ABSTRACT
Human infections by avian influenza A(H7N9) virus entail substantial morbidity and mortality. Treatment of infected patients with the neuraminidase (NA) inhibitor oseltamivir was associated with emergence of viruses carrying NA substitutions. In the NA inhibition (NI) assay, R292K conferred highly reduced inhibition by oseltamivir, while E119V and I222K each caused reduced inhibition. To facilitate establishment of laboratory correlates of clinically relevant resistance, experiments were conducted in ferrets infected with virus carrying wild-type or variant NA genes recovered from the A/Taiwan/1/2013 isolate. Oseltamivir treatment (5 or 25 mg/kg of body weight/dose) was given 4 h postinfection, followed by twice-daily treatment for 5 days. Treatment of ferrets infected with wild-type virus resulted in a modest dose-dependent reduction (0.7 to 1.5 log10 50% tissue culture infectious dose [TCID50]) in nasal wash viral titers and inflammation response. Conversely, treatment failed to significantly inhibit the replication of R292K or E119V virus. A small reduction of viral titers was detected on day 5 in ferrets infected with the I222K virus. The propensity for oseltamivir resistance emergence was assessed in oseltamivir-treated animals infected with wild-type virus; emergence of R292K virus was detected in 3 of 6 ferrets within 5 to 7 days postinfection. Collectively, we demonstrate that R292K, E119V, and I222K reduced the inhibitory activity of oseltamivir, not only in the NI assay, but also in infected ferrets, judged particularly by viral loads in nasal washes, and may signal the need for alternative therapeutics. Thus, these clinical outcomes measured in the ferret model may correlate with clinically relevant oseltamivir resistance in humans.
IMPORTANCE This report provides more evidence for using the ferret model to assess the susceptibility of influenza A(H7N9) viruses to oseltamivir, the most prescribed anti-influenza virus drug. The information gained can be used to assist in the establishment of laboratory correlates of human disease and drug therapy. The rapid emergence of viruses with R292K in treated ferrets correlates well with the multiple reports on this NA variant in treated human patients. Our findings highlight the importance of the discovery and characterization of new antiviral drugs with different mechanisms of action and the use of combination treatment strategies against emerging viruses with pandemic potential, such as avian H7N9 virus, particularly against those carrying drug resistance markers.
INTRODUCTION
Antiviral medications play an important role in the management of influenza disease. Since early 2013, human infections with the avian influenza A(H7N9) virus in China have been associated with acute respiratory distress syndrome and have resulted in significant morbidity and mortality (1). Because the H7N9 virus is resistant to M2 blockers, patients have been treated with neuraminidase (NA) inhibitors (NAIs), among which oseltamivir is the most widely used. Notably, evidence for a therapeutic effect of oseltamivir treatment was reported in H7N9-infected patients during the 2013 outbreak (2–5). In early 2015, the first imported H7N9 cases in North America were described in two travelers returning from China; both patients were given oseltamivir and eventually recovered from the disease (6). However, influenza viruses can develop resistance to oseltamivir in patients receiving prophylaxis and/or treatment, as was shown with the seasonal H1N1 and H1N1pdm09 viruses carrying H274Y (H275Y in N1 numbering; for consistency, N2 numbering is used throughout the paper) (7, 8), or through spontaneous changes (9–11). Further, among oseltamivir-resistant variants, E119V, R292K, and D197N/E substitutions have been commonly detected in H3N2 and type B viruses (12–15).
Assessment of influenza virus oseltamivir susceptibility is not straightforward. Unlike M2 blockers, changes conferring NAI resistance are subtype specific. For instance, the H274Y substitution is an established oseltamivir resistance marker for H1N1 subtype viruses (16); however, it does not confer resistance in H3N2 subtype viruses (17). In addition, there is a lack of correlation between the drug susceptibility phenotypes determined in cell culture and in vivo (18). Surveillance laboratories utilize NA inhibition (NI) assays to determine the NAI concentration needed to inhibit 50% of viral NA activity (IC50). Using collective data from seasonal influenza viruses, a median IC50 is determined for each type/subtype, and this IC50 is used to determine the fold change in the IC50 for a particular strain. In accordance with the guidance provided by the WHO Antiviral Working Group (19), influenza A viruses showing reduced (a 10- to 100-fold increase in the IC50) or highly reduced (>100-fold increase in the IC50) inhibition by an NAI are further analyzed to identify the molecular marker(s) conferring the elevated IC50. However, there is no established cutoff IC50 or fold difference to discriminate drug-sensitive from drug-resistant viruses. Thus, the terms reduced inhibition and highly reduced inhibition used to describe the outcome of the NI assay are not synonymous with reduced susceptibility or resistance of the virus.
Presently, only influenza A viruses of H1N1 and H1N1pdm09 subtypes that show highly reduced inhibition in NI assays and carry H274Y are reported to the WHO Global Influenza Surveillance and Response System as oseltamivir-resistant viruses (16). Other viruses exhibiting reduced or highly reduced inhibition are not reported as resistant due to insufficient evidence of resistance to NAIs in humans. Many factors, such as a balance of hemagglutinin receptor-binding affinity, NA receptor-cleaving activity, and virus virulence, have been shown to affect susceptibility to NAIs (20, 21). Uncertainties in interpreting laboratory results necessitate detailed characterization of the drug susceptibility of emerging viruses, such as the H7N9 subtype.
In the early stage of the 2013 outbreak, A/Taiwan/1/2013 (H7N9) virus (Taiwan/1) was recovered from a patient who underwent two courses of oseltamivir treatment (2, 4). In a single isolate, we identified four virus variants differing by a single amino acid in the NA compared to the wild-type (WT) virus (22). When tested in the fluorescent NI assay with oseltamivir carboxylate, these viruses exhibited a wide range of IC50s. The R292K NA variant showed a >10,000-fold increase in the IC50 and was defined as a highly reduced (>100-fold) inhibited virus. The 3 NA variants, E119V, I222R, and I222K, were defined as viruses showing reduced inhibition. However, it was unknown which laboratory result, if any, should be interpreted as detection of oseltamivir resistance that may be of clinical importance.
Ferrets are considered the preferred surrogate animal model for human infection with influenza viruses (23, 24) and assessment of antiviral drug susceptibility (25, 26); other animal models have also been used to evaluate the NAI susceptibility of zoonotic viruses (27, 28). The aim of the study reported in this paper was to assess the oseltamivir susceptibility of the H7N9 WT virus (no changes in the NA) and three NA variants, carrying R292K, E119V, or I222K, in the ferret model. Our findings provide information that will facilitate establishing laboratory correlates of clinically relevant resistance in H7N9 viruses.
MATERIALS AND METHODS
Ethics statement.Animal experiments were conducted in strict compliance with the guidelines of the CDC Institutional Animal Care and Use Committee (IACUC) in association with the PHS Policy, the Animal Welfare Act (U.S. Department of Agriculture [USDA]), and the Guide for Animal Care and Use of Laboratory Animals. The animal protocol was approved by the CDC IACUC. All procedures were performed under enhanced animal biosafety level 3 conditions.
Viruses and sequence analysis.The isolation of WT and NA variants (R292K, E119V, and I222K) from the A/Taiwan/1/2013 (H7N9) isolate and their full characterization, including the NA sequence, were previously reported (22).
Virus inoculation, clinical signs, and oseltamivir treatment in ferrets.Male ferrets (Mustela putorius furo) aged 3 to 5 months (Triple F Farms, PA, USA) and serologically negative by hemagglutination inhibition (HI) assay for currently circulating influenza A(H1N1)pdm09, A(H3N2), and type B viruses were used in this study. The ferrets were housed in individual cages and monitored for at least 3 days for acclimation and to establish baseline body temperatures prior to the start of the study. Clinical signs of illness (activity level, nasal and ocular discharge, sneezing, and, if present, diarrhea), body weight, and temperature were recorded daily throughout the 14-day study. The temperature was measured twice daily by subcutaneous implantable temperature transponders (Bio Medic Data Systems, DE, USA).
Intranasal inoculation (10 ferrets per virus) using 106 50% tissue culture infectious doses (TCID50) of virus diluted in sterile phosphate-buffered saline (PBS) (0.5-ml total volume) was performed under anesthesia, induced by intramuscular administration of a ketamine-xylazine-atropine mixture (25 mg/kg of body weight, 2 mg/kg, and 0.05 mg/kg, respectively). Oseltamivir phosphate was kindly provided by F. Hoffmann-La Roche, Ltd. (Basel, Switzerland). Ten ferrets infected with the same virus were divided into three treatment groups (placebo control, n = 4; oseltamivir, 5 mg/kg/dose, n = 3, and 25 mg/kg/dose, n = 3). Placebo (15% fructose solution in PBS) or oseltamivir, diluted in PBS supplemented with 15% fructose, was administered orally starting at 4 h postinfection and continued twice daily for 5 consecutive days (total, 11 regimens). The 5-mg/kg dose yields a systemic drug exposure equivalent to that of the recommended 75-mg dose (twice daily) of oseltamivir treatment in human adults (29).
Nasal-wash collection and sample processing.Nasal washes were collected daily (under anesthesia) for 10 days postinfection (p.i.) by flushing nostrils with 1 ml (total) of sterile PBS and further processed for determination of infectious viral titers using a TCID50 assay (limit of detection, 1.6 log10 TCID50/ml). Inflammatory-cell counts and total protein concentrations were measured as previously described (22). Pyrosequencing was performed to detect NA changes at the R292, E119, and I222 positions as previously described (22).
Statistics.Statistical analyses were performed using the GraphPad Prism 5 program (GraphPad Software, CA, USA), and the statistically significant difference level was set at an α value equal to 0.05 (P < 0.05). The unpaired two-tailed Student t test was applied to evaluate the statistical significance between viral titers and other analyzed parameters of placebo-treated and oseltamivir-treated animals. The total amount of virus shedding was estimated by calculating the area under the curve (AUC) using the virus titers determined at different days p.i.
RESULTS
Effect of oseltamivir treatment on H7N9 virus replication and signs of disease in ferrets.The nasal washes from infected ferrets were collected daily to assess the effect of oseltamivir treatment on virus replication and inflammatory responses. In placebo-treated animals, WT, R292K, E119V, and I222K viruses replicated to comparably high titers in the upper respiratory tract (URT), with peak titers ranging from 5.7 to 6.4 log10 TCID50/ml (Fig. 1). Overall, the durations of viral shedding were similar in the placebo-treated groups, regardless of the virus used. A minor difference was observed on day 7, when two of four I222K virus-infected ferrets were positive for virus shedding while only one animal from each of the placebo-treated groups shed virus. No virus shedding was detected beyond 7 days p.i. in all groups.
Effect of oseltamivir treatment on viral replication in the URT of WT (A and B), R292K (C and D), E119V (E and F), and I222K (G and H) virus-infected ferrets. (A, C, E, and G) Placebo versus 5 mg/kg/dose. (B, D, F, and H) Placebo versus 25 mg/kg/dose. Animals (n = 3 or 4/group) were intranasally inoculated with 106 TCID50, and nasal washes were collected daily for 10 days. Only two animals or fewer shed virus at 7 days p.i. *, statistically significant difference. The lower limit of detection was 1.6 log10 TCID50/ml. The error bars indicate standard deviations.
Oseltamivir treatment had no effect on the duration of virus shedding. Nevertheless, a modest dose-dependent antiviral effect was detected in the WT-infected animals. The low dose of oseltamivir reduced viral titers by 0.7 to 1.1 log10 TCID50 at 3 (P = 0.033), 4 (P = 0.004), and 5 (P = 0.006) days p.i. (Fig. 1A), while the higher oseltamivir dose led to 0.9 to 1.5 log10 TCID50 reduction on 4 days—days 1 (P = 0.037), 3 (P = 0.004), 4 (P = 0.001), and 5 (P = 0.005) (Fig. 1B). Conversely, there was no statistically significant reduction for the R292K virus titers at any time points (Fig. 1C and D) (P > 0.19 for both drug doses). We observed a slight decrease in E119V virus titers after oseltamivir treatment; however, the reduction was not statistically significant (P > 0.11 for both doses) (Fig. 1E and F). Small but statistically significant differences between I222K virus titers of drug-treated and placebo-treated animals were detected at 5 and 6 days p.i. Specifically, we detected a 1.4 log10 TCID50 reduction on day 5 for animals given the lower dose of oseltamivir (P < 0.0001) (Fig. 1G) and a 0.5 to 1.6 log10 TCID50 reduction on days 5 (P = 0.006) and 6 (P = 0.049) for those that received the higher drug dose (Fig. 1H).
Next, we measured the individual AUCs to determine the reduction in overall viral titers in treated versus nontreated animals. The results were consistent with reduced replication of WT virus due to the oseltamivir treatment; 12% (P = 0.017) and 17% (P = 0.002) reductions with the lower and higher oseltamivir doses, respectively (Table 1). The replication of the I222K virus showed an 9 to 11% decrease in the presence of oseltamivir (P < 0.04 for both the lower and higher doses). A much smaller difference between the placebo and oseltamivir groups was observed for the E119V (7% reduction) and R292K (0 to 7% reduction) viruses. Overall, the AUCs did not differ substantially between the two oseltamivir doses for all viruses.
Effects of oseltamivir treatment on viral replication and inflammatory response in A/Taiwan/1/2013 (H7N9)-infected ferrets
The ferrets exhibited only mild signs of disease in this study, despite the use of a high inoculation dose, and this result is consistent with our previous observations (22). Although sporadic sneezing was noted, no other observable clinical signs were seen, including changes in body weight and temperature (data not shown), in any of the animals, regardless of the virus used and the treatment group.
Effect of oseltamivir treatment on virus-induced inflammation in the ferret upper respiratory tract.The inflammatory response was assessed by measuring the number of inflammatory cells and total protein concentrations in ferret nasal washes collected at 0, 1, 3, 5, 7, and 9 days p.i. The upregulation patterns of these two parameters in placebo-treated animals were similar to that in our previous report (22). The inflammatory-cell counts, presumably consisting of sloughed-off epithelial cells and resident and recruited immune cells, such as dendritic cells, macrophages, neutrophils, and lymphocytes, in nasal washes peaked at 5 days p.i. with WT, 3 days p.i. with R292K, 5 days p.i. with E119V, and 7 days p.i. with I222K virus (Fig. 2). Both the lower and higher oseltamivir doses significantly reduced the numbers of inflammatory cells at 7 and 9 days p.i. with WT virus (P < 0.02 for all) (Fig. 2A). At 7 days p.i., a small yet statistically significant reduction (P = 0.041) in inflammatory-cell counts was also observed after treatment with the lower, but not the higher, oseltamivir dose in I222K virus-infected animals (Fig. 2D).
Effect of oseltamivir treatment on inflammatory-cell counts in the URT of WT (A), R292K (B), E119V (C), and I222K (D) virus-infected ferrets. Animals (n = 3 or 4/group) were intranasally inoculated with 106 TCID50, and nasal washes were collected daily for 10 days. *, statistically significant difference. The average value at 0 days p.i. from placebo-treated animals was used as a baseline. The error bars indicate standard deviations.
Elevated protein concentrations were detected in nasal washes of ferrets infected with WT or E119V virus (Fig. 3A and C), peaking at 7 days p.i. (∼2.5-fold higher than baseline). In contrast, infection with the R292K and I222K viruses did not elevate protein levels. Oseltamivir treatment, particularly the higher dose, significantly lowered total protein concentrations in nasal washes of WT virus-infected ferrets on days 5 and 7 (Fig. 3A) (P < 0.05 for all). A reduction of protein concentrations in oseltamivir-treated animals infected with E119V virus was found (Fig. 3C); however, the difference was not statistically significant (P > 0.05). Overall, infection with WT or E119V virus induced comparable degrees of inflammation, which were higher than that seen in R292K- or I222K-infected animals. Elevated inflammatory parameters upon WT virus infection were significantly reduced in response to oseltamivir treatment; conversely, no statistically significant changes were detected in oseltamivir-treated animals infected with any of the 3 NA variant viruses.
Effect of oseltamivir treatment on total protein concentration in the URT of WT (A), R292K (B), E119V (C), and I222K (D) virus-infected ferrets. Animals (n = 3 or 4/group) were intranasally inoculated with 106 TCID50, and nasal washes were collected daily for 10 days. *, statistically significant difference. The average value at 0 days p.i. from placebo-treated animals was used as a baseline. The error bars indicate standard deviations.
Emergence of oseltamivir resistance.To assess the propensity for resistance to emerge in H7N9 virus under drug pressure in the ferret model, we extracted viral RNA from nasal washes of WT virus-infected animals collected at 3 to 7 days p.i. and analyzed NA sequences at residues 292, 119, and 222 by pyrosequencing. In the group treated with the lower oseltamivir dose, 1 of 3 animals shed virus containing R292K at 6 days p.i. However, no shedding of the virus was detected on the following days (Table 2). At the higher oseltamivir dose, 2 of 3 ferrets shed the R292K virus; it was detected at 5 and 6 days p.i. in 1 animal and at 6 and 7 days p.i. for the second animal. No shedding was detected beyond 7 days p.i. Emergence of E119V and I222K viruses was not detected in any collected nasal washes.
Detection of R292K substitution in the upper respiratory tracts of WT-infected and oseltamivir-treated ferretsa
DISCUSSION
Our study provides the first experimental evidence for a therapeutic effect of oseltamivir in ferrets infected with A(H7N9) virus. Although the drug effect was modest, significant decreases in both virus replication and inflammatory response were detected in the URT of ferrets infected with the Taiwan/1 WT virus. Further, the results suggest that each NA amino acid substitution studied here affects not only susceptibility in the NI assay, but also the outcome of oseltamivir treatment in ferrets. The therapeutic effect of oseltamivir was diminished in ferrets infected with the R292K variant. In a previous study, we demonstrated resistance of Taiwan/1 R292K virus to oseltamivir treatment in mice (30). This result was consistent with the outcome of another study in which mice infected with a reassortant H7N9 virus carrying the NA gene from A/Shanghai/1/2013 with the R292K substitution developed illness despite NAI treatment (31).
Both E119V and I222K NA variants were characterized as having reduced inhibition by oseltamivir; however, the fold increases in the IC50 were different: 84-fold for E119V and 32-fold for I222K. In the present study, there was a reduction in the therapeutic effect of oseltamivir following infection with either of the two viruses. However, the viral titers in oseltamivir-treated ferrets infected with I222K virus were statistically lower at 5 days p.i. We did not test the I222R NA variant in ferrets; however, based on comparable IC50s (22) and structural/biochemical similarities to K at this position, 222R is also likely to produce similar effects on oseltamivir treatment in ferrets. Therefore, all substitutions detected in the virus isolated from the Taiwanese patient should be considered markers of clinically relevant oseltamivir resistance.
Current options for therapeutic treatment of H7N9 influenza virus infection are limited due to universal resistance of these viruses to M2 blockers (32). Oseltamivir is the most widely available NAI, and emergence of resistance to the drug poses a problem for clinical care. If a virus carrying R292K, E119V, I222K, or I222R was detected in a clinical setting, which NAI, if any, would be best for treatment? The E119V virus was normally inhibited by zanamivir, peramivir, and laninamivir (22), although the zanamivir IC50 increased by 9-fold. Virus with I222K or I222R showed a 6- to 14-fold increase in the IC50s of the three NAIs and thus may not be fully susceptible to this class of drugs. Moreover, the R292K substitution caused highly reduced inhibition by peramivir and reduced inhibition by zanamivir and laninamivir. Notably, viral lung titers of R292K virus-infected mice were significantly reduced when treated with zanamivir (31); this effect was limited in oseltamivir-treated mice. It is unknown whether treatment of ferrets with zanamivir would produce a similar outcome. Molecular markers for zanamivir resistance are not well established, which makes it challenging to postulate the resistance phenotypes for this NAI.
The R292K substitution in H7N9 viruses is the most commonly reported marker associated with highly reduced inhibition by oseltamivir in the NI assay (32, 33). R292K has been detected 1 to 9 days after initiating treatment (3, 4, 32). It is therefore not surprising that we detected the emergence of R292K in WT virus-infected animals treated with oseltamivir 5 days after the first treatment. The rapid emergence of viruses with R292K in treated ferrets correlates well with the multiple reports on this NA variant in treated human patients (34). Because viral shedding did not last long, a more appropriate model, such as immunocompromised ferrets (35), is needed to investigate whether prolonged treatment would result in selection of an R292K-dominant virus population in the host and/or emergence of additional molecular markers. We found that a higher oseltamivir dose appeared to increase the emergence of this variant NA, a finding consistent with a previous study demonstrating the emergence of oseltamivir-resistant seasonal influenza virus in human volunteers receiving the highest drug dose (21). Notably, when the treatment concluded, the presence of R292K declined gradually, and the majority of viruses with this substitution were undetectable by 7 days p.i. Similarly, it was shown in a drug-free competitive-mixture study that WT H7N9 virus overgrew the R292K variant analyzed in nasal washes of infected ferrets (36).
Detection of resistance (i.e., presence of R292K) in this study was identified only if ≥50% of the virus population had R292K. Nevertheless, the emergence of NA variants in oseltamivir-treated ferrets in this study occurred more frequently in H7N9-infected animals than in similar studies conducted with the H1N1pdm09 virus (37) and highly pathogenic avian influenza H5N1 viruses (38). The high titers produced by H7N9 viruses and the reduced efficacy of oseltamivir provide favorable conditions for resistance emergence. It must be taken into consideration that, because we used a plaque-purified clone of the virus isolate recovered from the treated patient, the presence of a very minor subpopulation with the R292K in the WT virus population cannot be completely ruled out. This condition could expedite the emergence of resistance in ferrets infected with the WT clone and treated with oseltamivir. Analysis of the virus inoculum by more sensitive methods, such as deep sequencing (39, 40), would offer an essential verification of the virus stock purity.
ACKNOWLEDGMENTS
We thank Ho-Sheng Wu, Feng-Yee Chang, and Ming-Tsan Liu from the Taiwan CDC for sharing the A/Taiwan/1/2013 (H7N9) virus isolate. We acknowledge Peter Eworonsky for excellent assistance with animal care. Oseltamivir phosphate was kindly provided by F. Hoffmann-La Roche, Ltd. (Basel, Switzerland).
Funding was made available through the U.S. CDC Influenza Division with partial funding by the interagency agreement between the Biomedical Advanced Research and Development Authority (BARDA) and the U.S. CDC.
The findings and conclusions of this report are those of the authors and do not necessarily represent the views of the funding agency.
We declare that we have no conflict of interest.
FOOTNOTES
- Received 8 December 2014.
- Accepted 23 February 2015.
- Accepted manuscript posted online 4 March 2015.
- Address correspondence to Larisa V. Gubareva, lgubareva{at}cdc.gov.
Citation Marjuki H, Mishin VP, Chesnokov AP, De La Cruz JA, Davis CT, Villanueva JM, Fry AM, Gubareva LV. 2015. Neuraminidase mutations conferring resistance to oseltamivir in influenza A(H7N9) viruses. J Virol 89:5419–5426. doi:10.1128/JVI.03513-14.
REFERENCES
- Copyright © 2015, American Society for Microbiology. All Rights Reserved.
