The characteristics and antigenic properties of recently emerged subclade 3C.3a and 3C.2a human influenza A(H3N2) viruses passaged in MDCK cells

Background Two new subclades of influenza A(H3N2) viruses became prominent during the 2014‐2015 Northern Hemisphere influenza season. The HA glycoproteins of these viruses showed sequence changes previously associated with alterations in receptor‐binding properties. To address how these changes influence virus propagation, viruses were isolated and propagated in conventional MDCK cells and MDCK‐SIAT1 cells, cells with enhanced expression of the human receptor for the virus, and analysed at each passage. Methods Gene sequence analysis was undertaken as virus was passaged in conventional MDCK cells and MDCK‐SIAT1 cells. Alterations in receptor recognition associated with passage of virus were examined by haemagglutination assays using red blood cells from guinea pigs, turkeys and humans. Microneutralisation assays were performed to determine how passage‐acquired amino acid substitutions and polymorphisms affected virus antigenicity. Results Viruses were able to infect MDCK‐SIAT1 cells more efficiently than conventional MDCK cells. Viruses of both the 3C.2a and 3C.3a subclades showed greater sequence change on passage in conventional MDCK cells than in MDCK‐SIAT1 cells, with amino acid substitutions being seen in both HA and NA glycoproteins. However, virus passage in MDCK‐SIAT1 cells at low inoculum dilutions showed reducing infectivity on continued passage. Conclusions Current H3N2 viruses should be cultured in the MDCK‐SIAT1 cell line to maintain faithful replication of the virus, and at an appropriate multiplicity of infection to retain infectivity.

. To extend these recent observations, we have examined the propagation of H3N2 viruses from subclades 3C.2a and 3C.3a in conventional MDCK cells and in MDCK-SIAT1 cells, and analysed sequence changes acquired on culture in the HA and NA glycoproteins.
We further assessed whether amino acid substitutions acquired during propagation altered the antigenic and receptor-binding properties of the new subclades of virus.

| Clinical specimens
The clinical specimens used in this study were collected during the

2013-2014 influenza season and provided by the WHO National
Influenza Centres in Hong Kong SAR, China, and Buenos Aires, Argentina.

| Virus isolation and propagation
Hundred microlitres per well of clinical specimens were inoculated in parallel on confluent cell monolayers of MDCK-parent or MDCK-SIAT1 cells in 24-well tissue culture plates and incubated at room temperature for 1 hour, and then, 1 mL of DMEM supplemented with 2 μg/mL TPCK-treated trypsin (Sigma Cat no T1426) was added.
Further passages were performed using 200 μL inocula of cell culture supernatant per well, both undiluted and at higher dilution, from the previous passage.

| Haemagglutination assays (HAA)
HAA were performed according to standard methods 10 using suspensions of guinea pig, human (both 1.0% v/v) or turkey (0.75% v/v) red blood cells (RBC), and HAA titres were determined in the absence or presence of 20 nM oseltamivir carboxylate in the diluent 6 .

| Neuraminidase activity quantification
Sialidase activity of each virus was measured using 2′-(4-methylum belliferyl)α-D-N-acetyl neuraminic acid (MUNANA; Sigma) 11 . The dilution factor of the virus required to give 50% conversion of the total MUNANA substrate, at an initial concentration of 60 μM, over 60 minutes at 37°C was taken as a relative measure of the sialidase content of the sample.

| Infection of conventional MDCK and MDCK-SIAT1 cells by recent H3N2 viruses
In

| Analyses of the antigenic properties of viruses in subclades 3C.3a and 3C.2a
As

| Reduced plaque size or abortive infections after further passages at high multiplicity of infection (MOI) in MDCK-SIAT1 cells
During the course of the MN assays, it was observed that when un-  notably so by the third passage ( Figure S1), the NA activity was less markedly reduced when virus was passed at the higher multiplicity.
The NA activity to virus infectivity ratio served as a proxy for the particle to infectivity ratio, and the ratio increases observed implied that some form of defective particles were being generated on passage at high multiplicity in MDCK-SIAT1 cells.  (Table S2).    tivity despite low infectivity as a proxy for the particle to infectivity ratio. To overcome this problem, we submit that current H3N2 viruses should be cultured in the MDCK-SIAT1 cell line to maintain faithful replication of the virus, and at an appropriate MOI to retain infectivity and avoid the generation of defective viruses. This is critical if HI and MN assays are to remain the primary, universally employed methods for antigenic characterisation of H3N2 viruses.