INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Influenza viruses have been shown to be quite mutable and adaptable (17, 18). High mutation rates that result in sequence changes in the two major surface antigens, hemagglutinin (HA) and neuraminidase (NA), are responsible for the frequent influenza epidemics in the human population (11). An investigational live influenza virus vaccine (FluMist) which contains three live, cold-adapted H1N1, H3N2, and B influenza viruses is being developed for influenza prevention (1). These vaccine viruses are 6/2 reassortants, in which the HA and NA genes are derived from the circulating wild-type viruses and the remaining six genes (PB2, PB1, PA, NP, M, and NS) are derived from the cold-adapted master donor strains, A/Ann Arbor/6/60 (CA-A) and B/Ann Arbor/1/66 (CA-B). The cold-adapted master donor strains were derived by H. F. Maassab at the University of Michigan by passaging wild-type viruses in primary chick kidney cells through progressively lower temperatures, resulting in viruses that replicated efficiently at 25°C (9, 10). The cold-adapted master donor strains exhibit several attenuation characteristics in cell culture and in a ferret animal model (9, 10). A global surveillance system governed by the Centers for Disease Control and Prevention and the World Health Organization was established to monitor the drift of the antigenic structures of emerging influenza viruses. The accrued information provides public health authorities with the basis of selecting the HA and NA genes from the appropriate wild-type viruses for incorporation into the influenza virus vaccine (3, 4). Therefore, the six genes from the cold-adapted master donor strains ensure the attenuation and the HA and NA genes from the wild-type viruses confer the protective immunity against contemporary influenza strains. Since genetic stability supports and maintains the attenuation and immunogenicity, it is important to monitor the conservation of the 6/2 genotype of FluMist after vaccination.

During the 1996 to 1997 influenza season, the efficacy of FluMist was studied in a multicenter, double-blind, and placebo-controlled clinical trial enrolling 1,602 children from 15 to 71 months of age. A 93% overall vaccine efficacy against culture-confirmed influenza was observed (1). Some nasal and throat swab specimens were collected during the 2-week postvaccination period. This report describes the detection, genotyping, and sequence analysis of the culture-positive samples from these specimens.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Clinical samples. As listed in Table 1, samples from 17 study participants were analyzed in this report. These samples represent all available nasal and throat swab specimens collected within a 2-week period after vaccination, for which the local clinical laboratory reported the isolation of influenza virus. The samples were obtained primarily from 1 of the 10 investigative sites based on clinical symptoms. No wild-type influenza viruses were circulating in the community at this time. The swab specimens were inoculated into rhesus monkey kidney (RhMK) cells at the clinical trial site for viral isolation. Both the culture supernatants and the remaining swab specimens were sent to Aviron for further laboratory analysis. For participant 6336, two samples corresponding to two different collection dates were provided.

Viruses. Each vaccine virus was a 6/2 reassortant derived from the two parental viruses, a cold-adapted master donor strain and a circulating wild-type virus. Parental viruses for the vaccine used in this clinical trial were CA-A and CA-B, the two cold-adapted master donor strains, and A/Texas/36/91 (H1N1) (A/TX), A/Wuhan/359/95 (H3N2) (A/WH), and B/Ann Arbor/1/94 (B/AA), the three wild-type viruses. Virus stocks were prepared in specific-pathogen-free eggs (SPAFAS, Inc., Preston, Conn.) as previously described (6).

Viral RNA genome isolation. A 200-µl aliquot of each culture supernatant and viral stock was extracted to isolate the virus genomic RNA. RNA was isolated by incubating the samples with proteinase K (0.7 µg/µl) and sodium dodecyl sulfate (0.5%) at 56°C for 10 min, followed by a phenol-chloroform extraction. The extracted RNA was resuspended in 100 µl of H₂O. The swab specimens were of limited volume, ranging from 10 to 220 µl; RNA was resuspended in 20 µl of H₂O after extraction.

Oligonucleotide synthesis. Oligonucleotides were synthesized with an ABI 392 oligonucleotide synthesizer (PE Applied Biosystems, Foster City, Calif.). Three fluorescent dyes (6-FAM [blue], HEX [green], and NED [yellow]) were used to label oligonucleotides at the 5' terminus. The fluorescent dye-labeled oligonucleotides were either synthesized at Aviron or purchased from PE Applied Biosystems.

Multiplex RT-PCR detection. It is important to implement rigorous contamination control in clinical diagnostic applications using PCR. We have outfitted a PCR laboratory comprised of three segregated parts: a reagent room, a template preparation room, and an amplification and analysis room. In addition to this physical segregation, standard operation procedures are established and the PCR process is conducted in a directional and irreversible flow from the reagent room to the template preparation room and then to the amplification room. Both positive and negative controls were included along with the samples for every reaction. Six pairs of PCR primers were designed to achieve specific amplification of the HA and NA genes of the H1N1, H3N2, and B influenza viruses. The primers were designed in such a way that all six PCR products from all three viruses could be generated in a single reverse transcriptase (RT)-PCR. A 10-µl aliquot of extracted RNA was used in an RT-PCR that was conducted with the same procedure as for each primer pair, following that of the Perkin-Elmer GeneAmp RNA PCR Kit with AmpliTaq Gold polymerase (PE Applied Biosystems). PCR conditions were 12 min at 95°C, 5 cycles of 1 min at 95°C, 1 min at 55°C, and 1 min at 72°C, followed by 30 cycles of 1 min at 95°C, 1 min at 60°C, and 1 min at 72°C. The multiple products resulting from a single reaction mixture were easily distinguished because they were different lengths and they were labeled with different fluorescent dyes. Gel electrophoresis was run on an ABI 377 automated sequencer (PE Applied Biosystems). Table 2 lists the primer sequences and their positions in the virus genomes.

RT-PCR and RFLP genotyping. As described previously (13), a 10-µl aliquot of extracted RNA was used in an RT-PCR to generate product from each individual gene for subsequent restriction digestion, followed by restriction fragment length polymorphism (RFLP) analysis. The RT-PCR was conducted by following the Qiagen Taq PCR handbook (Qiagen, Inc., Valencia, Calif.). PCR conditions were 1 min at 94°C, 2 min at 55°C, and 3 min at 72°C, which was repeated for 30 cycles. A 1% agarose gel electrophoresis was used to analyze the digested and nondigested products. The HA gene of influenza A was typed by the generation of RT-PCR products with subtype-specific primers. Table 3 lists the primers used in the RFLP analysis.

Multiplex RT-PCR and F-SSCP genotyping. The fluorescent single-strand conformation polymorphism (F-SSCP) gel electrophoresis was performed as previously described (2), except for the following modifications: (i) an ABI 377 automated sequencer (PE Applied Biosystems) with an external refrigerated water circulator (Neslab Instruments, Inc., Portsmouth, N.H.) was used, (ii) the gel was run at a 20°C setting for 6 h, and (iii) the GENESCAN-500 ROX (PE Applied Biosystems) was used as an internal standard. The GENESCAN-500 ROX was a red-labeled DNA marker and was added to each sample. The marker bands in each sample could be used to normalize the lane-to-lane variation (2). Data were collected and analyzed by GENESCAN software (PE Applied Biosystems).

Nucleic acid sequencing and analysis. A 10-µl aliquot of extracted RNA was used in a multiplex RT-PCR to generate multiple gene products for sequencing. The RT-PCR products were sequenced with an ABI PRISM dRhodamine Terminator Cycle Sequencing Ready Reaction kit and run on either an ABI 373 stretch or ABI 377 automated sequencer (PE Applied Biosystems). Sequence analysis was performed with MacDNASIS (Hitachi Software, San Bruno, Calif.).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Multiplex RT-PCR detection. The nasal and throat swab specimens were cultured to isolate viruses, and the presence of influenza A or B virus was reported by the local clinical laboratory. The culture supernatants and the remaining swab specimens were sent to Aviron. Further subtyping of influenza A viruses using HA type specific antiserum and phenotypic analysis of cold adaptation (ca) and temperature sensitivity (ts) were performed on the cultured viruses by the Aviron clinical testing laboratory (ACTL). These results are summarized in Table 1 (S. Pennathur, personal communication). For participant 6599, the original culture supernatant was consumed for the subtyping and phenotyping analysis. An attempt was made at ACTL to reculture the virus with the remaining swab specimen, but no influenza virus was recovered. The amount of viable influenza virus contained in the swab specimen was presumed to be inadequate for culturing due to the long storage and freezing-thawing procedures.

View larger version (28K): [in this window] [in a new window]   FIG. 1.   Gel file of fragment analysis of multiplex RT-PCR products of HA and NA genes. Blue bands denote H3 and N2, yellow bands denote H1 and N1, and green bands denote HA and NA of influenza B. H2 was labeled with blue dye. In each pair of bands with the same color, the slower moving band was HA and the faster moving band was NA. The size markers are shown in red. The product sizes are listed in Table 2.

RT-PCR and RFLP genotyping. As shown in Fig. 1, there were seven H3N2 and 13 B isolates, and one sample (participant 6599) contained no detectable virus. RT-PCR and RFLP genotyping was performed separately on the influenza A and influenza B viruses. The participant 6599 sample was included in the influenza B group because it was previously reported to contain influenza B virus by the local clinical laboratory.

View larger version (164K): [in this window] [in a new window]   FIG. 2.   A 1% agarose gel showing the RFLP analysis of the M gene of influenza B virus. A nondigested set of products was loaded on the left (NO RE), and the digested corresponding products were loaded on the right (PvuII). Solid triangles highlight the digested patterns of samples 6372 and 6336B.

Multiplex RT-PCR and F-SSCP genotyping. The 10 influenza B virus-containing swab specimens identified by the multiplex RT-PCR detection (Table 1) were multiplexed with HA, M, and NA genes and subjected to F-SSCP genotyping as shown in Fig. 3. The results demonstrated that all of the M genes were derived from CA-B and all of the HA genes were derived from B/AA. All of the NA genes, except for one of the samples from participant 6336 (Fig. 3, sample 6336B), were shown to be derived from B/AA. The mobility exhibited in an F-SSCP genotyping is sequence dependent, and a single nucleotide variation can result in a definite mobility shift (2, 12). By sequence analysis (see below), the NA gene in sample 6336B was derived from B/AA, but it contained one point mutation. Unlike RFLP genotyping, F-SSCP genotyping demonstrated that the M genes in samples 6372 and 6336B were derived from CA-B. Further sequence analysis confirmed this conclusion (see below).

View larger version (37K): [in this window] [in a new window]   FIG. 3.   Gel file of F-SSCP analysis of multiplex RT-PCR products of HA, M, and NA genes of influenza B virus. The RT-PCR products generated in separate reactions for individual genes of the two parental viruses are shown in the first six lanes. The slower moving green bands denote the HA gene, the blue bands denote the M gene, and the faster moving green bands denote the NA gene. Since only one strand was labeled, double bands associated with each gene most likely represented two conformations exhibited by a single molecule. White dots highlight the NA gene of sample 6336B, which exhibited different mobility from those of both CA-B and B/AA.

Sequence analysis. Various degrees of sequence heterogeneity exist among influenza strains. Directly comparing the nucleic acid sequences between vaccine and parent viruses can provide the absolute identity of the genotype. To efficiently generate DNA templates for sequencing, multiplex RT-PCR was designed to generate short DNA fragments (~300 bp) of multiple genes in a single reaction. The RNA extracted from the culture supernatants was used in the multiplex RT-PCR for sequencing. Using the nonlabeled set of primers listed in Table 2, the NP, HA, and NA genes of six influenza A isolates and the M, HA, and NA genes of 11 influenza B isolates were amplified. The corresponding gene fragments of the parental viruses were also similarity amplified for sequencing. Samples from participant 6148 of the influenza A group and participants 6126 and 6584 of the influenza B group were not sequenced because the culture supernatants of these samples were consumed during the earlier experiments. No swab specimens remained for sequencing.

View larger version (132K): [in this window] [in a new window]   FIG. 4.   An alignment of four electropherograms showing the PvuII restriction site. Green, blue, black, and red colors denote A, C, G, and T nucleotides, respectively. The underline marks the position of the restriction site. The sequence enclosed in the parentheses contained nucleotide changes from the PvuII recognition sequence.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The FluMist vaccine appeared to be genotypically stable after replication in the human host. As summarized in Table 1, all viruses that had the complete genotyping analysis maintained the 6/2 genotype. They also maintained the ca and ts phenotypes (Table 1), suggesting that they were shed vaccine viruses. All available culture-positive samples, as determined by the local clinical laboratory, that were obtained during the 2-week postvaccination period were studied in this report. As seen in Table 1, no circulating wild-type virus was detected, no H1N1 vaccine virus was detected, and no distinct shedding pattern or duration was observed for the H3N2 and B vaccine viruses. A single nucleotide mutation did occur in each of the M and NA genes of the influenza B virus from participant 6336. The M gene of the virus from participant 6372 contained vaccine virus with and without a point mutation. This approach to postvaccination monitoring provides a comprehensive strategy for direct virus detection, subtyping, analysis of virus shedding patterns, genotyping, and assessment of sequence variations.

It was not surprising that we did not detect wild-type viruses or reassortants between vaccine and wild-type viruses because no wild-type viruses were yet circulating in the community during the period that clinical specimens were collected. In another study, we did detect and genotype wild-type viruses from specimens collected when wild-type influenza virus was circulating in the community (data not shown). In this report, over 11,800 nucleotides representing small regions of the NP, HA, and NA genes of six influenza A isolates and the M, HA, and NA genes of 11 influenza B isolates were deduced. Three nucleotide changes were found. One nucleotide change in the NA gene that did not alter the codon was found in isolate 6336B. One nucleotide change that would direct a Ser-to-Leu codon change in the M gene was also found in isolate 6336B. The M gene of isolate 6327 contained a mixture with and without the same mutation observed in isolate 6336B. We are currently planning full genome sequencing on large numbers of viruses isolated in a prospective study of shedding and transmission. Previously, single gene studies (14) provided evidence that PB2, PB1, PA, and M genes were required for the attenuation of the cold-adapted donor strain of influenza A virus. However, information about the specific nucleotide changes attributable to attenuation was limited (7) and confounded by extragenic suppression and gene constellation effects (15, 16). Therefore, comprehensive full-genome sequencing is necessary initially.

From the technical point of view, several issues should be noted when analyzing clinical samples. RhMK cells were used to culture the virus from the nasal and throat swab specimens, and the culture supernatants were subsequently analyzed. The RhMK cell culture step might reflect the replication potential of the vaccine viruses in the swab specimens. Although no H1N1 shedding was detected in culture supernatants or swab specimens in this study, more extensive analysis is required to determine the shedding pattern, duration, and quantity of the vaccine viruses. RNA virus populations are composed of quasispecies, and the landscape of quasispecies is fluid to environmental changes (8). Additional viral replication in cell culture might cause new mutations or reflect a different landscape of the quasispecies. In this study, sequencing was performed with the culture supernatants. The mutation observed in the NA gene of the participant 6336B sample appeared to be present in the swab specimen as well, as evidenced by the mobility shift observed in the F-SSCP genotyping (Fig. 3). It is not known whether the mutation observed in the M gene of the participant 6336B sample occurred after replication in cell culture. Nevertheless, sequences of quasispecies were observed in the participant 6372 sample, as evidenced by observing mixed M gene sequences, with and without the point mutation (Fig. 4). A direct analysis of the swab specimen without the cell culture step should offer a much better assessment of the sequence variations of the vaccine viruses after replication in the human host.

Two important keys for a successful RT-PCR and RFLP genotyping are primer design and restriction enzyme selection. A good primer design ensures that the amplification of gene fragments is specific and efficient. The selection of more than one restriction enzyme for the analysis of each gene allows the demonstration of more than one distinct restriction pattern for the gene by RFLP genotyping. Using RFLP, any nucleotide mutation that occurs at the restriction site alters the genotype identification. Nucleotide mutations that occur outside the restriction site cannot be assessed. As shown in Fig. 2, RFLP data suggested a possible mixture of wild-type and vaccine viruses or a reversion of the M gene of the vaccine virus to the wild type in two samples. In fact, sequence analysis proved that both of these viruses were vaccine strains.

We have previously reported the utility of a multiplex RT-PCR and F-SSCP genotyping method for analyzing influenza virus reassortants (2). The result in Fig. 3 demonstrates that this method is also applicable to clinical specimens. The multiplex RT-PCR and F-SSCP genotyping offers a different perspective. F-SSCP genotyping is based on the principle that, under nondenaturing conditions, single-stranded DNA assumes unique conformations that are sequence dependent (2, 12). With this method, a single nucleotide mutation that occurs within the PCR fragment causes the fragment to exhibit a different mobility (2). As seen in Fig. 3, the mobility of the NA gene fragment of sample 6336B was different from that of both CA-B and B/AA. Sequence analysis revealed that it was B/AA but contained one point mutation. Genotyping by multiplex RT-PCR and F-SSCP is efficient and cost-effective, with improvements made since our earlier report (2). In our experiences, the newer ABI 377 sequencer increases throughput. With the attachment of an external cooler, gel electrophoresis run time has been reduced from 26 to 6 h, and the single-stranded conformation is better maintained at the lower temperature during runs.

The sample sizes of clinical specimens are often small, and they often contain small amounts of virus. Throughout this study, the multiplexing scheme was designed for direct detection, genotyping, and sequencing. In this way, the information that can be obtained from the irreplaceable clinical specimens is maximized.