Virological influenza surveillance during flu season 2016/17
Virological influenza surveillance data in the Shanghai port were collected on a weekly basis. From February 2016 to September 2017, a total of 64 swab samples were collected from passengers of different countries, including 41 passed through Asia (25 Hong Kong and 12 Southeast Asia, especially), 16 passed through Europe, 7 passed through the America, and 7 passed through Oceania.
Influenza A (H3N2) virus activity increased from the 44th week of 2016, peaked in the 1st week of 2017 and decreased afterwards. The highest proportion of A(H3N2) was observed in summer (28/64, 43.7%), followed by winter (22/64, 34.3%) which outnumbering in spring and fall (11/64, 21.8%).
Phylogeny relationships of circulating A(H3N2) viruses during flu season 2016/17
Of the 610 genetically characterized viruses, 546 were provided from GISAID EpiFlu databases. All 64 HA genes sequenced by Shanghai port belonged to the H3N2 3C.2a clade. This clade also included vaccine strain A/HongKong/4801/2014, supporting the vaccine recommendation in 2016-2018 northern hemisphere influenza season by WHO. Among the 64 viruses, the majority (n=20, 31.2%) belonged to the subclade 3C.2a.1 represented by A/Singapore/INFIMH-16-0019/2016. The proportions for other subclades was 26.5% (3C.2a.2, n=17), 25% (3C.2a.3, n=16) and 6.2% (3C.2a.4, n=4) (Fig. 1).
Individual clades of influenza A (H3N2) are typically defined by amino acid substitutions that occur as they diversify from parental strains. Analysis of HA sequences indicated co-circulation of multiple variants in clade 3C.2a, all variants within subclade 3C.2a.1 shared four substitutions N121K, N171K, I406V and G484E. Three additional substitutions were observed in 3C.2a.1 subcluster: S92R and H311Q in cluster I, G479E in cluster II. Variants 3C.2a.3 shared N121K/E and S144K (I58V and S219R in cluster I and T135K and R150K in cluster II), Variants 3C.2a.2 were characterized by T131K and R142K substitutions and variants 3C.2a.4 were characterized by D53N, R142G, S144R, I182T and Q197H (Fig. 2).
There were more 3C.2a.1 variants identified from samples collected in summer (n=11) than in winter (n=7). This subclade was further divided into two homogenous sub-clusters (cluster I and II; Fig. 1). The strains from cluster I were concentrated in the winter, and the cluster II strains were persisted more common in the summer months. Most of the subgroup 3C.2a.3 Viruses had happened in summer. And we also found that there was no prominent summer or winter trend of viruses clustered in 3C.2a.2.
The clade pattern of circulating A(H3N2) influenza viruses, 2016/17
To analyze the geographical distribution of A(H3N2) in China, 31 provinces were classified into six regions on the basis of geographic proximity: North (Beijing), East-coastal (Shanghai), East-inland (Anhui), South-coastal (Guangdong), South-inland (Guizhou), Northeast (Jilin), Northwest (Shanxi) and West (Sichuan). According to our phylogenetic analysis, the A(H3N2) number of the above six regions be counted (Fig. 3A). The Proportions for A(H3N2) in these regions were 5%, 43%, 7.2%, 19%, 4%, 7% ,5% and 7%, respectively. Interestingly, higher epidemic waves of influenza A (H3N2) were observed in Eastern and Southern in China coastal area, and we presumed that higher density of trade contributed to it.
The genetic diversity results (Fig. 3B) indicated that the diversity increased in the East and South, especially coastal cities, Shanghai and Guangzhou. Both cities covered the all clades and subcaldes of the current A (H3N2). 3C.2a.3 (60%) was dominant in Guangzhou, with a small proportion of 3C.2a.4 (10%), 3C.2a (3%), 3C.1(2%), 3C.3(5%), 3C.2a.1-I(10%) and 3C.2a.1(5%). In contrast, 3C.2a.1, 3C.2a.1-I and 3C.2a.2 were the major subcluster in Shanghai, with proportions of 19.21%, 32.36 and 30.34%, respectively. 3C.2a.3-II (4%), 3C.1 (1%), 3C.3(2%) and 3C.2Aa(1%) were also detected in this region. The diversity of the clade pattern and dominant clade in these two coastal cities matched well with the trends of the current global influenza A(H3N2), likely because of the higher density of migration and subtropical monsoon climate.
Prediction of glycosylation sites in A(H3N2) viruses during flu season 2016/17
There were two models of predicted glycosylation sites in the HA proteins of the A(H3N2) clade 3C.2a : 12 potential glycosylation sites (N8ST, N22GT, N38AT, N45SS, N63CT, N126WT, N133GT, N158YT, N165VT, N246ST, N285GS and N483GT) and 11 potential glycosylation sites (N8ST, N22GT, N38AT, N45SS, N63CT, N126WT, N133GT, N158YT, N165VT, N246ST and N285GS). All of Shanghai port virus strains in Clade 3C.2a.1 had 11 potential glycosylation sites, and the rest of the other clade had 12. Comparing to the vaccine strains 2016/17 A/HongKong/4801/2014 (N8ST, N22GT, N38AT, N45SS, N63CT, N126WT, N133GT, N165VT, N246ST, N285GS and N483GT), Clade 3C.2a.1 virus were loss of potential glycosylation site 483(NGT), the virus of Clade 3C.2a.2, Clade 3C.2a.3 and Clade 3C.2a.4 were gain of potential glycosylation site 158(NYT).
Estimation of vaccine efficacy for A (H3N2)
To assess the effect of the accumulated mutations in the HA1 domain on predicted vaccine efficacy in a given year, the Pepitope method was used to evaluate how closely the vaccine strain resemble the circulating strain (Table 1). Theoretically, when p epitope in the dominant epitope is higher than 0.19, the vaccine efficacy becomes negative19,20. For the 2016/17 season, the pepitope between A/HongKong/ 4801/2014 vaccine strain and A/H3N2 strains showed a dominant mutation in epitope B (160,194) and the p epitope of 0.095, which indicated the VE against those strains was 50.06% (E=23.53% of 47%, Pepitope = 0) of that of a perfect match with the vaccine strain. In epitopes A, the p epitope of 0.105 (substitutions: 131, 142) for A(H3N2) strains predicted a vaccine efficacy against these strains of 44.68% (E = 21% of 47%, Pepitope = 0) of that of a perfect match with the A/HongKong/ 4801/2014 vaccine strain.