Evaluation of the enzyme-linked immunosorbent assay (ELISA) and other serological tests for the diagnosis of toxoplasmosis

Bulletin of the World Health Organization, 58 (1): 99-105 (1980) Evaluation of the enzyme-linked immunosorbent assay (ELISA) and other serological tests for the diagnosis of toxoplasmosis Y. CARLIER,1 D. BOUT,2 J. P. DESSAINT,1 A. CAPRON,3 F. VAN KNAPEN,4 E. J. RUITENBERG,4 R. BERGQUIST,5 &G. HULDT5 The enzyme-linked immunosorbent assay (ELISA) was evaluated in human toxoplas- mosis in three laboratories using theirownprocedures. Thesame batch ofserum samples was investigated in the three laboratories. ELISA results were compared by statistical analysis both with one another and with those of the dye test (DT), immunofluorescence (IF), complement fixation test (CFT), and indirect haemagglutination (IHA). Highly significant correlations were obtained between the three laboratories with ELISA using two different antigens and enzyme conjugates. The correlations between ELISA and the otherserologicaltestsshowedthefollowingsequence: CFT>IF>IHA >DT. Highly significant correlations were obtained between ELISA using anti-e-chain and anti- total immunoglobulin conjugates. The agreement in discrimination between sera with low and high antibody levels was goodfor all the different ELISA techniques but discrimination between positive and negative sera depended rather on the ELISA procedure used. Enzyme-linked immunosorbent assay (ELISA) has proved to be a technique of interest for antibody detection in body fluids (1). One of the main advan- tages of this method, apart from its sensitivity, is the possibility of mechanizing nearly all the steps in the assay, thus enabling many sera to be studied in a short time (2-4). Moreover, the reagents used are stable and, although the enzymes are expensive, only minute quantities are used, so that the cost per assay is low (5). Good results have already been reported in parasitic diseases (6-9). In temperate countries, the most fre- quent parasitic disease in humans is toxoplasmosis; congenital infection with Toxoplasma gondii may be quite severe, causing chorioretinitis and hydrocepha- lus or microcephaly as late effects. Acquired toxo- plasmosis, however, rarely gives clinical symptoms and a simple, rapid, and reliable serological method is therefore required for diagnosis. Many immuno- logical techniques have been tested for this pur- pose-the dye test (DT), immunofluorescence (IF), complement fixation test (CFT), and indirect haemag- glutination (IHA)-but these tests are not easily automated for large-scale screening. I Docteur en Medecine, Centre d'lmmunologie et de Biologie Parasitaire (CIBP), Institut Pasteur, 59000 Lille, France. 2 Docteur en Pharmacie, CIBP. 3Professeur a la Faculte de Medecine, Directeur du CIBP. 4 Veterinarian, National Institute of Public Health, Bilthoven, Netherlands. 5 Physician, Department of Parasitology, National Bacterio- logical Laboratory, Stockholm, Sweden. The application of ELISA to the diagnosis of toxoplasmosis has been described previously by various authors (5, 6, 10-14). In the present study, three laboratories, all with experience in enzyme- immunoassays and in the serology of toxoplasmosis, compared the results they obtained with ELISA with those obtained by other serological techniques. For ELISA, the same selected sera and two different soluble Toxoplasma antigens and conjugates were used by the three laboratories. Each laboratory fol- lowed its own procedure. MATERIALS AND METHODS Toxoplasma antigens In Lille (France) antigens were prepared from the RH strain of Toxoplasma gondii. Mice were infected intraperitoneally with 3 x 106 parasites and 5 x 106 Ehrlich cells. Three days later, 4-6 x 108 Toxoplasma were harvested in two peritoneal washings from each mouse, using Hanks-Wallace medium. The parasites were washed in the same medium by centrifugation. Toxoplasma for IF were resuspended in sucrose solution and then freeze-dried. A suspension con- taining 75-200 parasites per microscope field (magni- fication x 400) was used. To prepare the soluble anti- genic extract used for IHA, CFT, and ELISA, the parasites were suspended in distilled water, frozen and thawed, sonicated, and centrifuged at 10 000 g for 1 3929 - 99 - 100 Y. CARLIER ET AL. hour at 4°C. The supernatant fluid, which was dia- lysed against distilled water and lyophilized, is referred to as soluble toxoplasma antigenic extract. In Bilthoven (Netherlands) Toxoplasma antigens for ELISA were prepared from mice similarly infected with 5 x 107 parasites of the RH strain. After washing in phosphated-buffered saline (PBS), the parasites were suspended in distilled water and repeatedly frozen quickly in isopentane (-70°C), then thawed at 37°C. The suspension was then centrifuged at 30 000g for 1 hour at 4°C and the supernatant was used as a soluble antigenic extract. The total protein content was measured according to the Biuret method. For IF, cryostat sections of infected-mouse brain tissue were used as antigen, the mice being infected 2 days before by cerebral inoculation with the RH strain of toxo- plasma (15). In Stockholm, IF was performed using a suspension of extracellular Toxoplasma prepared according to Bodner et al., using 5 x 104 organisms per spot (16). Sera One hundred sera from the serum bank of the Centre d'Immunologie et de Biologie Parasitaire (Institut Pasteur de Lille) were lyophilized and distri- buted randomly between the three laboratories. These sera were classified as indicated in Table 1 using DT, IHA, and CFT (see page 102). The L1 limit was used to Table 1. Classification of the sera using the results of the three serological tests that correlated most closely with the ELISA results No. Inter- of group Groups sera DT IHA CFT limits 1. negative 20 < 1/50 < 1/80 < 1/4 <Llc sera a 11. positive serab69 >1/50 >1/80 >1/4 > L c 111. sera with low 23 1/50- 1/80- 1/4- : Ld antibody titre 1/1250 1/640 1/32 IV. sera with 9 >1/1250 >1/640 >1/32 >L2d high anti- body titre a Groups I and 11 did not include 11 sera, the serological findings on which did not allow their safe classification into one group, but these sera were used for the correlation study. b Group II included groups IlIl and IV but also sera that could not be classified in these groups because the 3 serological results did not fit into the same class. c L, is the limiting value between negative and positive sera. d L2 is the limiting value between sera with low and high antibody titre. divide the sera into negative or positive groups for toxoplasmosis. The L2 limit discriminated between sera with a low antibody titre and those with a higher titre. Serological tests In all the three laboratories, IF was carried out using the procedure described by Ambroise Thomas (17). Commercial fluorescein isothiocyanate (FITC), labelled antihuman immunoglobulins, and/or anti- IgG were obtained, respectively, from Institut Pasteur Production (Paris), Nordic (Tilburg), and National Bacteriological Laboratory (Stockholm). The IF detection of specific IgM antibodies against Toxo- plasma, using the method of Remington et al. (18), was done in Lille using antihuman IgM-FITC conju- gate from Institut Pasteur Production (Paris). CFT was carried out in Lille according to the classi- cal Kolmer technique, using the same soluble Toxoplasma antigenic extract as used in ELISA. IHA was performed in Lille using the same soluble antigenic extract (19), the serum either remaining untreated or being treated with 2-13-mercaptoethanol 1 mol/litre. DT was done in Stockholm using the procedure described by Sabin & Feldman (20). ELISA ELISA was carried out in all three laboratories by the method of Engvall & Perlmann (21) with some modifications (see technical details Table 2). The anti- genic extracts were coated on disposable polystyrene tubes. Before assay, the tubes were washed 3 times and emptied by suction. The sera under examination were diluted 1:500 and incubated in the antigen-coated tubes. After 3 washings, horseradish peroxidase- labelled sheep antihuman immunoglobulins (or anti- IgG or anti-IgM), obtained commercially from Institut Pasteur Production (Paris) or prepared according to Lanner et al. (22), were added at a suit- able dilution and incubated. Excess conjugate was then removed by thorough washing and the substrate was added. The amount of peroxidase fixed to the tubes was measured in a spectrophotometer at 405 nm (orthodianisidine) or at 450 nm (5-aminosalicylic acid) using the substrate solution as the zero. Results for serum samples and blank were expressed in optical density (OD) values. Several assays were done for each sample and the mean was calculated. Statistical analysis The logarithmic transformations of the titres and the arithmetic values ofOD were used. The normality of the distribution in each test was verified. Statistical analysis was performed in Lille using multiple linear TOXOPLASMOSIS: SEROLOGICAL TESTS FOR DIAGNOSIS 101 Table 2. Technical details concerning the ELISA technique used in the three laboratories8 Lille Bilthoven Stockholm Antigen coating concentration Lille Ag Lille Ag Lille Ag (mg/litre) 3 (dry weight) 25 (protein) 2 (protein) Bilt. Ag Bilt. Ag 5 (dry weight) 5 (protein) buffer PBS PBS carbonate (0.01 mol/litre) (0.01 mol/litre) (0.05 mol/litre) pH 7.2 7.2 9.6 time 3 h 1 h overnight temperature 37 0C 37 0C room temp. storage 48h (40C) 40h (4°C) 48h (4OC) Washing fluid PBS + 0.2% Tap-water + PBS + 0.05% Tween 20 0.05% Tween 20 Tween 20 Serum sample dilution 1:500 1:128 1:500 buffer PBS+0.2% PBS+1% BSA PBS +0.05% Tween 20 Tween 20 incubation time 4h 2h 5h incubation room 37 0C room temperature temperature temperature HRP-labelled antihuman 1g sheep serum dilution 1:2000 1:1000(Lille Ag) 1:1000 1:1500(Bilt. Ag) buffer PBS PBS + 0.05% PBS + 0.05% Tween 20 + Tween 20 4% BSA incubation overnight 1 h overnight time incubation 4 OC 37 0C room temperature temperature Enzyme reaction substrate H202+ ortho- H202 + H,O, + dianisidine 5-aminosalicylic 5-aminosalicylic acid acid incubation 1 h 1 h 15 min time incubation room room room temperature temperature temperature temperature stop HCI none NaOH a PBS = phosphate-buffered saline; BSA = bovine serum albumin. regression. The goodness of fit was estimated by analysis of variance using a Wang 2200 computer. Owing to the large number of samples tested, a con- fidence level equal to or less than 0.001 was considered as significant for correlations between the tests. The estimated average values of ELISA were calculated from the multiple regression equation. Using these values, the percentages of co-negativity and co- positivity (ELISA versus DT, IHA, and CFT, the tests that had the highest correlations with ELISA) were calculated in order to evaluate the ELISA results. RESULTS Correlations between all tests without distinction between immunoglobulin classes Results obtained for ELISA using an antitotal human immunoglobulins conjugate and other sero- logical tests were analysed in all possible combi- nations. (a) Correlations between ELISA in the three lab- oratories. The ranges ofOD values obtained for nega- tive and positive reference sera in the three labora- tories are presented in Table 3. Correlations between Table 3. ELISA OD values of blank (without serum) and a range of three negative or positive reference sera Negative Positive Conjugates ELISA8 Blank reference reference sera sera ELL 0.09 0.10 - 0.20 1.20 - 1.25 ELB 0.06 0.05- 0.20 1.20- 1.23 Anti-total EBL 0.010 0.12 - 0.15 0.95- 1.22 lg EBB 0.035 0.10 - 0.13 1.92 - 2.32 ESL 0.004 0.01 - 0.02 0.06 - 0.21 Anti-lgG ELL 0.07 0.06- 0.20 1.20 - 1.22 ELB 0.06 0.06- 0.18 1.04- 1.10 9 For abbreviations, see footnote (b) p. 102. the results obtained in the three laboratories with ELISA, under different technical conditions and using two different antigen preparations are shown in Table 4. Highly significant positive correlations were obtained (0.71r<,0.93; P<0.001). The assay per- formed in Bilthoven, using Bilthoven antigen, showed the highest correlation coefficient with the other ELISA results (0.83<r<0.93; P<0.001). (b) Correlations between classical serological tests. The good correlations between classical serological tests were confirmed (0.75<r<0.91; P<0.001). 102 Y. CARLIER ET AL. Table 4. Correlation coefficients between ELISA results from the three different laboratories (87( df ( 98; 1.2 x 10-11 (P 42.10x 10-8) ELISA° ELL ELB EBL EBB ESL 0.83 0.71 0.77 0.84 EBB 0.88 0.83 0.93 EBL 0.76 0.82 ELB 0.79 8 For abbreviations, see footnote (b) below. Table 5. Correlation coefficients between different ELISA techniques and the other serological tests (83 ( df (4 98; 4.5 x 10-o1° P< 5.3 x 10-6) ELL8 ELB EBL EBB ESL IFL 0.68 0.70 0.54 0.62 0.54 IFB 0.81 0.80 0.67 0.76 0.68 IFS 0.80 0.79 0.62 0.72 0.66 DT 0.76 0.72 0.59 0.69 0.65 CFT 0.83 0.78 0.70 0.80 0.72 IHA 0.76 0.67 0.57 0.70 0.68 8 For abbreviations, see footnotes (a) and (b) below. (c) Correlations between ELISA andthe othersero- logical tests. Correlations between each serological test and ELISA (all laboratories) showed the follow- ing sequence: CFT>IFB>IFS>IHA>DT>IFLa (Table 5). This agrees with the results obtained pre- viously (6, 11-13) in which an excellent correlation between ELISA, IF, IHA, and DT was also found. It may therefore be assumed that ELISA detects anti- bodies synthesized early (as do IF and DT) as well as those appearing later, after the complement-depen- dent lysis of Toxoplasma, i.e., antibodies against the soluble antigens of the parasite as detected by CFT or IHA. The correlations of the ELISA results using the different techniques with those of all serological tests were as follows: ELL>ELB>EBB>ESL>EBL.b In terms of correlation coefficient, ELL and ELB were performed according to the same technique, but using different antigens. This indicates the need to stan- dardize the technique. EBB was intermediate in this a IFB, IFS, and IFL = immunofluorescence performed in Bilthoven, Stockholm, and Lille, respectively. b ELL, ELB = ELISA performed in Lille, using Lille antigen and Bilthoven antigen, respectively; EBB, EBL = ELISA performed in Bilthoven, using Bilthoven antigen and Lille antigen, respectively; ESL = ELISA performed in Stockholm using Lille antigen. classification, which could explain why it showed the highest correlation with the other ELISA results, as indicated in Table 4. Correlations between the tests exploring antibody classes (a) Correlations between ELISA and serological tests detecting IgG antibodies. The correlation coef- ficients between ELISA and IF in the determination of IgG class antibodies and between ELISA and IHA after 2-p-mercaptoethanol treatment of the sera were significant (0.64<r<0.76; P<0.001) (Table 6). No significant difference was seen between the two anti- gens used in ELISA for the determination of IgG class antibodies. Moreover, highly significant correlations were obtained between ELISA using anti-Y-chain or antitotal immunoglobulin conjugates (0.76<r<0.97; P<0.001). Therefore the cheaper enzyme-labelled antitotal human immunoglobulin conjugate can be regarded as preferable. Table 6. Correlation coefficients between different ELISA techniques for total lg and ELISA and other serological tests for IgG (90< df <98; 4.8 x 10-'3< P< 2.8 x 10-7) ELL IgG ELB IgG ELL total Ig 0.97 0.85 ELB total lg 0.78 0.89 IF IgG 0.64 0.71 IHA 2/3MEa 0.76 0.69 B 2fiME = 2-p-mercaptoethanol. (b) Correlations between ELISA and IF for the study ofIgM class antibodies. Only 12 sera with IgM antibodies detected by IF were studied and the results with ELISA were poor. This requires further study. Agreement between ELISA and other serological test values The regression coefficients were calculated between each ELISA technique and each of the other sero- logical tests. Estimated OD values ofELISA were then calculated from the multiple regression equation for each combination of serological tests, and some of these calculated values are shown in Table 7. Three serological tests presented higher regression coef- ficients with all the ELISA: CFT, IHA, and DT. They were therefore chosen to classify the sera as indicated in Table 1. This classification permitted the compari- son between the experimental ELISA values for each serum and the estimated values (L, and L2); it also TOXOPLASMOSIS: SEROLOGICAL TESTS FOR DIAGNOSIS Table 7. Representative OD values of ELISA, estimated from multiple regression equations using serological test results IFL8 IFB IFS DT CFT IHA ELL ELB EBL EBB ESL 150 128 10 10 4 80 0.20 0.15 0.01 0.06 0.01 450 256 40 250 8 160 0.39 0.31 0.17 0.37 0.05 1350 512 80 250 16 320 0.52 0.44 0.29 0.58 0.06 1350 1024 80 1250 32 640 0.70 0.56 0.46 0.89 0.10 4000 1024 160 1250 64 640 0.72 0.64 0.53 0.98 0.10 4000 2048 160 1250 64 640 0.73 0.65 0.55 0.99 0.10 12000 2048 320 6250 64 1280 0.90 0.76 0.65 1.19 0.14 36000 4096 640 6250 128 2560 1.03 0.89 0.77 1.40 0.15 a For abbreviations, see footnotes (a) and (b), p. 102. allowed the calculation of co-positivity, co-negativity, co-negativity, it is likely that in this case ELISA gives and agreement between these three serological tests some false positive results. On the other hand, EBB and the ELISA values, as indicated in Table 8. and EBL using a serum dilution of 1:128, which is The agreement was good for discrimination of sera lower than for the other ELISA (1:500), gave a better with low and high levels of antibodies (L2 = score for discrimination between sera with low or 87.5-100%), the different ELISA techniques being higher antibody titre than for the separation into posi- classified as follows: ELL >EBB>EBL>ELB >ESL. tive and negative sera. This could be explained by the Co-negativity was slightly higher than co-positivity, sigmoid shape of the dose-response curve generally which suggested that ELISA may give some false obtained for different dilutions of the same serum, negative results. For discrimination between positive which leads to overestimation of negative or low anti- and negative sera, agreement (L1 = 77.7-95.5%) body titres and an underestimation of sera with higher gave the following classification: ELL>ELB>EBB> antibody titres (R. Bergquist et al., unpublished ESL >EBL. Co-positivity being generally higher than results). Table 8. Agreement between ELISA, DT, CFT, and IHA Co-positivity Co-negativity Agreement ELISAa Limits No. % No. % No. % L, 0.20 67/69 97.1 18/20 90.0 85/89 95.5 ELL L2 0.70 9/9 100 23/23 100 32/32 100 L, 0.20b 59/69 85.5 20/20 100 79/89 88.8 ELB L2 0.56 7/9 77.7 22/23 95.6 29/32 90.6 L, 0.15b 52/63 82.5 11/18 61.1 63/81 77.7 EBL L2 0.46 5/7 71.4 20/20 100 25/27 92.6 L, 0.13b 62/63 98.4 7/18 38.8 69/81 85.1 EBB L2 0.89 6/7 85.7 20/20 100 26/27 96.3 L1 0.03b 62/69 90 11/20 55 73/89 82 ESL L2 0.11 5/9 55.5 23/23 100 28/32 87.5 a For abbreviations, see footnote (b), p. 102. b The value considered as L, was that of the negative reference serum when the calculated ELISA OD was inferior to this measured value. 103 Y. CARLIER ET AL. DISCUSSION The most interesting result of this study is the excellent correlations obtained between the ELISA techniques using two different antigens in the three laboratories. However, this study underlines the necessity to choose a standardized technique for the performance of the ELISA test, the main parameters of which have been systematically studied (Carlier et al., unpublished results). The use of the linear part of the curve and of several dilutions of each sera with suspected high antibody titre appears mandatory (R. Bergquist et al., unpublished results). Moreover, since OD values obtained for a given serum vary from one procedure to another, it would be useful to use international units instead of OD for comparative clinical studies. The two Toxoplasma soluble antigenic extracts were efficient for detecting IgG antibodies but less efficient for detecting IgM antibodies. In conclusion, this study indicated that ELISA is a reproducible technique for the detection of anti- Toxoplasma antibodies, which appears to be just as satisfactory as the other serological tests. Being sensi- tive and simple to automate and standardize, ELISA may be considered a suitable method for serodiagnosis and particularly for serological screening of toxo- plasmosis. ACKNOWLEDGEMENTS The authors are grateful to Mrs A. Baudry, Mrs B. Dutoit, Mrs A. Caron, Mrs J. Cornette, Mrs S. 0. Panggabean, and Mr L. Hornez for their excellent technical assistance and to Mrs C. Colson for her contribution to the preparation of this manuscript. They are also grateful for the help and encouragement of Dr J. C. Dugimont and Dr P. Wattre. RESUME EVALUATION DU TITRAGE AVEC IMMUNOADSORBANT LIt A UNE ENZYME (ELISA) PAR RAPPORT A D'AUTRES tPREUVES S-ROLOGIQUES POUR LE DIAGNOSTIC DE LA TOXOPLASMOSE Pour evaluer ELISA en tant que methode de serodia- gnostic de la toxoplasmose humaine, trois laboratoires A Lille, Bilthoven et Stockholm ont effectue ce type de titrage sur un meme lot de serums, mais dans des conditions tech- niques differentes (tableau 2). En outre, deux antigenes solubles de Toxoplasma gondii-et des conjugues enzyma- tiques differents-ont et utilises dans chacun des labora- toires de Lille et Bilthoven, ce qui a conduit A classer les resultats des epreuves sous les sigles suivants: ELL, ELB: epreuve ELISA effectuee A Lille au moyen de l'antigene de Lille et de l'antigene de Bilthoven, respectivement; EBB, EBL: ELISA effectuee A Bilthoven au moyen de l'antigene de Bilthoven et de l'antigene de Lille, respectivement; enfin ESL: ELISA effectuee A Stockholm au moyen de l'antigene de Lille. Une analyse statistique (regression lineaire multiple) a ensuite e faite pour comparer les resultats d'ELISA dans ces differentes series d'epreuves entre eux et avec ceux des techniques classiques: epreuve par coloration (dye-test: DT), immunofluorescence (IF), reaction de fixation du complement (RFC) et hemagglutination indirecte (IHA). Des correlations positives hautement significatives ont e obtenues entre les resultats des epreuves ELISA dans les trois laboratoires (0,71 r.0,93; P<0,001). Les corr&- lations entre chacune des epreuves serologiques classiques et l'ensemble des ELISA se sont classees comme suit: RFC> IF>IHA>DT. Les correlations entre chaque serie d'epreuves ELISA et l'ensemble des epreuves serologiques classiques ont permis d'etablir le classement suivant: ELL>ELB>EBB>ESL>EBL. En ce qui concerne la detection de la classe d'anticorps, les correlations entre ELISA et les epreuves IF et IHA ont e significatives pour les IgG (0,64<r<0,76; P<0,001). Si l'on ne considere que les epreuves ELISA, aucune difference significative n'est apparue en fonction de l'antigene utilise. En outre, des correlations hautement significatives ont e etablies pour les ELISA utilisant soit le conjugue anti-immunoglobuline totale-dont le cout est moins eleve-soit le conjugue anti- chaine'Y (0,76(r(0,97; P<0,0001). Pour la detection des IgM en revanche, des resultats mediocres ont e obtenus avec ELISA, mais les experiences n'ont porte que sur 12 serums contenant des IgM decelees par la technique d'immunofluorescence et d'autres etudes sont encore n6cessaires. Des valeurs estimees d'ELISA (exprimees en densite optique:OD) sat e calculees A partir de 1'equation de regression multiple pour chaque serie d'epreuves serolo- giques portant sur des serums de titres differents (tableau 7). Les trois epreuves serologiques presentant les plus forts coefficients de regression avec tous les ELISA, soit CFT, IHA et DT, ont ete en consequence utilisees pour classer les serums analyses en trois groupes: negatifs, positifs avec titre eleve, et positif avec faible titre (tableau 1). Cette classifi- cation a permis de comparer, pour chaque groupe de serums, les valeurs experimentales d'ELISA avec les valeurs estimees, ainsi que d'etablir les taux de co-positivite, co- negativite et concordance pour l'ensemble de ces trois epreuves serologiques comparees A ELISA. Cette derniere 104 TOXOPLASMOSIS: SEROLOGICAL TESTS FOR DIAGNOSIS 105 comparaison donne la relation suivante: ELL>EBB> EBL>ELB >ESL. A la valeur limite entre serums a faible taux et a taux eleve d'anticorps (L2), la concordance a e constatee dans 87,5 a 100% des cas selon le laboratoire (tableau 8). Les taux de co-negativite sont, pour cette meme valeur, legerement superieurs A ceux de co-positivite, ce qui suggere qu'ELISA peut donner quelques resultats fausse- ment negatifs. A la valeur limite entre serums positifs et negatifs (L1 ), la concordance-constat&e dans 77,7 A 95,5% des cas-peut s'exprimer ainsi: ELL>ELB>EBB>ESL> EBL; les taux de co-positivite etant generalement plus eleves que ceux de co-negativite, on peut penser que, dans ce cas egalement, ELISA peut donner quelques resultats fausse- ment negatifs. Le resultat le plus interessant de cette etude est 1'excellente correlation obtenue entre les epreuves ELISA executees dans les trois laboratoires, avec deux antigenes et conjugues enzymatiques differents. ELISA apparalt comme une tech- nique reproductible et aussi satisfaisante pour la detection d'anticorps anti- Toxoplasma que les autres epreuves s&ologiques. Etant sensible, facile a automatiser et a standardiser, elle represente donc une methode pratique de serodiagnostic, en particulier pour le depistage de la toxo- plasmose. REFERENCES 1. WIsDoM, B. Enzyme-immunoassay. Clinical chemistry, 22: 1243-1255 (1976). 2. DUGIMONT, J. C. ET AL. Essais d'automatisation de l'ELISA (Enzyme linked immunosorbent assay). In: Organisation des Laboratoires-Biologie et prospec- tive, Paris, Expansion Scientifique Franqaise, 1975, pp. 627-629. 3. RUITENBERG, E. J. ET AL. Mechanization of the enzyme-linked immunosorbent assay (ELISA) for large scale screening of sera. Journal of immunological methods, 16: 351-359 (1977). 4. RUITENBERG, E. J. & BROSI, B. J. M. Automation in enzyme immunoassay. Scandinavian journal of immu- nology, 8: (Suppl. 7): 63-72 (1978). 5. BOUT, D. ET AL. Critical evaluation of ELISA for diagnosis of parasitic diseases. In: Peeters, H., ed. Pro- ceedings of XXIVth Colloquium on Protides of the Biological Fluids, Oxford, Pergamon Press, 1976, pp. 775-779. 6. BOUT, D. ET AL. Immunodiagnosis of human parasitic diseases by the enzyme-linked immunosorbent assay. In: Feldmann, ed., INSERM Symposium on Immuno- enzymatic Technique, Amsterdam, North Holland Publishing Company, 1975, pp. 175-182. 7. VOLLER, A. ET AL. Enzyme immunoassays for parasitic diseases. Transactions of the Royal Society of Tropical Medicine and Hygiene, 70: 98-106 (1976). 8. RUITENBERG, E. J. ET AL. Enzyme immunoassay for the serodiagnosis of parasitic infections. Biomedicine, 26: 311-314 (1977). 9. Le titrage avec immunoadsorbant lie a une enzyme (ELISA). Bulletin of the World Health Organization, 55: 557-568 (1977). 10. DUGIMONT, J. C. ET AL. Apport des methodes immuno- enzymologiques au diagnostic de masse et a la surveillance de la toxoplasmose humaine. In: S6rologie de l'infection toxoplasmique, en particuliera son debut: methodes et interpretation des resultats, Lyon, Fonda- tion Merieux, 1975, pp. 83-95. 11. VOLLER, A. ET AL. A microplate enzyme immunoassay for toxoplasma antibody. Journal ofclinicalpathology, 29: 150-153 (1976). 12. WALLS, K. W. ET AL. Use of the enzyme-linked immunosorbent assay (ELISA) and its microadaptation for the serodiagnosis of toxoplasmosis. Journal of clinical microbiology, 5: 273-277 (1977). 13. RUITENBERG, E. J. & KNAPEN, F. van. The enzyme- linked immunosorbent assay and its application to para- sitic infections. Journal of infectious diseases, 136: 267-273 (1977). 14. KNAPEN, F. van & PANGGABEAN, S. 0. Detection of circulating antigen during acute infections with Toxo- plasma gondii by enzyme linked immunosorbent assay. Journal of clinical microbiology, 6: 545-547 (1977). 15. RUITENBERG, E. J. ET AL. Toepassing van de immuno- fluorescentie methode in de serodiagnostiek van toxo- plasmose. Volksgezondheid verslagen en mededlingen, 33: 227-230 (1971). 16. BODNER, S. J. ET AL. The purification of Toxoplasma gondii antigen from mouse peritoneal exudate. Trans- actions of the Royal Society of Tropical Medicine and Hygiene, 66: 530 (1972). 17. AMBROISE-THOMAS, P. Immunofluorescence in the diagnosis, therapeutic follow up and seroepidemiologi- cal studies of some parasitic diseases. Transactions of the Royal Society of Tropical Medicine and Hygiene, 70: 107-112 (1976). 18. REMINGTON, J. S. ET AL. IgM antibodies in acute toxo- plasmosis; prevalence and significance in acquired cases. Journal of laboratory and clinical medicine, 71: 885 (1968). 19. WATTRE, P. ET AL. Evaluation des methodes d'hemag- glutination indirecte dans les toxoplasmoses humaines et experimentales. In: S&rologie de l'infection toxo- plasmique, en particuliera son debut: methodes et inter- pretation des resultats, Lyon, Fondation Mrieux, 1975, pp. 49-60. 20. SABIN, A. B. & FELDMAN, H. A. Dyes as micro- chemical indicators of a new immunity phenomenon affecting a protozoon parasite (Toxoplasma). Science, 108: 660-663 (1948). 21. ENGVAL, E. & PERLMANN, P. Enzyme-linked immuno- sorbent assay (ELISA). Quantitative assay of immuno- globulin G. Immunochemistry, 8: 871-874 (1971). 22. LANNER, M. ET AL. Purification of enzyme labelled conjugate by affinity chromatography. In: Hoffman- Ostenhof et al., ed., Proceedings of the International Symposium on Affinity Chromatography, Vienna, September 1978. Oxford, Pergamon Press, 1978, pp. 237-241.