Genetic characterization of influenza A(H3N2) viruses from 2014 to 2017 in Yantai, east of China

Specimen collection

Respiratory samples were collected from April 2014 to March 2017 among patients with an influenza-like illness (ILI) (sudden onset of fever ≥38 °C combined with respiratory symptoms such as cough, sore throat and absence of other diagnoses) who sought medical attention at two sentinel surveillance hospitals in Yantai. The mean number of outpatients in the two sentinel surveillance hospitals is about 300 per day. The Yantai area is located in eastern China, covering an area of more than 14,000 km², with a population of about 7.06 million population. A total of 5183 swabs were collected with the informed consent of all patients or guardians of children. These swabs were placed in virus transport media tubes immediately and sent to the Yantai Center for Disease Control and Prevention within 24 h.

Ethical statement

All procedures performed in studies involving human participants were in accordance with the ethical standards of the Institutional and/or National Research Committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Informed consent was obtained from all individual participants included in the study.

Virus isolation and subtype identification

These clinical specimens were isolated in Madin-Darby canine kidney (MDCK) cells, and the presence of virus in the culture was confirmed by hemagglutination assay. The subtype of influenza, which was isolated with a minimum of eight hemagglutination units, was determined by hemagglutination inhibition (HAI) assay with the standard antigen and antiserum from the Chinese Nation Influenza Center, according to the surveillance protocol published by the Ministry of Health (MoH) of China [8]. HAI assays were performed in V-bottom 96-well microtiter plates with the standard antigen controls, 1% human type O erythrocytes controls and negative controls using 1% suspension of human type O erythrocytes. The standard serum was tested with two-fold serial dilutions from 1:10 to 1:1280. The HAI titer was the reciprocal of the highest dilution of the antiserum that cause complete inhibition of hemagglutination. A virus isolate was recognized as a particular subtype when it reacted with one standard antiserum at a four-fold or greater HAI titer than when it reacted with other standard antisera. A total of 45 cultures confirmed as positive for influenza A/H3N2 were selected based on their distribution over time (14 in the 2014/2015 season, 10 in the 2015/2016 season, 21 in the 2016/2017 season) for the HA and NA gene analysis.

Nucleic acid extraction and sequencing

Viral RNA was extracted using the RNeasy Mini Kit (Qiagen 52904, Germany) according to the manufacturer’s instructions, and the HA and NA sequences were subsequently amplified by reverse transcription polymerase chain reaction (RT-PCR) using a PrimeScript One Step RT-PCR Kit (Qiagen210212, Germany). The primers and cycling conditions were used for amplification as described before [9]. RT-PCR products were sequenced in both directions using the Sanger method. Sequences were submitted to the GenBank database with the accession numbers MF593700-MF593789.

Phylogenetic analysis

Sequences from the vaccine and reference strains (the accession numbers in Table 1) were retrieved from the Global Initiative on Sharing All Influenza Data (GISAID) databases for phylogenetic analyses. The phylogenetic trees were generated with the Neighbor-Joining method, using the MEGA 7.0.26 software package (Philadelphia, PA, USA). The Kimura 2-parameter model was used and the reliability of the trees was provided by bootstrap analysis of 1000 replications. The potential N-linked glycosylation sites were predicted using the NetNGlyc 1.0 server, with a threshold value of >0.5.

Estimation of vaccine efficacy using the pepitope model

The potential VE of influenza A/H3N2 was evaluated using the pepitope model, which measures the antigenic distance between the vaccine strains and circulating strains and has a better correlation with vaccine VE than antisera HAI assay or phylogenetic analyses [10]. Pepitope is calculated by the fraction of substituted amino acid in the dominant epitope of HA [11]. The association between VE efficacy for influenza A(H3N2) and pepitope is given by VE=−2.47×pepitope+0.47 [12], [13], [14]. If pepitope=0, VE will be 47% as a perfect match between vaccine and the circulating strains [10].

The activity of influenza viruses

From April 2014 to March 2017, 5183 samples from ILI patients were tested for influenza by virus isolation. Based on the HAI assay results, 8.95% (464/5183) were positive for influenza viruses: 36% (n=167) were typed as influenza B, while 64% (n=297) were influenza A. The confirmed influenza A viruses comprised 114 (38%) A(H1N1) pdm09 and 183 (62%) A(H3N2). In each surveillance year, there were at least two subtypes co-circulating in the Yantai area, and a seasonal pattern of influenza activity was observed. Influenza A(H3N2) appeared in August 2014, lasted about 5 months and peaked in November 2014. From December 2014, the A(H1N1)pdm09 influenza virus was observed, persisted until April 2015 and replaced the influenza A(H3N2) gradually. The influenza A(H3N2) viruses circulated at low level in 2015/2016 influenza season, and became the dominant subtype again until November 2016. The prevalence of A(H3N2) persisted until March 2017, and was replaced by A(H1N1) pdm09 (Figure 1).

Figure 1:

Month-wise distribution of influenza subtypes from April 2014 to Mar 2017 seasons.

Hemagglutination inhibition (HAI) assay results

During the study period, a total of 183 isolates were identified as influenza A(H3N2) through HAI assay. One hundred and twenty one isolates, 14 isolates and 48 isolates were identified in the 2014/2015 season, 2015/2016 season and 2016/2017 season, respectively. In the 2014/2015 and 2015/2016 seasons most of the viruses were recognized (115/121, 95.04% in 2014/2015 and 12/14, 85.71% in 2015/2016) by the corresponding standard antisera at titers four-fold and ≥eight-fold lower than the homologous titer. However, in the 2016/2017 season, almost all of the isolates (42/48, 87.5%) from Yantai area were recognized well by the standard antisera at titers ≤two-fold reduced compared to the homologous titers of the antisera (Table 2).

Phylogenetic analysis

The HA phylogenetic analysis was conducted on 45 studied isolates along with 15 HA genes of the northern hemisphere vaccine and reference strains, and the NA phylogenetic analysis was performed on the NA gene of the 45 studied isolates and three northern hemisphere vaccine strains.

In comparison with A/Texas/50/2012 (vaccine strain for 2014/2015), the average nucleotide and amino acid similarities on the HA of the 2014/2015 season were estimated to be 98.62% and 97.36%. While the similarities of nucleotide and amino acid between the A(H3N2) strains from the 2015/2016 seasons and A/Switzerland/9715293/2013 (the vaccine for 2015/2016) were 98.3% and 97.4%, respectively. When compared to A/Hong Kong/4801/2014 (the vaccine strain for 2016/2017), the average HA nucleotide and amino acid identities among the 2016/2017 strains were 99.2% and 98.9%, respectively.

For NA of the influenza A(H3N2), the average similarities of nucleotide and amino acid between strains isolated from Yantai city and the given vaccine strains were 99.27% and 99.56% in the 2014/2015 season, 98.77% and 98.72% in the 2015/2016 season, and 98.47% and 97.92% in the 2016/2017 season, respectively.

The HA phylogenetic tree showed that all of the influenza A(H3N2) strains from the April 2014 to March 2017 seasons belonged to the genetic clade 3C and shared mutations N145S (epitopeA), V186G (epitopeB), P198S (epitopeB) and F219S (epitopeD). Most strains (12/14) of the 2014/2015 season fell into the subgroup 3C.3a and possessed the mutations A138S (epitopeA), R142G (epitopeA), F159S (epitopeB) and T128A (epitope), while strains isolated from the 2015/2016 and 2016/2017 seasons clustered in 3C.2a characterized by N144S (epitopeA), F159Y (epitopeB), K160T (epitopeB) and Q311H (epitopeC). Then, the 3C.2a strains underwent further evolution to yield two subclades. The strains form the 2015/2016 season possessed mutations N171k (epitopeD), and the 2016/2017 strains were characterized by T131K (epitopeA) and R142K (epitopeA) (Table 3, Figure 2).

Figure 2:

Phylogenetic tree comparing the HA gene of A/H3N2 Yantai isolates and vaccine strains.

No mutation was observed in the HA receptor binding sites (positions 98Y, 136T, 153W, 183H, 190D, 221P and 227E). The HA proteins of the A(H3N2) clade 3C.3a strains had 12 potential glycosylation sites (8, 22, 38, 45, 63, 122, 133, 144, 165, 246, 285 and 483). Comparing to the 3C.3a strains, the A128T mutation which resulted in the gain of a potential glycosylation site occurred in the clade 3C.2a. The N144S substitution caused the loss of a potential glycosylation site, and the 158 glycosylation site was observed in 3C.2a strains due to the mutations F159Y and K160T.

Compared to the NA sequence of A/Texas/50/2012, most strains in the 2014/2015 season carried the substitutions E221D(13/14) and I392T(10/14), and those strains in the 2015/2016 and 2016/2017 seasons had substitutions T267K(28/31), S245N(30/31), S247T(30/31), I380V(30/31) and P468H(28/31) (Table 3).

The NA proteins of A(H3N2) circulating in the Yantai area in the 2014/2015 season had eight potential glycosylation sites (61, 70, 86, 146, 200, 234, 329 and 367), while the strains in the 2015/2016 and 2016/2017 seasons had an additional N-linked glycosylation site due to the substitutions S245N.

No substitution was observed in the catalytic sites (118, 151, 152, 224, 276, 292, 371 and 406) or framework sites supporting the catalytic residues (119, 156, 178, 179, 198, 222, 227, 274, 277, 294 and 425) in the strains circulating in the Yantai area [15]. Most importantly, all of the analyzed strains had no the oseltamivir resistance substitutions E119V and R292k [16].

Evaluations of vaccine efficacy

To estimate the effect of the accumulated amino acid substitutions in the HA1 domain on predicted VE in a given year, the pepitope model was used to evaluate how well the vaccine strains match the circulating strain. Theoretically, the VE becomes negative when the pepitope is higher than 0.19 [17]. For the 2014/2015 season, the pepitope between A/Texas/50/2012 and A/H3N2 strains circulating in Yantai area showed a mean value of 0.20 (dominant epitope=B, substitutions: 128, 159, 186, 198), which indicated a negative VE against those strains present that year of −5.18% (E=−2.44% of 47%, pepitope=0). For the 2015/2016 season, the mean pepitope value between the vaccine strains and A/H3N2 strains in Yantai area was 0.158 (dominant epitope=A; substitutions: 138, 142, 144), which suggested the VE against those strains was 16.97% (E=7.97% of 47%, pepitope=0.198) of that of a perfect match with the A/Switzerland/9715293/2013 vaccine strain, while the mean pepitope value for the 2016/2017 season was 0.11 (dominant epitope=A; substitutions: 131, 142), which indicated the VE against those strains was 42.05% (E=19.76% of 47%) of that of a perfect match with the A/Hong Kong/4801/2014 vaccine strain (Table 4).