Journal of Medical Virology 38:147-151 (1992)
Rapid Detection of Respiratory Viruses Using Mixtures of Monoclonal Antibodies on Shell Vial Cultures Jurjen Schirm, Dirk S. Luijt, Geke W. Pastoor, Johan M. Mandema, and F. Peter Schroder Regional Public Health Laboratory M.S., D.S.L., G.W.P., F.P.S.), and Department of Pediatrics (J.M.M.),University Hospital, Groningen, The Netherlands Eleven hundred and thirty-three clinical specimens submitted to the laboratory for diagnosis of respiratory virus infections were tested by direct immunofluorescence (DIF) for respiratory syncytial virus (RSV),by shell vial culture, and by conventional cell culture. The shell vial cultures were stained with 8 different monoclonal antibodies both 1 day and 3-7 days after inoculation. In order to limit the cost and the workload, mixtures of monoclonal antibodies were used. Coverslips with HEp-, cells were incubated with a mixture of FITC-labeled monoclonal antibody to RSV and nonlabeled monoclonal antibody to adenovirus. When no RSV positive IF staining was observed after the first incubation step, the same coverslip was incubated once more with FITClabeled anti-mouse antibody. A positive reaction at this stage indicated the presence of adenovirus. Similarly, cultures of tertiary monkey kidney cells were investigated with a mixture of two FITC-labeled monoclonals to the influenza viruses A and B and three nonlabeled monoclonals to the parainfluenza viruses 1, 2 and 3. If influenza virus or parainfluenza virus was detected, the exact type was determined by staining different parts of a duplicate coverslip. Shell vial cultures for cytomegalovirus (CMV) were always performed separately on human embryonic lung fibroblasts. Using this approach, we detected RSV (n = 248), CMV (n = 42), parainfluenza virus (n = 311, influenza virus (n = 281, and adenovirus (n = 6), in most cases after only one day of culture. For RSV, the sensitivity of the shell vial method was too low (74%) to allow omission of DIF (sensitivity 95%). For the other viruses, the "shell vial/monoclonal antibody mixture" approach was very attractive, being rapid, very specific (>%'Yo), and also very sensitive (probably
>%YO).
0 1992 Wiley-Liss, Inc.
0 1992 WILEY-LISS, INC.
KEY WORDS: rapid diagnosis, shell vials, immunofluorescence, antibody mixtures, (para-)influenza viruses, respiratory syncytial virus
INTRODUCTION Clinical specimens submitted to diagnostic laboratories for diagnosis of respiratory virus infections are usually investigated by conventional cell culture methods. The detection of respiratory syncytial virus (RSV) is often performed more rapidly using nonculture methods such as direct immunofluorescence (DIF1or enzyme immunoassays or by rapid shell vial culture techniques or both [Johnston and Siegel, 1990; Morris et al., 1990; Takimoto et al., 1991; Smith et al., 19911. It has been reported that in more than 20% of specimens from patients with a presumptive diagnosis of RSV infection, other viruses are found, notably (para-)influenza viruses, cytomegalovirus (CMV), rhinovirus and, to a lesser extent, adenovirus, herpes simplex virus (HSV) and enteroviruses [Blanding et al., 1989; Waner et al., 19901. Most of these viruses, except for rhinovirus, can quite easily be detected in conventional cell culture systems, although culture of CMV, adenovirus and (para-)influenza viruses may take 1-4 weeks. In most studies, the sensitivity of rapid nonculture methods for respiratory viruses other than RSV is reported to be low: generally lower than 80% for the (para-)influenza viruses and even lower than 50% for the adenoviruses [Ray et al., 1987; Stokes et al., 1988; Mills et al., 1989; Stout et al., 1989; Takimoto et al., 19911. Accepted for publication March 31,1992. Address reprint requests to Dr. J. Schirm, Regional Public Health Laboratory, Van Ketwich Verschuurlaan 92, 9721 SW Groningen, The Netherlands.
Schirm et al.
148 The shell vial culture technique is used now universally for the rapid detection of, among others, CMV [Gleaves et al., 19851, and also, less often, for the detection of influenza viruses [Stokes e t al., 1988; Mills et al., 19891 and adenovirus [August et al., 19871. We describe the use of shell vial cultures for the detection of the influenza viruses A and B, the parainfluenza viruses 1, 2 and 3, adenovirus, and CMV in respiratory specimens. In order to limit the workload, mixtures of monoclonal antibodies were used.
MATERIALS AND METHODS Clinical Specimens Clinical specimens were obtained from all patients whose primary complaint was a respiratory illness. The majority of these patients were infants with bronchiolitis or pneumonia and/or a presumptive diagnosis of RSV infection. Some hospitalized children without respiratory complaints were included for epidemiological reasons. Most specimens were nasopharyngeal aspirates, pharyngeal washings or sputum samples. In addition, a limited number of throat swabs or other respiratory specimens was investigated. The specimens were submitted to the laboratory in virus transport medium, except for those few specimens (no swabs) t h a t arrived in the laboratory within a n hour. All specimens were diluted tenfold in virus transport medium or in phosphate buffered saline (PBS),mixed vigorously, and centrifuged at 800 x g for 5 min. The sediment was smeared onto a microscope slide, air dried, and fixed with acetone a t -20°C for 10 min. The remaining cells were resuspended in the washing fluid and used for tissue culture. Monoclonal Antibodies The fluorescein isothiocyanate (FITC)-labeled monoclonal antibodies of Imagen (Novo Nordisk Diagnostics Ltd., Cambridge, U.K.) were used for the detection of RSV, influenza A virus, and influenza B virus. Monoclonal antibodies to the parainfluenza viruses 1 , 2 and 3 were obtained from Chemicon International Inc. (Temecula, CA). Adenovirus was detected by the groupspecific monoclonal antibody 1F-8 (Whittaker, Walkersville, MD) or 1-3 (obtained from Dr. J.C. de Jong, National Institute of Public Health and Environment [RIVM], Bilthoven, The Netherlands). Preliminary comparison of these two adenovirus specific monoclonals showed no major differences in sensitivity, specificity or immunofluorescence (IF) patterns. Two different monoclonal antibody preparations were also used for the detection of CMV. Initially, a mixture of the monoclonal antibodies C10 and C11 (obtained from Prof. T.H. The, University of Groningen, The Netherlands), both directed to the lower matrix protein pp65 [Grefte et al., 19911, was used. Later, we changed to the monoclonal antibody 9221 (directedto the CMV major immediate early antigen; obtained from DuPont de Nemours, 's-Hertogenbosch, Netherlands), which appeared to be
slightly more sensitive than the ClO/Cll mixture when used in the shell vial culture system (unpublished observation). All monoclonal antibodies were tested for cross-reactivities with the above respiratory viruses, except for rhinovirus and the enteroviruses. In addition, the monoclonal antibodies to CMV were also tested for crossreactions with HSV and varicella zoster virus. Only the parainfluenza monoclonals showed some intertypic cross reactivity, but in all cases, the specific monoclonal antibody gave a much stronger IF intensity than the cross-reacting monoclonal antibody.
Cell Culture Techniques Human embryonic lung fibroblasts (HEL, a locally produced cell line), HEp-2 cells, and tertiary monkey kidney (t-MK) cells (obtained from RIVM, Bilthoven, The Netherlands) were grown in conventional culture tubes and on coverslips in shell vials. Minimum essential medium with Hanks salts, supplemented with nonessential amino acids and foetal calf serum was used. The monolayers of t-MK cells were always treated with 1mg/ml of trypsin before use. Clinical specimens were inoculated into duplicate culture tubes with t-MK and HEL cells (0.3 ml each) and into duplicate shell vials with t-MK, HEp-2 and HEL cells (0.2 ml each). Inoculation of shell vials was enhanced by centrifugation at 700 x g for 45 min. All culture tubes and shell vials were incubated at 37°C. Twice a week, for up to two weeks after inoculation, the monolayers in the culture tubes were microscopically examined for cytopathic effects (CPE). Monolayers showing CPE were stained by hematoxylin-eosin or with monoclonal antibodies. In cultures found positive for adenovirus or enterovirus, the exact virus type was determined by neutralization. In addition, t-MK culture tubes were tested for hemadsorbing activity after 3-7 days and after 14 days using guinea pig erythrocytes. The identity of the virus in hemadsorption positive cultures was subsequently determined by hemadsorption inhibition with virus specific antisera. One day after inoculation, the first of each pair of duplicate CS was fixed by replacing the medium with 1 ml of acetone (without drying) and leaving at -20°C for 10 min. If the subsequent immunofluorescent (IF) staining was negative, the second CS was fixed and stained 2-6 days later. IF Staining All CS were stained in two steps of 30 min at 37°C each. CS with HEp-2 cells were first incubated with a mixture of FITC-labeled monoclonal antibody to RSV, dilution 1/4, and one of the nonlabeled monoclonal antibodies to adenovirus: monoclonal 1F-8 at a dilution of 1/2 or monoclonal 1-3 a t a dilution of MOO. When no RSV-specific I F staining (speckled) was observed after the first step, the CS was rinsed in PBS and subsequently incubated with FITC-labeled rabbit antimouse immunoglobulin (Dakopatts, Glostrup, Denmark), di-
Monoclonal Antibody Mixtures on Shell Vials
149
TABLE I. ComDarison of Different Shell Vial Cultures* Virus(es) RS-virus Influenza AIB Parainfluenza-1,2,3 Adenovirus Cvtomegalovirus
t-MK
Cell tvDe HEP-2
+ ++ ++ +
++ + + ++
+
*
HEL
+ 2 + ++ ++
*IF staining was performed after one day of shell vial culture with different cell types. The degrees of IF intensity were expressed as: + + very strong; + moderate; weak.
*
lution 1/30. When adenovirus was detected after this staining step, usually showing nuclear IF, the virus was further typed by neutralization in the conventional culture method, if possible. CS with t-MK cells were first incubated with a mixture of two FITC-labeled monoclonal antibodies to the influenza viruses A and B, both at a dilution of 114, and the three nonlabeled monoclonal antibodies to the parainfluenza viruses 1, 2 and 3, each at a dilution of MOO. If no influenza virus-specific I F staining (nuclear or cytoplasmic) was observed after the first step, the CS was incubated once more with the FITC-labeled rabbit antimouse immunoglobulin in order to reveal the possible presence of one of the parainfluenza viruses (cytoplasmic IF). If influenza virus (A or B) or parainfluenza virus ( 1 , 2 , or 3) was detected on the first t-MK CS, the exact virus type was determined by staining different parts of the duplicate CS with the separate virus specific monoclonal antibodies. If only the second CS was IF-positive, typing was performed, if possible, on virus recovered from the conventional tube cell culture. CS with HEL cells were at first stained only with one of the CMV-specific monoclonal antibody preparations, both a t dilution 1/10, and subsequently with the FITClabeled rabbit antimouse immunoglobulin. Smears for detection of RSV by DIF were stained with the undiluted FITC-labeled monoclonal antibody to RSV for 15 min at 37°C.
RESULTS In a preliminary study, we tested the growth of the viruses and the performance of the monoclonal antibodies on shell vial cultures with different cell types (Table I). Because of the similar cell type preferences, we chose to combine the detection of the influenza viruses A and B and the parainfluenza viruses 1, 2 and 3 using mixtures of the appropriate monoclonal antibodies on t-MK cells. The IF patterns of adenovirus and CMV on HEL cells were usually quite similar: homogenous nuclear. For that reason, and since at that time, we did not have a n FITC-labeled monoclonal antibody to either CMV or adenovirus, we chose to combine the detection of adenovirus with the detection of RSV using shell vials with HEp-2 cells. Eleven hundred and thirty-three clinical specimens were investigated between December 1989 and September 1991. In all cases, DIF on RSV was combined
TABLE 11. Comparison of Shell Vial Method to Other Tests in 1,133 Clinical Samples All viruses
Test resulta Shell vial Shell vial Shell vial Shell vial
+ /other tests + + /other tests - /other tests + -
/other tests
-
238 54 63 778
RSV 177 10 61 885
Otherb viruses 61
44 2 1.026
”“Other test”: DIF in case of RSV and conventional cell culture in case of the other viruses. bThe nonRSV viruses tested on shell vials were the influenza viruses A and B, the parainfluenza viruses 1,2 and 3, CMV and adenovirus. For a minority of the specimens, no IF staining was performed for CMV (120 specimens) or adenovirus (177 specimens). None of these specimens were found to be positive by the conventional cell culture method.
with shell vial cultures of t-MK cells and HEp-2 cells and conventional tube cell cultures with t-MK cells and HEL cells. The CS with t-MK cells were always stained and further investigated as described above. The CS with HEp-2 cells were always stained with the monoclonal antibody to RSV, in most cases ( 9 5 6)~as a mixture with the monoclonal antibody to adenovirus. Shell vial cultures with HEL cells for the detection of CMV were performed for 1,013of the specimens. Using this approach, 354 positive samples were identified, containing RSV (n = 248), influenza virus A or B (n = 28), one of the parainfluenza viruses (n = 31), adenovirus ( n = 61, CMV ( n = 42), and other viruses (n = 9) which were only detectable by conventional cell culture in our system. These other viruses consisted of rhinovirus ( n = 3), coxsackie virus type A9 ( n = 21, echovirus type 7 ( n = 11,echovirus type 18 (n = l),and HSV type 1 (n = 2). The exact type of two of the adenoviruses could be determined, and was type 2 in both cases. Ten of the positive samples contained two different viruses, notably CMV + RSV (n = 81, adenovirus type 2 + RSV (n = l),and CMV + parainfluenza type 3 ( n = 1). In general, the shell vial approach appeared to be quite efficient: of the complete group of “shell vial viruses”, 292 (82%) of the total of 355 viruses were found on shell vials, whereas 301 (85%)were found by other methods, i.e., direct immunofluorescence (DIF) on smears in case of RSV and conventional cultures and/or hemadsorption for the other viruses (Table 11). However, the shell vial results were quite different for RSV, compared to the other viruses. Of the 248 specimens found positive for RSV, only 187 (75%)were detected by the shell vial method, and 238 (96%) by DIF. Only 10 RSV positive samples were missed by DIF. In all these 10 cases, the DIF could not be interpreted since the smears contained hardly any cells. In contrast, of the 107 specimens found positive for one of the other viruses, 105 (98%)were detected by the shell vial method and only 63 (59%)by the conventional culture method. In this group, the viruses missed by the shell vial method were one CMV, and one parainfluenza virus type 3 (see also Table 111).The 44 viruses missed by the conventional culture methods (Table 111) were either
Schirm et al.
150 TABLE 111. Results for Different Viruses Tested on Shell Vials
Virusa
RS-virus Influenza A Influenza B Influenza A/B Parainfluenza 1 Parainfluenza 2 Parainfluenza 3 P. influenza 11213 Adenovirus CMV
Total nonRSV
Total number of positives
Total
248 15 11 2 11 3 16 1 6 42 107
187 15 11 2 11 3 15 1 6 41 105
Viruses found by IFb on shell vials Confirmed‘ Not confirmed 177 13 11 0 8 3 15 0 4 7 61
10 2 0 2 3 0 0 1 2 34 54
“Influenza AIB = influenza virus A or B; Parainfluenza 11213 = parainfluenza virus 1 . 2 or 3. In these cases, the exact virus type could not further be determined since only the second coverslip was IF positive. bFor a minority of the specimens, no IF staining was performed for CMV (120 specimens}or adenovirus (177 specimens). None of these specimens were found to be positive by the conventional cell culture method. ‘IF on shell vials was confirmed by DIF in case of RSV and by conventional cell culture for the other viruses.
CMVs (n = 34) and/or viruses found on the second coverslip only, suggesting that the amount of virus in the sample was rather low. In general, more than 90% of influenza viruses and about 70% of the parainfluenza viruses, adenoviruses, and CMV were already detected on the first coverslip, that is, after only one day of shell vial culture.
DISCUSSION During the last few years, clinicians have become increasingly aware of the advantages of diagnosing RSV infections as quickly a s possible, both for therapeutic reasons [Hall et al., 19851and in order to prevent nosocomial infections [Krasinski et al., 19901. In addition, i t has also become clear that respiratory complaints of infants admitted to hospital may also be caused by numerous other viruses other than RSV [Blanding et al., 1989; Waner et al., 19901. These viruses may also spread nosocomially and thus should be detected fairly quickly as well. For RSV, the direct nonculture methods are faster and more sensitive than the slower culture methods, including shell vial cultures [Johnston and Siegel, 1990; this study]. This is especially true for clinical specimens obtained in the later stages of RSV infections, which contain relatively little viable RSV. This was clearly demonstrated in a n earlier study: 84% of RSV positive early specimens could be detected by shell vial cultures a s compared to only 53% of RSV positive follow-up specimens (unpublished results). This may also be the explanation for the relatively poor detection of RSV on shell vials in the present study: about onefourth of the RSV positive specimens were follow-up specimens. In contrast, direct nonculture methods for other respiratory viruses are generally still considered to be less successful, although positive reports have been pub-
lished for influenza A and B [Spada e t al., 1991; Waner et al., 19911. Influenza viruses and parainfluenza viruses are generally detected by hemadsorption on cell cultures, which may be found positive after only 2-3 days of culture, although much more time is generally required to detect the parainfluenza viruses [Minnich and Ray, 19871. Moreover, a positive result in this hemadsorption test always needs to be followed by another time-consuming step in order to identify the hemadsorbing virus. In contrast, with shell vial methods, the detection, identification, and typing of these viruses can be performed almost simultaneously within a few days (this study). Our shell vial system was technically somewhat complicated by the use of mixtures of monoclonal antibodies. Nevertheless, the present data clearly show that this approach resulted in a very sensitive method for the detection of (para-)influenza viruses, adenovirus, and CMV from respiratory specimens: 61 of the 63 viruses detected by the conventional culture method, usually taking 1-2 weeks, were detected by the shell vial culture method within a week, and often within one day. Moreover, the shell vial technique detected 44 additional viruses, which was CMV in most cases. This surplus of viruses was not surprising since, in our hands, CMV usually requires a longer conventional culture time than the two weeks used in the present study. Also, adenoviruses or other viruses present in small amounts (these were often found on the second CS only) may be difficult to detect within two weeks by conventional methods. Although this does not prove that all the additional 44 positive findings were specific, it is considered that most of them were specific: our monoclonal antibodies did not cross-react with laboratory strains of various tested viruses, and most of the additional CMVs were from patients which had been CMV-positive more than once. Even when only
Monoclonal Antibody Mixtures on Shell Vials
151
samples positive by conventional cultures or with DIF (in case of RSV) were considered to be true positives, the specificity of the shell vial culture approach would still be very high: 97% for CMV and a t least 99% for all the other viruses tested (calculated from the data in Table 111). For most of the individual nonRSV viruses, too few were detected to assess accurately the sensitivity of our shell vial system. Nevertheless, a rough estimation can be made: on the basis of the assumptions mentioned above the sensitivity would be >95% for the (para-)influenza viruses and adenovirus and at least 88% for CMV. Alternatively, when all positive findings are considered to be specific findings, the sensitivity would even be nearly 100%for all the nonRSV viruses. It is concluded that the “shell vial/monoclonal antibody mixture” approach is much more rapid than the conventional cell culture method and much more sensitive, without any loss of specificity. Moreover, the method fits easily into the daily routine presently performed in most virus diagnostic laboratories. In contrast, most rapid nonculture methods interrupt the daily practice in routine laboratories, where specimens are investigated preferentially in batches. This may even lead to batchwise, and thus less rapid, handling of the nonculture tests. In spite of this, it is clear that specimens suspected of containing RSV should be tested first by a sensitive rapid nonculture method for RSV (for instance, DIF). In addition, and perhaps only if the RSV test is negative, shell vial cultures with mixtures of monoclonal antibodies should be performed. Using this approach, well over 90% of the respiratory viruses present in these specimens will be detected within a few days.
REFERENCES August MJ, Warford AL ( 1987): Evaluation of a commercial monoclonal antibody for detection of adenovirus antigen. Journal of Clinical Microbiology 25:2233-2235.
B
’
Grefte JMM, van der Giessen M, van der Gun BTF, van Son WJ, The TH (1991):The predominant viral antigen present in the periferal blood leucocytes during a n active cytomegalovirus (CMV)infection is the lower matrix protein pp65. In Landine MP (ed):“Progress in Cytomegalovirus Research.” Amsterdam: Elsevier, pp 233-236. Hall CB, McBride JT, Gala CL, Hildreth SW, Schnabel KC (1985): Ribavirin treatment of respiratory syncytial viral infection in infants with underlying cardiopulmonary disease. Journal of the American Medical Association 254:3047-3051. Johnston SLG, Siege1 CS (1990):Evaluation of direct immunofluorescence, enzyme immunoassay, centrifugation culture, and conventional culture for the detection of respiratory syncytial virus. Journal of Clinical Microbiology 28:239&2397. Krasinski K, Lacouture R, Holzman RS, Waithe E, Bonk S, Hanna B (1990):Screening for respiratory syncytial virus and assignment to a cohort a t admission to reduce nosocomial transmission. J . Pediatrics 116:894-898. Mills RD, Cain KJ, Woods GL (19891:Detection of influenza virus by centrifugal inoculation of MDCK cells and staining with monoclonal antibodies. Journal of Clinical Microbiology 27:2505-2508. Minnich LL, Ray CG (1987):Early testing of cell cultures for detection of hemadsorbing viruses. Journal of Clinical Microbiology 25:421422. Morris DJ, Semple D 11990): Rapid detection of respiratory syncytial virus in nasopharyngeal aspirates by direct immunofluorescence using monoclonal antibodies. Serodiagnosis a n Immunotherapy in Infectious Disease 453-57. Ray CG, Minnich LL (1987): Efficiency of immunofluorescence for rapid detection of common respiratory viruses. Journal of Clinical Microbiology 25:355-357. Smith MC, Creutz C, Huang YT (1991):Detection of respiratory syncytial virus in nasopharyngeal secretions by shell vial technique. Journal of Clinical Microbiology 29:463465. Spada B, Biehler K, Chegas P, Kaye J , Riepenhoff-Talty M (1991): Comparison of rapid immunofluorescence assay to cell culture isolation for the detection of influenza A and B viruses in nasopharyngeal secretions from infants and children. Journal of Virological Methods 33:305-310. Stokes CE, Bernstein JM, Kyger SA, Hayden FG (1988):Rapid diagnosis of influenza A and B by 24-h fluorescent focus assays. Journal of Clinical Microbiology 26:1263-1266. Stout C, Murphy MD, Lawrence S, Julian S (1989): Evaluation of a monoclonal antibody pool for rapid diagnosis of respiratory viral infections. Journal of Clinical Microbiology 27:448-452. Takimoto S, Grandien M, Ishida MA, Pereira MS, Paiva TM, Ishimaru T, Makita EM, Martinez CHO (1991): Comuarison of enzvmelinked immunosorbent assay, indirect immun;fluorescence aisay, and virus isolation for detection of respiratory viruses in nasopharyngeal secretions. Journal of Clinical Microbiology 29470474. Waner JL, Whitehurst NJ, Todd SJ, Shalaby H, Wall LV (1990): Comparison of directigen RSV with viral isolation and direct immunofluorescence for the identification of respiratory syncytial virus. Journal of Clinical Microbiology 28:480483.
a
fluorescence testing, Journal of Clinical Microbiology 27:1438-
~
~
Rapid Detection of Respiratory Viruses Using Mixtures of Monoclonal Antibodies on Shell Vial Cultures Jurjen Schirm, Dirk S. Luijt, Geke W. Pastoor, Johan M. Mandema, and F. Peter Schroder Regional Public Health Laboratory M.S., D.S.L., G.W.P., F.P.S.), and Department of Pediatrics (J.M.M.),University Hospital, Groningen, The Netherlands Eleven hundred and thirty-three clinical specimens submitted to the laboratory for diagnosis of respiratory virus infections were tested by direct immunofluorescence (DIF) for respiratory syncytial virus (RSV),by shell vial culture, and by conventional cell culture. The shell vial cultures were stained with 8 different monoclonal antibodies both 1 day and 3-7 days after inoculation. In order to limit the cost and the workload, mixtures of monoclonal antibodies were used. Coverslips with HEp-, cells were incubated with a mixture of FITC-labeled monoclonal antibody to RSV and nonlabeled monoclonal antibody to adenovirus. When no RSV positive IF staining was observed after the first incubation step, the same coverslip was incubated once more with FITClabeled anti-mouse antibody. A positive reaction at this stage indicated the presence of adenovirus. Similarly, cultures of tertiary monkey kidney cells were investigated with a mixture of two FITC-labeled monoclonals to the influenza viruses A and B and three nonlabeled monoclonals to the parainfluenza viruses 1, 2 and 3. If influenza virus or parainfluenza virus was detected, the exact type was determined by staining different parts of a duplicate coverslip. Shell vial cultures for cytomegalovirus (CMV) were always performed separately on human embryonic lung fibroblasts. Using this approach, we detected RSV (n = 248), CMV (n = 42), parainfluenza virus (n = 311, influenza virus (n = 281, and adenovirus (n = 6), in most cases after only one day of culture. For RSV, the sensitivity of the shell vial method was too low (74%) to allow omission of DIF (sensitivity 95%). For the other viruses, the "shell vial/monoclonal antibody mixture" approach was very attractive, being rapid, very specific (>%'Yo), and also very sensitive (probably
>%YO).
0 1992 Wiley-Liss, Inc.
0 1992 WILEY-LISS, INC.
KEY WORDS: rapid diagnosis, shell vials, immunofluorescence, antibody mixtures, (para-)influenza viruses, respiratory syncytial virus
INTRODUCTION Clinical specimens submitted to diagnostic laboratories for diagnosis of respiratory virus infections are usually investigated by conventional cell culture methods. The detection of respiratory syncytial virus (RSV) is often performed more rapidly using nonculture methods such as direct immunofluorescence (DIF1or enzyme immunoassays or by rapid shell vial culture techniques or both [Johnston and Siegel, 1990; Morris et al., 1990; Takimoto et al., 1991; Smith et al., 19911. It has been reported that in more than 20% of specimens from patients with a presumptive diagnosis of RSV infection, other viruses are found, notably (para-)influenza viruses, cytomegalovirus (CMV), rhinovirus and, to a lesser extent, adenovirus, herpes simplex virus (HSV) and enteroviruses [Blanding et al., 1989; Waner et al., 19901. Most of these viruses, except for rhinovirus, can quite easily be detected in conventional cell culture systems, although culture of CMV, adenovirus and (para-)influenza viruses may take 1-4 weeks. In most studies, the sensitivity of rapid nonculture methods for respiratory viruses other than RSV is reported to be low: generally lower than 80% for the (para-)influenza viruses and even lower than 50% for the adenoviruses [Ray et al., 1987; Stokes et al., 1988; Mills et al., 1989; Stout et al., 1989; Takimoto et al., 19911. Accepted for publication March 31,1992. Address reprint requests to Dr. J. Schirm, Regional Public Health Laboratory, Van Ketwich Verschuurlaan 92, 9721 SW Groningen, The Netherlands.
Schirm et al.
148 The shell vial culture technique is used now universally for the rapid detection of, among others, CMV [Gleaves et al., 19851, and also, less often, for the detection of influenza viruses [Stokes e t al., 1988; Mills et al., 19891 and adenovirus [August et al., 19871. We describe the use of shell vial cultures for the detection of the influenza viruses A and B, the parainfluenza viruses 1, 2 and 3, adenovirus, and CMV in respiratory specimens. In order to limit the workload, mixtures of monoclonal antibodies were used.
MATERIALS AND METHODS Clinical Specimens Clinical specimens were obtained from all patients whose primary complaint was a respiratory illness. The majority of these patients were infants with bronchiolitis or pneumonia and/or a presumptive diagnosis of RSV infection. Some hospitalized children without respiratory complaints were included for epidemiological reasons. Most specimens were nasopharyngeal aspirates, pharyngeal washings or sputum samples. In addition, a limited number of throat swabs or other respiratory specimens was investigated. The specimens were submitted to the laboratory in virus transport medium, except for those few specimens (no swabs) t h a t arrived in the laboratory within a n hour. All specimens were diluted tenfold in virus transport medium or in phosphate buffered saline (PBS),mixed vigorously, and centrifuged at 800 x g for 5 min. The sediment was smeared onto a microscope slide, air dried, and fixed with acetone a t -20°C for 10 min. The remaining cells were resuspended in the washing fluid and used for tissue culture. Monoclonal Antibodies The fluorescein isothiocyanate (FITC)-labeled monoclonal antibodies of Imagen (Novo Nordisk Diagnostics Ltd., Cambridge, U.K.) were used for the detection of RSV, influenza A virus, and influenza B virus. Monoclonal antibodies to the parainfluenza viruses 1 , 2 and 3 were obtained from Chemicon International Inc. (Temecula, CA). Adenovirus was detected by the groupspecific monoclonal antibody 1F-8 (Whittaker, Walkersville, MD) or 1-3 (obtained from Dr. J.C. de Jong, National Institute of Public Health and Environment [RIVM], Bilthoven, The Netherlands). Preliminary comparison of these two adenovirus specific monoclonals showed no major differences in sensitivity, specificity or immunofluorescence (IF) patterns. Two different monoclonal antibody preparations were also used for the detection of CMV. Initially, a mixture of the monoclonal antibodies C10 and C11 (obtained from Prof. T.H. The, University of Groningen, The Netherlands), both directed to the lower matrix protein pp65 [Grefte et al., 19911, was used. Later, we changed to the monoclonal antibody 9221 (directedto the CMV major immediate early antigen; obtained from DuPont de Nemours, 's-Hertogenbosch, Netherlands), which appeared to be
slightly more sensitive than the ClO/Cll mixture when used in the shell vial culture system (unpublished observation). All monoclonal antibodies were tested for cross-reactivities with the above respiratory viruses, except for rhinovirus and the enteroviruses. In addition, the monoclonal antibodies to CMV were also tested for crossreactions with HSV and varicella zoster virus. Only the parainfluenza monoclonals showed some intertypic cross reactivity, but in all cases, the specific monoclonal antibody gave a much stronger IF intensity than the cross-reacting monoclonal antibody.
Cell Culture Techniques Human embryonic lung fibroblasts (HEL, a locally produced cell line), HEp-2 cells, and tertiary monkey kidney (t-MK) cells (obtained from RIVM, Bilthoven, The Netherlands) were grown in conventional culture tubes and on coverslips in shell vials. Minimum essential medium with Hanks salts, supplemented with nonessential amino acids and foetal calf serum was used. The monolayers of t-MK cells were always treated with 1mg/ml of trypsin before use. Clinical specimens were inoculated into duplicate culture tubes with t-MK and HEL cells (0.3 ml each) and into duplicate shell vials with t-MK, HEp-2 and HEL cells (0.2 ml each). Inoculation of shell vials was enhanced by centrifugation at 700 x g for 45 min. All culture tubes and shell vials were incubated at 37°C. Twice a week, for up to two weeks after inoculation, the monolayers in the culture tubes were microscopically examined for cytopathic effects (CPE). Monolayers showing CPE were stained by hematoxylin-eosin or with monoclonal antibodies. In cultures found positive for adenovirus or enterovirus, the exact virus type was determined by neutralization. In addition, t-MK culture tubes were tested for hemadsorbing activity after 3-7 days and after 14 days using guinea pig erythrocytes. The identity of the virus in hemadsorption positive cultures was subsequently determined by hemadsorption inhibition with virus specific antisera. One day after inoculation, the first of each pair of duplicate CS was fixed by replacing the medium with 1 ml of acetone (without drying) and leaving at -20°C for 10 min. If the subsequent immunofluorescent (IF) staining was negative, the second CS was fixed and stained 2-6 days later. IF Staining All CS were stained in two steps of 30 min at 37°C each. CS with HEp-2 cells were first incubated with a mixture of FITC-labeled monoclonal antibody to RSV, dilution 1/4, and one of the nonlabeled monoclonal antibodies to adenovirus: monoclonal 1F-8 at a dilution of 1/2 or monoclonal 1-3 a t a dilution of MOO. When no RSV-specific I F staining (speckled) was observed after the first step, the CS was rinsed in PBS and subsequently incubated with FITC-labeled rabbit antimouse immunoglobulin (Dakopatts, Glostrup, Denmark), di-
Monoclonal Antibody Mixtures on Shell Vials
149
TABLE I. ComDarison of Different Shell Vial Cultures* Virus(es) RS-virus Influenza AIB Parainfluenza-1,2,3 Adenovirus Cvtomegalovirus
t-MK
Cell tvDe HEP-2
+ ++ ++ +
++ + + ++
+
*
HEL
+ 2 + ++ ++
*IF staining was performed after one day of shell vial culture with different cell types. The degrees of IF intensity were expressed as: + + very strong; + moderate; weak.
*
lution 1/30. When adenovirus was detected after this staining step, usually showing nuclear IF, the virus was further typed by neutralization in the conventional culture method, if possible. CS with t-MK cells were first incubated with a mixture of two FITC-labeled monoclonal antibodies to the influenza viruses A and B, both at a dilution of 114, and the three nonlabeled monoclonal antibodies to the parainfluenza viruses 1, 2 and 3, each at a dilution of MOO. If no influenza virus-specific I F staining (nuclear or cytoplasmic) was observed after the first step, the CS was incubated once more with the FITC-labeled rabbit antimouse immunoglobulin in order to reveal the possible presence of one of the parainfluenza viruses (cytoplasmic IF). If influenza virus (A or B) or parainfluenza virus ( 1 , 2 , or 3) was detected on the first t-MK CS, the exact virus type was determined by staining different parts of the duplicate CS with the separate virus specific monoclonal antibodies. If only the second CS was IF-positive, typing was performed, if possible, on virus recovered from the conventional tube cell culture. CS with HEL cells were at first stained only with one of the CMV-specific monoclonal antibody preparations, both a t dilution 1/10, and subsequently with the FITClabeled rabbit antimouse immunoglobulin. Smears for detection of RSV by DIF were stained with the undiluted FITC-labeled monoclonal antibody to RSV for 15 min at 37°C.
RESULTS In a preliminary study, we tested the growth of the viruses and the performance of the monoclonal antibodies on shell vial cultures with different cell types (Table I). Because of the similar cell type preferences, we chose to combine the detection of the influenza viruses A and B and the parainfluenza viruses 1, 2 and 3 using mixtures of the appropriate monoclonal antibodies on t-MK cells. The IF patterns of adenovirus and CMV on HEL cells were usually quite similar: homogenous nuclear. For that reason, and since at that time, we did not have a n FITC-labeled monoclonal antibody to either CMV or adenovirus, we chose to combine the detection of adenovirus with the detection of RSV using shell vials with HEp-2 cells. Eleven hundred and thirty-three clinical specimens were investigated between December 1989 and September 1991. In all cases, DIF on RSV was combined
TABLE 11. Comparison of Shell Vial Method to Other Tests in 1,133 Clinical Samples All viruses
Test resulta Shell vial Shell vial Shell vial Shell vial
+ /other tests + + /other tests - /other tests + -
/other tests
-
238 54 63 778
RSV 177 10 61 885
Otherb viruses 61
44 2 1.026
”“Other test”: DIF in case of RSV and conventional cell culture in case of the other viruses. bThe nonRSV viruses tested on shell vials were the influenza viruses A and B, the parainfluenza viruses 1,2 and 3, CMV and adenovirus. For a minority of the specimens, no IF staining was performed for CMV (120 specimens) or adenovirus (177 specimens). None of these specimens were found to be positive by the conventional cell culture method.
with shell vial cultures of t-MK cells and HEp-2 cells and conventional tube cell cultures with t-MK cells and HEL cells. The CS with t-MK cells were always stained and further investigated as described above. The CS with HEp-2 cells were always stained with the monoclonal antibody to RSV, in most cases ( 9 5 6)~as a mixture with the monoclonal antibody to adenovirus. Shell vial cultures with HEL cells for the detection of CMV were performed for 1,013of the specimens. Using this approach, 354 positive samples were identified, containing RSV (n = 248), influenza virus A or B (n = 28), one of the parainfluenza viruses (n = 31), adenovirus ( n = 61, CMV ( n = 42), and other viruses (n = 9) which were only detectable by conventional cell culture in our system. These other viruses consisted of rhinovirus ( n = 3), coxsackie virus type A9 ( n = 21, echovirus type 7 ( n = 11,echovirus type 18 (n = l),and HSV type 1 (n = 2). The exact type of two of the adenoviruses could be determined, and was type 2 in both cases. Ten of the positive samples contained two different viruses, notably CMV + RSV (n = 81, adenovirus type 2 + RSV (n = l),and CMV + parainfluenza type 3 ( n = 1). In general, the shell vial approach appeared to be quite efficient: of the complete group of “shell vial viruses”, 292 (82%) of the total of 355 viruses were found on shell vials, whereas 301 (85%)were found by other methods, i.e., direct immunofluorescence (DIF) on smears in case of RSV and conventional cultures and/or hemadsorption for the other viruses (Table 11). However, the shell vial results were quite different for RSV, compared to the other viruses. Of the 248 specimens found positive for RSV, only 187 (75%)were detected by the shell vial method, and 238 (96%) by DIF. Only 10 RSV positive samples were missed by DIF. In all these 10 cases, the DIF could not be interpreted since the smears contained hardly any cells. In contrast, of the 107 specimens found positive for one of the other viruses, 105 (98%)were detected by the shell vial method and only 63 (59%)by the conventional culture method. In this group, the viruses missed by the shell vial method were one CMV, and one parainfluenza virus type 3 (see also Table 111).The 44 viruses missed by the conventional culture methods (Table 111) were either
Schirm et al.
150 TABLE 111. Results for Different Viruses Tested on Shell Vials
Virusa
RS-virus Influenza A Influenza B Influenza A/B Parainfluenza 1 Parainfluenza 2 Parainfluenza 3 P. influenza 11213 Adenovirus CMV
Total nonRSV
Total number of positives
Total
248 15 11 2 11 3 16 1 6 42 107
187 15 11 2 11 3 15 1 6 41 105
Viruses found by IFb on shell vials Confirmed‘ Not confirmed 177 13 11 0 8 3 15 0 4 7 61
10 2 0 2 3 0 0 1 2 34 54
“Influenza AIB = influenza virus A or B; Parainfluenza 11213 = parainfluenza virus 1 . 2 or 3. In these cases, the exact virus type could not further be determined since only the second coverslip was IF positive. bFor a minority of the specimens, no IF staining was performed for CMV (120 specimens}or adenovirus (177 specimens). None of these specimens were found to be positive by the conventional cell culture method. ‘IF on shell vials was confirmed by DIF in case of RSV and by conventional cell culture for the other viruses.
CMVs (n = 34) and/or viruses found on the second coverslip only, suggesting that the amount of virus in the sample was rather low. In general, more than 90% of influenza viruses and about 70% of the parainfluenza viruses, adenoviruses, and CMV were already detected on the first coverslip, that is, after only one day of shell vial culture.
DISCUSSION During the last few years, clinicians have become increasingly aware of the advantages of diagnosing RSV infections as quickly a s possible, both for therapeutic reasons [Hall et al., 19851and in order to prevent nosocomial infections [Krasinski et al., 19901. In addition, i t has also become clear that respiratory complaints of infants admitted to hospital may also be caused by numerous other viruses other than RSV [Blanding et al., 1989; Waner et al., 19901. These viruses may also spread nosocomially and thus should be detected fairly quickly as well. For RSV, the direct nonculture methods are faster and more sensitive than the slower culture methods, including shell vial cultures [Johnston and Siegel, 1990; this study]. This is especially true for clinical specimens obtained in the later stages of RSV infections, which contain relatively little viable RSV. This was clearly demonstrated in a n earlier study: 84% of RSV positive early specimens could be detected by shell vial cultures a s compared to only 53% of RSV positive follow-up specimens (unpublished results). This may also be the explanation for the relatively poor detection of RSV on shell vials in the present study: about onefourth of the RSV positive specimens were follow-up specimens. In contrast, direct nonculture methods for other respiratory viruses are generally still considered to be less successful, although positive reports have been pub-
lished for influenza A and B [Spada e t al., 1991; Waner et al., 19911. Influenza viruses and parainfluenza viruses are generally detected by hemadsorption on cell cultures, which may be found positive after only 2-3 days of culture, although much more time is generally required to detect the parainfluenza viruses [Minnich and Ray, 19871. Moreover, a positive result in this hemadsorption test always needs to be followed by another time-consuming step in order to identify the hemadsorbing virus. In contrast, with shell vial methods, the detection, identification, and typing of these viruses can be performed almost simultaneously within a few days (this study). Our shell vial system was technically somewhat complicated by the use of mixtures of monoclonal antibodies. Nevertheless, the present data clearly show that this approach resulted in a very sensitive method for the detection of (para-)influenza viruses, adenovirus, and CMV from respiratory specimens: 61 of the 63 viruses detected by the conventional culture method, usually taking 1-2 weeks, were detected by the shell vial culture method within a week, and often within one day. Moreover, the shell vial technique detected 44 additional viruses, which was CMV in most cases. This surplus of viruses was not surprising since, in our hands, CMV usually requires a longer conventional culture time than the two weeks used in the present study. Also, adenoviruses or other viruses present in small amounts (these were often found on the second CS only) may be difficult to detect within two weeks by conventional methods. Although this does not prove that all the additional 44 positive findings were specific, it is considered that most of them were specific: our monoclonal antibodies did not cross-react with laboratory strains of various tested viruses, and most of the additional CMVs were from patients which had been CMV-positive more than once. Even when only
Monoclonal Antibody Mixtures on Shell Vials
151
samples positive by conventional cultures or with DIF (in case of RSV) were considered to be true positives, the specificity of the shell vial culture approach would still be very high: 97% for CMV and a t least 99% for all the other viruses tested (calculated from the data in Table 111). For most of the individual nonRSV viruses, too few were detected to assess accurately the sensitivity of our shell vial system. Nevertheless, a rough estimation can be made: on the basis of the assumptions mentioned above the sensitivity would be >95% for the (para-)influenza viruses and adenovirus and at least 88% for CMV. Alternatively, when all positive findings are considered to be specific findings, the sensitivity would even be nearly 100%for all the nonRSV viruses. It is concluded that the “shell vial/monoclonal antibody mixture” approach is much more rapid than the conventional cell culture method and much more sensitive, without any loss of specificity. Moreover, the method fits easily into the daily routine presently performed in most virus diagnostic laboratories. In contrast, most rapid nonculture methods interrupt the daily practice in routine laboratories, where specimens are investigated preferentially in batches. This may even lead to batchwise, and thus less rapid, handling of the nonculture tests. In spite of this, it is clear that specimens suspected of containing RSV should be tested first by a sensitive rapid nonculture method for RSV (for instance, DIF). In addition, and perhaps only if the RSV test is negative, shell vial cultures with mixtures of monoclonal antibodies should be performed. Using this approach, well over 90% of the respiratory viruses present in these specimens will be detected within a few days.
REFERENCES August MJ, Warford AL ( 1987): Evaluation of a commercial monoclonal antibody for detection of adenovirus antigen. Journal of Clinical Microbiology 25:2233-2235.
B
’
Grefte JMM, van der Giessen M, van der Gun BTF, van Son WJ, The TH (1991):The predominant viral antigen present in the periferal blood leucocytes during a n active cytomegalovirus (CMV)infection is the lower matrix protein pp65. In Landine MP (ed):“Progress in Cytomegalovirus Research.” Amsterdam: Elsevier, pp 233-236. Hall CB, McBride JT, Gala CL, Hildreth SW, Schnabel KC (1985): Ribavirin treatment of respiratory syncytial viral infection in infants with underlying cardiopulmonary disease. Journal of the American Medical Association 254:3047-3051. Johnston SLG, Siege1 CS (1990):Evaluation of direct immunofluorescence, enzyme immunoassay, centrifugation culture, and conventional culture for the detection of respiratory syncytial virus. Journal of Clinical Microbiology 28:239&2397. Krasinski K, Lacouture R, Holzman RS, Waithe E, Bonk S, Hanna B (1990):Screening for respiratory syncytial virus and assignment to a cohort a t admission to reduce nosocomial transmission. J . Pediatrics 116:894-898. Mills RD, Cain KJ, Woods GL (19891:Detection of influenza virus by centrifugal inoculation of MDCK cells and staining with monoclonal antibodies. Journal of Clinical Microbiology 27:2505-2508. Minnich LL, Ray CG (1987):Early testing of cell cultures for detection of hemadsorbing viruses. Journal of Clinical Microbiology 25:421422. Morris DJ, Semple D 11990): Rapid detection of respiratory syncytial virus in nasopharyngeal aspirates by direct immunofluorescence using monoclonal antibodies. Serodiagnosis a n Immunotherapy in Infectious Disease 453-57. Ray CG, Minnich LL (1987): Efficiency of immunofluorescence for rapid detection of common respiratory viruses. Journal of Clinical Microbiology 25:355-357. Smith MC, Creutz C, Huang YT (1991):Detection of respiratory syncytial virus in nasopharyngeal secretions by shell vial technique. Journal of Clinical Microbiology 29:463465. Spada B, Biehler K, Chegas P, Kaye J , Riepenhoff-Talty M (1991): Comparison of rapid immunofluorescence assay to cell culture isolation for the detection of influenza A and B viruses in nasopharyngeal secretions from infants and children. Journal of Virological Methods 33:305-310. Stokes CE, Bernstein JM, Kyger SA, Hayden FG (1988):Rapid diagnosis of influenza A and B by 24-h fluorescent focus assays. Journal of Clinical Microbiology 26:1263-1266. Stout C, Murphy MD, Lawrence S, Julian S (1989): Evaluation of a monoclonal antibody pool for rapid diagnosis of respiratory viral infections. Journal of Clinical Microbiology 27:448-452. Takimoto S, Grandien M, Ishida MA, Pereira MS, Paiva TM, Ishimaru T, Makita EM, Martinez CHO (1991): Comuarison of enzvmelinked immunosorbent assay, indirect immun;fluorescence aisay, and virus isolation for detection of respiratory viruses in nasopharyngeal secretions. Journal of Clinical Microbiology 29470474. Waner JL, Whitehurst NJ, Todd SJ, Shalaby H, Wall LV (1990): Comparison of directigen RSV with viral isolation and direct immunofluorescence for the identification of respiratory syncytial virus. Journal of Clinical Microbiology 28:480483.
a
fluorescence testing, Journal of Clinical Microbiology 27:1438-
~
~
