Zbl. Bakt. Hyg., I. Abt, Orig. A 255, 448-455 (1983)

ELISA for IgM and IgG Antibodies Against Tick-borne

Encephalitis Virus: Quantification and Standardization of Results

FSME-Enzymantikorpertest: Quantifizierung und Standardisierungder Ergebnisse

HANNS HOFMANN, FRANZ X. HEINZ, and HEIDE DIPPE

Institute of Virology, University Vienna, Austria (Direktor: Prof. Dr. Christian Kunz)

With 1 Figure' Received January 3, 1983

Abstract

A method for standardizing specific antibody determination by ELISAis presented. Tick­borne encephalitis (TBE)-ELISA results are expressed in arbitrary units by comparing ex­tinction values of test sera (tested in triplicate at a fixed dilution) with the extinction-curveof a standard serum run in the same test. The standard serum was given an arbitrary valueof 1000 Vienna Units, and the corresponding units for the test serum are read from thestandard curve. As a standard for IgM antibody determination, a pool of sera was takenfrom patients with recent TBE at about the time of maximal IgM antibody production.This IgM-standard was used for a three-layer as well as for a four-layer ELISA. The IgGstandard for detection of IgG antibodies in the three-layer ELISA consisted of sera frompersons with a past infection. To exclude low nonspecific results only values higher then50 Vienna Units are regarded as undoubtedly positive. This method for standardizingELISA gives reproducible results, even when plates coated on different days with differentbatches of antigen were used.

Zusammenfassung

Die Extinktionswerte der Testsera, die in dreifachem Ansatz, aber nur in einer Ver­diinnung gepruft worden waren, wurden mit der Extinktionskurve eines Standardserumsverglichen, das im selben Testansatz gepruft worden war. Dem Standardserum wurde einWert von 1000 Wiener Einheiten (Vienna Units) zugeordnet, so dag die entsprechendenEinheiten des Testserums leicht an der Standardkurve abgelesen werden konnten. AlsStandard fiir die IgM-Antikorperbestimmung diente ein Serumgemisch von Patienten mitfrischer FSME urn den Hohepunkt der IgM-Antikorperbildung. Dieser IgM-Standardwurde sowohl fiir den 3-Schichten- als auch fiir den 4-Schichten-ELISA verwendet. AlsStandard fur die IgG-Antikorperbestimmung im 3-Schichten-ELISAdiente ein Gemisch vonSera von Patienten mit langer zuriickliegender Friihsommermeningoenzephalitis. Urn nied­rige unspezifische Ergebnisse auszuschalten, wurden erst Werte iiber 50 Einheiten als sicher

TBE-ELISA Standardization 449

positiv gewertet.Eskonnte gezeigtwerden, daIS die Methode gut reproduzierbare Ergebnisseliefert, selbst wenn Platten verwendet wurden, die an verschiedenen Tagen und mit ver­schiedenen Antigenpraparationen beschichtet worden waren.

Introduction

The ELISA test, as used to demonstrate the presence of specific IgM antibodiesin body fluids, is gaining more and more in practical significance in the field ofdiagnostic virology. This test has already been established as a valuable tool forthe diagnosis of a number of viral infections. The quantification and standardizationof results however still appears to be problematic. Most authors avoid any quanti­fication and only report whether a serum is positive or negative for IgM or IgGantibodies. Occasionally the extinction value is used as a measure of the amountof antibody present. This method allows only a gross quantitative differentiationamong sera in the same test. Many authors quantify specific antibodies determiningthe titer according to the methods of classical serology; but this has many dis­advantages, not the least of such is the high expense. Feigner (1978, 1982) expressesthe antibody content in a serum as the multiple of normal activity (MONA). Thebest method appears to be a comparison of test sera with a standard serum (Leinikkiand Passila, 1977; Leinikki et al., 1978; Malvano et al., 1982). This method is dis­cussed here, using the TBE-ELISA as an example, whereby two different and in­dependent test principles were used.

Materials and Methods

1. Sera: The test and standard sera used in the test were all samples from various hos­pitals in Austria which had been sent to the Institute of Virology for virus diagnosis.

2. Three-layer ELISA (test A): For this test, which has been described elsewhere (Hof­mann et al., 1978), the antigen is bound to the microtiter plate. After addition of patientserum dilutions, the plates are incubated for 2 h at 37°C and after washing with PBS thealkaline phosphatase-linked anti-human IgM or anti-human IgG serum from swine (OrionDiagnostica, Helsinki, Finland) is added, followed by another incubation for 2 h at 37°C.The plates are washed again, substrate is added, and after incubation for 30 min at roomtemperature the absorbance is read on a photometer (Titertek, Multiskan).

3. Four-layer ELISA (test B): This test has also been described elsewhere (Roggendorfet al., 1981; Heinz et al., 1981). An anti-s-specific antiserum is bound to the microtiterplate. After incubation with test serum for 2 h at 37°C, the plates are washed and theantigen is added. Incubation takes place over night at 4°C, followed by a washing andincubation for 2 h at 37°C with horseradish peroxidase-linked anti-TBE-IgG from rabbit.After a 30 min reaction period with the substrate the results are again read on the Multiskan.

Results

1. Antibody determination by means of a standard serum

A pool consisting of 20 sera taken from patients with recent TBE at about thetime of maximum IgM-antibody content was used as TBE-IgM antibody standard.The TBE-IgM-antibody concentration of this pool represents therefore a mean value

450 H.Hofmann, F.X.Heinz, and H.Dippe

of specific IgM-antibodie s present in patients' sera at the peak of the IgM response.This pool, which also contains specific IgG-antibodies, was arbitrarily given a valueof 1000 antibody units - designated by us as Vienna Units.

As a standard for IgG-antibody determination a serum pool from patients in theconvalescent phase at the approximate time of highest IgG antibody productionwas used. Again an arbitrary value of 1000 Vienna Units was assigned to thisstandard.

The standards were frozen at - 20°C in small portions. For the establishment ofstandard curves in each individual test the stand ard sera were diluted in twofoldsteps in a pool of negative sera resulting in sera containing 500, 250, 125 etc. ViennaUnits. All sera thus obt ained were tested in triplicate at a dilution of 1: 100 fortest A and 1: 1000 for test B. A standard curve was drawn from the mean valueof the extinctions obtained. Fig. 1 shows three typical standard curves for the IgG­antibody determination and for the IgM-antibody determin ation in test A and intest B.

1.5

to

0.5

IgG IgM(A) 19M (B)

4 16 63 250 1000 4 16 63 250 1000 4 16 63 250 10008 32 125 500 8 32 125 500 8 32 125 500

Vienna Units

Fig. 1. Sta ndard curves for 3 layer test (IgM and IgG) and 4 layer rest,

2. Determination of the optimal dilution

In the described method a serum is to be tested in only one dilution in tr iplicate.Th is dilution should allow the determination of even small amounts of specific ant i­bodies. If this dilution is too high there is a danger for loss of low antibody con­centrations; on the other hand, if it is too low, unspecificserumfactors may simulatea positive result. To determine the optimal dilution, the mixtures of the standardserum (with negative serum) as well as 10 test sera were tested for test A in dilution s1: 50, 1: 100 and 1: 200 respectively and for test B in dilutions 1:500, 1: 1000 and1: 2000 respectively. From that experiment a dilution of 1:100 in test A and 1: 1000in test B was chosen for all further tests.

TBE-ELISA Standardization 451

3. Antibody determination in sera with high antibody concentrations(> 1000 Vienna Units)

If a serum contains a high amount of antibody (> 1000 V.U.) the mean extinctionvalue will lie outside the standard curve and can therefore not be determined. Insuch a case the serum is also tested at a 10 fold higher dilution (1: 1000 in test A)

Table 1. Results obtained in 3-layer ELISA using different dilutions of test sera

SerumNo.

dilution 1: 100IgG IgM

dilution 1: 1000 (corrected results)IgG IgM

123456789

310360670760

> 1000> 1000> 1000> 1000> 1000

85460110950

> 1000> 1000> 1000> 1000

20

500440990750

3600280069007250

> 10000

160360360790

25002000

> 10000> 10000

20

Table 2. Results of TBE-negative sera expressed in Vienna Units (cut off)

Serum 3-layer ELISA -l-layer ELISAIgG IgM IgM

TBE excluded 1 < 15 < 15 < 102 16 < 15 <103 < 15 < 15 < 104 16 < 15 <105 16 < 15 < 106 20 < 15 <107 16 < 15 < 108 19 17 < 109 16 < 15 < 10

10 < 15 < 15 < 10

Coxsackie B- 11 < 15 < 15 <10Meningitis 12 < 15 < 15 < 10

13 < 15 < 15 <1014 < 15 < 15 <1015 < 15 < 15 < 1016 < 15 < 15 < 1017 43 < 15 1018 < 15 < 15 < 1019 < 15 < 15 52020 < 15 < 15 < 1021 < 15 < 15 1022 < 15 < 15 <10

452 H.Hofmann, F.X.Heinz, and H.Dippe

and the resulting value is multiplied by 10. As can be seen in Table 1, good correla­tion, suitable for diagnostic purposes, was found between results obtained at adilution of 1: 100 and 1: 1000. This is at least valid for the 3-layer ELISA (test A).

4. Determination of a lower limit (cutoff)

Very low extinction values may indicate the presence of specific antibodies, butmay also represent unspecific and therefore false positive results. For this reasonsera from 12 patients having a Coxsackie-B virus infection and no TBE, as wellas 10 sera from patients with suspected but not diagnosed TBE were examined. Asshown in Table 2 only 1 serum from each group had antibody values about ap­proximately 20 V.U. in test A, whereas the others were beyond the detection limitof 15 V.U.

In order to eliminate shomewhat stronger unspecific reactions we suggest a cutoffvalue of 50 V.U. which gave satisfactory results in our 3 years experience with thistest.

The same sera were also tested in the 4-layer ELISA. With one exception all 22sera showed a completely negative result of less than 10 V.U. The exception wasa patient with diagnosed Coxsackie-B infection. Values of 500 Units were foundby repeated testing, although this serum was IgM- and IgG-antibody negative inthe 3-layer ELISA nor did it contain rheumatoid factor. As it is known that serataken very early in the course of TBE also contain some IgG- in addition toIgM-antibodies, this result has to be considered as false positive. Our 3 years ex­perience with the 4-layer ELISA in routine practice however shows clearly, that thistype of false positive result is exceedingly rare. The cutoff of 50 Units establishedfor the 3-layer ELISA is also valid for the 4 layer test. This value is clearly exceededin the presence of specific IgM-antibodies.

5. Reproducibility of the method

To test the reproducibility of the method, 6 sera were examined on 5 differentdays, using plates which had been coated on different days. As can be seen inTables 3 and 4, the variation coefficient was on the average about 25%.

Table 3. Reproducibility of TBE IgM ELISA (3-layer test)

Serum individual values" mean standard coefficientdeviation of variation

1 180 390 190 250 220 246 85 35%2 70 75 70 115 85 83 19 23%3 500 300 325 370 460 391 86 22%4 145 185 150 150 125 151 22 15%5 250 345 265 310 292 43 15%6 390 390 325 290 360 351 44 12%

* The individual values were obtained on different days with different batches of plates.

TBE-ELISA Standardization 453

Table 4. Reproducibility of TBEIgG ELISA

Serum individual values" mean standard- coefficientdeviation of variation

1 135 120 180 100 133 34 26%2 350 300 340 510 310 362 85 23%3 525 255 285 640 360 413 164 39%4 405 230 365 320 225 309 80 26%5 340 225 345 290 260 292 51 17%6 850 490 610 650 670 654 130 20%

" The individual values were obtained on different days with different plates.

Discussion

The ELISA now has an established place in viral diagnosis. For clinical purposesit is often sufficient to report whether or not IgM- and/or IgG-antibodies againsta certain agent are present. For certain clinical problems it is however necessaryto give a quantitative estimate of the antibody concentration, for example whenthe time of onset of infection is unclear and has to be determined by means of thecourse of antibody formation. For this purpose many authors simply report theextinction values (Veytorp et al., 1979; Heck et al., 1980); others determine theantibody titer according to the methods of classical serology (Schmitz et al., 1977;Chia and Spence, 1979; Denoyelet al., 1980; Forghani and Schmidt, 1980; Nicolai­Scholtenet al., 1980; Kleinmanet al., 1981 j Murphy et al., 1981).The former methodis relatively inexact, the latter has proved to be complicated, expensive and lackingreproducibility. For these reasons it appears most favorable to compare the testserum to a standard serum in the same test, as many authors have already done intheir ELISA or RIA. However all these methods omit the fact that the slopes of thedilution curves of test sera and the reference sample are not completely parallel.In our opinion this fact can be neglected as far as clinical diagnosis is concerned.Comparing the 3 above mentioned methods for the assay of toxoplasmosis IgG­antibodies, Malvano and coworkers (1982) also come to the conclusion that thesmallest variation (from test to test; from laboratory to laboratory j with the useof new reagents, by changes of concentration in the reagents) is obtained when resultsare reported as units which are read on a standard curve run in the same test.

Our investigation has shown that this method is well suited for the standardiza­tion of TBE ELISA. The reproducibility of the values obtained - at least for the3-layer ELISA- can be considered to be good (see Tables 3 and 4). With an averagevariation coefficient of 25%, a twofold increase in units can be considered to be asignificant rise in titer - even when comparing tests run on different days withdifferent batches of plates.

It must be emphasized that the above suggested methods are not devoid of prob­lems. The specific antibody content of a serum expressed in units is, for example,dependent on the concentration of other interfering serum factors. In the 3-layerELISA IgG antibodies can cause a lower binding of IgM antibodies due to compe­titive inhibition and in the 4-layer ELISA the determination of specific IgM anti-

454 H.Hofmann, F.X.Heinz, and H.Dippe

bodies depends on the content of total IgM. Both of these errors however, can bekept at a minimum by the' method described, because the reference serum for theIgM-antibody test represents a pool of several patient sera, so that the total IgMcorresponds to the mean normal value. Furthermore, this serum contains only 400V.U. of specific IgG antibodies. It must be further pointed out that the above men­tioned interference is reduced, the higher the test serum is diluted. Therefore in somesera we found slightly higher values in the 3-layer ELISA by testing a dilution of1: 1000 with subsequent correction of the results than we did in testing a dilutionof 1: 100, as can be seen in Table 1. However it must be mentioned that the differ­ences in results in the 4-layer ELISA after examination of a dilution of 1: 10000(with correction by the factor 10) and 1: 1000 were more pronounced (The corre­sponding results are not summarized in a table). In the testing of negative sera wefound in the 4-layer ELISA one serum which gave a false positive result. This isexplainable theoretically if a serum contains the rheumatoid factor. Either therheumatoid factor is bound in a complex with specific IgG antibodies by anti-aantibodies, whereby the antigen binds itself to the specific IgG antibodies, or therheumatoid factor is directly adsorbed on the plastic surface and binds the enzymelinked TBE antibodies from rabbit with its anti-IgG activity. However it is interest­ing that our serum contains no TBE-IgG-antibodies nor a rheumatoid-factor in theLatex test. Thus this single false positive result cannot yet be explained.

In summary the described method is very suitable to estimate the concentrationof specific IgM and IgG antibodies against TBE virus in serum. Such a quantitativeresult may be of significance for the diagnosis of recent cases of TBE, but mostimportant is also allows the quantitative analysis of the antibody response aftervaccination.

References

Cbia, W.K. and L. Spence: Quantitative determination of Cytomegalovirus IgG antibodyby enzyme-linked immunosorbent assay (ELISA). Canad. J. Microbiol. 25 (1979) 1082­1086

DenoyeI, G., A. Gaspar, and Ch. Nouyrigat: Enzyme Immunoassay for Measurement ofAntibodies to Herpes SimplexVirus Infection: Comparison with Complement Fixation,

Immunofluorescent-Antibody and Neutralization Techniques. J. Clin, Microbiol. 11(1980) 114-119

Pelgner, P.: Stepless Antibody Determination with the Stick-ELISA Technique. ResultsExpressed as Multiple of Normal Activity (MONA). Zbl. Bakt. Hyg., LAbt. Orig. A 242

(1978) 100-105Felgner, P.: Ein Verfahren zur quantitativen Bewertung von Antikorpern mit dem Enzym­

Immunoassay. Lab. med. 6 (1982) 253-257Forghani, B. and N.].Schmidt: Humoral Immune Response to Virions and Dense Bodies

of Human Cytomegalovirus Determined by EnzymeImmunofluorescense Assay.J. Med.Virol. 6 (1980) 119-127

Gerlich, W.H. and W.Liier: Selective Detection of IgM-Antibody Against Core Antigenof the Hepatitis B Virus by a Modified Enzyme Immune Assay. J. Med. Virol. 4 (1979)227-238

Heck, F.C., ]. D. Williams, and G.Pruett: Interpretation of Spectrophotometric AbsorbanceValues to Define Results of Enzyme-Linked Immunosorbent Assays. J. Clin. Microbial.11 (1980) 398-401

Heinz, F. X., M. Roggendorf, H. Hofmann, Ch. Kunz, and F.Deinhardt: Comparison of

TBE-ELISA Standardization 455

Two Different Enzyme Immunoassays for the Detection of IgM Antibodies against Tick­Borne Encephalitis Virus in Serum and CSF. J. Clin. Microbiol. 14 (1981) 141-146

Hofmann, H., W.Frisch-Niggemeyer, and F.Heinz: Rapid diagnosis of Tick-Borne ence­phalitis (TBE) by means of enzyme linked immunosorbent assay (ELISA). J. gen. Virol.42 (1979) 505-511

Karsh, l., S.P. Halbert, E. Klima, and A. D. Steinberg: Quantitative Determination of Rheu­matoid Factor by an Enzyme Labelled Immunoassay. J. Immunol. Meth. 32 (1980) 115­126

Kleiman, M. B., C. K. L. Blackburn, S.E. Zimmerman, and M. L. V. French: Comparison ofEnzyme-Linked Immunosorbent Assay for Acute Measles with Hemagglutination Inhibi­tion, Complement Fixation and Fluorescent-Antibody Methods. J. Clin, Microbiol. 14(1981) 147-152

Leinikki, P. O. and S. Passila: Quantitative, Semiautomated, Enzyme-Linked Immunosor­bent Assay for Viral Antibodies. J. infect. Dis. 136 (Suppl.) (1977) 294-299

Leinikki, P.O., I. Shekarchi, P. Dorsett, and ]. L. Sever: Enzyme-Linked ImmunosorbentAssay Determination of Specific Rubella Antibody Levels in Micrograms of Immuno­globulin G per Milliliter of Serum in Clinical Samples. J. Clin, Microbiol. 8 (1978) 419­423

Loon, A.M. van, ]. Th.M. van der Logt, and ]. van der Veen: Enzyme-Linked Immuno­sorbent Assay for Measurement of Antibody Against Cytomegalovirus and RubellaVirus in a Single Serum Dilution. J. Clin. Path. 34 (1981) 665-669

Malvano, R., A.Boniolo, M.Dovis, and M.Zannio: ELISA for Antibody Measurement:Aspect Related to Data Expression. J. Immunol. Meth. 48 (1982) 51-60

Murphy, B. R., M. A. Phelan, D. L. Nelson, R. Yarchoan, E. Tierney, D. W. Alling, and R. M.Chanock: Hemagglutinin-Specific Enzyme-Linked Immunosorbent Assay for Antibodiesto Influenza A and B Viruses. J. Clin. Microbiol. 13 (1981) 554-560

Nicolai-Scholten, M. E., R. Ziegelmaier, F. Behrens, and W. Hopken: The Enzyme-LinkedImmunosorbent Assay (ELISA) for Determination of IgG and IgM Antibodies After In­fection with Mumps Virus. Med. Microbiol. Immunol. 168 (1980) 81-90

Roggendorf, M., F. Heinz, F. Deinhardt, and Ch. Kunz: Serological Diagnosis of Acute Tick­Borne Encephalitis by Demonstration of Antibodies of the IgM Class. J. Med. Virol. 7(1981) 41-50

Schmitz, H., W.-W.Doerr, Tr.Kampa, and A. Vogt: Solid-Phase Enzyme Immunoassay forImmunoglobulin M Antibodies to Cytomegalovirus. J. Clin. Microbiol. 5 (1977) 629-634

Vejtorp, M., E.Fanoe, and ].Leerhoy: Diagnosis of Postnatal Rubella by the Enzyme­Linked Immunosorbent Assay for Rubella IgM and IgG Antibodies. Acta path. micro­bioI. scand. B 87 (1979) 155-160

Yolken, R. H.: Enzyme-linked Immunosorbent Assay (ELISA): A Practical Tool for RapidDiagnosis of Viruses and Other Infections Agents. Yale J.BioI. Med. 53 (1980) 85-92

Ziola, B., O. Meurman, M.- T. Matikainen, A. Salmi, and]. L. Kalliomiiki: Determination ofHuman Immunoglobulin M Rheumatoid Factor by a Solid-Phase RadioimmunoassayWhich Uses Human Immunglobulin G in Antigen-Antibody Complexes. J. Clin, Micro­bioI. 8 (1978) 134-141

Prof. Dr. Hanns Hofmann, Institute of Virology, Kinderspitalgasse 15, A-1095 Vienna,Austria

30 Zbl. Bakt. Hyg., I. Abt. Orig. A 255