an antigen prepared from rabbit corneas infected lack of standardization which prevents rigorous,

Transcription

1 JOURNAL OF CLINICAL MICROBIOLOGY, Feb. 1980, p Vol. 11, No /80/ /06$02.00/0 Enzyme Immunoassay for Measurement of Antibodies to Herpes Simplex Virus Infection: Comparison with Complement Fixation, Immunofluorescent-Antibody, and Neutralization Techniques GERARD A. DENOYEL,* ALEXANDRE GASPAR, AND CHRISTIAN NOUYRIGAT Service des Virus, Institut Pasteur de Lyon, Lyons Cedex 2, France An enzyme immunoassay (EIA) was developed for detecting antibody to herpes simplex virus, and the results were compared with those of complement fixation, indirect immunofluorescent-antibody, and plaque reduction neutralization tests. Test sera showed very little nonspecific reactivity even at a starting dilution as low as 1:10. EIA results showed excellent correlation with results obtained by the neutralization test, with an average gain in sensitivity of EIA proved very useful in detecting current herpes simplex virus infection, and antibody appeared in all cases soon after clinical onset. EIA appears to be a rapid, sensitive, and specific method for routine demonstration of herpes simplex virus antibody in a clinical setting. Enzyme immunoassays (EIAs) have been introduced as an alternative to radioimmunoassay for the detection of soluble antigens and for titration of antibodies in biological fluids (5, 6). Since its introduction, EIA has been applied to serodiagnosis of viral diseases in humans (17). Although results have been encouraging, problems still exist concerning the nature of the support, the antigen, and its fixation. The classic technique (5) involves passive adsorption of antigen onto a plastic support in an alkaline medium. Prior testing is necessary to select the most satisfactory type and shape of support to be used (2). Even then, quality variations from one lot of plates to another can hamper reproducibility of the test. Antigen fixation can be improved by covalent binding of antigen to bovine serum albumin with glutaraldehyde (16). Another approach is simply to fix antigen to the plastic support with formaldehyde (11). The nature of the antigen employed can influence the reliability of the reaction. Gilman and Docherty (8) emphasized the necessity of using a partially purified herpes simplex antigen. This reduced excessive background activity which rendered interpretation of results impossible. Under these conditions they obtained a good correlation between complement fixation (CF) and EIA results. EIA gave antibody titers 100- to 200-fold higher than those obtained by CF. However, results were not reliable when sera were tested at a starting dilution below 1:256. This high level of nonspecificity makes it difficult to compare EIA with other techniques currently used for serodiagnosis of 114 herpetic infections. Vestergaard et al. (16) used an antigen prepared from rabbit corneas infected with herpes simplex virus type 1 (HSV-1), solubilized in detergent, and subjected to ion-exchange chromatography to recover a glycoprotein fraction. When this antigen was covalently fixed to plastic, background activity was considerably reduced and there was excellent correlation between EIA and plaque reduction neutralization results, with EIA giving an appreciable increase in antibody titer. Despite the problems indicated above and the lack of standardization which prevents rigorous, overall interpretation of the results, EIA is considered to be a highly sensitive, specific, and easily performed test. In light of the results obtained by the abovementioned authors, it seemed of interest to test an immunoenzymatic technique for routine serodiagnosis of HSV disease. We used an antigen that was extracted from infected cells in alkaline medium and fixed to a plastic support with formaldehyde. The results of EIA have been correlated with those of other techniques routinely performed in our laboratory: CF, indirect fluorescent-antibody, and plaque reduction neutralization. MATERIALS AND METHODS Sera. A total of 164 sera were tested. First, 110 sera from pregnant women aged 18 to 36 years were collected between weeks 12 and 14 of pregnancy. Sera were immediately randomized and dispensed aseptically for independent use in the various serological tests. Care was taken to avoid hemolysis and bacterial

2 VOL. 11, 1980 contamination. The remaining 54 sera were collected from subjects undergoing active herpesvirus infection. Twelve paired sera came from HSV- 1-infected individuals (eight gingivostomatitis cases and four meningoencephalitis cases). Six paired sera were from patients undergoing active cytomegalovirus (CMV) infection and nine paired sera were from those with active varicella-zoster (VZV) disease. The causal virus was isolated from the subjects, and CF seroconversions to the homologous antigen were observed. Sera were stored at -20 C before use in HSV EIA. Cells. MRC-5 (10) human embryonic lung fibroblasts (passages 25 to 29) were used for virus and antigen production. Cells were grown in Eagle basic medium containing 0.5% sodium bicarbonate and 10% fetal calf serum. After HSV infection, cells were maintained in the same medium without serum. Viruses. The MacIntyre strain of HSV-1 and the MS strain of HSV-2 were obtained from the American Type Culture Collection. Stock suspensions prepared on MRC-5 cells were titrated by plaque assay (13). The titers were: HSV-1, 1065 plaque-forming units per ml; HSV-2, 10"' plaque-forming units per ml. CF test. CF antibody for HSV was assayed by the standard Microtiter procedure used in our laboratory (14). HSV antigen from infected Vero cells (Institute Virion, Ltd.) was obtained from Dynatech Laboratories, Rungis, France. It was used at a working dilution of 1:6. Indirect fluorescent-antibody method. MRC-5 cells were grown directly on flame-sterilized Cooke Microprint slides (10 spots per slide). A 0.05-ml volume of cell suspension (250,000 cells per ml) in Eagle basic medium with 10% fetal calf serum was deposited on each spot. The slides were placed in a sterile petri dish and incubated in a 5% CO2 atmosphere for 2 h at 36 C. After this time, Eagle basic medium was added (20 ml per dish), and incubation was continued until cells were confluent (within 48 h). The cells were then infected with HSV-1 at approximately 2 plaque-forming units per cell. After 24 to 36 h, cells exhibiting a 75% cytopathic effect were washed three times with 0.01 M phosphate-buffered saline (PBS) (ph 7.4), fixed in acetone for 15 min at 4 C, and stored at -60 C until used for testing. Control slides of uninfected MRC-5 cells were prepared in parallel. Twofold dilutions (1:10 to 1:2,560) of sera were applied to the infected cells (0.025 ml per spot). Each dilution was run in duplicate. Control titrations (1:10 to 1:160) of sera on uninfected cells were performed simultaneously. The staining procedure was the same as that described by El Falaky et al. (4). Fluorescein-labeled rabbit anti-human immunoglobulins (Dako Immunoglobulins, Copenhagen, Denmark) were used at a working dilution of 1:50. The serum titer was expressed as the reciprocal of the highest dilution giving evident green fluorescence on at least 50% of the cells. A Zeiss microscope equipped with a halogen-quartz lamp (Osram, 12 V), an interference filter, and a 515-nm barrier filter was used to read the reactions. Neutralization test. The plaque reduction test was performed as described by Schmidt et al. (13), using MRC-5 cells in Costar cluster dishes (model 3506; Costar, Cambridge, Mass.), with 200 plaque- HSV EIA 115 forming units of virus and approximately 10 hemolytic units of complement (1) per well in serum-virus mixtures. HSV-neutralizing antibody titers were expressed as the highest dilution of serum producing 50% or greater reduction in plaque count. EIA procedure. (i) Antigen production. Cells were grown in disposable roller bottles (850 cm2, Corning Glass Works, Corning, N.Y.). The monolayer was infected with HSV-1 or HSV-2 cell-free viral suspensions at approximately 2 plaque-forming units per cell. When the cytopathic effect was complete, the medium was discarded. Cells were washed in Hanks basic salt solution, scraped off the glass, centrifuged at 650 x g for 10 min, suspended in 0.1 M glycine-nacl- NaOH buffer (ph 10) (3 ml for the harvest of each bottle), and then sonicated (3 x 1 min) with an Alcatel- Pons apparatus type S (tip T 44c, output control 4). After centrifugation at 10,000 rpm for 10 min (30.2 rotor, Beckman Instruments), the antigen (supernatant fluid) was removed and stored at -80 C until used for testing. Control antigen was obtained from uninfected MRC-5 cells treated in the same way. Antigens were used at optimal concentration as determined by block titration against six human immune sera and six antibody-negative sera (as shown by plaque reduction test). The working dilution was one giving maximum reactivity with immune sera and a negligible reaction with negative sera. For the test, the antigen dilution was 1:400, corresponding to approximately 120,ug of protein per ml, as determined by ultraviolet absorption using bovine serum albumin as a standard. (ii) Microplate sensitization. Standard polystyrene flat-bottomed microplates (Cooke Microtiter M 29 AR, Dynatech, France) were usually used. For comparison, flat-bottomed Linbro IS-FB 96 plates, round-bottomed Cooke M 24 A plates, Linbro IS-MRC 96 plates, and plates treated for tissue culture (Cooke M 29 ART, Linbro IS-FB 96 TC) were also tested. An optimal dilution of HSV or control antigen in PBS was added to the wells (0.1 ml per well) and allowed to dry at 37 C. Then 0.1 ml of a 10% Formalin solution in PBS was added to each well, and plates were held at 40C for 15 min. Plates were washed twice with distilled water. The wells were filled (0.3 ml) with PBS containing 0.5% bovine serum albumin fraction V and held at 40C for 18 h. After two additional washings with PBS containing 0.05% Tween 20, the plates were used in tests. (iii) EIA reagents. Horseradish peroxidase-conjugated goat anti-human immunoglobulin G was obtained from a commercial source (Cappel Laboratories, Inc., Cochranville, Pa.). The optimal dilution for EIA as determined in preliminary titrations was 1:400. To detect enzyme activity by color change, a solution containing H2O2 and 5-amino-2-hydroxybenzoic acid was prepared as described by Ruitenberg et al. (12). (iv) EIA test. Samples (0.1 ml) of serial twofold dilutions of sera (1:10 to 1:5,120) in PBS were added to sensitized wells. Plates were placed at 37 C in a moist chamber. Each serum was tested against HSV and control antigen. After 1 h, plates were washed three times (5 min per wash) with PBS-Tween, and 0.1 ml of optimally diluted conjugate was added to

3 116 DENOYEL, GASPAR, AND NOUYRIGAT each well. Incubation and washings were repeated as described above. Enzyme substrate solution (0.2 ml) was added to each well, and the plates were placed at 20 C for 30 min in the dark. The enzyme reaction was stopped by adding 1 drop of 4 N NaOH. The amount of brownish color was rapidly measured directly through the wells in a Vernon PHI-4 spectrophotometer at 450 nm. The endpoint of the reaction was reached when the absorbance with HSV antigen was the same as that with control antigen. Statistical analysis. Coefficients of correlation (r) were calculated for the different serological techniques. RESULTS Reaction with control antigen. The levels of light absorption obtained by mixing only the substrate with either control or HSV antigen were always virtually identical: they were negligible and were considered as the base line of the test. Only six sera reacted with control antigen, yielding titers between 1:10 and 1:80. Of these, four sera were hemolyzed and two were contaminated with bacteria. Background reactivity increased in parallel with increases in the concentration of antigen and conjugate, demonstrating the necessity for carefully determining optimal concentrations of these reagents in preliminary tests. Correlation with other serological tests. Figure 1 compares results obtained with EIA to those demonstrated by CF (A), indirect fluorescent antibody (B), and plaque reduction neutralization (C) for 110 sera from asymptomatic, pregnant women. These techniques correlated significantly with EIA, which proved to be from 2- to 16-fold more sensitive. Thirteen sera were negative (<1:5) by CF. Of these, 11 had neutralizing titers of <1:10, and 9 were negative (<1:10) by indirect fluorescent antibody and EIA. It thus appears that the basic lack of sensitivity of CF may lead to false-negative results (15). Two sera were positive by EIA although they had titers of <1:10 by neutralization. It is difficult to know whether these sera were truly positive as detected by the increased sensitivity of EIA, or whether EIA titers represent a nonspecific reaction, even though these sera did not react with control antigen (<1:10). ETA specificity. To test specificity, paired sera from patients with active VZV and CMV infections were tested in the HSV EIA reaction. The paired sera all showed a greater than four- FIG. 1. Comparison of HSV EIA antibody titers with titers obtained by CF (A), fluorescent-antibody (IFA) (B), and plaque reduction neutralization (NEUT) (C). Sera tested were from 110 women without evidence of current HSV infection. r = Coefficient of correlation: CF, 0.50; fluorescent antibody, 0.83; neutralization, Be. < 40- LL ( w L1. 4. l A B a S a I0 13 a2 S I s 4 a A 2 5 C CF 4 to 1I 2 5 I I 1 5 I/ s0 0 I FA J. CLIN. MICROBIOL. I 2 2 s * S /1 I/ * / <to NEUT //

4 VOL. 11, 1980 fold CF antibody increase to the corresponding VZV or CMV antigen, and in all cases, the causal virus was isolated (Table 1). No cross-reactions were observed with CMV-positive sera either in HSV CF or HSV EIA. Only one pair of sera from a patient with herpes zoster showed an appreciable increase in HSV EIA antibody (no. 7: titers 1:40 to 1:320), whereas the HSV CF test showed antibody titer rises in all cases, ranging from two- to fourfold dilution. Although encouraging, these results should be confirmed on a greater number of sera. All sera yielded similar titers whether tested against HSV-1 or HSV-2 antigen. EIA reproducibility. To test reproducibility, 35 sera with titers between <1:10 and 1:640 were tested several times by different persons working independently and unaware of each other's findings. In all cases, titers obtained for each serum were remarkably reproducible with an error not greater than two dilutions, but mostly one dilution. Greatest reproducibility was seen for titers of <1:10 and for low to moderate titers (1:10 to 1:160). TABLE 1. Specificity of HSV EIAa Infecting HSV antibody ti- Patient Infecting vi- virus CF ters no. rus antibody titers CF EIA 1 CMV CMV < CMV CMV < CMV < CMV , VZV VZV < VZV VZV < VZV < VZV <5 < VZV < VZV < VZV <5 < ahsv antibody response in six patients with current CMV infections and nine patients with active VZV infections. HSV EIA 117 Influence of the plastic support. The 35 sera used above were tested in parallel by the same technician on various types of plates (see Materials and Methods). Results were the same with, at most, a one-dilution difference for flatbottomed plates. When plates treated for tissue culture were used, results were not as reliable due to an increased background level. With round-bottomed plates, reading of wells was irregular due to the design of the spectrophotometer used. However, visual examination was feasible using our enzyme-substrate system, and it yielded valid results. Storage of sensitized plates. Plates were coated, treated with formaldehyde, and then stored at -20 C. The 35 sera used for reproducibility tests were assayed successively eight times in the course of 1 month by the same technician, each time with a series of stored plates. Results were remarkably consistant, and variations in titers did not exceed a onefold dilution. In 280 assays, the same titers were obtained 148 times, a one-dilution difference was obtained 116 times, and a two-dilution difference was found 16 times. Negative titers (<1:10) were TABLE 2. Reliability of HSV EIA for serodiagnosis of current HSV infections particularly consistent. Eight separate testings of three sera with titers <1:10 resulted in only three weakly positive reactions (1:10). Reliability of EIA for diagnosis of active HSV infection. Table 2 compares results ob- Patient HSV antibody titers by Clinical ~~~~Days after Nu No.Age impression Neution onset 1 4 Stomatitis , Stomatitis 2 <5 < Stomatitis , Stomatitis 3 <5 < , Stomatitis 5 < , Stomatitis 3 < , Stomatitis , Stomatitis 3 <5 < , Encephalitis 3 < , Encephalitis ,280 2, Encephalitis , ,280 5, Encephalitis 4 < ,120

5 118 DENOYEL, GASPAR, AND NOUYRIGAT tained using CF, plaque reduction neutralization, and EIA for 12 paired sera from patients undergoing active HSV-1 infections. Increased titers of EIA as compared to CF were evident in that seven acute-phase sera had CF titers <1:5 and EIA titers ranging from 1:20 to 1:160. Only three acute-phase sera were negative (<1/10) by neutralization and positive in EIA, with titers of 1:20, 1:40, and 1:80. The last result was obtained with a serum which also reacted with the control antigen (1:40) due to bacterial contamination and therefore represented a nonspecific reaction. Although each technique used in this study was able to detect an active HSV infection, the EIA antibody titers were consistently higher. In addition, EIA detected antibodies before other tests did. DISCUSSION In our hands, EIA proved to be a sensitive reaction which demonstrated a low number of nonspecific reactions. As opposed to results of Gilman and Docherty (8), no appreciable background reactivity was detected even at a starting serum dilution as low as 1:10. However, as a measure of security, the initial serum dilution was fixed at 1:20 for future determinations. The only problems in interpreting the results occurred with sera that were hemolyzed or contaminated with bacteria. The problem of hemolysis was circumvented by careful collection of blood and rapid serum separation. Aseptic handling of sera and the addition of sodium azide resolved the problem of contamination. EIA correlated well with CF (r = 0.50; P < 0.001) and particularly with plaque reduction neutralization (r = 0.73; P < 0.001), and gave, in the latter case, an average gain in sensitivity of 1.65-fold. EIA also showed an excellent correlation with indirect fluorescent antibody (r = 0.83; P < 0.001), which uses the same methodological principle but a different means of visualization. The advantage of EIA is that it depends on a less subjective mode of interpretation when spectrophotometry is used. In addition, EIA can be read easily and reliably with the naked eye. Our results are in general agreement with those of Forghani et al. (7), who studied VZV EIA. Few nonspecific reactions were observed with sera of patients with VZV infection, and none was observed with those from CMV-infected individuals. A demonstrable EIA antibody response appeared soon after the clinical onset of active HSV infection. Nevertheless, it was possible to detect a clear seroconversion in all cases. However, it was difficult to determine a possible significant titer in a single serum. It could be suggested that titers of -1:2,560 might indicate a recent infection. J. CLIN. MICROBIOL. Antigen prepared as described under Materials and Methods could not distinguish between HSV-1 and HSV-2 antibodies. This was expected in light of the strong antigenic relationship between the two serotypes (9, 18), and the type determination of antibody was not the purpose of this work. Type-specific antibodies should therefore be sought by using a purified antigenic fraction, and this point deserves a further investigation. It is remarkable that other techniques, and especially CF, yielded a wide range of titers for the same EIA titer, particularly in sera with low and moderate titers. A possible explanation might be that after primary herpesvirus infection residual antibodies are heterogeneous, with variable reactivity among different subjects. Due to its sensitivity, EIA might detect nonavid antibodies more reliably than do other tests. It is also possible that the greater precision and objective interpretation of EIA result in less diversity among titers. As demonstrated elsewhere, different lots of plates can attach antigen in an irregular manner (3). For this reason, we investigated the fixation of antigen with formaldehyde (11). This required more time for preparing sensitized plates but also permitted us to use plates of different makes, shapes, and batches, except for plates treated for tissue culture which consistently gave high background reactivity. In our hands, EIA demonstrated desirable properties for a medical virological laboratory in terms of sensitivity, reproducibility and specificity, ease of performance and adaptability to daily routine work, use of reagents that are easy to prepare or to obtain, and objective reading of results. However, to ensure valid results, it was essential to use correctly collected sera, to carefully pretitrate both antigen and conjugate, and to wash plates thoroughly to avoid background reactivity which interfered with valid reading. Taking these precautions into consideration, we feel that EIA is currently a good routine tool for serodiagnosis of herpetic infections. ACKNOWLEDGMENTS We thank Florian Horodniceanu, Susan Michelson, and Ewald Edlinger, Pasteur Institute, Paris, France, for encouragement and advice. We particularly thank Y. Jacquet and W. Beal for excellent technical assistance. LITERATURE CITED 1. Benyesch-Melnick, M Human cytomegalovirus, p In E. H. Lennette, E. H. Spaulding, and J. P. Truant (ed.), Manual of clinical microbiology, 2nd ed. American Society for Microbiology, Washington, D.C. 2. Bidwell, D. E., A. Bartlett, and A. Voller Enzyme immunosorbent assay for viral diseases. J. Infect. Dis. 136(Suppl.):S274-S Chessum, B. S., and J. R. Denmark Inconstant ELISA. Lancet i: 161.

6 VOL. 1 l, El Falaky, I. H., B. F. Vestergaard, and A. Hornsleth IgG, IgA and IgM antibody responses in patients with genital and non-genital herpetic lesions determined by the indirect fluorescent antibody method. Scand..J. Infect. Dis. 9: Engvall, E., K. Jonsson, and P. Perlmann Enzyme-linked immunosorbent assay. II. Quantitative assay of protein antigen, immunoglobulin G, by means of an enzyme-labeled antigen and antibody coated tubes. Biochim. Biophys. Acta 251: Engvall, E., and P. Perlmann Enzyme-linked immunosorbent assay, ELISA. III. Quantitation of specific antibodies by enzyme-linked antiimmunoglobulin in antigen-coated tubes. J. Immunol. 109: Forghani, B., N. J. Schmidt, and J. Dennis Antibody assays for varicella-zoster virus: comparison of enzyme immunoassay with neutralization, immune adherence hemagglutination, and complement fixation. J. Clin. Microbiol. 8: Gilman, S. C., and J. J. Docherty Detection of antibodies specific for herpes simplex virus in human sera by the enzyme-linked immunosorbent assay. J. Infect. Dis. 136(Suppl.):S286-S Hayward, G. S., N. Frenkel, and B. Roizman Anatomy of herpes simplex virus DNA: strain differences and heterogeneity in the locations of restriction endonuclease cleavage sites. Proc. Soc. NatI. Acad. Sci. U.S.A. 72: Jacobs, J. P., C. M. Jones, and J. P. Baille Characteristics of a human diploid cell designated HSV EIA 119 MRC-5. Nature (London) 227: Lewis, V. J., W. L. Thacker, and S. H. Mitchell Enzyme-linked immunosorbent assay for chlamydial antibodies. J. Clin. Microbiol. 6: Ruitenberg, E. J., P. A. Steerenberg, B. J. M. Brossi, and J. Buys Serodiagnosis of Trichinella spiralis infections in pigs by enzyme-linked immunosorbent assays. Bull. W.H.O. 51: Schmidt, N. J., E. H. Lennette, and R. L. Magoffin Immunological relationship between herpes simplex and varicella-zoster viruses demonstrated by complement fixation, neutralization and fluorescent antibody tests..j. Gen. Virol. 4: Sever, J. L Application of a microtechnique to viral serological investigations. J. Immunol. 88: Smith, J. W., J. F. Peutherer, and F. 0. MacCallum The incidence of herpes virus hominis antibody in the population. J. Hyg. 65: Vestergaard, B. F., P. C. Grauballe, and H. Spangaard Titration of herpes simplex virus antibodies in human sera by the enzyme-linked immunosorbent assay (ELISA). Acta Pathol. Microbiol. Scand. Sect. B 85: Voller, A., D. Bidwell, and A. Bartlett Microplate enzyme immunoassays for the immunodiagnosis of virus infections, p In N. R. Rose and H. Friedman (ed.), Manual of clinical immunology. American Society for Microbiology, Washington, D.C. 18. Wildy, P Antigens of herpes simplex virus of oral and genital origin. Cancer Res. 33: