2009年新型流感病毒A(H1N1)的基因變異、擇選、病毒性狀改變與流行病學特徵、防疫政策及其流行愈趨重要之相關性

Abstract

Newly emerged triple reassortant 2009 pandemic influenza A (pH1N1) viruses were detected in the United States (US) and Mexico in March-April, 2009 and then rapidly spread worldwide. The overall objective of this study was to investigate the association between molecular and phenotypic dynamic changes of pH1N1 viruses and epidemiological characteristics and Taiwan’s public health interventions for better prevention/control of novel influenza viruses in the future. The specific aims were: 1) to compare viral sequence variations in nucleotides (NTs) and amino acids (AAs) of hemagglutinin (HA) and neuraminidase (NA) of pH1N1 isolated in Taipei and Kaohsiung metropolitans at pre-peak, on-going peak and post-peak of the 2009-2010 epidemic, 2) to analyze the association between the spreading of Taipei’s HA-E374K mutants and epidemiological characteristics and public health interventions, and 3) to explore the selection mechanisms of E374K, including viral biophenotypic changes and co-mutations in the other genes for better fitness. A cross-sectional study was performed, using 196 pH1N1 virus strains (168 Taipei’s and 28 Kaohsiung’s) from June, 2009 to October, 2010. The viruses were passaged twice in the Madin-Darby Canine Kidney (MDCK) cells and viral nucleic acids were amplified by reverse transcription–polymerase chain reaction, (RT-PCR). The NTs and AAs of 196 HA and 40 NA genes were analyzed their viral antigenic sites, receptor binding, N-linked glycosylation sites and drug resistance genes. Strain variations in viral antigenicity used hemagglutination inhibition (HI) test. Tempo-spatial analyses of 118 pH1N1 strains of Taipei’s patients with their residential district-specific population density used the Morans’s I to measure presence of E374K cluster by global spatial clustering analysis and to further examine where were local spatial clusters by local indicators of spatial association (LISA). The association between E374K and epidemiological characteristics (age, gender, population density of the districts, and spatial clustering), or at different periods after 3 strategies of interventions (use of antiviral drug, class suspension and vaccination), or clinical severity was analyzed by univariate and multiple logistic regression analyses. Lastly, to elucidate selection mechanisms for the fitness of E374K better than E374, viral antigenicity changes, replication ability and the co-mutation of the six internal viral genes were compared, using the full-length sequences (PB2, PB1, PA, NS, M, NP) of 30 Taiwanese pH1N1 strains collected from the Influenza Virus Resource, National Center of Biotechnology Information (NCBI), USA. Global trends in increasing E374K mutants were also examined using NCBI sequence data in different countries. The results revealed that the cumulated numbers of AA changes in HA and NA were higher in the post–peak than those in the pre-peak period of the epidemic [HA: 6.7% (1/15) vs 74.6% (47/63)，p<0.0001； NA: 36.84% (7/19) vs 61.9% (13/21)，p=0.21)]. Detail analyses identified two mutants persistently circulated with increasing percentages. One mutant, HA-S203T located at antigenic site Ca, was firstly detected at 21th week, 2009 and became dominant before week 34 (86.84%, 33/38), and totally replaced after week 35 (100%, 136/136), suggesting that the S203T mutant emerged and increased viral frequency in foreign countries in early pandemic before it entered Taiwan. The other mutant, E374K located at the stalk region of HA2 was firstly found at week 34 in Taipei and rose as a major circulated strain at post-peak of the epidemic (64.65%, 64/99). In addition, 14.94% and 3.44% of 174 isolates had one and two amino acids changes in the four antigenic sites, respectively but they did not persist through all the epidemic periods. Only 6 strains had variations at receptor binding sites (4 at 220-loop and 2 at 190-helix) and another 2 strains showed variations in the loss of one N-linked glycosylation site of HA (2.56%, 2/78). The NA of 40 strains retained all N-linked glycosylation sites without H275Y mutation responsible for Tamiflu resistance. Taken together, most of the pH1N1 had conserved antigenicity, N-linked glycosylation sites of HA and NA and variations in HA (S203T, Q293H, D222G, N125D and R205K) were not associated with clinical severity. The unique adaptive E374K mutant was first detected at 3 weeks before the epidemic peak in Taipei and 6 weeks later (40th week) in Kaohsiung and then increased significantly higher in the post-peak than those in the pre-peak period of the epidemic [64.65% (64/99) vs 9.28% (9/97), p＜0.0001] in both cities. The frequency of E374K reached to 85.7% (96/112) in 2010-2011 winter [wild type E374: 1.8% (2/112), E374G:12.5%, (14/112)].The E374K was not associated with clinical severity [mild vs severe cases: 34.8% (37/116) vs 37.3% (19/52), p=0.6]. The tempo-spatial spreading of E374K mutants was more concentrated during the post–peak (41st-52nd week) in seven districts of Taipei City. Multivariate logistic regression analysis confirmed that higher odds ratios (OR) occurred in later time periods (OR=1.53, p <0.001) and in areas with spatial clustering (OR=4.57, p=0.047), after controlling age and population density. After the three major interventions, the E374K variant did not disappear but was even associated with increasing percentages after the usage of Tamiflu since August 1, 2009 [0% (0/17) vs 40.78% (73/179), p<0.001]. Such a phenomenon was not found in other mutations in the four antigenic sites [29.4% (5/17) vs 17.8% (28/157), p=0.32]. During the 2nd peak of class suspension (week 41-45), the E374K reached 90% (9/10) with tempo-spatial clusters within weeks of 41-52. Finally, these E374K mutants increased after vaccination (22.9%, 32/140 vs 72.3%, 41/56, p<0.001) with persistently high frequency through 10 months post-vaccination on November 1 16, 2009. Vaccination also significantly elevated Sa mutants (2.4%, 3/127 vs 23.4%, 11/47, p<0.001). To investigate the mechanism of the survival of E374K in human, 7 E374K strains were firstly tested for HI or MNt antibodies, using the three anti-pH1N1-HI(+) serum samples from human, sheep and pig. No significant difference in sero-titers between E374K and E374 (≦2 fold), indicated that E374K did not survive through immune escape. The growth curve of E374K in MDCK cells showed a similar pattern to that of E374 without significant difference. The replication advantage of E374K needs to be further tested in human respiratory tract cells. Lastly, co-mutation analyses revealed that 8 E374K viruses isolated from June, 2009 to October, 2010 had 100% (8/8) unique co-mutations at T257A of PB1 but such a co-mutation was totally replaced by other sites [PA: N321K (81.82%, 9/11), A343T (54.55%, 9/11); M: V89I (81.82%, 9/11); PB1: I397M (54.55%. 6/11), I435T (63.64%. 7/11)] in E374K viruses obtained from November 1, 2010 to February 2011. All co-mutations were absent in E374 viruses. Taken together, the Taiwan’s finding on temporal increase in E374K percentage from this study was consistent with observations in several high population countries (Singapore, UK, China and India). It is very likely that E374K evolved through natural evolution under low selection pressure and obtained evolutionary advantages at specific sites with temp-spatial clusters of cases in areas with high population density, possibly through co-mutations in other genes and thus facilitating better viral replication capability in human respiratory cells and fast human-to-human transmission to become a dominant mutant. Future efforts need to increase sample size and examine the E374 replication in different human respiratory cells for further confirmation. This is the first study examining viral changes during the naive phase of a pandemic of influenza through integrated virological/serological/clinical surveillance, tempo-spatial analysis, and intervention policies. Our results enlighten to carefully monitor amino acids of HA and NA and co-mutations in other segments of pandemic influenza viruses isolated at exponential/peak phases in areas with high cluster cases.