Generation and Characterization of Recombinant = Influenza A (H1N1) Viruses Harboring Amantadine Resistance = Mutations

Abstract

The emergence of = resistance to=20 amantadine in influenza A viruses has been shown to occur rapidly = during=20 treatment as a result of single-amino-acid substitutions at = position 26,=20 27, 30, 31, or 34 within the transmembrane domain of the = matrix-(M)-2=20 protein. In this study, reverse genetics was used to generate and=20 characterize recombinant influenza A (H1N1) viruses harboring = L26F, V27A,=20 A30T, S31N, G34E, and V27A/S31N mutations in the M2 gene. In = plaque=20 reduction assays, all mutations conferred amantadine resistance, = with drug=20 concentrations resulting in reduction of plaque number by 50%=20 (IC₅₀s) 154- to 3,300-fold higher than those seen for = the wild=20 type (WT). M2 mutants had no impairment in their replicative = capacities in=20 vitro on the basis of plaque size and replication kinetics = experiments. In=20 addition, all mutants were at least as virulent as the WT in=20 experimentally infected mice, with the highest mortality rate = being=20 obtained with the recombinant harboring a double V27A/S31N = mutation. These=20 findings could help explain the frequent emergence and = transmission of=20 amantadine-resistant influenza viruses during antiviral pressure = in the=20 clinical setting.

Influenza A viruses are still associated = with=20 significant rates of morbidity and mortality during winter = epidemics in=20 both industrial and developing countries. These viruses are also=20 responsible for occasional pandemics with a devastating impact on = human=20 health and global economy (6). Therefore, the control of influenza A virus = infections=20 remains a major international clinical goal. The adamantane = compound=20 amantadine was the first antiviral shown to inhibit the = replication of=20 influenza A viruses by blocking the ion channel activity of the = matrix=20 (M)-2 protein and preventing viral uncoating (14). Amantadine has been used for over 3 decades = in the=20 treatment and prevention of influenza A virus infections. Its use = is=20 associated with some benefits in terms of duration of illness and = severity=20 of symptoms (6). However, the emergence of drug-resistant = viruses has been=20 shown to occur rapidly during treatment as a result of = single-amino-acid=20 substitutions at position 26, 27, 30, 31, or 34 within the = transmembrane=20 domain of the M2 protein (9, 12). It has been also reported that two or more M2 = mutant=20 virus variants could be recovered from amantadine-treated children = and=20 immunocompromised patients (2, 9, 12). A recent report showed that H3N2-infected = individuals=20 shed more amantadine-resistant variants than their H1N1 = counterparts (11). Moreover, specific amantadine-resistance = mutations seem=20 to be subtype dependent, with the S31N and V27A mutations = predominating in=20 H3N2 and H1N1 viruses, respectively (11).

Drug-resistant M2 mutants of the H3N2 = subtype do=20 not show impaired growth characteristics or altered virulence in = ferrets=20 (13). However, to our knowledge, the effects of M2 = mutations=20 on the growth properties and virulence of H1N1 viruses have not = been=20 studied. Furthermore, a systematic evaluation of various M2 = mutations by=20 use of a uniform genetic background (i.e., in recombinant viruses) = has not=20 been performed.

The aim of the present study = was to use=20 a reverse genetics system to assess the effects of various single = and=20 double mutations within the transmembrane domain of the M2 protein = of=20 influenza A (H1N1) viruses on the resistance phenotype to = amantadine, in=20 vitro replication kinetics, and virulence in = mice.

MATERIALS AND METHODS

Plasmids.

A set of plasmids allowing the rescue of = the=20 recombinant influenza A/WSN/33 (H1N1) virus was used for = generating all M2=20 mutants (3). The reverse genetics system includes eight = influenza=20 virus RNA-coding transcription plasmids (pPOLI-PA, -PB1, -PB2, = -NP, -HA,=20 -NA, -M, and -NS) and polymerase and nucleoprotein expression = plasmids=20 (pCAGGS-PA, -PB1, -PB2, -NP) kindly provided by Peter Palese = (Mount Sinai=20 School of Medicine, New York, N.Y.). The M2 protein of influenza = A/WSN/33=20 virus already contains an asparagine residue at position 31 = conferring=20 natural resistance to amantadine (5). Thus, the original pPOL-M plasmid was first = mutated to a=20 susceptible variant (the so-called wild type [WT]) by = incorporating an=20 N31S substitution by use of appropriate primers and a QuickChange=20 site-directed mutagenesis kit (Stratagene, La Jolla, Calif.). The = resulted=20 pPOLM_31S was then used for incorporation of single = mutations,=20 i.e., L26F, V27A, A30T, and G34E, whereas the original=20 (pPOL-M_31N) plasmid was used for incorporation of the=20 additional V27A mutation to generate the double mutant=20 V27A/S31N.

Rescue of recombinant = viruses.

To generate the recombinant viruses, 1 = =CE=BCg each of the=20 12 plasmids was first transfected into 293T (human embryonic = kidney) cells=20 followed by viral amplification in Madin-Darby canine kidney cells = essentially as reported previously (3). The M2 genes of all recombinant viruses were = sequenced=20 after reverse transcription-PCR amplification from cell culture=20 supernatants to verify the presence of the desired = mutations.

Determination of = amantadine-resistance=20 phenotype and replication capacities.

Susceptibility of = recombinant=20 viruses to amantadine was determined using a plaque reduction = assay in=20 Madin-Darby bovine kidney (MDBK) cells. The number of plaques = formed at=20 different drug concentrations was recorded, and the drug = concentration=20 resulting in reduction of plaque number by 50% (IC₅₀) = was=20 calculated. The replicative capacity of the different recombinant = viruses=20 was evaluated on the basis of the virus plaque size generated in = the=20 absence of drug and by replication kinetics experiments. Briefly, = MDBK=20 monolayers were infected with recombinant viruses at a = multiplicity of=20 infection (MOI) of 0.001 and supernatants were collected at 12, = 24, 36,=20 48, and 72 h postinfection and assayed for numbers of PFU in MDBK=20 cells.

Experimental infection. =

The virulence of the = recombinant=20 viruses was evaluated in 7- to 8-week-old female BALB/c mice = (Charles=20 River Canada, Lasalle, Quebec, Canada). The animals were housed = three per=20 cage and kept under conditions which prevented cage-to-cage = infections.=20 Groups of 16 mice were inoculated intranasally, under isoflurane=20 anesthesia, with 10³ PFU of WT or mutant recombinant = viruses in=20 40 =CE=BCl of phosphate-buffered saline (PBS). One group of mice = was inoculated=20 with phosphate-buffered saline and served as sham-infected = negative=20 control. Mice were monitored daily during 14 days postinoculation = for=20 clinical signs (loss of body weight, abnormal behavior, and = ruffled fur)=20 and mortality. The deaths were recorded daily and the mean day of = death=20 (MDD) was determined.

Statistical = evaluations.=20

In animal studies, = mortality rates=20 were analyzed by using the Fisher's exact test whereas the = percentages of=20 weight loss and MDDs were analyzed by using the nonparametric = Mann-Whitney=20 U rank-sum test. Virus titers generated in the in vitro = replication=20 kinetics experiments were analyzed by Student's t test = with=20 Satterthwaite correction. A P value of = =E2=89=A40.05 was=20 considered significant.

RESULTS

Generation of = influenza A=20 (H1N1) virus recombinant M2 mutants.

As part of this study, = seven=20 recombinant influenza A viruses, including the WT, five single = mutants=20 (L26F, V27A, A30T, S31N, and G34E), and one double (V27A/S31N) = mutant,=20 were rescued in the WSN/33 (H1N1) background. Sequencing of the M = genes=20 from recovered viruses revealed that all M2 mutations were = correctly=20 incorporated and that no additional changes were = present.

Susceptibility to amantadine = of=20 recombinants.

As shown in Table 1,=20 all single- and double-recombinant M2 mutants exhibited resistance = to=20 amantadine when assessed by plaque reduction assay. The = IC₅₀=20 values for these mutants were 154- to 3,300-fold higher than that = for the=20 recombinant WT virus. The IC₅₀ values for the double=20 (V27A/S31N)-recombinant mutant were 10-fold and 2-fold lower than = those=20 for the V27A and S31N single mutants, respectively.

TABLE=20 1.

In vitro = and in=20 vivo characteristics of recombinant influenza A (H1N1) M2=20 mutants

Replicative = capacity and=20 virulence of recombinants.

Recombinant mutants = generated plaques=20 of sizes similar to those seen with the recombinant WT, with the = exception=20 of L26F and A30T mutants, whose plaque size was slightly reduced = (by=20 1.5-fold) compared to the results seen with the WT (Fig. 1).=20 Also, all recombinants showed similar replication kinetics in a = 72-hour in=20 vitro experiment, with no significant differences in virus titers = at each=20 time point (Fig. 2).

FIG.=20 1.

Virus = plaque size=20 of recombinant WT A/WSN/33 (H1N1) virus and various M2 = mutants.=20 Recombinant WT and M2 mutant (L26F, V27A, A30T, S31N, G34E, = and=20 V27A/S31N) viruses generated in the supernatant of = transfected 293T=20 cells were propagated in 6-well plates=20 (more ...)

FIG.=20 2.

Replication=20 kinetics of recombinant WT A/WSN/33 (H1N1) virus and various = M2=20 mutants. MDBK cells were infected with the recombinant WT = and M2=20 mutant (L26F, V27A, A30T, S31N, G34E, and V27A/S31N) viruses = at an=20 MOI of 0.001. At the indicated times postinfection,=20 (more ...)

An inoculum of = 10³ PFU=20 resulted in clinical signs of illness (body weight loss, reduced = vitality,=20 and ruffled fur) in mice; these results appeared by day 4 = postinoculation=20 with all tested recombinants. As shown in Table 1,=20 there were no significant differences in the mortality rates of = the=20 different M2 mutants compared to the WT results, with the = exception of the=20 double mutant (V27A/S31N), which was associated with an increased=20 mortality rate (87.5 versus 50% for the WT). Similarly, the double = mutant=20 and the S31N mutant were associated with increased weight loss on = day 5=20 compared to the WT results. The lowest MDD (4.71 days) was = obtained with=20 the double mutant (V27A/S31N), whereas there were no significant=20 differences between the results for the WT and those for the = remaining=20 mutant groups (Table 1).

DISCUSSION

Despite the development of a new class of=20 anti-influenza virus agents, the neuraminidase inhibitors, use of=20 amantadine and its congener rimantadine still constitutes an = important=20 antiviral strategy for the control of influenza A virus epidemics; = this=20 strategy would have even more significance in the context of an = eventual=20 pandemic. Indeed, amantadine has an excellent inhibitory effect on = different subtypes of influenza A viruses in addition to its low = cost,=20 good chemical stability, and excellent bioavailability. On the = other side,=20 a major concern has been the rapid emergence of = amantadine-resistant=20 viruses. Indeed, despite the fact that naturally occurring=20 amantadine-resistant influenza A viruses are rarely found about = (0.8% in=20 the general population [16]), rates of resistance exceeding 30% and up to = 80% have=20 been found in H3N2 viruses after only a few days of therapy in = both=20 immunocompetent and immunocompromised subjects (6, 9, 12). Moreover, drug-resistant M2 mutants (H3N2 = subtype) are=20 transmissible and appear to retain good replication capacities and = virulence in ferrets (13).

For unknown reasons, amantadine resistance = appears to=20 emerge less frequently during amantadine therapy in influenza A = (H1N1)=20 viruses than in A (H3N2) strains (11). In a study in Japan, the most frequent M2 = mutation=20 conferring amantadine resistance in H1N1 viruses occurred at = position 27=20 (V27A) whereas the S31N mutation predominated in the H3N2 subtype = (11). Other drug-resistant M2 mutations at = positions 26=20 (L26F), 31 (S31N), and 34 (G34E) were recovered from H1N1 clinical = isolates at a lower frequency (11, 12). In addition, a few uncommon M2 mutations such = as A29V=20 and I43V were also detected in H1N1 viruses recovered from=20 amantadine-treated children, although the role of these mutations = in drug=20 resistance remains to be confirmed (12).

In this=20 study, we aimed to assess the effects of known amantadine = resistance M2=20 mutations on drug resistance, in vitro replicative capacities, and = virulence in an influenza A (H1N1) virus (A/WSN/33) by use of a = reverse=20 genetics system. The rescue of our mutants from the same genetic=20 background precluded the confounding effects of gene constellation = and=20 additional changes in other genes. A total of seven recombinant = viruses,=20 including the WT and five single and one double M2 mutants, were = rescued.=20 As expected, all mutants exhibited high-level resistance to = amantadine,=20 with =E2=89=A5100-fold reductions (range, 154- to 3,300-fold) in = susceptibility=20 compared to the WT results (Table 1).=20 Interestingly, the levels of resistance of the double V27A/S31N = mutant=20 were lower than that of each of the corresponding single mutants = by=20 factors of 10- and 2-fold, respectively (Table 1).=20 Thus, the combination of these two M2 mutations did not confer a=20 synergistic effect in term of resistance. The differences seen = between the=20 different mutants in the levels of resistance could be related to = the=20 variable effects of these mutations on the M2 protein activity as=20 previously described (4). For example, M2 substitutions at positions 26, = 30, 31,=20 and 34 were found to decrease the M2 activity of influenza=20 A/chicken/Germany/34 (H7N1) virus whereas substitutions at = position 27=20 increased such activity (4). Furthermore, the I27T mutation was shown to = suppress the=20 attenuating effect of the S31N mutation on M2 activity in the = double=20 (I27T/S31N) mutant (4). Thus, although the M2 activity of our = recombinants was=20 not determined, it is likely that the different mutations = evaluated in=20 this study had different effects on the levels of M2 protein=20 activity.

There was almost no impact of the different = M2=20 mutations on viral fitness, as assessed on the basis of the size = of viral=20 plaques generated in the absence of drug and of virus titers = following=20 MDBK infection at a low MOI (Fig. 1=20 and 2).=20 These findings are in agreement with a previous report showing = that=20 influenza A/H3N2 recombinant viruses can undergo multiple cycles = of=20 replication without M2 ion channel activity (15). In that study, recombinants containing the M2 = gene of=20 A/Udorn/307/72 (H3N2) with V27T, A30P, and S31N mutations were = found to=20 replicate as efficiently as the recombinant WT in Madin-Darby = canine=20 kidney cells. Also, a recombinant mutant which had no detectable = M2 ion=20 channel activity due to a deletion in the transmembrane domain of = the=20 protein (M2-del_29-31) exhibited replication efficiency = in vitro=20 similar to that of the WT virus (15). Similarly, the virulence of our recombinant = M2 mutants=20 was at least as high as that of the WT virus when evaluated in=20 experimentally infected BALB/c mice (Table 1).=20 Of note, the inoculation of mice with viruses harboring the = V27A/S31N=20 mutations resulted in a significantly lower MDD and higher = mortality rate=20 compared to the recombinant WT virus results. In addition, both = the double=20 mutant and the single S31N mutant were associated with increased = weight=20 loss compared to the WT virus. It is therefore possible that the = S31N=20 mutation could confer an advantageous virulence impact for the = A/WSN/33=20 virus, which is a natural S31N variant. Of interest, the same = mutation has=20 also been recently reported in highly pathogenic H5N1 viruses = infecting=20 birds and humans in Asia (10). Other amantadine- or rimantadine-resistant = human and=20 avian influenza viruses have also shown a virulence pattern = similar to=20 those seen with their respective WT viruses in various animal = models (1, 7, 13).

In=20 conclusion, the results of our study suggest that in similarity to = what=20 has been reported with viruses of other subtypes, including H3N2, = H2N2,=20 and H5N2 viruses (1, 7, 13), M2 mutations conferring resistance to = amantadine do not=20 compromise the replicative capacity or virulence of H1N1 viruses. = Such=20 results support the clinical relevance of M2 mutants, particularly = for=20 immunocompromised patients, with whom amantadine-resistant = influenza=20 viruses have been associated with fatal infections (2, 9). However, the reason(s) that such = amantadine-resistant=20 influenza A virus variants have not been associated with a major = epidemic=20 remain(s) to be clarified. Finally, the effects of M2 mutations = are in=20 sharp contrast with those of neuraminidase mutations associated = with=20 resistance to neuraminidase inhibitors, which were found to = compromise=20 more severely the viral fitness and virulence of influenza A = viruses (8). Altogether, these results may explain the = difference in=20 the frequency of drug resistance between these two classes of=20 anti-influenza virus agents in the clinic.

Acknowledgments

This study was supported by the = Canadian=20 Institutes of Health Research and GlaxoSmithKline Canada.

We are grateful to Peter Palese = (Mount Sinai=20 School of Medicine, New York, N.Y.) and Ervin Fodor (Sir William = Dunn=20 School of Pathology, University of Oxford, Oxford, United Kingdom) = for=20 providing viral plasmids.