4 DISCUSSION
H6N6 subtype AIV is widely prevalent in poultry, and its host range has expanded to mammals, such as swine. Undoubtedly, it has become an endemic disease of domestic fowl and domestic animals. Here, three chicken-originated H6N6 subtypes of AIV were multiple reassortment virus, and their gene segments were derived from group-Ⅱ (ST2853-like) of Eurasian lineages. Terrestrial bird may have an intermediate host role in the cross-species transmission of influenza virus from birds to humans (Lam et al., 2013; Malik Peiris, 2009; Yang et al., 2017). At the same time, the molecular epidemiological investigation showed that H6N6 subtype AIV has been prevalent in terrestrial bird chickens (H. Wu et al., 2016). Therefore, the H6H6 virus from chickens may acquire the potential to infect humans. The switch of receptor-binding preference from avian-like SA2,3-Gal to human-like SA-2,6Gal is an key factor in avian influenza virus crossing interspecies barriers and efficiently transmitting to humans. The receptor-binding domains in the head of HA of influenza virus can specifically recognize and bind to avian-like SA2, 3-Gal and/or human-like SA2, 6-Gal receptor; yet, the molecular mechanism of receptor-binding preference switch in different avian influenza virus subtypes needs to be further elucidated. HA of H5N1 virus with N224K/Q226L mutations had a key role in switching receptor-binding preference from avian-like SA2,3-Gal to human-like SA-2,6Gal (Imai et al., 2012). HA with Q226L and G186V mutations in the H7N9 virus could result in virus binding to the human-like receptor (Dortmans et al., 2013), and HA with S137N, E190V, and G228S mutations of the H6N1 virus is essential in the process of acquiring the ability to recognize human-like virus receptors (Ni, Kondrashkina, & Wang, 2015; F. Wang et al., 2015).
In this study, three strains of H6N6 virus, 224, 226, 228, 137 138, and 190 of HA receptor binding domains, have no mutation. However, P186T, H156R, S263G mutations, an amino acid deletion at 158 of HA were found in the ZZ346 strain. Interestingly, receptor-binding analysis by HA assay indicated that the ZZ346 strain of the H6N6 virus could bind to avian-like SAα-2,3Gal and human-like SAα-2,6Gal receptors. Some studies reported that HA with G186V mutations in H7N9 was able to bind to the human-like receptor (Dortmans et al., 2013). A combination of HA (H156N, S263R) and PA (I38 M) mutations might enhance the virulence of the virus in mice (Tan et al., 2014), suggesting that substitution of 186, 156, 263 of HA might be related to binding to human-like SAα-2,6Gal receptors and effectively replicate in mammals. Therefore, we speculate that the chicken-originated ZZ346 virus strain of H6N6 virus with P186T, H156R, S263G mutations of HA could bind to avian-like SAα-2,3Gal and human-like SAα-2,6Gal receptors. Mutation or deletion of NA was observed during the adaptation of viruses to a new host, thus suggesting it crosses the host restriction (Hughes, McGregor, Suzuki, Suzuki, & Kawaoka, 2001). Recently, amino acid deletion in NA’s stalk region in some H6N6 subtype AIV was found (Li et al., 2019). In this study, the 11aa deletion in the stalk region of NA located at positions 59-69 was found in the ZZ346 virus strain but not in JX20490 and ZZ1923 strains. The functional balance between HA and NA is crucial to the survival of the virus. HA affects virus binding to host cell, and NA affects progeny virus particles releasing from the host cell. Only when the two cooperate to reach the best balance state can the virus effectively replicate in the host cell, causing host infection onset (Gen et al., 2013). T271K, E627K, D701N of PB2 can enhance the polymerase activity, which favors enhancing the pathogenicity and transmission of the H6N6 virus in mammals. However, no substitution of these sites of PB2 of H6N6 virus was observed in this study, which was consistent with published papers (G. Wang et al., 2014; Zhang et al., 2011).
It remains unclear whether the ZZ346 strain of the H6N6 virus that can bind to avian-like SAα-2,3Gal and human-like SAα-2,6Gal receptors could replicate and infect in mammals and humans. Accordingly, we selected the three chicken-originated H6N6 strains to inoculate BALB/c mice and human lung tissues. In the mice infection experiment, some strains of H6N6 influenza viruses were able to infect mice. After inoculating mice with ZZ346 strain, the virus and viral antigen NP were detected in the trachea and lung. In addition, antibodies in serum were detected on day 14 post-inoculation, indicating that the virus could effectively replicate in mice, thus causing infection. Regarding the JX20490 strain, NP protein was detected in the trachea and bronchial epithelial cells but not in the lungs. In addition, no virus was detected in the tracheal, bronchus, and alveolar tissue. The results showed that chicken-originated ZZ346 strain of H6N6 could directly infect mice without prior adaptation, which is consistent with a previous study(H. Wu et al., 2016). Next, human lung tissue was inoculated with three strains of chicken-originated H6N6 subtype AIV in vitro . Lung tissue inoculated with ZZ346 strain showed local cell necrosis; virus and viral antigen NP were detected in the tissue. In contrast, virus and NP protein were not detected in the lung tissue after inoculation with JX20490 and ZZ1923 strains. Therefore, these data suggested that the ZZ346 strain was effectively replicated in the human lung without prior adaptation. At the same time, it provided evidence that P186T, H156R, S263G mutation, and amino acid deletion at 158 of HA were conducted to the switch of the H6N6 virus from binding to avian-like SAα-2,3Gal receptor to binding human-like SAα-2,6Gal and avian-like SAα-2,3Gal receptor. Based on the serum H6 antibody of the exposed population(Kayali G, 2010; Xin et al., 2015), our results indicated that the H6N6 virus acquired the potential to infect humans. 11aa deletion (59-69) in the stalk region of NA was found in ZZ346 strain, which also occurred in H5N6 virus-infected with human beings(Jin et al., 2017), speculating that the deletion at positions 59-69 in the stalk region of NA may be related to the infection of H6N6 in mammals, especially in humans. In this study, ZZ346 strain with HA variation (P186T, H156R, S263G mutation, and amino acid deletion at 158) and the 11aa (59-69) deletion in the stalk region of NA directly infected mice and effectively replicated in human lung tissue without prior adaption.
The first case of human infection with H6N1 avian influenza was reported on May 20, 2013, in Taiwan, China (Wei et al., 2013). The emergence of human infection cases with H6N1 indicated the unpredictability of influenza virus transmission and novel viruses’ potential threat. Some studies reported that the influenza virus causing pandemic was generated from avian-human (or-swine) influenza A virus reassortments (Smith et al., 2009), but the avian influenza virus involved in the reassortments is not necessarily a highly pathogenic avian influenza virus. Moreover, mild symptoms caused by low pathogenic virus infection can be easily ignored, increasing the chances of virus spread, adaptive mutation, and reassortment. Currently, the prevention and control of influenza pandemic are mainly focusing on the subtypes of bird flu viruses H5N1 and H7N9, which cause severe human diseases and deaths. However, due to unpredictability and the inadequacy of influenza virus-related knowledge, we cannot predict which subtype of influenza A virus will cause the next influenza pandemic. Although the H6N6 virus is a low pathogenic virus, it is widely prevalent in poultry. It repeatedly infects swine, which has the potential to evolve into a novel influenza virus infecting human beings. Therefore, this study confirmed that some chicken-originated H6N6 viruses might acquire the ability to recognize and bind to human-like receptors, thus increasing human infection risk. Our study emphasized the importance of continuous and intensive monitoring of the terrestrial bird H6N6 virus evolution to prevent transmission to humans.