Influenza A virus hemagglutinin mutations associated with use of neuraminidase inhibitors correlate with decreased inhibition by anti-influenza antibodies

Eight plasmids of the 8 gene segments of wild-type A/California/04/09 A(H1N1)pdm09 (CA/04) virus were kindly provided by Dr. Robert G. Webster at St. Jude Children’s Research Hospital, Memphis, TN. Recombinant viruses were generated by DNA transfection of 293 T cells, and the point mutations were inserted into the HA gene of wild-type virus using a Quickchange site-directed mutagenesis kit (Stratagene) [16]. Stock viruses were prepared in Madin-Darby canine kidney (MDCK) cells at 37 °C for 72 h and their entire HA and NA genes were sequenced to verify the presence of the desired HA1 mutations and the absence of any additional HA/NA substitutions. The recombinant viruses were designated according to their HA1 substitutions. All experimental work was performed in a biosafety level 2 laboratory approved for use with these strains by the U.S. Department of Agriculture and the U.S. Centers for Disease Control and Prevention.

Structural bioinformatics was used to determine if the NAI treatment-emergent HA1 mutations can be mapped to previously described H1N1 HA antigenic sites. The X-ray crystal structure of the CA/04 HA protein (PDB:3LZG) was downloaded from the protein databank and analyzed using MacPymol (DeLano Scientific LLC). This X-ray crystal structure was selected for our analysis because it is the same structure for which A(H1N1) epitopes were mapped previously [17]. The putative antigenic sites that are conserved between influenza A(H1N1) viruses and A(H1N1)pdm09 were identified based on comparison of their HA amino acid sequences and structures (Fig. 1). Each of the antigenic sites was mapped onto the HA1 trimer and colored black to distinguish these sites from the rest of the structure, which is colored in gray. NAI treatment-emergent HA1 mutations identified in cell culture or in clinical studies were mapped onto the structure using yellow or orange (for substitutions occurred in the RBS).

A hemagglutination inhibition (HI) assay was performed with 0.5% chicken red blood cells by a standard method [26]. We used goat and ferret antisera as well as human convalescent sera from individuals who were confirmed to be infected with A(H1N1)pdm09 virus and a panel of 4 mAbs to HA of the CA/04 strain. Human sera were collected from anonymous donors (i.e., Donor 1, Donor 2, and Donor 3) in the licensed BioLife plasma collection centers during 2009–2010 and these serum samples were commercially obtained from Baxter Inc. Four antibodies to CA/04 HA, mAbs 28665, 28666, 28667, and 28668, were obtained through BEI Resources, NIAID, NIH. Virus-neutralizing titers were determined by infection of MDCK cells and expressed as the reciprocal of the highest serum dilution that neutralized 50% of fifty 50% tissue culture infectious doses (TCID₅₀) of virus after incubation at 37 °C for 72 h.

To determine multistep growth curves for each virus in the presence or absence of oseltamivir carboxylate (5 nM = 5 effective concentration-50 for the CA/04 virus [27]), human lung epithelial Calu− 3 cells were inoculated via the apical side with the H1N1 viruses at a multiplicity of infection of 0.001 plaque-forming units (PFU)/cell. After incubation for 1 h, the cells were washed and overlaid with medium with or without 5 nM oseltamivir carboxylate, 0.3% bovine serum albumin, and 1 μg/ml l-(tosylamido-2-phenyl)ethylchloromethylketone-treated trypsin. The supernatants were collected at 24, 48, and 72 h post-infection and stored at − 70 °C until titration.

Virus yields were compared by analysis of variance (ANOVA) with Dunnett’s multiple comparisons test. A probability value of 0.05 was prospectively chosen to indicate that the findings were not the result of chance alone.

A group of HA substitutions that were identified in nonclinical NAI resistance-selection experiments or in clinical specimens from NAI treatment or surveillance studies were analyzed in this study (Table 1). Many of these mutations are described in the current FDA-approved NAI package inserts [12‐14]. In general, NAI-resistant substitutions are included in drug product labels based on clinical observation, cell culture selection, and phenotype data. HA substitutions are generally included as NAI treatment-emergent mutations if they meet one of the following criteria: (i) selected in cell culture in the presence of NAI; (ii) observed as treatment-emergent in more than one patient, (iii) observed as treatment-emergent in a single patient at positions identified in cell culture as impacting drug susceptibility; and (iv) observed in surveillance or baseline clinical study samples at positions identified in cell culture as impacting susceptibility. It is worth noting that HA/NA substitutions observed at baseline in clinical studies or in surveillance samples are included in the approved drug product labels because there are examples of circulating amino acid polymorphisms that clearly reduce susceptibility of influenza virus to antivirals. Examples include NA H275Y in the pre-2009-pandemic H1N1 lineage and M2 S31 N in the majority of currently circulating seasonal influenza viruses [28] as well as other viruses [29, 30]. However, it is likely that the impact of HA1 mutations on susceptibility to NAIs is highly strain- and target tissue-dependent. Therefore, some polymorphisms may not be selected by or alter susceptibility to NAIs in the currently circulating strains.

We mapped the amino acid substitutions located in the HA1 domain of A(H1N1) and/or A(H1N1)pdm09 viruses onto the H1 HA1 trimeric crystal structure (Fig. 1). These mutations were identified in nonclinical NAI resistance-selection experiments or clinical specimens from treatment or surveillance studies and are listed in the corresponding NAI drug package inserts (Table 1). We also visualized the HA structural association with previously identified antigenic sites [17]. Most of the NAI treatment-emergent H1 HA1 mutations mapped to the antigenic sites predicted to be important for immunity. These mutations included D125S, G155E, S162 N, and D222G (H1 numbering convention is used here and throughout the text). Substitutions at positions S183 and D222 were associated with the RBS and have been shown to impact escape from neutralizing antibodies [3, 31] (Fig. 1a, b). Substitution R208K mapped to the interior of the trimer below the surface-exposed globular domain distal to the HA1 antigenic sites.

We evaluated the replicative ability of the recombinant H1N1 mutants by assaying their virus yields in comparison to those of the parental virus after multiple replication cycles in Calu− 3 cells. As shown in Fig. 2a, the CA/04^G155E and CA/04^D222G viruses grew to significantly higher titers than the wild-type virus at 72 h post-infection (~ 1.1 log₁₀PFU/ml, P < 0.05). Both mutants also formed larger plaques than the parental CA/04 virus in MDCK cells (P < 0.05, data not shown). To determine if any of the NAI treatment-emergent HA1 mutations are associated with a drug-dependent phenotype, we examined the replication of the H1N1 mutants in the presence of 5 nM oseltamivir carboxylate in Calu− 3 cells (Fig. 2b). We observed that all of the HA1 mutations could rescue the weak growth capacity of the wild-type virus, thus masking any replication defect and increasing NAI resistance (~ 2.8-fold increase in viral titers compared to CA/04, P < 0.05).

To evaluate the impact of NAI treatment-emergent HA1 mutations on immune reactivity, we assessed the sensitivity of viruses containing selected HA1 substitutions to several anti-H1 mAbs and polyclonal antisera in HI and MN assays. HI of goat and ferret antisera raised against pandemic A(H1N1)pdm09 viruses was significantly reduced (~ 20-fold↓) against recombinant CA/04^G155E and CA/04^D222G viruses carrying the G155E and D222G substitutions, respectively, compared to the wild-type CA/04 virus (Table 2). The HI activity of mAbs 28665 and 28668, which target antigenic site Sa [32], was reduced > 48-fold. The S183P substitution resulted in ~ 4-fold drop in 28665 titers and ~ 2-fold drop in 28666 and 28667 titers. There were no significant changes in ferret antisera or mAb HI activity against V152I and S162 N. To further characterize the antigenic properties of the mutant viruses, we performed MN assays using mAbs and human serum samples against CA/04. Consistent with the HI results, the MN activity of mAbs 28665 and 28668 and human serum samples collected from three different donors was reduced by >170-fold against recombinant CA/04^G155E and CA/04^D222G compared to CA/04 (Table 3). Overall, our results indicate that NAI treatment-emergent HA1 mutations may alter the antigenic profiles of A(H1N1)pdm09 and result in decreased antigenicity of the HA protein. However, this effect is selective because not all substitutions resulted in demonstrable changes in the HI and MN assays.