BRIEF REPORT • JID 2004:190 (1 November) • 1627

B R I E F R E P O R T

Influenza Viruses Resistantto the Antiviral Drug Oseltamivir:Transmission Studies in Ferrets

M. Louise Herlocher,1 Rachel Truscon,1 Stephanie Elias,1 Hui-Ling Yen,1

Noel A. Roberts,2 Suzanne E. Ohmit,1 and Arnold S. Monto1

1Department of Epidemiology, School of Public Health, University of Michigan,Ann Arbor; 2Roche Discovery, Welwyn Garden City, Herts, United Kingdom

Three type A influenza viruses, each of which has a distinctneuraminidase-gene mutation and is resistant to the neur-aminidase inhibitor oseltamivir, have been isolated. Previ-ously, in the ferret model, an R292K mutant of a type A(H3N2) virus was not transmitted under conditions in whichthe wild-type virus was transmitted. This model was used toinvestigate whether the E119V mutant of a type A (H3N2)virus and the H274Y mutant of a type A (H1N1) virus wouldbe transmitted under similar circumstances. Both mutantviruses were transmitted, although the H274Y mutant re-quired a 100-fold-higher dose for infection of donor ferretsand was transmitted more slowly than was the wild type.Both the mutant and the wild-type viruses retained theirgenotypic characteristics.

The neuraminidase (NA) inhibitors, oseltamivir and zanamivir,

are a newer class of antiviral agent licensed for the prevention

and treatment of type A and type B influenza virus infections

[1–3]. The 2 NA inhibitors differ structurally, resulting in os-

eltamivir being orally bioavailable whereas zanamivir is not;

however, both bind to the catalytic site of viral NA and com-

petitively inhibit this key function [1].

As with other antiviral agents, the emergence of influenza

viruses with reduced susceptibility to the NA inhibitors is pos-

sible during treatment with either oseltamivir or zanamivir. In

clinical studies, 3 known mutations in the NA gene of influenza

Received 13 February 2004; accepted 10 May 2004; electronically published 28 September2004.

Presented in part: Options for the Control of Influenza V, Okinawa, Japan, 7–11 October2003 (abstract W05–1).

Financial support: Hoffmann–La Roche.A.S.M. has received research support from Hoffmann–La Roche and has served as a

consultant to Roche and GlaxoSmithKline.Reprints or correspondence: Dr. Arnold S. Monto, Dept. of Epidemiology, School of Public

Health, 109 Observatory, University of Michigan, Ann Arbor, MI 48109 (asmonto@umich.edu).

The Journal of Infectious Diseases 2004; 190:1627–30� 2004 by the Infectious Diseases Society of America. All rights reserved.0022-1899/2004/19009-0014$15.00

type A viruses with phenotypic expression of reduced suscep-

tibility have been identified after exposure to oseltamivir [4–

6]. The more common R292K (argininerlysine) substitution

occurs at the functional site of the N2 NA; the E119V (glutamic

acidrvaline) substitution occurs within the framework of the

N2 NA; and the H274Y (histidinertyrosine) substitution oc-

curs within the N1 NA and interacts with a functional site.

Previous studies have suggested that these mutant viruses may

have some degree of compromised infectivity and transmissi-

bility, compared with the respective wild-type viruses [7–9,];

however, the degree of compromise may vary by mutation.

In ferrets, a model of influenza transmission has been estab-

lished in which influenza infection in donors can be transmitted

to susceptible contacts [10, 11]; ferrets provide the best animal

model of human influenza, given the 2 species’ similarity in both

receptor type (ferrets can be infected without virus adaptation)

and the course of infection [12]. In a previous study of the

transmission of the R292K mutation, we reported that infection

by and transmission of influenza virus with this mutation does

not occur under conditions in which the wild-type virus readily

infects and is transmitted [13]. We now report the results of

similar studies, in which infection and transmission of the E119V

and the H274Y variants are evaluated in ferrets.

Materials and methods. Influenza viruses representing the

distinct NA mutations were isolated during clinical studies em-

ploying oseltamivir as treatment. One of these viruses was A/

Wuhan-like (H3N2), which had the E119V mutation and was

isolated during a clinical trial against naturally acquired infec-

tions during the 1997–98 influenza season; the second was A/

New Caledonia–like (H1N1), which had the H274Y mutation

and was isolated during a clinical treatment trial during the

2000–01 season. Paired wild type (wt) and mutant (MUT) strains,

isolated from each study subject’s pretreatment and last post-

treatment positive nasal-swab specimens, respectively, were

originally isolated in primary monkey kidney cell culture and

were amplified and plaque purified in MDCK cells, and the

NA genotype was determined; stock wt and MUT viruses for

each pair were provided by Hoffmann–La Roche, and no fur-

ther amplification was performed.

Before they were included in the transmission studies, 6-

week-old ferrets (Marshall Farms) were confirmed to be se-

ronegative (titer, !1:10) for the respective influenza viruses with

which they subsequently were infected, by hemagglutination-

inhibition assay. Before comparative transmission studies, chal-

lenge doses were determined by minimum-infectious-dose ti-

tration of both wt viruses (i.e., A/Wuhan and A/New Caledonia)

at Swinburne U

niversity of Technology on Septem

ber 5, 2014http://jid.oxfordjournals.org/

Dow

nloaded from

1628 • JID 2004:190 (1 November) • BRIEF REPORT

Table 1. Comparisons of infectivity and transmissibility of wild-type (wt) and mutant (MUT) virus pairs, for2 neuraminidase (NA) genotypes isolated during studies of ferrets treated with oseltamivir.

Pretreatment specimenand last positive specimenfrom same subject

Infectious dose,dilution of stock

Donor ferrets’infection status(i.e., infectivity)

Recipient ferrets’infection status

(i.e., “transmissibility”)

Sequenceconfirmation

of NA genotype

A/Wuhan-like (H3N2)wt (E119) 1.0 � 10�6a 4 of 4 11 of 11 wtMUT (E119V) 1.0 � 10�6b 4 of 4 11 of 11 MUT

A/New Caledonia–like (H1N1)wt (H274) 1.5 � 10�6c 4 of 4 12 of 12 wtMUT (H274Y)

Initial inoculation 1.5 � 10�5d 0 of 4 (at day 7) 0 of 12 (at day 7)Rechallenge, day 7 1.5 � 10�3d 4 of 4 12 of 12 MUT

a 6.0 log10 TCID50.b 6.2 log10 TCID50.c 7.0 log10 TCID50.d 6.0 log10 TCID50.

in ferrets. In brief, for each wt virus, donor ferrets were chal-

lenged with a single dose (inoculated intranasally) of diluted

(duplicate dilutions 10�5, 10�6, and 10�7) stock wt virus received

from Hoffmann–La Roche. The stock preparations were un-

diluted harvest from amplification in MDCK cells with known

titers (TCID50 values). For each wt virus, each of 6 donor ferrets

(2 per dilution) was housed with 2 susceptible contact ferrets,

and the maximal dilution that resulted in transmission to both

recipient ferrets was used in the studies comparing the trans-

mission of wt virus to that of MUT virus.

For each wt -MUT pair in the comparative studies, 2 groups

(1 wt and 1 MUT) of 4 donor ferrets each were inoculated

intranasally with 0.5 mL of specified challenge doses (equivalent

dilutions) of either wt virus or MUT virus (0.25 mL/nostril),

and each donor ferret was immediately housed with 3 suscep-

tible contact ferrets. wt virus– and MUT virus–infected ferrets

were kept in separate rooms. The course of infection in donor

(14 days) and recipient (17 days) ferrets was monitored by

daily assessment of temperature (normal temperature for ferrets

is 101.5�F; temperatures �102.8�F were considered to indicate

fever) and collection of nasal washes, which were tested im-

mediately by rapid influenza diagnostic tests (Directigen Flu A;

Becton Dickinson); recipient ferrets were removed from cages

when they tested as influenza positive, to minimize recipient-

to-recipient transmission. Nasal washes were also batch pro-

cessed for determination, by hemagglutination assay, of virus

titer in MDCK cells [13]. Virus titers for each influenza-positive

time point were calculated by the Spearman-Karber equation;

virus titers in wt virus– and MUT virus–infected ferrets at days

1–14 after infection were compared by analysis of the area un-

der the curve (AUC), by Student’s t test [13].

Seroconversion of donor and recipient ferrets was considered

to have occurred when there was a �4-fold increase in antibody

titer between preinfection and day 28 after infection, as deter-

mined by standard hemagglutination-inhibition assay; both the

proportion of donor and recipient ferrets with �4-fold in-

creases in titers and the mean fold titer increases were calculated

and compared for wt -MUT virus pairs, by x2 test and Student’s

t test, respectively. Virus isolates from nasal washes were con-

firmed as being either wt or MUT, by sequence analysis of the

NA-gene product, performed at Professional Genetics Labo-

ratories (now Capio Diagnostics) in Uppsala, Sweden.

Results. A/Wuhan-like (E119V) virus: As shown in table 1,

all donor ferrets infected with either wt virus or MUT virus (4

ferrets each) were positive for either wt virus or MUT virus,

respectively, with no infection-time delay in those exposed to

MUT virus (i.e., donor ferrets were positive at day 2 after in-

oculation, in both the wt group and the MUT group); likewise,

all contact ferrets exposed to either wt virus– or MUT virus–

infected donor ferrets were positive for either wt virus or MUT

virus, respectively, with no transmission-time delay in those

exposed to MUT virus (i.e., contact ferrets were positive at day

4 after inoculation of donor ferrets, in both the wt-virus group

and the MUT-virus group). All viruses isolated from donor and

recipient ferrets were confirmed, by sequence analysis, as being

the appropriate wt genotype or MUT genotype. Fever occurred

sporadically in both wt -virus and MUT-virus donor and re-

cipient ferrets, being detected at least once in 3 (75%) of 4 wt-

virus donor ferrets and in 5 (45%) of 11 wt-virus recipient

ferrets, compared with 2 (50%) of 4 MUT-virus donor ferrets

and 1 (9%) of 11 MUT-virus recipient ferrets; the number of

days of fever in ferrets exposed to wt virus was greater than

that in ferrets exposed to MUT virus (�2 days vs. 1 day, re-

spectively; for all comparisons). All wt-virus and MUT-P 1 .05

virus donor and recipient ferrets seroconverted to A/Wuhan

(H3N2), with similar mean fold increases in antibody titer

(table 2); also, the mean AUC values did not significantly differ

(1) between donor and recipient ferrets, either in the wt-virus

group or in the MUT-virus group, or (2) between wt-virus and

at Swinburne U

niversity of Technology on Septem

ber 5, 2014http://jid.oxfordjournals.org/

Dow

nloaded from

BRIEF REPORT • JID 2004:190 (1 November) • 1629

Table 2. Mean fold increases, in hemagglutinin (HA) titer, in-dicating seroconversion to infecting influenza virus, and meanarea under the curve (AUC), for wild-type (wt) and mutant (MUT)donor and recipient ferrets.

A/Wuhan-likeE119V

A/New Caledonia–likeH274Y

Fold increase in HA titera

wtDonor ferrets 7.3 (2.5) 6.0 (0.8)Recipient ferrets 7.7 (1.3) 6.4 (0.8)

MUTDonor ferrets 6.3 (1.0) 6.0 (1.2)Recipient ferrets 7.5 (0.9) 5.8 (0.6)

AUCb

wtDonor ferrets 17.4 (4.4) 20.5 (6.4)Recipient ferrets 17.0 (4.1) 12.4 (2.0)

MUTDonor ferrets 16.8 (2.5) 13.1 (2.7)Recipient ferrets 17.6 (4.0) 12.1 (2.2)

NOTE. Data are mean (SD).a Between preinfection serologic specimens and 28-days-after-infection se-

rologic specimens, determined by standard HA-inhibition assay.b Calculated from viral titers obtained on day 1–day 14 after inoculation of

donor ferrets.

MUT-virus donor ferrets and wt-virus and MUT-virus recipient

ferrets.

A/New Caledonia–like (H274Y) virus: All 4 donor ferrets in-

fected with wt virus were positive for wt virus at day 2 after

inoculation; in contrast, at day 7 after inoculation, 0 of 4 donor

ferrets exposed to a similar titer of MUT virus were positive

for MUT virus (table 1); at day 7 after inoculation, the latter

4 donor ferrets were rechallenged with a 100-fold-higher dose

of MUT virus, and all 4 had productive infections with MUT

virus, with a slight delay in time (day 3 after inoculation) to

productive infection. All contact ferrets exposed to donor fer-

rets receiving the original dose of wt virus and all contact ferrets

exposed to donor ferrets receiving the increased dose of MUT

virus after rechallenge had productive infections with either wt

virus or MUT virus, respectively. However, for those exposed

to MUT virus, time to infection was slightly delayed (positive

at day 5, 6, or 7 after inoculation of MUT-virus donor ferrets,

compared with day 4 or 5 for wt-virus recipient ferrets). Again,

both wt virus and MUT virus shed by donor and recipient

ferrets maintained their characteristics when sequence analysis

was performed. Fever occurred sporadically in both wt-virus

and MUT-virus donor and recipient ferrets and was detected

at least once in all (100%) of the 4 wt-virus donor ferrets and

in 6 (50%) of 12 wt-virus recipient ferrets, compared with 2

(50%) of 4 MUT-virus donor ferrets and 7 (58%) of 12 MUT-

virus recipient ferrets; the number of days of fever in ferrets

exposed to wt virus was similar to that in ferrets exposed to

MUT virus (1 or 2 days; for all comparisons). All wt-P 1 .05

virus and MUT-virus donor and recipient ferrets seroconverted

to A/New Caledonia (H1N1), with similar mean fold increases

in antibody titer (table 2); also—and as in the case of the results

for A/Wuhan (E119V)—the mean AUC values did not signif-

icantly differ (1) between the donor and recipient ferrets, in

either the wt-virus group or the MUT-virus group, or (2) be-

tween wt-virus and MUT-virus donor ferrets and wt-virus and

MUT-virus recipient ferrets (table 2).

Discussion. Viruses resistant to the NA inhibitors have

been infrequently isolated from individuals treated with osel-

tamivir and have been documented only once in treatment with

zanamivir, in an immunocompromised individual [14]. The

highest frequency of isolation of viruses resistant to oseltamivir

occurred in the placebo-controlled pediatric trial of oseltamivir

in young children [6]. In that study, 10 (5.5%) of the treated

children shed virus with an NA mutation (virus from 8 of these

individuals had the R292K mutation, virus from 1 had the

E119V mutation, and virus from 1 had the H274Y mutation)

[6]. This finding is in sharp contrast to the situation in which

the M2 inhibitors, amantadine and rimantadine, are used in

treatment, where ∼30% of isolates have mutations known to

be associated with phenotypic expression of resistance [4, 15].

The M2 inhibitor–resistant viruses are also known to be fully

fit—that is, to infect and transmit in a manner similar to that

of the wild type viruses [4, 15].

Not only has resistance been less frequently encountered with

the NA inhibitors, but the consequences of that resistance may

be of less concern [4]. At least some of the resistant mutants

are less fit, with reduced ability to replicate and transmit [5, 7,

8, 13]. In a previous study employing the ferret model and

using methods identical to those used in the present report,

infection and transmission of the R292K mutant (A/Sydney-

like [H3N2]) did not occur under conditions in which the wild

type from which it was derived was readily transmitted [13].

In contrast, the results in the present study indicate that the

E119V mutant was transmitted as was the H274Y mutant; how-

ever, the latter required a 100-fold-higher challenge dose to

infect donor ferrets and was transmitted more slowly than was

the wt. In both cases, the transmitted MUT virus retained the

amino acid substitution associated with resistance. Interesting-

ly, in previous studies, using cell culture and animal models

and not involving transmission, the H274Y mutant has been

shown to be compromised in its replicative ability [8]. No

comparable data on the E119V mutant are available; however,

other mutations at amino acid position 119 are known to have

an adverse effect on the stability of NA [5, 9].

What are the consequences of the finding that some NA

inhibitor–resistant variants can be transmitted, at least in the

ferret model? The most common variant, the R292K-mutant

virus, was impaired in both infectivity and transmissibility [13];

this mutation is at the functional site of NA, which would

at Swinburne U

niversity of Technology on Septem

ber 5, 2014http://jid.oxfordjournals.org/

Dow

nloaded from

1630 • JID 2004:190 (1 November) • BRIEF REPORT

indicate substantial impairment of virus fitness [5]. In contrast,

the mutation in the least-impaired virus (E119V) occurs within

the framework of the N2 NA and therefore is not directly

involved in enzyme activity. If such viruses are transmitted, it

is uncertain whether, over time, they could predominate over

susceptible strains; however, because such emergence could be

an unpredictable consequence in countries—such as Japan—

that may use the largest amounts of drugs, the Neuraminidase

Inhibitor Susceptibility Network has been established, to mon-

itor, in collaboration with the World Health Organization and

on a population scale, whether such an emergence is occurring

[4]. Continued isolation of resistant variants at low frequency

remains likely, especially in treated immunocompromised in-

dividuals [14]. For these reasons, systematic monitoring must

be continued.

References

1. Gubareva LV, Kaiser L, Hayden FG. Influenza virus neuraminidaseinhibitors. Lancet 2000; 355:827–35.

2. Hayden FG, Osterhaus ADME, Treanor JJ, et al. Efficacy and safety ofthe neuraminidase inhibitor zanamivir in the treatment of influenzavirus infections. N Engl J Med 1997; 337:874–80.

3. Treanor JJ, Haydne FG, Vrooman PS, et al. Efficacy and safety of theoral neuraminidase inhibitor oseltamivir in treating acute influenza.JAMA 2000; 283:1016–24.

4. Zambon M, Hayden FG. Position statement: global Neuraminidase In-hibitor Susceptibility Network (NISN). Antiviral Res 2001; 49:147–56.

5. McKimm-Breschkin JL. Resistance of influenza viruses to neuramini-dase inhibitors—a review. Antiviral Res 2000; 47:1–17.

6. Whitley RJ, Hayden FG, Reisinger KS, et al. Oral oseltamivir treatmentof influenza in children. Pediatr Infect Dis J 2001; 20:127–33.

7. Carr J, Ives J, Kelly L, et al. Influenza virus carrying neuraminidasewith reduced sensitivity to oseltamivir carboxylate has altered prop-erties in vitro and is compromised for infectivity and replicative abilityin vivo. Antiviral Res 2002; 54:79–88.

8. Ives JAL, Carr JA, Mendel DB, et al. The H274Y mutation in the influenzaA/H1N1 neuraminidase active site following oseltamivir phosphate treat-ment leave virus severely compromised both in vitro and in vivo. AntiviralRes 2002; 55:307–17.

9. Colacino JM, Chirgadze NY, Garman E, et al. A single sequence changedestabilizes the influenza virus neuraminidase tetramer. Virology 1997;236:66–75.

10. Herlocher ML, Truscon R, Fenton R, et al. Assessment of developmentof resistance to antivirals in the ferret model of influenza virus infec-tion. J Infect Dis 2003; 188:1355–61.

11. Herlocher ML, Elias S, Truscon R, et al. Ferrets as a transmission modelfor influenza: sequence changes in HA1 of type A (H3N2) virus. J In-

fect Dis 2001; 184:542–6.

12. Reuman PD, Keely S, Schiff GM, James N. Assessment of signs ofinfluenza illness in the ferret model. J Virol Methods 1989; 24:27–34.

13. Herlocher ML, Carr J, Ives J, et al. Influenza virus carrying an R292Kmutation in the neuraminidase gene is not transmitted in ferrets. An-tiviral Res 2002; 54:99–111.

14. Jackson HC, Roberts N, Wange ZM, Belshe R. Management of influ-enza: use of new antivirals and resistance in perspective. Clin DrugInvestig 2000; 20:447–54.

15. Hayden FG, Sperber SJ, Belshe RB, Clover RD, Hay AJ, Pyke S. Re-covery of drug-resistant influenza A virus during therapeutic use ofrimantadine. Antimicrob Agents Chemother 1991; 35:1741–7.

at Swinburne U

niversity of Technology on Septem

ber 5, 2014http://jid.oxfordjournals.org/

Dow

nloaded from