Detection of oseltamivir‐resistant zoonotic and animal influenza A viruses using the rapid influenza antiviral resistance test

Abstract Mutations in the influenza virus neuraminidase (NA) that cause reduced susceptibility to the NA inhibitor (NAI) oseltamivir may occur naturally or following antiviral treatment. Currently, detection uses either a traditional NA inhibition assay or gene sequencing to identify known markers associated with reduced inhibition by oseltamivir. Both methods are laborious and require trained personnel. The influenza antiviral resistance test (iART), a prototype system developed by Becton, Dickinson and Company for research use only, offers a rapid and simple method to identify such viruses. This study investigated application of iART to influenza A viruses isolated from non‐human hosts with a variety of NA subtypes (N1‐N9).


| INTRODUC TI ON
Zoonotic and animal influenza A viruses pose a significant threat to public health; they can cause severe disease in humans with little protection afforded by seasonal vaccination due to antigenic differences. 1 NAIs are routinely used to treat individuals infected with influenza viruses, regardless of subtype, and the oseltamivir is the most commonly prescribed anti-influenza therapeutic. Antiviral resistance can emerge in nature or following treatment with NAIs through changes to the surface antigen NA that affect neuraminidase inhibitor (NAI) binding. Such changes may cause resistance to one or more NAIs. 2 While NA gene sequence analysis is often used to screen viruses for established markers of resistance, genetic analysis cannot identify viruses carrying new molecular markers, or assess the degree of reduced susceptibility. Thus, phenotypic NAI assays are commonly used to assess viral susceptibility to NAIs. 3 In these assays, virus is diluted to a targeted level of NA activity and tested against serially diluted NAI to determine an IC 50 , the drug concentration needed to inhibit 50% of NA activity. To report the results for seasonal influenza A viruses, the fold change of the test virus is calculated by comparison to a reference IC 50 value, either a subtype-specific median or the IC 50 of a control virus lacking the NA change. 4 However, this approach cannot be readily applied to testing and reporting of | 523 HODGES Et al.
non-seasonal influenza virus susceptibility to NAIs because of difficulty of acquiring and testing large numbers of each distinct subtype and wide range of genetic lineages within each subtype. Moreover, NAI results require careful interpretation, as laboratory correlates of clinically relevant resistance have not been established, except for viruses carrying an N1 NA with the H275Y substitution. 5 Infections caused by viruses displaying reduced inhibition (RI) or highly reduced inhibition (HRI) phenotypes may be more difficult to control by therapeutic intervention, which can lead to prolonged illness and virus shedding. 6 Simple and rapid assays that can be used by surveillance laboratories and in clinical settings are needed to detect viruses with reduced susceptibility to NAIs. As previously reported, the prototype influenza antiviral resistance test (iART), developed by BD Technologies (BARDA Contract HHSO100201300008C), is able to phenotypically detect seasonal influenza viruses that display RI/HRI by oseltamivir. 7 This assay compares influenza-specific sialidase (NA) activity with and without a single drug concentration, requires only 1 hour, and does not need extensive training to carry out. Here, we present similar findings for zoonotic and animal influenza viruses.

| COMPARISON OF IART TO NAI A SSAY
To verify the ability of iART to efficiently detect NA enzymatic activity and inhibition by oseltamivir of various subtypes (N1 through N9), a variety of zoonotic and animal influenza viruses were tested. This included viruses (n = 45) isolated from wild birds, poultry, a domestic cat, and zoonotic human infections propagated in MDCK cells or fertilized chicken eggs (Table 1). NA sequence analysis did not identify known or suspected markers of resistance to oseltamivir (Table S1). Viruses were tested using both the fluorescence-based NAI and iART assays, as previously described. 4 All virus isolates were found to be susceptible to inhibition by oseltamivir in the iART assay (R-factor ≤0.70). In the NAI assay, all calculated IC 50 values were in the nanomolar/subnanomolar range; some differences among subtypes were observed, as expected, with the greatest IC 50 value observed for N8 viruses and the lowest for N2 viruses ( Table 1). The median IC 50 for all subtypes (calculated using an average IC 50 for each subtype) was determined to be 0.48 nmol/L (Table S2). Using the median IC 50 , the fold change was calculated for each isolate. As expected, all tested viruses were determined to be normally inhibited (NI) by oseltamivir, and, therefore, susceptible to this drug, according to the criteria implemented by the Expert Working Group on Antiviral Susceptibility for the WHO Global Influenza Surveillance and Response System 5 (<10-fold increase compared to the median IC 50 ). The data from the gold standard NAI assay showed good correlation with the results obtained using iART, verifying the test's ability to detect NA enzymatic activity and inhibition by oseltamivir for non-seasonal influenza viruses.
To verify that iART was able to detect reduced susceptibility to oseltamivir of avian and zoonotic viruses, nine virus isolates with NA amino acid substitutions known to affect oseltamivir susceptibility were tested by both the NAI and iART assays ( Table 2). Calculated IC 50 values were compared to control viruses that lacked the NA substitution, as well as to the median IC 50 value calculated above.
The median IC 50 fold change calculation is necessary when a matching wild-type virus is not available or a virus with an unknown NA sequence is tested. The method of fold change did not change the interpretation for eight of nine viruses (Table 2). One isolate ( Table 2, A/Vietnam/HN30408/2005 clone 1) was interpreted as having RI using the fold change determined with the control virus IC 50 , normal inhibition (NI) using the fold change determined with the median IC 50 , and an R-factor that was below the pre-set threshold of 0.70 (0.57). Two viruses ( Table 2,  A wide range of R-factors were observed, which correlated with the range of fold differences determined by NAI assay ( Figure S1).
Viruses with the highest R-factors (ie, >4.0) were also identified as having HRI by the NAI assay. Viruses with RI or fold change values near the 10-fold cutoff had R-factors near the 0.70 threshold.
These results demonstrated that any virus reported as resistant by iART would have RI/HRI by NAI. Non-resistant viruses, particularly those with elevated R-factors, also showed some reduced inhibition by oseltamivir. With further testing and refinement of the R-factor threshold, iART may be able to differentiate between RI and HRI viruses in the future. Alternatively, any specimen with an R-factor above 0.50 could be flagged for sequence analysis and additional testing in the NAI assay. None of the wild-type type viruses shown in Table 1 or seasonal viruses reported previously would be flagged as having potentially reduced susceptibility using a lower threshold for type A viruses. 7

| RECOMB INANT N9 PROTEIN S WITH K NOWN MARK ER S OF RI/HRI BY OS ELTAMIVIR
Amino acid substitutions known to reduce susceptibility to oseltamivir E119V, I222K/R, H274Y, R292K, and R371K (N2 numbering) have been detected in the NA of A(H7N9) viruses isolated from humans. 8 In addition, I222T was detected in an A(H7N9) virus isolated from a non-human primate after oseltamivir treatment. 9 To determine whether iART is able to identify NA with these changes as resistant to oseltamivir, the respective recombinant N9 (rN9) proteins were generated using the A/Shanghai/2/2013 NA as a backbone, as previously described. 10 The use of recombinant protein allows testing of amino acid changes that reduce enzymatic activity in addition to reducing susceptibility to NAIs, including R292K (R289K in N9 straight numbering), the most commonly identified NA change detected in H7N9 human cases. The R-factors of rN9 proteins carrying substitutions E119V, I222K/R, H274Y, R292K, or R371K categorized them as resistant to oseltamivir and correlated with NAI assay outcomes ( Criteria for interpreting NAI assay results based on IC 50 fold increase compared with the control virus/median IC 50 value: normal inhibition (NI) <10-fold, reduced inhibition (RI) 10-to 100-fold, and highly reduced inhibition (HRI) >100-fold. d R-factor: ratio of chemiluminescent signal intensity generated by viral NA activity with and without inhibitor (ie, oseltamivir carboxylate). R-factor interpretation based on pre-set cutoff for influenza A (resistance is ≥0.70).
IC 50 values ( Figure S1); all rN9 with R-factors above 2.0 were identified as having HRI by the NAI assay. The rN9 protein with I222T was identified as non-resistant by iART. In the NAI assay, the fold change conferred by this substitution was below the threshold of 10, further confirming the correlation between the two assays.

| IART VS NAI A SSAY UNDER LOW PH COND ITI ON S (PH 5. 3 VS 6 . 8)
As mentioned above, R292K is the most commonly reported NA marker in oseltamivir-treated patients infected with A(H7N9) viruses.
In addition, this change is also known to reduce enzymatic activity, making detection of drug resistance difficult using the standard NAI assay due to insufficient activity for testing or wild-type activity masking resistance. 11 It was previously reported that detection of R292K viruses could be improved by NAI testing at an acidic pH. 12 To confirm this finding, testing was performed on a highly patho- While the gold standard NAI assay continues to be the assay of choice for surveillance laboratories, it is cumbersome and requires highly trained personnel. iART provides an alternative, simple method for detecting oseltamivir-resistant viruses using a small and portable device with built-in software for data interpretation.
Viruses detected by iART with elevated R-factors can be flagged for genetic analysis and comprehensive phenotypic evaluation. This design and ease of use may allow oseltamivir susceptibility testing in locations currently unable to carry out the NAI assay.