A SOLID-PHASE FOR THE DETERMINATION OF BACTERIAL-SPECIFIC WITHIN DIFFERENT IMMUNOGLOBULIN CLASSES: APPLICATION TO BOVINE BRUCELLA ABORTUS ANTIBODIES D. Levieux, G. Bezard

To cite this version:

D. Levieux, G. Bezard. A SOLID-PHASE RADIOIMMUNOASSAY FOR THE DETERMINA- TION OF BACTERIAL-SPECIFIC ANTIBODIES WITHIN DIFFERENT IMMUNOGLOBULIN CLASSES: APPLICATION TO BOVINE BRUCELLA ABORTUS ANTIBODIES. Annales de Recherches Vétérinaires, INRA Editions, 1978, 9 (3), pp.523-530. ￿hal-00901034￿

HAL Id: hal-00901034 https://hal.archives-ouvertes.fr/hal-00901034 Submitted on 1 Jan 1978

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. A SOLID-PHASE RADIOIMMUNOASSAY FOR THE DETERMINATION OF BACTERIAL-SPECIFIC ANTIBODIES WITHIN DIFFERENT IMMUNOGLOBULIN CLASSES: APPLICATION TO BOVINE BRUCELLA ABORTUS ANTIBODIES

D. LEVIEUX

G. BEZARD.

Station de Pathologie de la Reproduction, I,N.R.A., Centre de Recherches de Tours, B.P. 7, Nouzilly, 37380 Monnaie, France.

Résumé

DOSAGE RADIOIMMUNOLOGIQUE POUR LA DETERMINATION DES DIFFERENTES CLASSES D’ANTICORPS SPECIFIQUES D’UNE BACTERIE : APPLICATION AUX ANTICORPS BOVINS CONTRE BRUCELLA ABORTUS. ― Un dosage radioimmunologique (DRI) en phase solide a été mis au point pour le titrage des différentes classes d’anticorps spécifiques d’une bactérie. Des tubes en verre sont sensibilisés par sédimentation, séchage et fixation au méthanol de corps bactériens de Brucella abortus. Les anticorps de sérum bovin qui se fixent sur les bactéries sont mis en évidence par l’addition d’un sérum de lapin spécifique d’une classe ou sous-classe d’immunoglobulines bovines puis d’anticorps de mouton anti- immunoglobulines de lapin marqués par !2e1. Cette technique en trois temps a été choisie car très sensible (< 0,3 picogramme d’anticorps/ml), donnant la meilleure discrimination entre la radioactivité spécifique et le bruit de fond et permettant d’éviter le marquage de plusieurs sérums de lapin anti-immunoglobulines bovines. Différents paramètres ont été étudiés : nature et préparation des tubes, quantité et fixation des Brucella, température et temps d’incubation, volume d’incubation, concentration ces réactifs. Ce DRI et la technique de fixation du complément (FC) ont été appliqués à des sérums prélevés à intervalles régu- liers sur des génisses infectées expérimentalement par t3rucella abortus pendant la gestation. Un sérum de lapin anti-lgG1 bovines a été utilisé dans le DRI car cette sous-classe est la plus active dans la F.C. La cinétique d’apparition des anticorps anti-Brucella est très signifi- cativement corrélée pour les deux techniques (r = 0,956 ; P < 0,001), démontrant ainsi la validité de ce DRI. Ce test est 10000 fois plus sensible que la F.C.

Introduction rement of antibodies against bacterial an- tigens depend upon secondary effects of Assays presently available for the measu- the primary - interaction, such * Present address : Laboratoire des Maladies as or complement activation, Nutritionnelles, C.R.Z.V. de Theix, LN.R.A., and do not always estimate the total anti- 63110 Beaumont, France. body content of the antiserum. In addition, such tests do not lend themselves to the Antigen was washed twice before use in study of the different classes of antibodies phosphate buffered saline (PBS) pH 6.9 and, involved. for the RIA only, twice in distilled water. This may, for instance, be achieved by of each class 1) purification immunoglobulin Albumin diluent by physico-chemical means, 2) assessing antibody activity of each class. Quantitative Albumin diluent was made by adding purification of each class is difficult and human serum albumin (HSA) to PBS at a consequently this procedure has a limited concentration of 5 mg/ml. value only. determine the na- Alternatively, one may Sera ture of immunoglobulins released by disso- ciation from antigen-antibody complexes but Bovine sera for the determination of im- dissociation may be far from complete (Read munoglobulins with antibody activity against et al., 1974). Brucella abortus were obtained at repeated In a different approach, the concentration intervals from seven heifers experimentally of each immunoglobulin class is determined, infected by Brucella abortus, strain 544, in by means of single radial , the 6th month of gestation (Plommet et al., before and after removal of its antibody 1976). Blood was collected from the jugular active portion (Nash and Heremans, 1969). vein, allowed to clot and the serum stored The accuracy of the measurements, essen- at -20°C. Before use in the RIA all the tially that of the immunodiffusion technique, sera were heat inactivated (56°C, 30 mn). limits this method to sera with high antibody Reference bovine serum « B1 », used for titers. purification of specific igG1 or IgG2 anti- Recently, a radioimmunoassay (RIA) has bodies, was a pool of 20 sera from brucel- been developed to measure antibodies losis-aborted cows with very high serological against Brucella abortus in bovine serum titers. by competition between immunoglobulins of the test sample and radiolabelled active Preparation of antisera immunoglobulins (Chappel et al., 1976). As described this method does not differentiate Rabbit antisera to bovine immunoglobulin the activity of the immunoglobulin classes classes and subclasses were obtained as and postulates that all the antibodies to be previously described (Levieux, 1974). antiserum to rabbit was tested have the same specificities as the Sheep igG1 pre- with radio-labelled one ; this is invalidated by the pared by immunisation 15 mg of puri- results given. fied IgGl (Diethylaminoethyl cellulose chro- After unsuccessful trials with different matography) for four weeks in Freund’s inhibition or indirect immu- complete adjuvant as multiple intradermal nofluorescent techniques, we propose a new injections. solid-phase RIA to measure specific anti- bodies within different immunoglobulin clas- Purification of antibodies ses, which avoids repeated centrifugations Bovine antibodies to Brucella were which are time consuming and difficult to antigen perform, without loss of material, for Bru- obtained as follows : 300 ml of the standar- cella antigen. dized antigen were incubated with 30 ml of heat inactivated serum B1. After cooling in an ice-bath the bacterial cells were cen- trifugated and repeatedly washed until the Materials and methods optical density (E !28°, ) of the supernatant was below 0.05. Antibodies were eluted with Preparation of antigen glycine-HC1 buffer (0.2 M, pH 2.6) for 15 mi- The antigen was a gift from Laboratories nutes at room temperature and the pH Roger Bellon (Neuilly, France) as a standar- adjusted to 7.0 with 1M dipotassium hydrogen dized Brucella abortus agglutination concen- phosphate. After filtration through 0.2 p,m trate (Alton and Jones, 1967). Antigen dilu- Millipore filters, the protein concentration tions quoted below refer to this concentrate. was determinated from absorbance measu- rements (E 1 280 = 13.5). The purification of on Sephadex G25 column (25 x 1 cm) equi- librated with PBS 0.1 M 7.2 the IgGl and IgG2 subclass antibodies was pH containing 5 HSA and the fractions checked for performed as previously described (Levieux, mg/ml 1974) using Sephadex G200 and DEAE cel- radioactivity in a Packard y counter. lulose chromatography. Purity of the pre- was checked immunoelectro- parations by Radioimmunoassay phoretic analysis both with rabbit anti-whole bovine serum and monospecific antisera for Fixation of Brucella antigen on glass igG1 or IgG2. The percentage of active tubes : 50 I-tl of a 1/10 dilution of the washed antibodies was tested by determination in standardized bacterial suspension were de- (Mancini et al., 1965) posited at the bottom of glass tubes and of the concentration of each immunoglobulin allowed to dry overnight at 37°C. The dried preparation before and after absorption with Brucella were then fixed with 0.5 ml metha- an excess of Brucella. Each preparation nol (100 °/o) for 10 minutes and the excess was more than 90 p. 100 active. reagent was poured off. Immediately before Sheep antibodies to rabbit IgG were pre- use the Brucella antigen was rehydrated for pared from the sheep antiserum by affinity 10 minutes with albumin diluent and excess chromatography on glutaraldehyde-polyme- reagent was again poured off. rized rabbit and IgG (Avrameas Ternynck, A typical three-step RIA was set up in 1969). quadruplicate as follows : 0.2 ml of a 10-4 dilution of the serum to be tested in albumin lodination of antibodies diluent or ten fold dilutions of specific immunoglobulins were incubated with the Sheep anti-rabbit [gG antibodies and bo- rehydrated antigen at room temperature for vine IgGl or IgG2 anti-Brucella antibodies 24 hours. The tubes were washed twice were labelled with izsl (Amersham) by the with 2 ml of cold albumin diluent, allowed chloramine T method (Greenwood et al., to drain on filter paper and incubated with 1963) modified as follows. One mCi (5 jjU) 0.2 ml of a 10-4 dilution of monospecific was used to label 10 !Lg of immunoglobulin rabbit anti-bovine immunoglobulins at room (10 ¡Jo1). in PBS 0.5 M pH 7.2. The reaction temperature for another 24 hours. After was allowed to proceed for 10 sec at room two washes with albumin diluent the tubes temperature and then stopped by addition were incubated with 0.2 ml of radiolabelled of 50 ¡.¡.g of sodium metabisulfite (10 !LI). sheep anti-rabbit IgG for the final 24 hours, After addition of 50 !g of potassium iodide washed twice with albumin diluent and (50 !ti), the mixture was chromatographied counted for radioactivity. Serological tests

The C.F.T. was performed by a micro-me- thod (Renoux et al., 1971) using cold fixa- tion.

Results

Optimisation of the

The first step involved testing with a one- step RIA using radiolabelled bovine anti- Brucella IgGl incubated in Brucella sensi- tized tube. The non-specifically absorbed radioactivity was determined by incubation of the sensitized tubes with iodinated sheep antibodies for rabbit IgG. Our final criteria for optimisation were to increase specific binding and decrease standard error.

- Nature and preparations of the test tube : glass and polystyrene tubes were compared for their ability to retain fixed Brucella after repeated washings. The same quantity and radioactivity of this labelled tracer was used. Fig. 1 shows that the glass tubes fixed bacterials cells more effi- ciently than polystyrene tubes.

- Cleaning of the glass tubes before fixation of Brucella : comparison was made between detergent (RBS 25, boiling or at room temperature), sulfochromic acid or ethanol (95 p. 100). Table 1 shows that boiling detergent or sulfochromic acid give the best fixation with standard error inferior to the one obtained with ethanol.

- Quantity of Brucella : the range tested covered dilutions ranging from 1/1 to 1/1000 of the standardized suspension, in 50 >1. Optimum results were obtained for the 1/10 dilution.

- Fixation of Brucella : comparison was made between : overnight air-drying, over- night air-drying plus ethanol or methanol fixation, overnight drying in methanol. The second method was the most efficient.

- Temperature and incubation time for fixation of antibodies : Fig. 2 shows that maximum fixation of antibodies occurs for 72 hours of incubation at 4°C, and 24 hours at 20°C or 37°C. For incubation longer than 48 hours at 37°C loss of radioactivity due to bacterial growth was observed. - Incubation volume : Table 2 shows Application of the three-step RIA results obtained for the same quantity of for the determination of Brucella labelled Brucella specific antibodies incu- specific IgGl antibodies synthetized bated with 0.1 ml to 1 ml for the same incu- after experimental infection bation time. Optimum results were obtained with minimal volumes but 0.2 ml was finally This subclass of IgG was chosen because chosen to minimize errors due to residual it is the predominant type of antibody impli- diluent not removed between incubations of specific antisera.

- Quantity of the monospecific rabbit anti- serum for bovine igG1, IgG2, IgAs, IgM : they were tested with the three-step RIA. Usually the different antisera were in the range of activity of a 10-3 to 10 4 dilution for 0.2 ml.

- Quantity of the radiolabelled sheep IgG antibodies for rabbit ]gG : Fig. 3 shows results obtained in the three-step RIA using rabbit sera specific for IgGl subclass. Incu- bation with 100,000 cpm gives the sharpest curve.

Reference curve

Fig. 4a shows a typical curve obtained with serial dilution of a Brucella specific igG1 preparation. Means and standard errors on quadruplicate analyses are plotted. This sigmoidal curve is straightened by plotting results on a log versus probit scale (Fig. 4b). The regression equation (Y = -0.401 log X + 6.596) obtained allows calculation of the antibody concentration of the samples (X) from the probit of per cent radioactivity fixed on the tubes (Y). Due to the very high sensitivity of this RIA system, less than 0.3 picogram antibody/ml can be titrated. cated in the . Fig. 5 shows IgG1 antibodies measured by RIA and Complement Fixation Tests on seven heifers experimentally infected by Brucella abortus strain 544. The kinetics of appearance of Brucella specific antibodies are in good agreement for these two tests (r = 0.956, P < 0.001) and this is good evidence for the validity of the proposed RIA. However slight differences occur in some sera within the first month after appearance of complement fixing anti- bodies, probably due to [gM antibodies which are active in Complement Fixation Test but not in the RIA using specific antisera for igG1 antibodies. The proposed RIA is 10,000 times more sensitive than the Complement Fixation Test.

Discussion

The assays presently available for the determination of specific bacterial antibodies within different immunoglobulin classes were not completely successful in our hands.

- Purification of each immunoglobulin class by physico-chemical means is very difficult, time consuming and can only be carried out on a small number of samples.

- Determination of immunoglobulins re- leased by dissociation of antigen-antibody complexes is particularly dependent of the complete dissociation of the complex. Inves- tigations with immuno-adsorption and elution for the purification of rabbit anti-Brucella abortus antibodies generally shows that less than 50 p. 100 of the antibody can be recovered (Read et al., 1974). The ease of elution on the other hand is largely deter- mined by antibody affinity which may be very different between the classes of anti- bodies present in different sera.

- Determination of each immunoglobulin class, by means of single radial immuno- diffusion, before and after removal of its antibody-active portion (Nash and Heremans, 1969) can be used only with high titer sera. For these authors the error is about 3 p. 100 ; we obtained 5 to 7 p. 100. The concentra- tion of igG1 in bovine serum is about 12 mg/ml and therefore it is impossible to measure less than 0.36 to 0.84 mg of specific antibody, that is, approximatively 360 to 840 is particularly dramatic for weak antisera international units, which is not so fre- such as antisera for IgG subclasses. Three quently found. different 1251-labelled antisera are needed in the RIA described by Parratt et al, for human - Radioimmunoassay using competition antibodies. The use of the three-step RIA between labelled anti-Brucella antibodies and described herein avoids the use of four antibodies in a test et al., sample (Chappel individually purified and labelled 1z51 rabbit postulates the same specificities for 1976) antibodies specific for each class or subclass the different antibodies. Results obtained by of bovine immunoglobulins. The sensitivity the autors invalide this 32 100 postulate : p. is greatly increased and allows better descri- of the tested sera were fixation complement mination between the final titer of the test RIA and 36 p. 100 sero- positive, negative serum and background. agglutination positive, RIA negative. The It can be seen from the reference curve RIA involves only specific fixation of anti- that levels between 10 bodies, and not secondary events as for (Fig. 4b) antibody ng and 25 mg can be measured from a complement fixation or seroagglutination single dilution of test This is tests which vary with different classes of sample. particularly useful in view of the fluctuation of antibodies. Therefore RIA may be some- very large antibodies titer. This curve times positive for sera negative in comple- sigmoidal may be transforma- ment fixation or seroagglutination, but never easily straightened by probit tion, statistical negative when these serological tests are allowing analysis. positive. Accepted for publication, May 26 th 1978. - The solid phase RIA reported here with Brucella sensitized tubes eliminates the repeated washings of bacterial antigen by centrifugation as performed by Parratt et al. Acknowledgments (1977) in their recently published paper. The centrifugation steps are time consuming and I am indebted to M. M. Blanc for his gene- difficult to perform without any loss of rous with various of this work bacteria, particularly with Brucella. The help aspects and for specialist advice. I should like to pellet adherence is dependent of the nature thank Dr R. Fensterbank for providing sera and quantity of antibodies contained in the and performing the Complement Fixation Test serum to be tested, and complete resuspen- and M. J. C. Fayet and M. P. Paccard for sion is often very difficult. performing the statistical analysis. The en- - Purification of antibodies by affinity couraging support by Dr M. Plommet, Direc- chromatography and their subsequent radio- tor of the Station of Reproduction Pathology labelling results in a loss of activity which is gratefully recognized.

Summary

A solid-phase radioimmunoassay (RIA) has been developed for quantitation of class-specific antibodies against bacteria. Brucella abortus cells were used to sensitize glass tubes after sedimentation, drying and methanol fixation. Bovine antibodies that attached to the bacteria were detected by binding of a rabbit antiserum specific for each class or subclass of bovine immunoglobulins followed by binding of 1231 sheep anti-rabbit immunoglobulin reagent. This three-step method was adapted as it was very sensitive (< 0.3 picogram antibody/ml), gave the best discrimination between sample and background counts and eliminated the need for several labeled rabbit anti-bovine immunoglobulin class-specific antisera. Parameters affecting quantitation of the assay were investigated : nature and preparation of the test tube, quantity and fixation of Brucella, temperature and incubation time, incubation volume, concentration of reagents. The RIA and the complement fixa- tion test (CFT) were applied to sera obtained at repeated intervals from seven heifers experimentaly infected with Brucella abortus during gestation. Rabbit antisera specific for bovine IgGl were used in the RIA because antibodies of this subclass are the predominant type implicated in the CFT. The kinetic of appearance of Brucella specific antibodies were very significantly correlated for the two tests (r = 0.956, P < 0.001) demonstrating the vali- dity of the proposed RIA. This test was 10,000 times more sensitive than CFT.

References

ALTON G.G., JONES L.M., 1967. Laboratory techniques in brucellosis. World Health Organisation Mono- graph Series, n° 55, Geneva. AVRAMEAS S., TERNYNCK T., 1969. The cross-linking of proteins with glutaraldehyde ant its use for the preparation of immunoabsorbents. immunochemistry 6, 53-66. CHAPPEL R.J., WILLIAMSON P., Mc NAUGHT D.J., DALLING M.J., ALLAN G.S., 1976. Radioimmuno- assay for antibodies against Brucella abortus: a new serological test for bovine brucellosis. J. Hyg. Camb. 77, 369-376. GREENWOOD F.C., HUNTER W.M., GLOVER J.S., 1963. The preparation of 1311-labelled human growth hormone of high specific radioactivity. Biochem. J. 89, 114-123. LEVIEUX D., 1974. Immunoglobulines bovines et Brucellose. I. Purification des immunoglobulines et preparation de leurs antis6rums sp6cifiques. Ann. Rech. Vet. 5, 329-342. MANCINI G., CARBONARA A.O., HEREMANS J.F., 1965. Immunochemical quantitation of by single radial immunodiffusion. Immunochemistry 2, 235-254. NASH D.R., HEREMANS J.F., 1969. A quantitative antibody-binding method for the determination of specific antibody within different immunoglobulin classes. Application to four Ig classes in the mouse. 17, 685-694. PARRATT D., NIELSEN K.H., WHITE R.G., 1977. Radioimmunoassay of IgM, IgG and IgA Brucella antibodies. Lancet (8021), 1075-1078. PLOMMET M., FENSTERBANK R., 1976. Vaccination against bovine brucellosis with a low dose of strain 19 administered by the conjunctival route. III. Serological response and immunity in the pregnant cow. Ann. Rech. Vet. 7, 9-23. READ R.S.D., COX J.C., WARD H.A., NAIRN R.C.. 1974. Conditions for purification of anti-Brucella antibodies by immunoadsorption and elution. Immunochemistry 11, 819-822. RENOUX G., PLOMMET M., PHILLIPPON A., 1971. Micror6actions d’agglutination et de fixation du compl6ment pour le diagnostic des Brucelloses. Ann. Rech. V6t. 2, 263-269.