A simplified and reliable enzyme-linked immunosorbent assay (ELISA) was applied to the detection .... were homogeneous for all the control wells of each plate;.
JOURNAL OF CLINICAL MICROBIOLOGY, Oct. 1986, p. 596-599 0095-1137/86/100596-04$02.00/0 Copyright © 1986, American Society for Microbiology
Vol. 24, No. 4
Simplified Enzyme-Linked Immunosorbent Assay for Specific Antibodies to Respiratory Syncytial Virus L. VAUR, H. AGUT, A. GARBARG-CHENON, G. PRUD'HOMME DE SAINT-MAUR, J. C. NICOLAS, F. BRICOUT* Laboratoire de Virologie, H6pital Trousseau, 75571 Paris Cedex 12, France
AND
Received 10 January 1986/Accepted 30 June 1986
A simplified and reliable enzyme-linked immunosorbent assay (ELISA) was applied to the detection of serum antibodies against respiratory syncytial virus (RSV). RSV-infected cells were fixed and dried on 96-well microtiter plates and kept at 4°C. The titers of reference sera were determined by endpoint dilution. A linear relation was found between the titers and the logarithm of absorbance values of sera diluted to 1:1,000 (r = 0.93, P < 0.001). Measurement of RSV antibodies wvas done by using a single serum dilution (1:1,000) in conjunction with a standard curve. A strong correlation was found between complement fixation and ELISA results (r = 0.89, P < 0.001). In addition, the ELISA method exhibited higher titers and a greater sensitivity than did complement fixation, although the applicability of the assay is limited with positive serum samples of low titer.
months without any problem regarding the reproducibility of results. Serum specimens. Sera were obtained from young children admitted to Trousseau Hospital and were stored at -20°C. The sera were tested for RSV antibodies by conventional CF assay (3). The immunofluorescence (IF) test was performed as previously reported (2). ELISA procedure. Before use, plates were washed three times with a 0.05% Tween 20 solution. Sera were diluted in phosphate-buffered saline containing 0.1% Tween and 1% bovine serum albumin (Sigma Chemical Co., St. Louis, Mo.). A 200-.ld sample of diluted serum was added to each well of antigen-coated plates and incubated at 4°C for 24 h. Plates were washed three times with 0.05% Tween solution. Anti-human immunoglobulin G, specific for -y chains, labeled with peroxidase (Institut Pasteur Production) was diluted in the same buffer as were sera, and 200 RI was added per well; the mixture was incubated for 1 h at 37°C. The plates were then washed six times with 0.05% Tween solution, after which 0.2 ml of o-phenylenediamine substrate (0.4 mg of o-phenylenediamine per ml in 0.1 M citrate phosphate buffer [pH 5.0] containing 0.015% hydrogen peroxide) was added per well. After about 1 min of incubation at room temperature, the reaction was stopped by adding 0.075 ml of 2 M H2SO4 per well, and the optical density at 492 nm (OD492) was measured by spectrophotometry. For each serum, the test was performed in three wells, two with infected cells and one with mock-infected cells as a control. The absorbance value was calculated as the difference between the mean absorbance value of infected cell wells and the value of the control well. The OD492 values were homogeneous for all the control wells of each plate; this background value never exceeded 0.15. Determination of ELISA titer by endpoint dilution. Sera were diluted twofold from 1:500 up to 1:128,000, and the absorbance value was determined for each dilution. Ten sera known to be negative by CF and IF tests were assayed by this method, and a mean curve was determined. This mean curve plus four standard deviations was arbitrarily defined as the reference curve of negative sera. The titer was the reciprocal of the endpoint dilution defined by the intersec-
For many years, complement fixation (CF) has been used as the basis for serological diagnosis of respiratory syncytial virus (RSV) infections. However, the use of CF presents some limits, especially for the detection of an increase in the amount of antibody in sera of young children (13). A neutralization test by plaque reduction has been used more recently (1, 13). This method exhibits a greater sensitivity, but is much too time-consuming to be used in routine
diagnosis. The enzyme-linked immunosorbent assay (ELISA) was developed to detect antibodies against members of the Paramyxoviridae family such as parainfluenza virus (6), measles virus (7), and mumps virus (8). Some authors have proposed the application of ELISA to RSV serology by using purified antigens (10), cell lysate antigens (9, 14), or a crude extract from infected cells (13). Good results were obtained in all cases. An ELISA method using infected cells coated onto plates as antigen has been described for detecting antibodies against measles virus (12). This method was very attractive, because it did not require any antigen purification or cellular treatment. We report the application of this simplified method to RSV serology. MATERIALS AND METHODS Infected cells. Vero cells were grown at 37°C in Eagle minimal essential medium (MEM) supplemented with 5% fetal calf serum (Seromned Biochrom). Cells were infected by RSV wild-type strain 8011245 which was previously isolated from a young child in Trousseau Hospital, Paris, France. After about 6 days, when cytopathic effect was maximal (the number of syncytia corresponded to about 70% of the infected cell monolayer), the cells were collected, washed twice in phosphate-buffered saline, suspended in phosphatebuffered saline at various concentrations, and plated into 96-well ELISA plates (Microtest Immuno; Nunc, Roskilde, Denmark). Mock-infected cells were treated in the same way. The plates were desiccated during 24 h at 37°C and stored at 4°C before assay. They could be used for at least 3 *
Corresponding author. 596
VOL. 24, 1986
ELISA FOR RSV SEROLOGY
0.D
597
08 0.7
0.6 0.5 0.4 0.3 0.2
0 5
10
20
40
80
160
320 640
Reciprocal of serum dilution (
1280 10 2
FIG. 1. Endpoint dilution ELISA for RSV antibodies. Results for five positive sera are shown. vz'z. Negative sera. The lower boundary of the hatched area represents the mean curve for negative sera, whereas the boundary of the hatched area represents the reference curve for negative sera (see Materials and Methods). ELISA titers (arrows) were determined as described in Materials and Methods. A, 1:4,000; A, 1:8,000; *, 1:32,000; O, 1:128,000; *, 1:128,000.
tion between the curve of the tested serum and the reference curve of negative sera (15).
RESULTS Determination of optimal parameters for ELISA. Several densities of cells (1,000 to 25,000 cells per well) were studied. A density of 5,000 cells per well was ultimately chosen. This cell concentration was the lowest which gave a strong absorbance signal and allowed easy and reproducible measurement of antibody titers. At lower concentrations, the absorbance signal was not strong enough to permit an accurate quantitative study. Four dilutions of antiglobulin were tested: 1:1,000, 1:2,000, 1:3,000, and 1:4,000. The last two dilutions provided weak signals, and later experiments were performed with a dilution of 1:2,000. Different positive sera which had been tested by CF were diluted twofold from 1:10 up to 1:1,280, and the ELISA was done by using previously defined optimal parameters for cell concentration and antiglobulin dilution. The results showed a prozonelike effect with a paradoxical increase of OD492 for dilutions from 1 to 1:640. Moreover, at the lowest dilutions, negative reference sera exhibited OD492 values of up to 0.15, and the range was too narrow to differentiate positive and negative sera in a significant way (data not shown). Consequently, it was decided to use sera dilutions higher than 1:500, the dilution which roughly corresponded to the disappearance of the prozonelike effect. Determination of ELISA titer by a single dilution of serum. To define a specific ELISA titer, the amounts of RSV antibodies in different positive sera were measured by endpoint dilution as previously described (15). OD492s decreased with serum dilution (Fig. 1), which was consistent with the absence of a prozonelike phenomenon. The titers of positive sera were measured as described in Materials and Methods (Fig. 1, arrows). The highest titers corresponded to the sera which provided the highest OD492s when tested at dilutions lower than the endpoint dilution. It could be therefore hypothesized that the OD492 at a single dilution correlated with the ELISA titer measured by endpoint dilution.
To check this hypothesis, the ELISA titers of 24 sera were measured by endpoint dilution and compared with the corresponding OD492s obtained with a dilution of 1:1,000 (Fig. 2). A strong correlation was found, with a regression coefficient of 0.93 (P < 0.001). We concluded that the ELISA titer of a serum could be obtained by referring the OD492 at a single dilution (1:1,000) to the standard linear curve (Fig.
2). Comparison of results obtained by ELISA and CF. The efficiency of CF and ELISA techniques was compared by using 96 sera from infants and young children who were clinically suspected to have RSV infection. A good correlation was found between ELISA titer determined by a single dilution (1:1,000) and CF titer (r = 0.89, P < 0.001). ELISA titers were about 400 times higher than CF titers. The ELISA test was more sensitive than CF: eight sera determined to be negative by CF were found to be positive by ELISA, and the presence of specific antibodies was confirmed by IF; no serum determined to be negative by ELISA was found to be
significantly positive by CF.
Specificity and sensitivity. Studies of paired sera from children with RSV primary infections showed an increase in 0.8 O.D
0.7-
0.6 0.5 0.4 0.3 0.20.1 o
r = 0.93
i~~~~~~~~ * *
/ 10
a
a
0
I
20
40
It
T
80 160 320 640 1280 ELISAtiter-( 10 2)
FIG. 2. Relationship between endpoint titers of 24 positive sera and their OD at a 1:1,000 dilution.
598
VAUR ET AL.
the amount of antibody in all cases. However, it might be argued that the ELISA test used could also detect high levels of antibodies to closely related viruses. The possibility of a cross-reaction with mumps virus was investigated by using 28 sera from children with mumps. Twenty of these sera were pairs exhibiting clear seroconversion to mumps virus. No correlation was found between mumps virus antibody titer (determined by CF) and RSV ELISA titer. Paired sera showed the same RSV antibody titer, independently of an increasing level of mumps virus antibodies (data not shown). All these results suggested that the antibodies detected by ELISA were specific to RSV. ELISA was found to be more sensitive than CF, as shown above. However, this single-dilution ELISA technique was not sensitive enough to measure very low titers of RSV antibodies. Some sera from infants were found to be negative both by the single-dilution ELISA method and by CF. When tested by IF, they were still negative at a dilution of 1:1,000 but appeared to be positive at lower dilutions (1:40 and 1:160). At these dilutions, the ELISA method did not provide clear results because of the prozonelike effect and because of high absorbance values obtained with negative reference sera, as previously mentioned. Most of the sera exhibiting a low titer of RSV antibodies were acute sera from infants with RSV infection; in all the sera from these infants during convalescence, RSV antibodies could be readily detected at a dilution of 1:1,000 by both ELISA and IF. DISCUSSION The simplified ELISA system reported in this paper appears to be a specific and sensitive method for the assay of serum immunoglobulin G antibodies to RSV. Virus-infected cells were used directly to coat plates, without any antigen purification or any freezing step. Preparation of the coated plates was rapidly and easily done. These plates can be used for several months after coating if they are stored at 4°C. From comparable data (12), plates are expected to exhibit similar stability when kept at room temperature. The determination of antibody titer by assay of a single serum dilution (1:1,000) in conjunction with a standard curve provided reproducible results. The risk of error was minimized by using two wells per serum. As previously reported (10, 13, 14), ELISA is more sensitive than CF for the detection of RSV antibodies. In addition, the test described here can be easily included into general ELISA procedures of serological analysis. The sensitivity and specificity of the test is dramatically reduced when testing sera at low dilutions. One contributing factor, the high OD492s obtained with negative reference sera, has been previously mentioned by other authors (10) and was probably caused by nonspecific absorption of immunoglobulins to crude cell antigens. Another factor, the prozonelike phenomenon, could be explained by either steric inhibition of specific immunoglobulin attachment, the presence of nonspecific inhibitors in human sera, or both. Consequently, the detection of very low titers of RSV antibodies, which requires low serum dilutions, is not possible with the simplified ELISA system, as reported by others (1, 10). However, the significance of low antibody levels in serum is controversial, particularly in the case of young infants, in which it may correspond to maternal antibodies (10, 13). A low titer of specific RSV antibodies in the serum during the acute phase of disease is therefore worthless for diagnosis unless a convalescent serum is tested simultaneoulsy. In pairs of sera from infants with acute RSV
J. CLIN. MICROBIOL.
infection, a significant increase in the level of specific antibodies was shown in all cases by the simplified ELISA method. Serological diagnosis was thus possible despite the fact that some acute sera provided false-negative results. The simplified ELISA might be improved by detection of RSV immunoglobulin M antibodies, but the presence of these antibodies has been mentioned in reinfections (2) and cannot be considered to be specific of primary infections. In fact, isolation of virus by culture and detection of virus antigen in respiratory secretions by IF (11) or by ELISA (4, 5) play a major role in the diagnosis of RSV infections. Keeping this idea in mind, we conclude that the ELISA system described here is much more convenient than CF as a complementary diagnosis procedure. In addition, this method could be helpful in testing the immune response in vaccine and epidemiological investigations. Our results strengthen previous observations regarding the possibility of directly using cells infected by viruses of the Paramyxoviridae family as antigens in ELISA diagnostic systems (12). Studies of a similar method are now under way for the detection of antibodies to mumps virus, and preliminary results are in agreement with those obtained with RSV. ACKNOWLEDGMENTS We are grateful to the Fondation pour la Recherche Medicale for the support given to our work. We are indebted to Katherine Kean and Paul Deny for critical reading of the manuscript.
LITERATURE CITED 1. Beishe, R. B., L. P. Van Vorris, M. A. Mufson, E. B. Buynak, A. A. McLean, and M. A. Hilleman. 1982. Comparison of enzyme-linked immunosorbent assay and neutralization techniques for measurement of antibody to respiratory syncytial virus: implications for parenteral immunization with live virus vaccine. Infect. Immun. 37:160-165. 2. Chomel, J. J., M. Aymard, J. P. Allard, and C. Bouvet. 1982. Le diagnostic rapide des infections a virus respiratoire syncytial (RS) par le titrage des IgM seriques (immunofluorescence indirecte). Ann. Virol. 133E:59-66. 3. Hamparian, V. V., A. Ketler, M. R. Hilleman, C. M. Reilly, L. McClelland, D. Cornfeld, and J. Stokes, Jr. 1961. Studies of acute respiratory illnesses caused by respiratory syncytial virus. I. Laboratory findings in 109 cases. Proc. Soc. Exp. Biol. Med. 106:717-722. 4. Hendry, R. M., and K. McIntosh. 1982. Enzyme-linked immunosorbent assay for detection of respiratory syncytial virus infection: development and description. J. Clin. Microbiol. 16:324-328. 5. Hornsleth, A., B. Friis, P. Andersen, and E. Brenoe. 1982. Detection of respiratory syncytial virus in nasopharyngeal secretions by ELISA: comparison with fluorescent antibody technique. J. Med. Virol. 10:273-381. 6. Julkunen, I. 1984. Serological diagnosis of parainfluenza virus infections by enzyme immunoassay with special emphasis on purity of viral antigens. J. Med. Virol. 14:177-187. 7. Kahane, S., V. Goldstein, and I. Sarov. 1979. Detection of IgG antibodies specific for measles virus by ELISA. Intervirology 12:39-46. 8. Leinikki, P. O., I. Shekarchi, N. Tzan, D. L. Madden, and J. L. Sever. 1979. Evaluation of enzyme-linked immunosorbent assay (ELISA) for mumps virus antibodies. Proc. Soc. Exp. Biol. Med. 160:363-367. 9. Meurman, O., 0. Ruuskanen, H. Sarkkinen, P. Hanninen, and P. Halonen. 1984. Immunoglobulin class-specific antibody response in respiratory syncytial infection measured by enzyme immunoassay. J. Med. Virol. 14:67-72. 10. Nandapalan, N., E. Routledge, and G. L. Toms. 1984. A enzymelinked immunosorbent assay (ELISA) for IgG and IgA antibod-
VOL. 24, 1986 ies to respiratory syncytial virus in low dilutions of secretions of human serum and secretions. J. Med. Virol. 14:285-294. 11. Pothier, P., J. C. Nicolas, G. Prudhomme De Saint-Maur, S. Ghim, A. Kazmierczak, and F. Bricout. 1985. Monoclonal antibodies against respiratory syncytial virus and their use for rapid detection of virus in nasopharyngeal secretions. J. Clin. Microbiol. 21:286-287. 12. Rice, G. P. A., P. Casali, and M. B. A. Oldstone. 1983. A new solid-phase enzyme-linked immunosorbent assay for specific antibodies to measles virus. J. Infect. Dis. 147:1055-1059. 13. Richardson, L. S., R. H. Yolken, R. B. Beishe, E. Camargo,
ELISA FOR RSV SEROLOGY
599
H. W. Kim, and R. M. Chanock. 1978. Enzyme-linked immunosorbent assay for measurement of serological response to respiratory syncytial virus infection. Infect. Immun. 20:660664. 14. Steinhoff, M. C., C. B. Hall, and K. C. Schnabel. 1980. Respiratory syncytial virus serology by a simplified enzyme-linked immunosorbent assay. J. Clin. Microbiol. 12:447-450. 15. Van Loon, A., J. T. M. Van der Logt, and J. Van der Veen. 1981. Enzyme-linked immunosorbent assay for measurement of antibody against cytomegalovirus and rubella virus in a single serum dilution. J. Clin. Pathol. 34:665-669.
