Abstract
A fluorescent focus identification assay (FFIDA) was developed for use in experimental studies and for quantitation of the components in a tetravalent live oral rotavirus vaccine. The assay utilizes four serotype-specific neutralizing monoclonal antibodies (MAb) to detect and quantify individual rotaviruses by immunofluorescence staining of fixed virus-infected monkey kidney cells. In mixed virus infections, all four MAb, W1 (serotype 1), 1C10 (serotype 2), R1 (serotype 3), and S4 (serotype 4), specifically stain the relevant homologous serotype without exhibiting any cross-reactivity against the other serotypes. Furthermore, the test is sensitive enough to differentiate at least twofold (0.3 log) differences in virus titer. The results of testing four individual experimental vaccine lots three or more consecutive times showed that all four lots contained similar proportions of the four vaccine strains as detected by the classical plaque neutralization identification test. The rapidity and efficiency of the FFIDA are desirable attributes that make it suitable for use in studies requiring identification and quantitation of one or more of the four major rotavirus serotypes.
Rotaviruses (RV) are the major cause of diarrhea in human infants (6). Gastroenteritis associated with infection by RV causes extensive morbidity in developed countries and significant mortality in less-developed countries (5). Worldwide, it is estimated that close to one million infant deaths occur as a result of rotaviral diarrhea and its sequelae (8). The development of an effective RV vaccine to reduce the morbidity and mortality of diarrheal disease in young children is a high priority of the World Health Organization. In collaboration with the National Institutes of Health, Wyeth Lederle Vaccines has developed a live tetravalent rotavirus vaccine (RV-TV) that is based on a modified Jennerian approach. The vaccine consists of four viruses—a rhesus rotavirus (RRV) (strain MMU18006) of VP7:3 and three rhesus-human reassortant viruses that are entirely rhesus except that 1 of 11 rhesus genes has been replaced by a human gene coding for VP7:1, -2, or -4, respectively (7). A number of clinical trials have demonstrated that the vaccine is highly effective in reducing the incidence of severe diarrhea as well as the number of infants requiring hospitalization in both developed and less-developed countries (2, 9, 10, 12). When vaccine lots are manufactured and released, the final product must contain the four component viruses at their proper titers. Initially, a plaque neutralization identification test (PN-ID) that employed four serotype-specific monoclonal antibodies (MAb) was used to establish the presence of each of the four component viruses in experimental lots. Since each MAb eliminated more than 98% of the homologous virus, a pool of any three of the four MAb selectively neutralized three viruses in the tetravalent vaccine, permitting only the fourth one to replicate. This approach is similar to methodology used for identifying the three component viruses of live oral poliovirus vaccine (1, 13). The plaque assay, however, is laborious and time-consuming, requiring many 60-mm-diameter petri dishes and, generally, 5 days for completion. A more-rapid virus identification test was needed to facilitate product release. To this end, a more-efficient assay for identifying each component virus in the tetravalent formulation was developed. The assay is based on the determination of serotype-specific fluorescent foci with anti-RV VP7 serotype-specific MAb to detect each virus in vaccine-infected monkey kidney cells. This study describes the assay and compares the results for four experimental vaccine lots with results generated by PN-ID.
MATERIALS AND METHODS
Viruses.
Vaccine lots designated A, B, C, and D; rotavirus monovalent concentrates, lots 1 (D×RRV), 2 (DS1×RRV), 3 (RRV), and 4 (ST3×RRV); and the human RV, Wa, DS1, and ST3, were used in this study. All human RV were originally received from A. Kapikian (National Institutes of Health, Bethesda, Md.) and were amplified in MA104 cells. Vaccine and the monovalent concentrates were produced at the Wyeth Lederle Vaccine Development Center in Marietta, Pa. The four tetravalent vaccine lots were formulated to contain an intended titer of 105 PFU/dose for each of the four vaccine strains, D×RRV (serotype 1), DS1×RRV (serotype 2), RRV (serotype 3), and ST3×RRV (serotype 4).
MAb and polyvalent rabbit RV antiserum.
Mouse ascites containing the G type-specific neutralizing monoclonal antibodies (MAb) designated W1 (anti-Wa VP7, serotype 1), 1C10 (anti-DS1 VP7, serotype 2), R1 (anti-RRV VP7, serotype 3), and S4 (anti-ST3RRV VP7, serotype 4) were used in this study. MAb 1C10 and MAb 60, another MAb directed against a linear epitope common to group A RV were received from H. Greenberg (Stanford University School of Medicine, Stanford, Calif.). The other three neutralizing MAb, W1, R1, and S4, were generated in our laboratory by using standard mouse hybridoma technology. BALB/c mice were immunized with CsCl gradient-purified triple-shelled rotavirus Wa, DS1×RRV, or ST3×RRV, and spleen cells from the immunized mice were subsequently fused with mouse myeloma cells (NS1). Polyclonal rabbit anti-RV serum was generated by repeatedly immunizing RV-naive rabbits with CsCl gradient-purified triple-shelled RV Wa strain (serotype 1). This rabbit antiserum cross-reacted with all four vaccine strains, D×RRV, DS1×RRV, RRV, and ST3×RRV.
A fluorescent focus assay for RV.
A fluorescent focus assay developed previously for the determination of FFU titers (fluorescent focus units) and serum antirotavirus neutralization titers was modified to enable serotyping of the four vaccine RV. Confluent rhesus monkey cells MA104 or MAE cells (a clone of MA104 cells obtained from Richard L. Ward, Division of Infectious Diseases, Children’s Hospital Medical Center, Cincinnati, Ohio) in 96-well microtiter plates (Costar catalog no. 3593) were infected with RV monovalent concentrates at a dilution that yielded about 500 FFU per 100 μl per well. After centrifugation at 1,000 × g for 1 h at room temperature to enhance virus absorption, the cells were washed once with serum-free Dulbecco’s modified Eagle medium and incubated in an atmosphere of 5% CO2 for 18 h. After incubation, the cells were fixed in cold 80% acetone for 15 to 20 min at −20°C and then air dried. Four sets of duplicate wells were reacted with one each of the four serotype-specific MAb at predetermined dilutions, followed by staining with biotin-avidin. Biotinylated goat anti-mouse immunoglobulin G (IgG) (for serotypes 1, 3, and 4 RV) or biotinylated goat anti-mouse IgA (for rotavirus stained by 1C10 of IgA isotype) was used to react the cells treated with MAb, and the cells were then incubated with fluorescein isothiocyanate-conjugated streptavidin. The titers of individual serotypes (FFU/ml) were determined by counting the number of fluorescent foci in the well with an inverted fluorescence microscope with suitable filters for fluorescein isothiocyanate at 100× magnification. For the total titer of the tetravalent vaccine, the cells in another duplicate set of wells were stained with either rabbit anti-Wa polyclonal serum or the MAb designated 60. Both reagents react similarly with all four RV vaccine strains. Four vaccine lots were tested at least three times each by this procedure.
Sensitivity and viral interference in the fluorescent focus virus identification assay (FFIDA).
A 2(4-1) factorial experimental design was adopted to test the sensitivity of the assay and to determine whether or not there was evidence of viral interference. The infectivities of the four RV vaccine strains, D×RRV, DS1×RRV, RRV, and ST3×RRV, were tested in eight different tetravalent formulations or pools. Each monovalent RV strain was formulated in the pools at a low or a high titer. The high levels of RV were targeted at 105 FFU/ml, calculated from preliminary FFU results. The low virus titer levels were obtained by dilution to one-third of the high levels. Each of the eight pools contained all four serotypes consisting of four high-titer stocks, four low-titer stocks, or a combination of two high- and two low-titer stocks. The eight formulations of virus were then tested to determine the titer of each serotype in each pool. The total virus titer was expected to equal the sum of the four individual titers, provided there was no interference with replication among the viruses. The data were analyzed by a Poisson regression model and a log link function with an overdispersion parameter by the SAS GENMOD procedure (for a generalized linear model).
RESULTS
Immunofluorescence staining specificity of anti-VP7 MAb against RV of different serotypes.
The specificity of the MAb used in fluorescence staining was first tested by checkerboard staining involving five anti-VP7 MAb and one polyclonal rabbit antiserum. The RV strains tested were RRV, Wa, DS1, and ST3 and the three rhesus × human RV vaccine strains, D×RRV, DS1×RRV, and ST3×RRV. The results showed that each serotype-specific MAb stained only its human homologous serotype or the reassortant that carried the VP7 gene of the same serotype (Table 1). No cross-reactivity was observed with any of the MAb. For example, W1, the MAb specific for serotype 1 VP7, stained only Wa or D×RRV, the two RV strains carrying the VP7 gene of human serotype 1 RV. The anti-serotype 4 MAb, S4, stained only serotype 4 RV, ST3 and ST3×RRV. MAb 60, in contrast, recognizes a common linear epitope of VP7 and stained all seven group A RV. The rabbit anti-RV polyclonal serum also stained all of the RV strains.
TABLE 1.
Specificity of anti-RV MAb used in FFIDA
| MAb | Viral antigen (serotype) | Fluorescence staining results against indicated RV strain
|
||||||
|---|---|---|---|---|---|---|---|---|
| Wa | D×RRV | DS1 | DS1×RRV | RRV | ST3 | ST3×RRV | ||
| W1 | VP7 (1) | + | + | − | − | − | − | − |
| 1C10 | VP7 (2) | − | − | + | + | − | − | − |
| R1 | VP7 (3) | − | − | − | − | + | − | − |
| S4 | VP7 (4) | − | − | − | − | − | + | + |
| 60 | VP7 (linear epitope of group A RV) | + | + | + | + | + | + | + |
| Rabbit antiserum | VP6 (group A-specific antigen) | + | + | + | + | + | + | + |
Test of RV replication interference in eight virus pools.
Titers of monovalent stocks utilized in the factorial experiment performed to determine the presence or absence of replication interference among the viruses are provided in Table 2. These titers were consistent with titers determined for individual serotypes in the eight virus pools (Table 3). The results showed that there was no interference or interaction between any of the serotypes in the tetravalent formulation, irrespective of the titers of the individual viruses. The sums and totals of each of the eight virus pools were also in excellent agreement (Table 3). Statistical analysis of the high to low virus ratio in the eight pools showed that there was no significant difference from the theoretical ratio of 3.0 (P > 0.5) for all four serotypes (Table 4). These results indicate that the test has virtually a 100% probability of differentiating a titer difference of threefold (or 0.5 log) and a 97% probability of differentiating a titer difference of twofold (or 0.3 log) between any two samples of the same serotype.
TABLE 2.
Titers of individual monovalent virus stocks utilized in the factorial experimenta
| Virus stock | FFU/ml (104) for indicated level
|
|
|---|---|---|
| Highb | Lowc | |
| DRRV | 11.6 | 3.0 |
| DSIRRV | 14.1 | 4.9 |
| RRV | 11.0 | 3.5 |
| ST3RRV | 7.6 | 1.7 |
All stocks were diluted 1:4.
The high levels of the monovalent viruses were targeted at 10 × 104 FFU/ml.
The low levels of virus stocks were targeted at one-third of the high levels.
TABLE 3.
Individual RV titers in eight virus pools determined by FFIDA
| Virus pool | FFU/ml (104) for indicated stocka
|
|||||
|---|---|---|---|---|---|---|
| DRRV | DS1RRV | RRV | ST3RRV | Sumb | Totalc | |
| 1 | 2.58 (L) | 3.08 (L) | 3.42 (L) | 2.46 (L) | 11.54 | 9.86 |
| 2 | 8.85 (H) | 5.21 (L) | 3.64 (L) | 6.83 (H) | 24.53 | 23.52 |
| 3 | 3.25 (L) | 9.52 (H) | 3.08 (L) | 6.33 (H) | 22.18 | 21.10 |
| 4 | 9.02 (H) | 9.30 (H) | 3.75 (L) | 1.62 (L) | 23.69 | 24.86 |
| 5 | 2.91 (L) | 3.70 (L) | 7.06 (H) | 4.93 (H) | 18.60 | 23.74 |
| 6 | 9.24 (H) | 2.80 (L) | 8.34 (H) | 1.79 (L) | 22.17 | 24.19 |
| 7 | 4.54 (L) | 11.42 (H) | 10.36 (H) | 2.58 (L) | 28.90 | 21.50 |
| 8 | 7.17 (H) | 11.7 (H) | 9.58 (H) | 7.11 (H) | 35.56 | 32.48 |
L, low-titer stock; H, high-titer stock.
Sum of four individual titers.
Total titer of each virus pool as determined by reaction with a common MAb, 60.
TABLE 4.
Statistical analysis of virus ratios in eight virus pools
| RV | Avg FFU/ml (104) ± SE for indicated virus titer
|
Virus ratio between high and low titers | P value testing for virus ratioa | |
|---|---|---|---|---|
| High | Low | |||
| DRRV | 8.6 ± 0.55 | 3.3 ± 0.36 | 2.6 | 0.24 |
| DS1RRV | 10.5 ± 0.63 | 3.7 ± 0.36 | 2.8 | 0.62 |
| RRV | 8.8 ± 0.55 | 3.5 ± 0.35 | 2.5 | 0.17 |
| ST3RRV | 6.3 ± 0.47 | 2.1 ± 0.27 | 3.0 | 0.96 |
No significant difference from the tested ratio of 3.0, all P values are >0.05.
Distribution of four RV serotypes in tetravalent vaccine and comparison to results obtained by PN-ID.
Geometric mean FFU titers and the percentage distribution of the four serotypes in four experimental vaccine lots were determined, and the results were compared to results obtained previously by PN-ID. The percentages of individual serotypes as part of the sum determined by FFIDA versus PN-ID were similar but differed significantly for serotypes 3 and 4 (Table 5). Because available data were limited to one PN-ID test for each vaccine lot, no reliable correlation coefficient for the absolute titers between the two tests could be derived.
TABLE 5.
Geometric mean FFU versus PFU titers as a percentage of the sum of four RV vaccine strains in RV-TV
| Vaccine lot | Serotype | FFUa/ml (104) | % of sum ± SE | PFUb/ml (104) | % of sum ± SE |
|---|---|---|---|---|---|
| A | 1 | 8.3 | 19.4 ± 2.8 | 3.5 | 17.8 |
| 2 | 11.7 | 27.3 ± 3.5 | 5.8 | 29.4 | |
| 3 | 15.4 | 35.9 ± 4.1 | 5.2 | 26.4 | |
| 4 | 7.5 | 17.4 ± 2.6 | 5.2 | 26.4 | |
| Sumc | 42.9 | 100 | 19.7 | 100 | |
| Totald | 42.8 | 99.9 | 15.2 | 77.2 | |
| B | 1 | 6.4 | 17.5 ± 2.4 | 6.7 | 22.0 |
| 2 | 11.7 | 32.1 ± 3.6 | 10.5 | 34.5 | |
| 3 | 11.6 | 31.9 ± 3.6 | 6.4 | 21.1 | |
| 4 | 6.8 | 18.5 ± 2.5 | 6.8 | 22.4 | |
| Sum | 36.5 | 100 | 30.4 | 100 | |
| Total | 39.4 | 107.8 | 20.1 | 66.1 | |
| C | 1 | 10.3 | 18.0 ± 2.6 | 5.9 | 19.2 |
| 2 | 19.1 | 33.2 ± 3.9 | 8.4 | 27.4 | |
| 3 | 18.6 | 32.4 ± 3.9 | 7.1 | 23.1 | |
| 4 | 9.4 | 16.4 ± 2.4 | 9.3 | 30.3 | |
| Sum | 57.4 | 100 | 30.7 | 100 | |
| Total | 55.0 | 95.7 | 27.2 | 88.6 | |
| D | 1 | 12.4 | 23.8 ± 3.2 | 4.7 | 19.5 |
| 2 | 14.8 | 28.4 ± 3.6 | 8.7 | 36.1 | |
| 3 | 16.1 | 30.9 ± 3.8 | 5.0 | 20.7 | |
| 4 | 8.8 | 16.9 ± 2.5 | 5.7 | 23.7 | |
| Sum | 52.1 | 100 | 24.1 | 100 | |
| Total | 49.6 | 95.2 | 20.7 | 85.9 | |
| Randome | 1 | 8.9 | 19.4 ± 2.5 | 5.9 | 22.2 ± 0.9 |
| Random | 2 | 14.1 | 30.5 ± 3.3 | 8.2 | 30.8 ± 2.1 |
| Random | 3 | 15.1 | 32.7 ± 3.5 | 5.9 | 22.2 ± 1.3f |
| Random | 4 | 8.1 | 17.5 ± 2.3 | 6.6 | 24.8 ± 1.8f |
| Sum | 46.2 | 100 | 26.6 | 100 | |
| Total | 47.2 | 102 | 20.4 | 79.4 |
Geometric mean FFU of more than three tests on each vaccine lot.
Based on previous PFU identification data of these RV-TV lots.
Sum of four individual titers.
Total titer of each vaccine lot determined by FFIDA or PN-ID.
Analysis of the four vaccine lots combined.
P < 0.05 by the z test comparing means of FFUs and PFUs.
DISCUSSION
Serotype-specific MAb have been used previously in enzyme-linked immunosorbent assays for serotyping of human RV (4, 11). In enzyme-linked immunosorbent assays, however, multiple serotypes cannot easily be distinguished in the same test. Accordingly, it was necessary to devise a method capable of quantifying individual serotypes in a sample containing a mixture of RV strains. To this end, an immunofluorescence staining method for the detection of RV in infected cultures was adopted (3). The results demonstrate that four serotype-specific MAb are highly specific in immunofluorescence staining of the homologous RV that carries the specific VP7 antigen and occurs without any cross-reactivity (Table 1). By using these serotype-specific MAb, it was possible to develop an FFIDA that could detect and quantify each of the four individual serotypes in RV-TV. Test results with eight different vaccine formulations showed that there was no interference among the four RV strains and that the FFIDA was capable of differentiating a twofold (or 0.3 log) difference in titer (Tables 2 to 4). Comparison of FFIDA data for four experimental tetravalent vaccine lots with historical PN-ID data for the same lots indicated similarity in the distribution of the viruses (Table 5). Additional studies, however, are needed to rigorously establish equivalence of the FFIDA with other RV assays.
FFIDA has several advantages over the PN-ID. First, it is simpler and more convenient. Since the FFIDA is performed in 96-well microtiter plates instead of 60-mm-diameter petri dishes, a large number of samples can be tested simultaneously. Second, it is relatively rapid. The results are obtained 2 days after cell infection compared to 5 days with a typical plaque neutralization assay. Third, the neutralizing capacity of MAb is not always 100% effective with a PN-ID test. Consequently, low levels of virus breakthrough can occur. Low-level cross-inhibition of other serotypes can also occur. These factors contribute to assay variability. The FFIDA, on the other hand, is more specific, since immunofluorescence staining provides an all-or-nothing signal, with no contribution due to cross-reactivity. Furthermore, the FFIDA can accommodate a large number of replicates, reducing intratest variability to a minimum. Because of the aforementioned advantages, FFIDA is highly suitable for identification and titration of individual RV vaccine serotypes, singly or in combination. With slight modifications, the test can also be used for a number of other applications, such as detection and serotyping of infectious virus shed by vaccinated infants and experimental animals, VP7 gene expression in transfected cells, and measurement of virus-specific neutralizing antibody.
REFERENCES
- 1.Benyesh-Melnick M, Melnick J L, Ramos A M. First International Conference on Live Poliovirus Vaccines, Washington, D.C., 22 to 26 June 1959. Scientific publication no. 44. Pan American Sanitary Bureau; 1959. Poliomyelitis infection rate among Mexican children fed attenuated poliovirus vaccines; pp. 272–285. [Google Scholar]
- 2.Bernstein D I, Glass R I, Rodgers G, Davidson B L, Sack D A. Evaluation of rhesus rotavirus monovalent and tetravalent reassortant vaccines in U.S. children. JAMA. 1995;273:1191–1196. [PubMed] [Google Scholar]
- 3.Butchaiah G. Infectivity assay of bovine rotavirus: evaluation of plaque and end-point method in comparison with immunofluorescent cell assay. Acta Virol. 1988;32:60–64. [PubMed] [Google Scholar]
- 4.Coulson B S, Unicomb L E, Pitson G A, Bishop R F. Simple and specific enzyme immunoassay using monoclonal antibodies for serotyping human rotaviruses. J Clin Microbiol. 1987;25:509–515. doi: 10.1128/jcm.25.3.509-515.1987. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.de Zoysa J R, Feachem G. Intervention for the control of diarrheal disease among young children: rotavirus and cholera immunization. Bull W H O. 1985;63:569–583. [PMC free article] [PubMed] [Google Scholar]
- 6.Kapikian A Z, Chanock R M. Rotavirus. In: Field B N, editor. Virology. New York, N.Y: Raven Press; 1985. pp. 863–906. [Google Scholar]
- 7.Kapikian, A. Z., Y. Hoshino, R. M. Chanock, and I. Peres-Schael. 1996. Efficacy of a quadrivalent rhesus rotavirus based human rotavirus vaccine aimed at preventing severe rotavirus diarrhea in infants and young children. J. Infect. Dis. 174(Suppl. 1):S65–S72. [DOI] [PubMed]
- 8.LeBoron C W, Lew J, Glass R I, Weber J M, et al. Annual rotavirus epidemic patterns in North America. JAMA. 1990;264:983–988. doi: 10.1001/jama.264.8.983. [DOI] [PubMed] [Google Scholar]
- 9.Perez-Schael I, Guntinas M J, Perez M, Pagone V, Rojas A M, Rosabel G, Cunto W, Hoshino Y, Kapikian A Z. Efficacy of the rhesus rotavirus-based quadrivalent vaccine in infants and young children in Venezuela. N Engl J Med. 1997;337:1181–1187. doi: 10.1056/NEJM199710233371701. [DOI] [PubMed] [Google Scholar]
- 10.Rennels M B, Glass R I, Dennehy P H, Berstein D I, Pichichero M E, Zito E T, Mark M E, Davidson B L, Kapikian A Z. Safety and efficacy of high-dose rhesus-human reassortant rotavirus vaccines—report of the National Multicenter Trial. Pediatrics. 1996;97:7–13. [PubMed] [Google Scholar]
- 11.Taniguchi K, Urasawa T, Morita Y. Direct serotyping of human rotavirus in stools by an enzyme-linked immunosorbent assay using serotype 1, 2, 3 and 4-specific monoclonal antibodies to VP7. J Infect Dis. 1987;155:1159–1166. doi: 10.1093/infdis/155.6.1159. [DOI] [PubMed] [Google Scholar]
- 12.Vesikari T. Presented at the 2nd Satellite Symposium of the 14th Annual Meeting of the European Society for Pediatric Infectious Diseases (ESPID), Elsinore, Denmark, 18 to 21 June 1996. 1996. Clinical experience with rotavirus vaccine in Finland. [Google Scholar]
- 13.World Health Organization. Expanded program on immunization. Manual for virological investigation of poliomyelitis. WHO/EPI/CDS/Polio/90.1. Geneva, Switzerland.