antibodies as capture- andror detector antibody can eliminate the interference of RF when human sera are tested by sandwich ELISA ŽLarsson et al., 1991.. Our finding of low level cross-reactivity in some chicken sera ŽFig. 2. suggests that the use of chicken antibodies may not always eliminate RF interference in bovine serology. Indirect ELISAs, similar to the one used in this study, have been used previously to detect IgA-RF in human sera ŽTeitsson and Valdimarsson, 1984; Jonsson et al., 1995.. It is therefore considered that the failure to detect IgA-RF reflects its absence from bovine sera, rather than a failure of IgA-RF to bind to solid phase antibody, due to inter-isotype competition with IgM- and IgG-RF. The association between RFs and arthritis in cattle has not been investigated, and rheumatoid arthritis is not recognised as a cause of lameness in cattle ŽBailey, 1985.. Therefore, the presence of RFs in bovine sera is most significant as a potential complicating factor in serological testing. As described above, RFs can cause a false signal in sandwich ELISAs by cross-linking solid phase capture- and liquid-phase detector-antibody in the absence of specific antigen. Non-specific reactions may also occur in indirect ELISAs designed to detect antigen-specific IgM, when the serum under test contains IgM-RF and antigen-specific IgG. In this case, IgM-RF bound to IgG that has reacted specifically with solid-phase antigen cannot be distinguished from a specific IgM-antigen interaction. This has been described in indirect ELISAs ŽUngar-Waron and Abraham, 1991b; Salonen et al., 1980; Tuokko, 1984., solid-phase radio-immunoassay ŽMeurman and Ziola, 1978. and indirect immuno-fluorescence ŽGispen et al., 1975; Hekker et al., 1979.. Removal of IgM-RF or antigen-specific IgG prior to testing are both effective methods of eliminating such reactions ŽLeinikki et al., 1978; Meurman and Ziola, 1978; Ungar-Waron and Abraham,
D.A. Graham et al.r Veterinary Immunology and Immunopathology 61 (1998) 239–250
249
1991b.. The use of reverse rather than indirect ELISAs also reduces the incidence of non-specific reactions ŽTuokko, 1984..
5. Conclusion We report the presence of IgM- and IgG-RF in bovine sera, with levels of both increasing with age. In addition, correlation between colostral status and IgG-RF levels in neonatal calves has been demonstrated. We highlight the importance of considering RFs as a possible source of non-specific reactions in veterinary serology, especially in relation to assays for antigen-specific IgM.
Acknowledgements This work has been partially funded by SVANOVA Biotech, Glunten, Uppsala Science Park, S-751 83 Uppsala, Sweden.
References Bailey, J.V., 1985. Bovine arthritides— classification, diagnosis, prognosis and treatment. In: Ferguson, J.G. ŽEd.., Symposium on Bovine Lameness and Orthopedics. The Veterinary Clinics of North America ŽFood Animal Practice., 1 Ž1. pp. 39–52. Butler, V.P., Vaughan, J.H., 1965. The reaction of rheumatoid factor with animal gamma-globulins: quantitative considerations. Immunology 8, 144–147. Capel, P.J.A., Gerlag, P.G.G., Hagemann, J.F.H.M., Koene, R.A.P., 1980. The effect of 2-mercaptoethanol on IgM and IgG antibody activity. J. Immunol. Methods 36, 77–80. Carter, S.D., Osborne, A.C., May, S.A., Bennett, D., 1995. Rheumatoid factor, anti-heat shock protein Ž65 kDa. antibodies and anti-nuclear antibodies in equine joint diseases. Eq. Vet. J. 27, 288–295. Gispen, R., Nagel, B., Brand-Saathof, B., De Graaf, S., 1975. Immunofluorescence test for IgM rubella antibodies in whole serum after absorption with anti-g Fc. Clin. Exp. Immunol. 22, 431–437. Halliwell, R.E.W., Werner, L.L., Baum, D.E., Newton, C.D., Wolfe, J.H., Schumacher, H.R., 1989. Incidence and characterisation of canine rheumatoid factor. Vet. Immunol. Immunopathol. 21, 161–175. Hekker, A.C., Brand-Saathof, B., Vis, J., Meijers, R.C., 1979. Indirect immunofluorescence test for detection of IgM antibodies to cytomegalovirus. J. Infect. Dis. 140, 596–600. Jonsson, T., Thorsteinsson, H., Arinbjarnarson, S., Thorsteinsson, J., Valdimarsson, H., 1995. Clinical implications of IgA rheumatoid factor subclasses. Ann. Rheum. Dis. 54, 578–581. Kahn, M.F., 1995. The history of rheumatoid factor. Revue de Rheumatisme 62, 513–517. Klaus, B.G.G., Bennett, A., Jones, E.W., 1969. A quantitative study of the transfer of colostral immunoglobulins to the newborn calf. Immunology 16, 293–299. Larsson, A., Sjoquist, J., 1988. Chicken antibodies: a tool to avoid false positive results by rheumatoid factor ¨ in latex fixation tests. J. Immunol. Methods 108, 205–208. Larsson, A., Karlsson-Parra, A., Sjoquist, J., 1991. Use of chicken antibodies in enzyme immunoassays to ¨ avoid interference by rheumatoid factors. Clin. Chem. 37, 411–414. Leinikki, P.O., Shekarchi, I., Dorsett, P., Sever, J.L., 1978. Determination of virus-specific IgM antibodies by using ELISA: elimination of false-positive results with protein A-sepharose absorption and subsequent IgM antibody assay. J. Lab. Clin. Med. 92, 849–857. McEwan, A.D., Fisher, E.W., Selman, I.E., Penhale, W.J., 1970. A turbidity test for the estimation of immune globulin levels in neonatal calf serum. Clin. Chim. Acta 27, 155–163.
250
D.A. Graham et al.r Veterinary Immunology and Immunopathology 61 (1998) 239–250
Meurman, O.H., Ziola, B.R., 1978. IgM-class rheumatoid factor interference in the solid-phase radioimmunoassay of rubella-specific IgM antibodies. J. Clin. Pathol. 31, 483–487. Moore, T.L., Osborn, T.G., Nesher, G., 1995. Immune complexes from sera of patients with juvenile rheumatoid arthritis reveal novel 40 and 60 kd bands. Clin. Exp. Rheumatology 13, 667–672. Mouritsen, S., 1986. Rheumatoid factors are anti-idiotypic antibodies against virus-induced anti-Fc receptor antibodies. Scand. J. Immunol. 24, 485–490. Neilsen, O.L., 1992. Detection of IgM rheumatoid factor in canine serum using a standardized enzyme-linked immunosorbent assay. Vet. Immunol. Immunopathol. 34, 139–147. Normansell, D.E., Young, C.W., 1975. The IgG specificity of anti IgG immunoglobulin rheumatoid factors. Immunochemistry 12, 187–188. Oppliger, I.R., Nardella, F.A., Stone, G.C., Mannik, M., 1987. Human rheumatoid factors bear the internal image of the Fc binding region of staphylococcal protein A. J. Exp. Med., 702–710. Peterson, C., Malone, C.C., Williams, R.C., 1995. Rheumatoid-factor-reactive sites on CH 3 established by overlapping 7-mer peptide epitope analysis. Mol. Immunol. 32, 57–75. Raeder, R., Boyle, M.D.P., 1991. Detection of rheumatoid-like factors in serum of chickens immunised with bacterial immunoglobulin binding proteins. J. Immunol. Methods 1991, 201–209. Roitt, M.I., 1991. Essential Immunology 7. Blackwell, Oxford, 356 pp. Salonen, E.-M., Vaheri, A., Suni, J., Wager, O., 1980. Rheumatoid factor in acute viral infections: interference with determination of IgM, IgG, and IgA antibodies in an enzyme immunoassay. J. Infect. Dis. 142, 250–255. Sasso, E.H., Barber, C.V., Nardella, F.A., Yount, W.J., Mannik, M., 1988. Antigenic specificities of human monoclonal and polyclonal IgM rheumatoid factors. J. Immunol. 140, 3098–3107. Sugii, S., 1994. Immunological cross-reactivity of mannan-binding proteins in bovine, chicken and human sera. J. Vet. Med. Sci. 56, 787–790. X Teitsson, I., Valdimarsson, H., 1984. Use of monoclonal antibodies and FŽab .2 enzyme conjugates for IgM, IgA and IgG rheumatoid factors. J. Immunol. Methods 71, 149–161. Tuokko, H., 1984. Comparison of non-specific reactivity in indirect and reverse immunoassays for measles and mumps immunoglobulin M antibodies. J. Clin. Microbiol. 20, 972–976. Ungar-Waron, H., Waron, M., Gluckman, A., Trainin, Z., 1985. Rheumatoid factors in natural retrovirus infections of the cat and cattle. J. Comp. Pathol. 95, 399–403. Ungar-Waron, H., Paz, R., Shkap, V., Hanna, B., Pipano, E., 1991a. IgM rheumatoid factor in cattle immunized against babesiosis. J. Vet. Med. B 38, 492–496. Ungar-Waron, H., Abraham, A., 1991b. Immunoglobulin M ŽIgM. indirect enzyme-linked immunosorbent assay and the involvement of IgM-rheumatoid factor in the sero-diagnosis of BHV-1 infection. Vet. Microbiol. 26, 53–63. Waron, M., Trainin, Z., Gluckman, A., Ungar-Waron, H., 1987. Reactivity of IgM-rheumatoid factors of human, feline, and bovine sera. J. Comp. Pathol. 97, 487–490.
Testing of bovine sera by ELISA for IgG, IgM and IgA rheumatoid factors D.A. Graham
a,)
, K.A. Mawhinney a , B.M. Adair a , M. Merza
b
a
Veterinary Sciences DiÕision, Department of Agriculture for Northern Ireland, Stoney Road, Stormont, Belfast BT4 3SD, UK b National Veterinary Institute, SVANOVA Biotech, Glunten, Uppsala Science Park, Uppsala S-751 83, Sweden Accepted 6 October 1997
Abstract Sera from 19 colostrum-deprived calves less than 1 week old, 24 colostrum-supplemented calves less than 1 week old, 36 3–5-month-old calves and 200 females greater than 9 months of age were tested by ELISA for the presence of IgM, IgG and IgA rheumatoid factors ŽRF.. An increasing level of IgM- and IgG-RF with age was found. IgG-RF levels in the colostrum-supplemented calves were significantly higher than in the non-supplemented calves Ž p - 0.001.. Individual IgG-RF values correlated with serum IgG levels, as determined by zinc sulphate turbidity testing Ž r s 0.59, p - 0.01.. No IgA-RF was detected. The cross-reactivity of IgM-RF with heterologous IgG was found to be greatest with rabbit IgG, followed by mouse and chicken IgG. The significance of rheumatoid factors in relation to diagnostic testing is discussed. q 1998 Elsevier Science B.V. Keywords: Bovine; ELISA; Rheumatoid factors
1. Introduction Rheumatoid factors ŽRFs. were initially described in human sera by Waaler in 1939 and subsequently by Rose in 1948 Žreviewed by Kahn, 1995., and they have been Abbreviations: BABG: Biotinylated anti-bovine IgG monoclonal antibody; BABM: Biotinylated anti-bovine IgM monoclonal antibody; COD: Corrected optical density; DMSO: Di-methyl sulphoxide; EPO: Extravidin peroxidase; GAMIgGŽFc.PO: Peroxidase conjugated goat anti-mouse IgGŽFc.; GARIgGPO: Peroxidase conjugated goat anti-rabbit IgG; 2-ME: 2-Mercaptoethanol; MAb: Monoclonal antibody; OD: Optical density; PBS: Phosphate-buffered saline; RFŽs.: Rheumatoid factorŽs.; RABIgGPO: Peroxidase-conjugated rabbit anti-bovine IgG antiserum; sd: Standard deviation; ZST: Zinc sulphate turbidity ) Corresponding author. Tel.: q44 1232 525749; fax: q44 1232 525787; e-mail: [email protected] 0165-2427r98r$19.00 q 1998 Elsevier Science B.V. All rights reserved. PII S 0 1 6 5 - 2 4 2 7 Ž 9 7 . 0 0 1 5 1 - 7
240
D.A. Graham et al.r Veterinary Immunology and Immunopathology 61 (1998) 239–250
extensively studied and characterised since. They are auto-antibodies of IgM, IgG and IgA isotypes associated with, but not exclusive to, patients with rheumatoid arthritis ŽTeitsson and Valdimarsson, 1984; Tuokko, 1984.. Serum RFs have a specificity for the Fc region of IgG1 , IgG2 and IgG4 , Žbut not IgG3 . ŽNormansell and Young, 1975., and exhibit a variable degree of cross-reactivity with IgG of other species ŽButler and Vaughan, 1965.. It has been shown that RFs share a common binding site on IgG with staphylococcal protein A ŽOppliger et al., 1987., and further mapping has identified the cg 2–cg 3 interface as containing the major antigenic determinants of polyclonal IgM-RF ŽSasso et al., 1988; Peterson et al., 1995.. It has been suggested that RFs represent anti-idiotype antibodies to bacterial and viral Fc receptors ŽMouritsen, 1986; Oppliger et al., 1987; Raeder and Boyle, 1991.. This is consistent with the Jerne network hypothesis, which proposes the existence of a series of immuno-regulatory idiotype–anti-idiotype networks ŽRoitt, 1991.. In comparison with the published literature on human RFs, relatively little information is available relating to species of veterinary interest. Rheumatoid factors have been reported in feline ŽUngar-Waron et al., 1985., canine ŽHalliwell et al., 1989; Neilsen, 1992. chicken ŽRaeder and Boyle, 1991; Sugii, 1994. and equine serum ŽCarter et al., 1995.. IgM-RF has been detected in cattle serum ŽUngar-Waron et al., 1985. and its reactivity with heterologous IgG reported ŽWaron et al., 1987.. IgM-RF levels in cattle have been shown to increase with age ŽUngar-Waron et al., 1991a., but be unaffected by infection with bovine leukaemia virus or use of a Babesia boÕis vaccine. Bovine sera have not previously been evaluated for the presence of IgG- or IgA-RF. In this paper, we describe the testing of a large number of bovine serum samples for IgG-, IgA- and IgM-RF, and examine the influence of age and maternally derived immunoglobulins on the levels detected. The implications of bovine RF for serological testing are discussed.
2. Materials and methods 2.1. BoÕine sera Four panels of bovine sera were examined. These consisted of sera from 19 colostrum-deprived calves ŽCD panel. and 24 colostrum-supplemented calves Žnon-CD panel. less than 1 week old, 36 3–5- month-old calves Ž3–5-month panel. and 200 heifers and cows, aged 9 months and older Ž) 9-month panel.. The latter panel was assembled by selecting 10 sera from each of 20 herds submitted for Brucella serology testing. The immunoglobulin levels in the sera of the CD and non-CD panels were measured by the zinc sulphate turbidity ŽZST. test, as previously described ŽMcEwan et al., 1970.. Briefly, 50 m l volumes of each serum were incubated at room temperature with 3 ml of zinc sulphate solution Ž208 mgrl in distilled water. and 3 ml distilled water, and the absorbance at 623 nm read after 1 h. Using a standard curve, the corrected absorbance of each serum was then expressed in ZST units Ž1 unit s 1 mgrml..
D.A. Graham et al.r Veterinary Immunology and Immunopathology 61 (1998) 239–250
241
2.2. Reagents 3X ,3X ,5X ,5X-tetramethylbenzidine ŽTMB., bovine, mouse, rabbit and chicken IgG, biotinylated antibovine-IgM ŽBABM. and -IgG ŽBABG. monoclonal antibodies, extravidin peroxidase ŽEPO., and peroxidase-conjugated goat anti-rabbit IgG ŽGARIgGPO. and anti-mouse IgGŽFc. ŽGAMIgGŽFc.PO. were purchased from Sigma-Aldrich, Poole, UK. 2.3. ELISA testing All ELISAs were performed on Immulon 1 flat-well microtitre plates ŽDynatech Laboratories, Chantilly, VA.. With the exception of stop solution Ž50 m lrwell., all volumes were 100 m lrwell. Plates were coated overnight at 48C, using 0.05 M carbonaterbicarbonate ŽpH 9.6. as coating buffer. The concentration of all immunological reagents, including coating immunoglobulins, was optimized prior to use. All dilutions were made in PTN Ž0.01 M PBS containing 2% wrv NaCl and 0.05% vrv Tween 80., and wells were washed with PBS-Tween Ž0.01 M PBS containing 0.05% wrw Tween 20. between each incubation. Substraterchromogen solution was prepared by adding 9 ml phosphate–citrate buffer ŽpH 5.0. and 2 m l 30% H 2 0 2 to 1 mg TMB dissolved in 1 ml DMSO. All incubations were done for 1 h at 378C, with the exception of the chromogenrsubstrate solution, which was incubated for 10 min before the reaction was stopped with 2.5 M H 2 SO4 . After stopping, plates were read at 450 nm on a Titertek Multiskan MCC plate reader ŽLabsystems, Helsinki, Finland.. 2.4. IgM-RF ELISA Serum samples were tested in duplicate at dilutions of 1r100 and 1r800 on plates coated with bovine IgG at 2 m grml. Bound IgM-RF was detected by successive incubations with BABM Ž1r35 000. and EPO Ž1r4000.. Controls on each plate consisted of duplicate wells to which PTN was added in place of serum ŽPTN control., or in which the serum and BABM liquid phases were replaced with PTN, and peroxidase-conjugated rabbit anti-bovine IgG ŽRABIgGPO. ŽNordic Immunology, Tilburg, Netherlands. Ž1r4000. was substituted for EPO Žcoating control.. The mean optical density ŽOD. of each serum at 1r100 and 1r800 was calculated and the corrected OD ŽCOD. obtained by subtracting the mean OD of the PTN control wells. 2.5. Pre-treatment of serum samples with 2-mercaptoethanol (2-ME) Pre-treatment with 2-ME was performed by incubating equal volumes of undiluted serum and 0.2 M 2-ME for 1 h at 378C. Twenty sera from the ) 9-month group were pre-treated in this way and re-tested by IgM-RF ELISA, as described above. 2.6. IgA-RF ELISA This followed the IgM-RF protocol, except that sera were tested at dilutions of 1r20 and 1r100, and the detection phases consisted of antibovine IgA MAb ŽVMRD, Pullman, WA. Ž1r200., followed by GAMIgGŽFc.PO Ž1r8000..
242
D.A. Graham et al.r Veterinary Immunology and Immunopathology 61 (1998) 239–250
2.7. Selection of heterologous coating IgG for IgG-RF ELISA Microtitre plates were coated in four blocks with bovine, rabbit, mouse and chicken IgG at 2 m grml, and 17 sera from the ) 9-month panel that gave a range of COD values when tested by IgM-RF ELISA were selected. These were then tested in duplicate at a dilution of 1r100 against each of these coating immunoglobulins, following the IgM-RF ELISA protocol. The heterologous IgG giving the highest degree of cross-reactivity, relative to bovine IgG, was selected and used as coating antibody for the IgG-RF ELISA Žrabbit IgG, see Section 3.. 2.8. IgG-RF ELISA Serum samples were tested in duplicate at dilutions of 1r20 and 1r100. Bound IgG-RF was detected using BABG Ž1r50 000.. GARIgGPO Ž1r5000. was substituted for RABIgGPO in the coating control wells. Otherwise, the protocol was as described for IgM-RF.
3. Results 3.1. ZST testing The ZST values of the CD panel ranged from 0–0.9 units Žmean s 0.2, standard deviation Žsd. s 0.3., confirming the colostrum-deprived status of this panel. In contrast, the ZST values of the non-CD panel ranged from 2–32 units Žmean s 14.5, sd s 8.7.. A ZST value of 20 units is considered to represent an adequate intake of colostrum ŽDr. D. Mackie, personal communication.. 3.2. IgM-RF ELISA The results of testing the serum panels at serum dilutions of 1r100 and 1r800 by IgM-RF ELISA are summarised in Table 1. An increase in the mean and range of IgM-RF COD values with age was detected when sera were tested at a dilution of 1r100 ŽFig. 1.. At this dilution, the differences between panels was significant
Table 1 Mean"one standard deviation of IgM-RF and IgG-RF ELISA COD values of 4 serum panels Panel
na
IgM-RF 1r100
CD Non-CD 3–5 months )9 months a b
19 24 36 200
b
0.043"0.025 0.013"0.011 0.173"0.120 0.592"0.287
Number of sera in each panel. Dilutions at which sera were tested.
IgG-RF 1r800
b
0.031"0.016 0.003"0.008 0.024"0.021 0.124"0.145
1r20
1r100
0.008"0.015 0.323"0.322 0.123"0.077 0.398"0.316
0.006"0.011 0.109"0.099 0.046"0.032 0.108"0.127
D.A. Graham et al.r Veterinary Immunology and Immunopathology 61 (1998) 239–250
243
Fig. 1. IgM-RF ELISA COD values for individual sera from 4 serum panels tested at a serum dilution of 1r100 Ž ns19 ŽCD., 24 Žnon-CD., 36 Ž3–5 months. and 200 Ž )9 months... Horizontal dashed lines indicate mean COD of each panel. Horizontal dotted lines indicate mean COD of each panel q1 sd Žnot shown for CD and non-CD panels..
Ž p - 0.001. in all cases. A reduced mean COD value for each panel was observed at a serum dilution of 1r800 ŽTable 1.. This reduction was significant Ž p - 0.001. for all except the CD panel Ž p s 0.81.. The mean signal in the coating and PTN control wells were 1.556 Ž"0.347. and 0.064 Ž"0.017., respectively. 3.3. Pre-treatment with 2-ME The mean IgM-RF COD values Ž"1 sd. of the 20 sera when tested at serum dilutions of 1r100 and 1r800 were 0.806 Ž"0.329. and 0.267 Ž"0.153., respectively. Pre-treatment with 2-ME significantly reduced the mean COD at both serum dilutions Ž p - 0.001. to 0.081 Ž"0.066. and 0.025 Ž"0.012., respectively.
244
D.A. Graham et al.r Veterinary Immunology and Immunopathology 61 (1998) 239–250
3.4. Selection of heterologous coating IgG for IgG-RF ELISA The mean IgM-RF COD of the 19 sera from the ) 9-month panel was greatest on wells coated with bovine IgG, followed by rabbit, mouse and chicken IgG ŽFig. 2.. The mean COD with bovine IgG was significantly greater Ž p - 0.001. than the mean value observed with any of the heterologous IgGs. The mean COD detected on rabbit IgG-coated wells was significantly greater than on chicken IgG-coated wells Ž p - 0.01.. The differences in the mean CODs between rabbit and mouse IgG and between mouse and chicken IgG were not significant Ž p s 0.081 and p s 0.336, respectively.. On the basis of these results, rabbit IgG at 2 m grml was chosen for coating IgG-RF ELISA plates.
Fig. 2. Mean IgM-RF corrected optical density ŽCOD. values of 19 sera from )9-month panel, tested in duplicate on wells coated at 2 m grml bovine, rabbit, mouse and chicken IgG. Horizontal dashed lines indicate mean COD of each panel. Horizontal dotted lines indicate mean COD of each panel q1 sd. ) Coating immunoglobulin.
D.A. Graham et al.r Veterinary Immunology and Immunopathology 61 (1998) 239–250
245
3.5. IgG-RF ELISA The results of testing the serum panels at serum dilutions of 1r20 and 1r100 by IgG-RF ELISA are summarised in Table 1. Fig. 3 shows the IgG-RF COD values of the panels when tested at a serum dilution of 1r20. At this serum dilution, an increase in the mean and range of IgG-RF COD values was detected with age. The means of the CD
Fig. 3. IgG-RF ELISA COD values for individual sera from 4 serum panels tested at a serum dilution of 1r20 Ž ns19 ŽCD., 24 Žnon-CD., 36 Ž3–5 months. and 200 Ž )9 months... Horizontal dashed lines indicate mean COD of each panel. Horizontal dotted lines indicate mean COD of each panel q1 sd Žnot shown for CD panel..
246
D.A. Graham et al.r Veterinary Immunology and Immunopathology 61 (1998) 239–250
Fig. 4. Correlation between IgG-RF ELISA COD values at a serum dilution of 1r20 and ZST units of the 24 sera of the non-CD panel. Ž r s 0.59, p- 0.01..
and ) 9-month panels were lowest and highest, respectively. However, the mean of the non-CD exceeded that of the 3–5-month panel. No significant difference was detected between the non-CD and ) 9-month panels Ž p s 0.28.. All other inter-panel comparisons indicated significant differences Ž p - 0.001.. The results of testing at a serum dilution of 1r100 yielded similar results, although the mean values were significantly reduced in all cases Ž p - 0.01., with the exception of the CD panel Ž p s 0.61.. At this dilution, the means of the non-CD and ) 9-month panels were almost identical Ž0.109 and 0.108, respectively.. Otherwise, all differences between panels were again significant Ž p F 0.01.. The mean signal in the coating and PTN control wells was 2.200 Ž"0.362. and 0.094 Ž"0.060., respectively. Regression analysis, by the least squares method, showed that IgG-RF values Ž1r20. of the non-CD panel were correlated with ZST levels ŽFig. 4. Ž r s 0.59, p - 0.01.. No correlation was found by this method between IgG-RF Ž1r20. and IgM-RF Ž1r100. values of the ) 9-month panel Ž r s 0.05, p s 0.47.. 3.6. IgA-RF ELISA The mean COD values at a serum dilution of 1r20 were extremely low Ž- 0.006.. Similar results were obtained when 50 sera from the ) 9-month panel were re-tested
D.A. Graham et al.r Veterinary Immunology and Immunopathology 61 (1998) 239–250
247
using a different anti-bovine IgA MAb ŽSerotec, Oxford, UK.. The mean signal in the coating and PTN control wells was 1.780 Ž"0.256. and 0.043 Ž"0.004., respectively.
4. Discussion Bovine IgM-RF has been reported previously ŽUngar-Waron et al., 1985., but we believe that this is the first time that bovine sera have been assayed for IgG- and IgA-RF. The strength of a specific signal should decrease as the serum test dilution increases; therefore, sera were tested at two different dilutions on each ELISA. The serum dilutions used in each ELISA were based on those used successfully by others to detect human RFs by indirect ELISA ŽTeitsson and Valdimarsson, 1984; Jonsson et al., 1995.. The IgG-RF and IgA-RF ELISAs were performed at lower serum dilutions than the IgM-RF ELISA to compensate for the use of rabbit IgG as coating immunoglobulin in the former, and the anticipated effect of inter-isotype competition in the latter. Ungar-Waron et al. Ž1991a. examined a total of 41 sera ranging from 2–4 week- to 9–10 year-old bovines and reported increasing levels of IgM-RF with age. Our findings, based on a much larger number of samples, and including neonatal calves of defined colostral status, confirm these results. IgM and to a lesser extent, IgG, are susceptible to reduction by 2-ME ŽCapel et al., 1980.. Pre-treatment of sera with 2-ME produced a reduction in IgM-RF COD value of approximately tenfold at both serum dilutions, confirming the specificity of the signal detected by IgM-RF ELISA. Low levels of IgG- and IgM-RF were detected in the CD panel. Idiotype networks, including low-affinity IgM with IgG-specificity ŽRoitt, 1991., have been described in foetal mice. The existence of similar idiotypic networks in the bovine foetus may be postulated as the source of these antibodies. The levels of IgG-RF detected in the non-CD panel were significantly higher than those in the CD panel. This is expected, given the higher levels of serum IgG present in these calves, as indicated by the ZST test. The correlation of IgG-RF values with the ZST reading of each sample indicates that IgG-RF is transferred to the neonate as part of the overall colostral IgG load. If IgM-RF is found in the IgM component of colostrum in the same way as IgG-RF is found in the IgG component, it would be expected that IgM-RF would be detected in the serum of colostrum-supplemented calves, given that IgM and IgG are absorbed from colostrum with equal efficiency ŽKlaus et al., 1969.. The finding that the IgM-RF levels of the non-CD panel were lower than those of the CD panel by a small, but significant, amount was therefore unexpected. It may be that the much higher levels of serum IgG in the calves of the non-CD panel results in a greater percentage of IgM-RF being complexed with serum IgG ŽMoore et al., 1995., and therefore unavailable for reaction with solid phase IgG. The increasing levels of IgM-RF and IgG-RF with age are in line with expectations. If RF antibodies are indeed auto-anti-idiotype antibodies, as predicted by the Jerne network hypothesis, then levels would be expected to increase as the immune system matures, and is exposed to an increasing range of antigens. Similarly, if RF antibodies are anti-idiotype antibodies produced in response to infection with organisms expressing
248
D.A. Graham et al.r Veterinary Immunology and Immunopathology 61 (1998) 239–250
Fc-binding proteins, an increased exposure to such organisms Žand therefore RF levels. with age is again to be expected. The wide range of IgM- and IgG-RF COD values seen in the - 9-month panel most probably reflects the wide age range of the cattle represented. The decrease in IgG-RF levels between the non-CD and 3–5-month groups reflects the lower level of maternally derived IgG-RF in the latter. The design of the IgG-RF ELISA required a system that allowed bound IgG-RF to be differentiated from solid phase IgG. The use of rabbit IgG as coating antibody, and an anti-bovine IgG-specific MAb as detector, made this possible. The presence of IgM-RF in sera being assayed for IgG-RF represented a possible source of error. Due to its pentavalent nature, IgM-RF can potentially bind to solid phase IgG and subsequently also capture the detecting antibody, generating a false signal ŽTeitsson and Valdimarsson, 1984.. The absence of a signal in the IgA-RF ELISA, which was performed at the same serum dilution, and which also used an IgG1-isotype detector MAb, and the lack of correlation between IgG- and IgM-RF results, indicate that the IgG-RF signal was specific. The greater cross-reactivity of bovine IgM-RF with rabbit than with mouse IgG has been reported previously ŽUngar-Waron et al., 1985; Waron et al., 1987., and our findings are consistent with these reports. We believe that this is the first time that chicken IgG has been included in this type of evaluation. Ž1988. have shown that human RF does not react with chicken Larsson and Sjoquist ¨ IgG, and based on this finding, the use of chicken antibodies as capture- andror detector antibody can eliminate the interference of RF when human sera are tested by sandwich ELISA ŽLarsson et al., 1991.. Our finding of low level cross-reactivity in some chicken sera ŽFig. 2. suggests that the use of chicken antibodies may not always eliminate RF interference in bovine serology. Indirect ELISAs, similar to the one used in this study, have been used previously to detect IgA-RF in human sera ŽTeitsson and Valdimarsson, 1984; Jonsson et al., 1995.. It is therefore considered that the failure to detect IgA-RF reflects its absence from bovine sera, rather than a failure of IgA-RF to bind to solid phase antibody, due to inter-isotype competition with IgM- and IgG-RF. The association between RFs and arthritis in cattle has not been investigated, and rheumatoid arthritis is not recognised as a cause of lameness in cattle ŽBailey, 1985.. Therefore, the presence of RFs in bovine sera is most significant as a potential complicating factor in serological testing. As described above, RFs can cause a false signal in sandwich ELISAs by cross-linking solid phase capture- and liquid-phase detector-antibody in the absence of specific antigen. Non-specific reactions may also occur in indirect ELISAs designed to detect antigen-specific IgM, when the serum under test contains IgM-RF and antigen-specific IgG. In this case, IgM-RF bound to IgG that has reacted specifically with solid-phase antigen cannot be distinguished from a specific IgM-antigen interaction. This has been described in indirect ELISAs ŽUngar-Waron and Abraham, 1991b; Salonen et al., 1980; Tuokko, 1984., solid-phase radio-immunoassay ŽMeurman and Ziola, 1978. and indirect immuno-fluorescence ŽGispen et al., 1975; Hekker et al., 1979.. Removal of IgM-RF or antigen-specific IgG prior to testing are both effective methods of eliminating such reactions ŽLeinikki et al., 1978; Meurman and Ziola, 1978; Ungar-Waron and Abraham,
D.A. Graham et al.r Veterinary Immunology and Immunopathology 61 (1998) 239–250
249
1991b.. The use of reverse rather than indirect ELISAs also reduces the incidence of non-specific reactions ŽTuokko, 1984..
5. Conclusion We report the presence of IgM- and IgG-RF in bovine sera, with levels of both increasing with age. In addition, correlation between colostral status and IgG-RF levels in neonatal calves has been demonstrated. We highlight the importance of considering RFs as a possible source of non-specific reactions in veterinary serology, especially in relation to assays for antigen-specific IgM.
Acknowledgements This work has been partially funded by SVANOVA Biotech, Glunten, Uppsala Science Park, S-751 83 Uppsala, Sweden.
References Bailey, J.V., 1985. Bovine arthritides— classification, diagnosis, prognosis and treatment. In: Ferguson, J.G. ŽEd.., Symposium on Bovine Lameness and Orthopedics. The Veterinary Clinics of North America ŽFood Animal Practice., 1 Ž1. pp. 39–52. Butler, V.P., Vaughan, J.H., 1965. The reaction of rheumatoid factor with animal gamma-globulins: quantitative considerations. Immunology 8, 144–147. Capel, P.J.A., Gerlag, P.G.G., Hagemann, J.F.H.M., Koene, R.A.P., 1980. The effect of 2-mercaptoethanol on IgM and IgG antibody activity. J. Immunol. Methods 36, 77–80. Carter, S.D., Osborne, A.C., May, S.A., Bennett, D., 1995. Rheumatoid factor, anti-heat shock protein Ž65 kDa. antibodies and anti-nuclear antibodies in equine joint diseases. Eq. Vet. J. 27, 288–295. Gispen, R., Nagel, B., Brand-Saathof, B., De Graaf, S., 1975. Immunofluorescence test for IgM rubella antibodies in whole serum after absorption with anti-g Fc. Clin. Exp. Immunol. 22, 431–437. Halliwell, R.E.W., Werner, L.L., Baum, D.E., Newton, C.D., Wolfe, J.H., Schumacher, H.R., 1989. Incidence and characterisation of canine rheumatoid factor. Vet. Immunol. Immunopathol. 21, 161–175. Hekker, A.C., Brand-Saathof, B., Vis, J., Meijers, R.C., 1979. Indirect immunofluorescence test for detection of IgM antibodies to cytomegalovirus. J. Infect. Dis. 140, 596–600. Jonsson, T., Thorsteinsson, H., Arinbjarnarson, S., Thorsteinsson, J., Valdimarsson, H., 1995. Clinical implications of IgA rheumatoid factor subclasses. Ann. Rheum. Dis. 54, 578–581. Kahn, M.F., 1995. The history of rheumatoid factor. Revue de Rheumatisme 62, 513–517. Klaus, B.G.G., Bennett, A., Jones, E.W., 1969. A quantitative study of the transfer of colostral immunoglobulins to the newborn calf. Immunology 16, 293–299. Larsson, A., Sjoquist, J., 1988. Chicken antibodies: a tool to avoid false positive results by rheumatoid factor ¨ in latex fixation tests. J. Immunol. Methods 108, 205–208. Larsson, A., Karlsson-Parra, A., Sjoquist, J., 1991. Use of chicken antibodies in enzyme immunoassays to ¨ avoid interference by rheumatoid factors. Clin. Chem. 37, 411–414. Leinikki, P.O., Shekarchi, I., Dorsett, P., Sever, J.L., 1978. Determination of virus-specific IgM antibodies by using ELISA: elimination of false-positive results with protein A-sepharose absorption and subsequent IgM antibody assay. J. Lab. Clin. Med. 92, 849–857. McEwan, A.D., Fisher, E.W., Selman, I.E., Penhale, W.J., 1970. A turbidity test for the estimation of immune globulin levels in neonatal calf serum. Clin. Chim. Acta 27, 155–163.
250
D.A. Graham et al.r Veterinary Immunology and Immunopathology 61 (1998) 239–250
Meurman, O.H., Ziola, B.R., 1978. IgM-class rheumatoid factor interference in the solid-phase radioimmunoassay of rubella-specific IgM antibodies. J. Clin. Pathol. 31, 483–487. Moore, T.L., Osborn, T.G., Nesher, G., 1995. Immune complexes from sera of patients with juvenile rheumatoid arthritis reveal novel 40 and 60 kd bands. Clin. Exp. Rheumatology 13, 667–672. Mouritsen, S., 1986. Rheumatoid factors are anti-idiotypic antibodies against virus-induced anti-Fc receptor antibodies. Scand. J. Immunol. 24, 485–490. Neilsen, O.L., 1992. Detection of IgM rheumatoid factor in canine serum using a standardized enzyme-linked immunosorbent assay. Vet. Immunol. Immunopathol. 34, 139–147. Normansell, D.E., Young, C.W., 1975. The IgG specificity of anti IgG immunoglobulin rheumatoid factors. Immunochemistry 12, 187–188. Oppliger, I.R., Nardella, F.A., Stone, G.C., Mannik, M., 1987. Human rheumatoid factors bear the internal image of the Fc binding region of staphylococcal protein A. J. Exp. Med., 702–710. Peterson, C., Malone, C.C., Williams, R.C., 1995. Rheumatoid-factor-reactive sites on CH 3 established by overlapping 7-mer peptide epitope analysis. Mol. Immunol. 32, 57–75. Raeder, R., Boyle, M.D.P., 1991. Detection of rheumatoid-like factors in serum of chickens immunised with bacterial immunoglobulin binding proteins. J. Immunol. Methods 1991, 201–209. Roitt, M.I., 1991. Essential Immunology 7. Blackwell, Oxford, 356 pp. Salonen, E.-M., Vaheri, A., Suni, J., Wager, O., 1980. Rheumatoid factor in acute viral infections: interference with determination of IgM, IgG, and IgA antibodies in an enzyme immunoassay. J. Infect. Dis. 142, 250–255. Sasso, E.H., Barber, C.V., Nardella, F.A., Yount, W.J., Mannik, M., 1988. Antigenic specificities of human monoclonal and polyclonal IgM rheumatoid factors. J. Immunol. 140, 3098–3107. Sugii, S., 1994. Immunological cross-reactivity of mannan-binding proteins in bovine, chicken and human sera. J. Vet. Med. Sci. 56, 787–790. X Teitsson, I., Valdimarsson, H., 1984. Use of monoclonal antibodies and FŽab .2 enzyme conjugates for IgM, IgA and IgG rheumatoid factors. J. Immunol. Methods 71, 149–161. Tuokko, H., 1984. Comparison of non-specific reactivity in indirect and reverse immunoassays for measles and mumps immunoglobulin M antibodies. J. Clin. Microbiol. 20, 972–976. Ungar-Waron, H., Waron, M., Gluckman, A., Trainin, Z., 1985. Rheumatoid factors in natural retrovirus infections of the cat and cattle. J. Comp. Pathol. 95, 399–403. Ungar-Waron, H., Paz, R., Shkap, V., Hanna, B., Pipano, E., 1991a. IgM rheumatoid factor in cattle immunized against babesiosis. J. Vet. Med. B 38, 492–496. Ungar-Waron, H., Abraham, A., 1991b. Immunoglobulin M ŽIgM. indirect enzyme-linked immunosorbent assay and the involvement of IgM-rheumatoid factor in the sero-diagnosis of BHV-1 infection. Vet. Microbiol. 26, 53–63. Waron, M., Trainin, Z., Gluckman, A., Ungar-Waron, H., 1987. Reactivity of IgM-rheumatoid factors of human, feline, and bovine sera. J. Comp. Pathol. 97, 487–490.