Journal of Medical Virology 38:283-287 (1992)
Detection of Respiratory Syncytial Virus in Acute Bronchiolitis in Infants Heather A. Cubie, J.M. Inglis, Eleanor E. Leslie, A.T. Edmunds, and B. Totapally Regional Virus Laboratory, City Hospital and Royal Hospital for Sick Children, Edinburgh, Scotland Direct detection of respiratory syncytial (RS) virus antigen i n nasopharyngeal secretions (NPS) provides the most rapid diagnostic test for RS infection, but more sensitive methods might be more beneficial in the study of virus shedding. RS virus RNA was extracted from cells stored at -70°C either in suspension with added RNAse inhibitor or as a pellet without inhibitor. The RNA was reverse transcribed, the resultant cDNA amplified by the polymerase chain reaction and detected by ethidium bromide staining after electrophoresis through agarose gel (RT-PCR). Of 217 specimens tested, 106 were positive b y antigen detection, 99 by RT-PCR, and 92 by virus isolation. In a series of 97 sequential NPS specimens from 15 infants in whom RS virus induced bronchiolitis was confirmed, antigen detection proved most sensitive in the first week after onset and RT-PCR detected most positive specimens in the second week. Storage of the cells as a pellet proved more satisfactory than storage as a suspension. A further round of amplification using nested primers increased the number of positive results obtained by RT-PCR. The sensitivity of antigen detection using directly labelled monoclonal antibody to RS virus was slightly higher than that of RT-PCR, but the specificity was slightly lower. o 1992 Wiley-Liss, Inc.
KEY WORDS: RS virus, antigen detection, RTPCR
INTRODUCTION Respiratory syncytial (RS) virus is associated with acute bronchiolitis in infants and is the commonest cause of lower respiratory tract infection requiring hospitalisation in this age group. Rapid diagnostic tests have proved to be invaluable in clinical management of individual patients because of the availability of tribavirin, a synthetic analogue of guanosine with antiviral activity against RS virus IHall et al., 19851. However nosocomial spread of RS virus is a major problem in paediatric wards and in immunocompromised patients [Englund et al., 19911 and sensitive tests of virus shedding would be useful not only to monitor the effective0 1992 WILEY-LISS. INC.
ness of treatment regimes but also to limit hospital transmission in long-stay patients such as those with underlying cardiac or pulmonary problems. The effectiveness of commercially available fluorescein labelled monoclonal antibodies to RS virus has already been demonstrated in providing a rapid diagnostic service, with results being available within a n hour of receipt of a n urgent specimen [Cubie et al., 1990I. Antigen detection proved more sensitive than detection of viral sequences by in situ hybridisation (ISH) [Cubie et al., 19911, but i t was considered that nucleic acid amplification techniques might prove more useful in the study of virus shedding in sequential specimens taken from patients, both treated and untreated.
MATERIALS AND METHODS Specimens NPS were received from hospitalised infants with bronchiolitis and examined directly for the presence of RS virus antigen (see below). When a positive specimen was identified, the ward was informed and a careful surveillance procedure set up, with attempts being made to obtain further NPS specimens daily during the first week and on days 10, 14, 21, and 28. Out-patient visits were arranged a s close as possible to the agreed dates. Ethical permission and parental consent for follow-up studies were obtained. Each sequential specimen was examined for the presence of RS virus antigen and inoculated into cell cultures the day it arrived. RT-PCR studies were carried out on samples stored for varying lengths of time a t -70°C a s a ) a suspension in phosphate buffered saline (PBS) to which placental RNAse inhibitor had been added or b) a cell pellet without the addition of RNAse inhibitor. Additional single specimens were received from other infants suspected of having respiratory virus infection. They were treated in a similar manner. Antigen Detection This was carried out on cytospin preparations using commercially available labelled monoclonal antibodies Accepted for publication May 12, 1992. Address reprint requests to Dr. H.A. Cubie, Regional Virus Laboratory, City Hospital, Greenbank Drive, Edinburgh EHlO 5SB, Scotland.
284
Cubie et al.
shaken gently and held on ice for a t least 1 h. It was centrifuged at 12,OOOg for 15 rnin and the supernate discarded. One milliliter of 75% ethanol was added and the tube re-centrifuged a t 9,OOOg for 10 min. The superVirus Culture nate was carefully removed and the tubes pulse-spun to Conventional cell cultures were inoculated with NPS enable the remaining supernate to be totally removed. in transport medium and examined twice weekly for 3 The tubes were upturned on filter paper and the pellets weeks for cytopathic effect [Cubie et al., 19911. HEp-2 dried for approximately 20 min. Extracted RNA was cells were used for the passage every 5-7 days of wild redissolved in 10 p1 sterile distilled water treated with type RS virus strains to act as positive controls, and diethyl pyrocarbonate (DEPC-DW). uninoculated HEp-2 cells were used as negative conReverse transcription. In a sterile 0.5 ml Eppendorf, trols. The cells from a single culture tube were washed 2 pl of primer pair -0.1 pg of each primer) and the 10 p1 twice with ice-cold Dulbecco’s balanced salt solution of RNA sample were mixed and heated a t 70°C for 10 and scraped into 1ml of solution in a n Eppendorftube. min. A master mix containing per sample 4 p1 of 5 x RT They were centrifuged to form a pellet (5 minutes; buffer, 2 pl dithiothreitol and 2 p1 of 0.01M deoxyribo10,000 rpm in a microfuge a t 4“C), the supernate was nucleotides was added. The enzyme reverse tranremoved and the pellet stored at -70°C until required. scriptase (MMLV RT; Superscript, Gibco BRL Ltd) was In this way control samples were always available and added at -20°C allowing 1 p1 per sample. The enzyme ready to use when NPS samples were tested. was allowed to act at 37°C in a water bath for 60 min, followed by chilling on ice. Reverse Transcription-Polymerase Chain Amplification. A master mix of 5 pl 1 0 PCR ~ Reaction (RT-PCR) buffer, 25 pl DEPC-DW and 0.4 pl DNA polymerase Primer sequences. The primers used were kindly from Thermus thermophilus (Tth, 5 u n i t d m l ; Hybaid provided by Dr. Pat Cane, Department of Biological Sci- Ltd. Teddington) per sample was made up and added to ences, University of Warwick, and had the following each RT product. Fifty microliters of paraffin oil were sequences [Cane and Pringle, 19911: added and the samples subjected to 30 cycles of denaturation, annealing, and extension (93°C for 1.5 min, 1938 55°C for 1.5 min, and 72°C for 1.5min) followed by one primer 1-5‘ GGA ACA AGT TGT TGA GGT TTA cycle of 93°C for 1.5 min, 55°C for 1.5 min, and 72°C for TGA ATA TGC 5 min in a Hybaid thermal cycler. 2215 A second round of amplification was carried out by primer 2-5’ CTT CTG CTG TCA AGT CTA GTA adding 10 pl of first round product to 1p1 primer pair 3 CAC TGT AGT and 4. This was heated a t 70°C for 10 min and cooled on These sequences come from part of the sequence of the ice and a mix of Tth in buffer added as before. The ‘N’ gene of the ‘A’ subtype of RS virus [Collins et al., samples were subjected to 30 cycles as before except 19851 and amplify a segment of 278 bases. Because of that a n annealing temperature of 45°C was used. the high degree of sequence homology of the ‘ N gene Agarose gel electrophoresis. Twelve microliters of (96% between subtypes ‘A’ and ‘B’ [Johnson and Col- amplified product were mixed with 2 pl of loading dye lins, 198911,these primers should amplify nearly all RS (bromophenol blue [Sambrook e t al., 19891) and added virus strains likely to be encountered. to a single slot in a slab gel consisting of 3% NuSieve Nested inner primers were synthesised by Oswel GTG agarose and 1%SeaKem GTG agarose (ICN Flow; DNA Service, Department of Chemistry, University of High Wycombe) and containing 5 pl ethidium bromide Edinburgh and had the following sequences: (10 mgimL), electrophoresed at a constant current of 40 mA for 1h and examined under a UV light source for 1983 the presence of bands of DNA of appropriate size. DNA primer 3-5’ TGA AGC AGG ATT CTA CCA TAT markers (1 kb DNA ladder and @X 174 RF DNAiHae 2160 I11 fragments; Gibco BRL, Life Technologies Ltd., Paisprimer 4-5‘ TTC AGC ATA TGC CTT TGC TGC ley) were added to each gel. RT-PCR RESULTS RT-PCR was carried out on stored cellular suspenExtraction. NPS cell suspensions in Eppendorf tubes were centrifuged a t 4°C to pellet the cells and the sions from routine first specimens submitted to the labsupernate removed. One milliliter of RNAzol B (Bio- oratory or from sequential samples obtained from 15 genesis Ltd., Bournemouth) was added to the pellet and children following confirmation of RS virus induced the tube shaken vigorously for 15 sec. One hundred bronchiolitis. One hundred twenty-three routine specimicroliters of chloroform were added, the mixture mens stored at -70°C in the presence of RNAse inhibishaken gently and kept on ice for 5 min then centri- tor for periods ranging from a few days to several weeks fuged at l2,OOOg for 15 rnin at 4°C. The aqueous top were tested and the results are shown in Table I. While layer was transferred to a sterile Eppendorf tube and 45 (68% of all positives) contained detectable RS viral a n equal volume of isopropanol added. The mixture was sequences by RT-PCR after storage, 49 (74%)had been
to RS virus (“Imagen” RS kit; Dako Diagnostics, Ltd., High Wycombe) as previously described [Cubie et al., 19911.
Detection of RS Virus in Infants
285
TABLE I. Comparison of RT-PCR, Antigen Detection, and Virus Isolation Using Stored Cell Suspensions From Nasopharyngeal Secretions”
RS FA RS FA RS I RS I
+ + -
RT-PCR+
RT-PCR-
Total
45 32 13 26 19
78 17 61 22 56
123 49 74 48 75
“RT-PCR = reverse transcription and polymerase chain reaction; RS FA = detection of antigen using fluorescein labelled antibody to RS virus; RS I = isolation of virus in cell culture.
positive by fluorescent antigen (FA) detection, and only 32 were positive by both methods. Similarly, 48 (73%) were positive by RS virus isolation but only 26 agreed with the RT-PCR result. Thus the overall detection rate was similar for all three methods, but both antigen detection and virus culture were carried out immediately on receipt of the specimen, while RT-PCR was carried out after storage. This may have reduced the apparent sensitivity of the RT-PCR technique and these results are discussed. A gel containing specimens amplified by RT-PCR is shown in Figure 1. PCR products are apparent for three specimens which were also positive by antigen detection and virus isolation (lanes 2,4, and 5) and a further two products from specimens from which no virus was isolated (lanes 1 and 3). No bands are apparent in two other NPS one of which appeared positive by antigen detection. An RS virus cell culture and uninoculated HEp-2 cells are included as positive and negative controls. During the period of the study, seven NPS yielded other viruses in culture and none of these gave a positive band in the RT-PCR. Ninety-seven of the 123 specimens came from 15 infants and were collected between 1 and 30 days after onset of symptoms. Comparison of the three detection methods in these specimens is shown in Figure 2. Virus detection was high in the first 6 days with 85% of specimens giving a positive result by a t least one of the methods, (12114 in the first 3 days, 19/23 during days 4-6). Antigen detection was the single most sensitive test at this stage. During the second week, more than 60% still gave positive results but more positives were detected by RT-PCR (21/40) than by antigen detection (17140) or culture (14140).Of the 15 infants from whom sequential samples were obtained, 5 were treated with tribavirin by small particle aerosol generator (SPAG) for 3 days, ranging from days 3 to 5 of symptoms to days 8-10 of symptoms. The duration of excretion of viable virus was comparable in both treated and untreated groups (approximately 8 days post onset). Antigen or nucleic acid was detected for slightly longer, particularly in the treated group (mean of 17.4 days post onset compared with mean of 13.9 days), but statistical analysis of such small groups was not possible. Because the sensitivity of the RT-PCR was lower than expected, changes in the preparation and storage of samples for RT-PCR were made in the current RS virus season.
Fig. 1. Agarose gel of RT-PCR products stained with ethidium bromide and photographed under UV light (abbreviations as in Table I). Lanes 1-7: Products from seven specimens of NPS. Lanes 6-9 HEp-2 cell extracts from RS virus infected and uninoculated cells, respectively. Lane 1 0 1 kb DNA ladder.
After routine cytospin preparation for antigen detection, the remaining cell suspension was centrifuged, the supernate removed and the cell pellet immediately stored at -70°C without the addition of RNAse inhibitor. Ninety four specimens were examined following this method and the results are presented in Table 11. Of 62 specimens positive by any one method, 57 (92%) were positive by FA detection, 52 (84%)by RT-PCR and 44 (71%)by virus isolation. Much closer agreement was obtained in this series with 50/52 (96%)of the RT-PCR positives being confirmed by another method or coming from patients from whom previous positive specimens had been obtained. A similar percentage of isolation positive results (95%;42144) were confirmed by another method, but only 49/57 (86%) of antigen positive results were confirmed. The sensitivity and specificity of these results is discussed. Details of the onset of symptoms were not available for all patients in this group, but from the information received, antigen and isolation positive results were obtained up to 9 days and RT-PCR positive results up to 12 days after the onset of symptoms. It was possible to re-amplify using inner primers 3 and 4, twenty-two specimens (15 of which are included in the analysis in Table 11) which had been negative by RT-PCR but positive by a t least one other method. Ten of these gave bands on agarose gels and migrated as
Cubie et al.
286
30
30
....... ....... ....... ....... ....... ....... ....... .......
1
25
25
20
20
15
15 10
10
5
0
~
1-3
4 - 6
7-9
10-14
15-21
...... ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... -.... ...... ...... ...... ...... ......
5
0 22->28
Days after onset of symptoms Fig. 2. Total number of positive RS virus results and number of specimens positive by each method a t various times after onset of symptoms (abbreviations as in Table I; n = No. of specimens).
TABLE 11. Comparison of RT-PCR, Antigen Detection, and Virus Isolation Using Stored Pelleted Cells From Nasopharvngeal Secretions”
RS FA RS FA RS I RS 1
RT-PCR+
RT-PCR
Total
52 43 9 33
42 14 28 11 31
94 57 37 44 50
+ +
19
~
~
‘Abbreviations as in Table I
expected for a product of 177 base pairs in length. The sensitivity and specificity of the different tests used are discussed below.
DISCUSSION The results demonstrate several difficulties in translating experimental research protocols into a diagnostic procedure. With a peak incidence of 26 NPS per day and repeat specimens from increasing numbers of infants initially entered into the study, the logistics of including RT-PCR into the daily routine of a diagnostic laboratory handling large numbers of relevant specimens was considerable and necessitated storage of the samples a t -70°C. However, storage of the extracted cells a s a suspension in the presence of RNAse inhibitor
I Bruce e t al., 19891 proved unsatisfactory. Agreement between antigen detection, isolation and RT-PCR results in the 123 specimens stored in this way was poor (Table I) and the RT-PCR results suggested a lower than expected sensitivity for this test. The problems of working with RNA are well appreciated and clinical specimens are rarely taken, sent, or stored under ideal conditions. Nevertheless, Zhang and Evans [ 19911 report a n RT-PCR method for influenza viruses made sensitive enough to cope with moderately degraded samples even without inhibition of ribonucleases. Despite this reduced sensitivity, the RT-PCR detected RS virus sequences in sequential specimens from infants with bronchiolitis and more positives were detected by this method in the second week after onset of symptoms (Fig. 2) than by antigen detection or by virus isolation. Recent experiments with cell pellets stored a t - 70°C without RNAse inhibitor gave more satisfactory results and were simpler to handle (Table 11). The RNAzol solution could be added directly to the frozen pellet at the start of the extraction procedure. In future studies of virus shedding, it is surmised that extraction of RNA directly from pelleted cells without - 70°C storage would give even more satisfactory results with greater and prolonged sensitivity. Extraction, reverse transcription, and amplification could be carried out within the working day and gel electrophoresis the following day.
Detection of RS Virus in Infants Increased sensitivity can be obtained by the use of nested primers in a second round of amplification and this was demonstrated in a group of 22 specimens (of which 15 are included in the analysis in Table 11)which appeared to be negative on the first round of amplification despite detection of antigen or virus. Ten were detected in gels using inner primers 3 and 4.Thus of the group of 62 specimens positive overall (Table 111, 54 (87%)were positive by RT-PCR including nested amplifications and 52 (96%)were confirmable by other methods. This compares with 57 (92%) positive by antigen detection of which on 51 (89%) were confirmable, and suggests a greater sensitivity for the antigen detection system. Diagnostic laboratories should be aware, however, of occasional false positive results with the monoclonal antibody. A second round of amplification adds considerably to the cost and time of the test, but nested primers are considered to provide a significant increase in specificity a s well as sensitivity [Clewley, 19891 and to reduce the risks of cross-contamination [Zhang and Evans, 19911. Furthermore, carefully chosen nested primers can permit the separate identification of subgroups ‘A’ and ‘B’ of RS virus [Cane and Pringle, 19911 and this extra stage could be very useful in epidemiologi.ca1 studies. It was hoped to study the difference in virus shedding in treated and untreated patients. However, although 15 infants were treated with tribavirin in the Royal Hospital for Sick Children in Edinburgh in 1989-90, only five required treatment during the 1990-91 season, the year of the study outlined above. From the limited data available on these infants and on 10 comparably aged controls, excretion of virus occurred for similar lengths of time, antigen and nucleic acid detection were only positive marginally longer in the treated group and the duration of symptoms was similar. However, the treated group had underlying disease including pulmonary dysplasia and congenital heart defects and were considerably more ill than the untreated group. It may be, therefore, that tribavirin had indeed reduced the markers of RS infection. While it would be unethical not to treat severely ill patients to show a difference in virus shedding, it would be worthwhile to compare virus shedding and detection of nucleic acid in treated and untreated hospitalised infants without underlying disease. Such a study is under consideration. While there is undoubtedly a place for molecular techniques such a s RT-PCR in the study of respiratory infections in children, we would reiterate our comment
287
[Cubie e t al., 19911 that antigen detection using commercially available fluorescein-labelled monoclonal antibodies to RS virus remains the most cost effective and rapid technique currently available for the diagnosis of RS virus induced lower respiratory tract infection.
ACKNOWLEDGMENTS We would particularly like to thank Dr. Pat Cane of the Department of Biological Sciences, University of Warwick for her helpful discussions concerning the RTPCR protocol and the kind provision of primers 1 and 2. We would also like to thank Margaret Devine of the Respiratory Laboratory, Royal Hospital for Sick Children, for coordinating the specimen collection and the parents of very small babies who willingly returned for a n increased number of out-patient visits. We are grateful to the Scottish Hospital Endowments Research Trust for funding this project and our thanks are due to Mrs. Avril Garvey for typing the manuscript. REFERENCES Bruce CB, Al-Nakib W, Almond JW, Tyrell DAJ (1989): Use of synthetic oligonucleotideprobes to detect rhinovirus RNA. Archives of Virology 105:179-187. Cane PA, Pringle CR (1991): Respiratory syncytial virus heterogeneity during an epidemic: Analysis by limited nucleotide sequencing (SH gene) and restriction mapping (N gene) Journal of General Virology 72:349-357. Clewley J P (1989): The polymerase chain reaction, a review of the practical limitations for human immunodeficiency virus diagnosis. Journal of Virological Methods 25:179-188. Collins PL, Anderson K, Langer SJ, Wertz GW (1985): Correct sequence for the major nucleocapsid protein mRNA of respiratory syncytial virus. Virology 146:69-77. Cubie HA, Winter GF, Leslie EE, Inglis JM (1990):Rapid detection of respiratory syncytial virus antigens in nasopharyngeal secretions. Journal of Virological Methods 27:121-124. Cubie HA, Inglis JM, McGowan A (1991): Detection of respiratory syncytial virus antigen and nucleic acid in clinical specimens using synthetic oligonucleotides. Journal of Virological Methods 34:27-35. Englund JA, Anderson LJ, Rhame FS (1991): Nosocomial transmission of RS virus in immunocompromised adults. Journal of Clinical Microbiology 29:115-119. 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 254:30473051. Johnson PR, Collins PL (1989):The 1B (NSZ), 1C (NS1)and N proteins of human respiratory syncytial virus (RSV)of antigenic subgroups A + B: Sequence conservation and divergence within RSV genomic RNA. Journal of General Virology 70:1539-1547. Sambrook J , Fritsch EF, Maniatis T (1989): Molecular Cloning: A Laboratory Manual. 2nd Edition. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press, Zhang W, Evans DH (1991): Detection and identification of human influenza viruses by polymerase chain reaction. Journal of Virological Methods 33:165-189.
Detection of Respiratory Syncytial Virus in Acute Bronchiolitis in Infants Heather A. Cubie, J.M. Inglis, Eleanor E. Leslie, A.T. Edmunds, and B. Totapally Regional Virus Laboratory, City Hospital and Royal Hospital for Sick Children, Edinburgh, Scotland Direct detection of respiratory syncytial (RS) virus antigen i n nasopharyngeal secretions (NPS) provides the most rapid diagnostic test for RS infection, but more sensitive methods might be more beneficial in the study of virus shedding. RS virus RNA was extracted from cells stored at -70°C either in suspension with added RNAse inhibitor or as a pellet without inhibitor. The RNA was reverse transcribed, the resultant cDNA amplified by the polymerase chain reaction and detected by ethidium bromide staining after electrophoresis through agarose gel (RT-PCR). Of 217 specimens tested, 106 were positive b y antigen detection, 99 by RT-PCR, and 92 by virus isolation. In a series of 97 sequential NPS specimens from 15 infants in whom RS virus induced bronchiolitis was confirmed, antigen detection proved most sensitive in the first week after onset and RT-PCR detected most positive specimens in the second week. Storage of the cells as a pellet proved more satisfactory than storage as a suspension. A further round of amplification using nested primers increased the number of positive results obtained by RT-PCR. The sensitivity of antigen detection using directly labelled monoclonal antibody to RS virus was slightly higher than that of RT-PCR, but the specificity was slightly lower. o 1992 Wiley-Liss, Inc.
KEY WORDS: RS virus, antigen detection, RTPCR
INTRODUCTION Respiratory syncytial (RS) virus is associated with acute bronchiolitis in infants and is the commonest cause of lower respiratory tract infection requiring hospitalisation in this age group. Rapid diagnostic tests have proved to be invaluable in clinical management of individual patients because of the availability of tribavirin, a synthetic analogue of guanosine with antiviral activity against RS virus IHall et al., 19851. However nosocomial spread of RS virus is a major problem in paediatric wards and in immunocompromised patients [Englund et al., 19911 and sensitive tests of virus shedding would be useful not only to monitor the effective0 1992 WILEY-LISS. INC.
ness of treatment regimes but also to limit hospital transmission in long-stay patients such as those with underlying cardiac or pulmonary problems. The effectiveness of commercially available fluorescein labelled monoclonal antibodies to RS virus has already been demonstrated in providing a rapid diagnostic service, with results being available within a n hour of receipt of a n urgent specimen [Cubie et al., 1990I. Antigen detection proved more sensitive than detection of viral sequences by in situ hybridisation (ISH) [Cubie et al., 19911, but i t was considered that nucleic acid amplification techniques might prove more useful in the study of virus shedding in sequential specimens taken from patients, both treated and untreated.
MATERIALS AND METHODS Specimens NPS were received from hospitalised infants with bronchiolitis and examined directly for the presence of RS virus antigen (see below). When a positive specimen was identified, the ward was informed and a careful surveillance procedure set up, with attempts being made to obtain further NPS specimens daily during the first week and on days 10, 14, 21, and 28. Out-patient visits were arranged a s close as possible to the agreed dates. Ethical permission and parental consent for follow-up studies were obtained. Each sequential specimen was examined for the presence of RS virus antigen and inoculated into cell cultures the day it arrived. RT-PCR studies were carried out on samples stored for varying lengths of time a t -70°C a s a ) a suspension in phosphate buffered saline (PBS) to which placental RNAse inhibitor had been added or b) a cell pellet without the addition of RNAse inhibitor. Additional single specimens were received from other infants suspected of having respiratory virus infection. They were treated in a similar manner. Antigen Detection This was carried out on cytospin preparations using commercially available labelled monoclonal antibodies Accepted for publication May 12, 1992. Address reprint requests to Dr. H.A. Cubie, Regional Virus Laboratory, City Hospital, Greenbank Drive, Edinburgh EHlO 5SB, Scotland.
284
Cubie et al.
shaken gently and held on ice for a t least 1 h. It was centrifuged at 12,OOOg for 15 rnin and the supernate discarded. One milliliter of 75% ethanol was added and the tube re-centrifuged a t 9,OOOg for 10 min. The superVirus Culture nate was carefully removed and the tubes pulse-spun to Conventional cell cultures were inoculated with NPS enable the remaining supernate to be totally removed. in transport medium and examined twice weekly for 3 The tubes were upturned on filter paper and the pellets weeks for cytopathic effect [Cubie et al., 19911. HEp-2 dried for approximately 20 min. Extracted RNA was cells were used for the passage every 5-7 days of wild redissolved in 10 p1 sterile distilled water treated with type RS virus strains to act as positive controls, and diethyl pyrocarbonate (DEPC-DW). uninoculated HEp-2 cells were used as negative conReverse transcription. In a sterile 0.5 ml Eppendorf, trols. The cells from a single culture tube were washed 2 pl of primer pair -0.1 pg of each primer) and the 10 p1 twice with ice-cold Dulbecco’s balanced salt solution of RNA sample were mixed and heated a t 70°C for 10 and scraped into 1ml of solution in a n Eppendorftube. min. A master mix containing per sample 4 p1 of 5 x RT They were centrifuged to form a pellet (5 minutes; buffer, 2 pl dithiothreitol and 2 p1 of 0.01M deoxyribo10,000 rpm in a microfuge a t 4“C), the supernate was nucleotides was added. The enzyme reverse tranremoved and the pellet stored at -70°C until required. scriptase (MMLV RT; Superscript, Gibco BRL Ltd) was In this way control samples were always available and added at -20°C allowing 1 p1 per sample. The enzyme ready to use when NPS samples were tested. was allowed to act at 37°C in a water bath for 60 min, followed by chilling on ice. Reverse Transcription-Polymerase Chain Amplification. A master mix of 5 pl 1 0 PCR ~ Reaction (RT-PCR) buffer, 25 pl DEPC-DW and 0.4 pl DNA polymerase Primer sequences. The primers used were kindly from Thermus thermophilus (Tth, 5 u n i t d m l ; Hybaid provided by Dr. Pat Cane, Department of Biological Sci- Ltd. Teddington) per sample was made up and added to ences, University of Warwick, and had the following each RT product. Fifty microliters of paraffin oil were sequences [Cane and Pringle, 19911: added and the samples subjected to 30 cycles of denaturation, annealing, and extension (93°C for 1.5 min, 1938 55°C for 1.5 min, and 72°C for 1.5min) followed by one primer 1-5‘ GGA ACA AGT TGT TGA GGT TTA cycle of 93°C for 1.5 min, 55°C for 1.5 min, and 72°C for TGA ATA TGC 5 min in a Hybaid thermal cycler. 2215 A second round of amplification was carried out by primer 2-5’ CTT CTG CTG TCA AGT CTA GTA adding 10 pl of first round product to 1p1 primer pair 3 CAC TGT AGT and 4. This was heated a t 70°C for 10 min and cooled on These sequences come from part of the sequence of the ice and a mix of Tth in buffer added as before. The ‘N’ gene of the ‘A’ subtype of RS virus [Collins et al., samples were subjected to 30 cycles as before except 19851 and amplify a segment of 278 bases. Because of that a n annealing temperature of 45°C was used. the high degree of sequence homology of the ‘ N gene Agarose gel electrophoresis. Twelve microliters of (96% between subtypes ‘A’ and ‘B’ [Johnson and Col- amplified product were mixed with 2 pl of loading dye lins, 198911,these primers should amplify nearly all RS (bromophenol blue [Sambrook e t al., 19891) and added virus strains likely to be encountered. to a single slot in a slab gel consisting of 3% NuSieve Nested inner primers were synthesised by Oswel GTG agarose and 1%SeaKem GTG agarose (ICN Flow; DNA Service, Department of Chemistry, University of High Wycombe) and containing 5 pl ethidium bromide Edinburgh and had the following sequences: (10 mgimL), electrophoresed at a constant current of 40 mA for 1h and examined under a UV light source for 1983 the presence of bands of DNA of appropriate size. DNA primer 3-5’ TGA AGC AGG ATT CTA CCA TAT markers (1 kb DNA ladder and @X 174 RF DNAiHae 2160 I11 fragments; Gibco BRL, Life Technologies Ltd., Paisprimer 4-5‘ TTC AGC ATA TGC CTT TGC TGC ley) were added to each gel. RT-PCR RESULTS RT-PCR was carried out on stored cellular suspenExtraction. NPS cell suspensions in Eppendorf tubes were centrifuged a t 4°C to pellet the cells and the sions from routine first specimens submitted to the labsupernate removed. One milliliter of RNAzol B (Bio- oratory or from sequential samples obtained from 15 genesis Ltd., Bournemouth) was added to the pellet and children following confirmation of RS virus induced the tube shaken vigorously for 15 sec. One hundred bronchiolitis. One hundred twenty-three routine specimicroliters of chloroform were added, the mixture mens stored at -70°C in the presence of RNAse inhibishaken gently and kept on ice for 5 min then centri- tor for periods ranging from a few days to several weeks fuged at l2,OOOg for 15 rnin at 4°C. The aqueous top were tested and the results are shown in Table I. While layer was transferred to a sterile Eppendorf tube and 45 (68% of all positives) contained detectable RS viral a n equal volume of isopropanol added. The mixture was sequences by RT-PCR after storage, 49 (74%)had been
to RS virus (“Imagen” RS kit; Dako Diagnostics, Ltd., High Wycombe) as previously described [Cubie et al., 19911.
Detection of RS Virus in Infants
285
TABLE I. Comparison of RT-PCR, Antigen Detection, and Virus Isolation Using Stored Cell Suspensions From Nasopharyngeal Secretions”
RS FA RS FA RS I RS I
+ + -
RT-PCR+
RT-PCR-
Total
45 32 13 26 19
78 17 61 22 56
123 49 74 48 75
“RT-PCR = reverse transcription and polymerase chain reaction; RS FA = detection of antigen using fluorescein labelled antibody to RS virus; RS I = isolation of virus in cell culture.
positive by fluorescent antigen (FA) detection, and only 32 were positive by both methods. Similarly, 48 (73%) were positive by RS virus isolation but only 26 agreed with the RT-PCR result. Thus the overall detection rate was similar for all three methods, but both antigen detection and virus culture were carried out immediately on receipt of the specimen, while RT-PCR was carried out after storage. This may have reduced the apparent sensitivity of the RT-PCR technique and these results are discussed. A gel containing specimens amplified by RT-PCR is shown in Figure 1. PCR products are apparent for three specimens which were also positive by antigen detection and virus isolation (lanes 2,4, and 5) and a further two products from specimens from which no virus was isolated (lanes 1 and 3). No bands are apparent in two other NPS one of which appeared positive by antigen detection. An RS virus cell culture and uninoculated HEp-2 cells are included as positive and negative controls. During the period of the study, seven NPS yielded other viruses in culture and none of these gave a positive band in the RT-PCR. Ninety-seven of the 123 specimens came from 15 infants and were collected between 1 and 30 days after onset of symptoms. Comparison of the three detection methods in these specimens is shown in Figure 2. Virus detection was high in the first 6 days with 85% of specimens giving a positive result by a t least one of the methods, (12114 in the first 3 days, 19/23 during days 4-6). Antigen detection was the single most sensitive test at this stage. During the second week, more than 60% still gave positive results but more positives were detected by RT-PCR (21/40) than by antigen detection (17140) or culture (14140).Of the 15 infants from whom sequential samples were obtained, 5 were treated with tribavirin by small particle aerosol generator (SPAG) for 3 days, ranging from days 3 to 5 of symptoms to days 8-10 of symptoms. The duration of excretion of viable virus was comparable in both treated and untreated groups (approximately 8 days post onset). Antigen or nucleic acid was detected for slightly longer, particularly in the treated group (mean of 17.4 days post onset compared with mean of 13.9 days), but statistical analysis of such small groups was not possible. Because the sensitivity of the RT-PCR was lower than expected, changes in the preparation and storage of samples for RT-PCR were made in the current RS virus season.
Fig. 1. Agarose gel of RT-PCR products stained with ethidium bromide and photographed under UV light (abbreviations as in Table I). Lanes 1-7: Products from seven specimens of NPS. Lanes 6-9 HEp-2 cell extracts from RS virus infected and uninoculated cells, respectively. Lane 1 0 1 kb DNA ladder.
After routine cytospin preparation for antigen detection, the remaining cell suspension was centrifuged, the supernate removed and the cell pellet immediately stored at -70°C without the addition of RNAse inhibitor. Ninety four specimens were examined following this method and the results are presented in Table 11. Of 62 specimens positive by any one method, 57 (92%) were positive by FA detection, 52 (84%)by RT-PCR and 44 (71%)by virus isolation. Much closer agreement was obtained in this series with 50/52 (96%)of the RT-PCR positives being confirmed by another method or coming from patients from whom previous positive specimens had been obtained. A similar percentage of isolation positive results (95%;42144) were confirmed by another method, but only 49/57 (86%) of antigen positive results were confirmed. The sensitivity and specificity of these results is discussed. Details of the onset of symptoms were not available for all patients in this group, but from the information received, antigen and isolation positive results were obtained up to 9 days and RT-PCR positive results up to 12 days after the onset of symptoms. It was possible to re-amplify using inner primers 3 and 4, twenty-two specimens (15 of which are included in the analysis in Table 11) which had been negative by RT-PCR but positive by a t least one other method. Ten of these gave bands on agarose gels and migrated as
Cubie et al.
286
30
30
....... ....... ....... ....... ....... ....... ....... .......
1
25
25
20
20
15
15 10
10
5
0
~
1-3
4 - 6
7-9
10-14
15-21
...... ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... -.... ...... ...... ...... ...... ......
5
0 22->28
Days after onset of symptoms Fig. 2. Total number of positive RS virus results and number of specimens positive by each method a t various times after onset of symptoms (abbreviations as in Table I; n = No. of specimens).
TABLE 11. Comparison of RT-PCR, Antigen Detection, and Virus Isolation Using Stored Pelleted Cells From Nasopharvngeal Secretions”
RS FA RS FA RS I RS 1
RT-PCR+
RT-PCR
Total
52 43 9 33
42 14 28 11 31
94 57 37 44 50
+ +
19
~
~
‘Abbreviations as in Table I
expected for a product of 177 base pairs in length. The sensitivity and specificity of the different tests used are discussed below.
DISCUSSION The results demonstrate several difficulties in translating experimental research protocols into a diagnostic procedure. With a peak incidence of 26 NPS per day and repeat specimens from increasing numbers of infants initially entered into the study, the logistics of including RT-PCR into the daily routine of a diagnostic laboratory handling large numbers of relevant specimens was considerable and necessitated storage of the samples a t -70°C. However, storage of the extracted cells a s a suspension in the presence of RNAse inhibitor
I Bruce e t al., 19891 proved unsatisfactory. Agreement between antigen detection, isolation and RT-PCR results in the 123 specimens stored in this way was poor (Table I) and the RT-PCR results suggested a lower than expected sensitivity for this test. The problems of working with RNA are well appreciated and clinical specimens are rarely taken, sent, or stored under ideal conditions. Nevertheless, Zhang and Evans [ 19911 report a n RT-PCR method for influenza viruses made sensitive enough to cope with moderately degraded samples even without inhibition of ribonucleases. Despite this reduced sensitivity, the RT-PCR detected RS virus sequences in sequential specimens from infants with bronchiolitis and more positives were detected by this method in the second week after onset of symptoms (Fig. 2) than by antigen detection or by virus isolation. Recent experiments with cell pellets stored a t - 70°C without RNAse inhibitor gave more satisfactory results and were simpler to handle (Table 11). The RNAzol solution could be added directly to the frozen pellet at the start of the extraction procedure. In future studies of virus shedding, it is surmised that extraction of RNA directly from pelleted cells without - 70°C storage would give even more satisfactory results with greater and prolonged sensitivity. Extraction, reverse transcription, and amplification could be carried out within the working day and gel electrophoresis the following day.
Detection of RS Virus in Infants Increased sensitivity can be obtained by the use of nested primers in a second round of amplification and this was demonstrated in a group of 22 specimens (of which 15 are included in the analysis in Table 11)which appeared to be negative on the first round of amplification despite detection of antigen or virus. Ten were detected in gels using inner primers 3 and 4.Thus of the group of 62 specimens positive overall (Table 111, 54 (87%)were positive by RT-PCR including nested amplifications and 52 (96%)were confirmable by other methods. This compares with 57 (92%) positive by antigen detection of which on 51 (89%) were confirmable, and suggests a greater sensitivity for the antigen detection system. Diagnostic laboratories should be aware, however, of occasional false positive results with the monoclonal antibody. A second round of amplification adds considerably to the cost and time of the test, but nested primers are considered to provide a significant increase in specificity a s well as sensitivity [Clewley, 19891 and to reduce the risks of cross-contamination [Zhang and Evans, 19911. Furthermore, carefully chosen nested primers can permit the separate identification of subgroups ‘A’ and ‘B’ of RS virus [Cane and Pringle, 19911 and this extra stage could be very useful in epidemiologi.ca1 studies. It was hoped to study the difference in virus shedding in treated and untreated patients. However, although 15 infants were treated with tribavirin in the Royal Hospital for Sick Children in Edinburgh in 1989-90, only five required treatment during the 1990-91 season, the year of the study outlined above. From the limited data available on these infants and on 10 comparably aged controls, excretion of virus occurred for similar lengths of time, antigen and nucleic acid detection were only positive marginally longer in the treated group and the duration of symptoms was similar. However, the treated group had underlying disease including pulmonary dysplasia and congenital heart defects and were considerably more ill than the untreated group. It may be, therefore, that tribavirin had indeed reduced the markers of RS infection. While it would be unethical not to treat severely ill patients to show a difference in virus shedding, it would be worthwhile to compare virus shedding and detection of nucleic acid in treated and untreated hospitalised infants without underlying disease. Such a study is under consideration. While there is undoubtedly a place for molecular techniques such a s RT-PCR in the study of respiratory infections in children, we would reiterate our comment
287
[Cubie e t al., 19911 that antigen detection using commercially available fluorescein-labelled monoclonal antibodies to RS virus remains the most cost effective and rapid technique currently available for the diagnosis of RS virus induced lower respiratory tract infection.
ACKNOWLEDGMENTS We would particularly like to thank Dr. Pat Cane of the Department of Biological Sciences, University of Warwick for her helpful discussions concerning the RTPCR protocol and the kind provision of primers 1 and 2. We would also like to thank Margaret Devine of the Respiratory Laboratory, Royal Hospital for Sick Children, for coordinating the specimen collection and the parents of very small babies who willingly returned for a n increased number of out-patient visits. We are grateful to the Scottish Hospital Endowments Research Trust for funding this project and our thanks are due to Mrs. Avril Garvey for typing the manuscript. REFERENCES Bruce CB, Al-Nakib W, Almond JW, Tyrell DAJ (1989): Use of synthetic oligonucleotideprobes to detect rhinovirus RNA. Archives of Virology 105:179-187. Cane PA, Pringle CR (1991): Respiratory syncytial virus heterogeneity during an epidemic: Analysis by limited nucleotide sequencing (SH gene) and restriction mapping (N gene) Journal of General Virology 72:349-357. Clewley J P (1989): The polymerase chain reaction, a review of the practical limitations for human immunodeficiency virus diagnosis. Journal of Virological Methods 25:179-188. Collins PL, Anderson K, Langer SJ, Wertz GW (1985): Correct sequence for the major nucleocapsid protein mRNA of respiratory syncytial virus. Virology 146:69-77. Cubie HA, Winter GF, Leslie EE, Inglis JM (1990):Rapid detection of respiratory syncytial virus antigens in nasopharyngeal secretions. Journal of Virological Methods 27:121-124. Cubie HA, Inglis JM, McGowan A (1991): Detection of respiratory syncytial virus antigen and nucleic acid in clinical specimens using synthetic oligonucleotides. Journal of Virological Methods 34:27-35. Englund JA, Anderson LJ, Rhame FS (1991): Nosocomial transmission of RS virus in immunocompromised adults. Journal of Clinical Microbiology 29:115-119. 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 254:30473051. Johnson PR, Collins PL (1989):The 1B (NSZ), 1C (NS1)and N proteins of human respiratory syncytial virus (RSV)of antigenic subgroups A + B: Sequence conservation and divergence within RSV genomic RNA. Journal of General Virology 70:1539-1547. Sambrook J , Fritsch EF, Maniatis T (1989): Molecular Cloning: A Laboratory Manual. 2nd Edition. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press, Zhang W, Evans DH (1991): Detection and identification of human influenza viruses by polymerase chain reaction. Journal of Virological Methods 33:165-189.
