1521-0111/90/3/188–198$25.00 http://dx.doi.org/10.1124/mol.116.105346MOLECULAR PHARMACOLOGY Mol Pharmacol 90:188–198, September 2016Copyright ª 2016 by The American Society for Pharmacology and Experimental Therapeutics

Pharmacological Characterization of the Spectrum of AntiviralActivity and Genetic Barrier to Drug Resistance of M2-S31NChannel Blockers

Chunlong Ma, Jiantao Zhang, and Jun WangDepartment of Pharmacology and Toxicology, College of Pharmacy, and BIO5 Institute, University of Arizona, Tucson, Arizona

Received May 26, 2016; accepted June 30, 2016

ABSTRACTAdamantanes (amantadine and rimantadine) are one of the twoclasses of Food andDrugAdministration–approved antiviral drugsused for the prevention and treatment of influenza A virusinfections. They inhibit viral replication by blocking the wild-type(WT) M2 proton channel, thus preventing viral uncoating. How-ever, their use was discontinued due to widespread drug re-sistance. Among a handful of drug-resistant mutants, M2-S31N isthe predominant mutation and persists in more than 95% ofcurrently circulating influenza A strains. We recently designed twoclasses of M2-S31N inhibitors, S31N-specific inhibitors andS31N/WT dual inhibitors, which are represented by N-[(5-cyclopropyl-1,2-oxazol-3-yl)methyl]adamantan-1-amine (WJ379)and N-[(5-bromothiophen-2-yl)methyl]adamantan-1-amine(BC035), respectively. However, their antiviral activities against

currently circulating influenza A viruses and their genetic barrier todrug resistance are unknown. In this report, we evaluated thetherapeutic potential of these two classes of M2-S31N inhibitors(WJ379 and BC035) by profiling their antiviral efficacy againstmultidrug-resistant influenza A viruses, in vitro drug resistancebarrier, and synergistic effect with oseltamivir. We found that M2-S31N inhibitors were active against several influenza A virusesthat are resistant to one or both classes of Food and DrugAdministration–approved anti-influenza drugs. In addition, M2-S31N inhibitors display a higher in vitro genetic barrier to drugresistance than amantadine. The antiviral effect of WJ379 wasalso synergisticwith oseltamivir carboxylate. Overall, these resultsreaffirm that M2-S31N inhibitors are promising antiviral drugcandidates that warrant further development.

IntroductionInfluenza viruses are the causative agents that lead to

annual seasonal influenza epidemics as well as sporadic, moredevastating influenza pandemics (Rappuoli and Dormitzer,2012; Monto and Webster, 2013). On average, influenza virusinfection leads to 36,000 deaths in the United States andapproximately 250,000–500,000 deaths globally each year(Thompson et al., 2003, 2004). However, current countermea-sures against influenza infection are curtailed by the emer-gence of drug-resistant strains. Resistance to the only orallybioavailable neuraminidase inhibitor, oseltamivir, has beencontinuously reported and was prevalent in the 2008–2009influenza season (Cheng et al., 2009; Samson et al., 2013). Theuse of M2 channel blockers (amantadine and rimantadine) isno longer recommended by the Centers for Disease Controland Prevention due to prevalent drug resistance (Bright et al.,2006; Fiore et al., 2008). Thus, novel antivirals are clearlyneeded. Toward this goal, we chose M2-S31N as the drug

target to develop the next generation of antiviral drugs. TheM2-S31N mutant is an ideal antiviral drug target because: 1)M2-S31N persists in more than 95% of currently circulatinginfluenza A viruses among humans, including oseltamivir-sensitive and oseltamivir-resistant strains (Dong et al., 2015),thus targeting M2-S31N is likely to yield broad-spectrumantiviral drugs; and 2) M2 is a validated antiviral drug target,and there is no such homolog proton channel encoded by thehuman genome, thus M2 inhibitors are expected to have highselectivity. As a homotetrameric proton-selective channel, onemutation in M2 actually results in four changes at the sametime, which leads to a profound impact on the structure andfunction of this very constricted channel. As a result, only avery limited number of M2 mutants confer both drugresistance and transmissibility among humans, of whichM2-S31N is the predominant mutant. This small set oftransmissible mutants suggests that M2 is a highly conserveddrug target compared with other viral proteins, rendering itan ideal drug target for the development of anti-influenzadrugs (Wang et al., 2015).M2-S31N mutant was traditionally tagged as an “undrug-

gable” target, and decades of traditional medicinal chemistrycampaign failed to yield a single hit compound (Wanka et al.,2013; Wang et al., 2015). Nevertheless, guided by information

This work was supported by the National Institutes of Health NationalInstitute of Allergy and Infectious Diseases [Grant AI119187], the PhRMAFoundation 2015 Research Starter Grant in Pharmacology and Toxicology, andthe start-up funds from the University of Arizona and the BIO5 Institute.

dx.doi.org/10.1124/mol.116.105346.

ABBREVIATIONS: BC035, N-[(5-bromothiophen-2-yl)methyl]adamantan-1-amine; CPE, cytopathic effect; DMEM, Dulbecco’s modified Eagle’smedium; FBS, fetal bovine serum; MDCK, Madin-Darby canine kidney; MOI, multiplicity of infection; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide; PBS, phosphate-buffered saline; PCR, polymerase chain reaction; ST6Gal I, b-galactoside a-2,6-sialytransferase I;TEVC, two-electrode voltage clamp; WJ379, N-[(5-cyclopropyl-1,2-oxazol-3-yl)methyl]adamantan-1-amine; WT, wild-type.

188

at ASPE

T Journals on February 21, 2022

molpharm

.aspetjournals.orgD

ownloaded from

gathered from previous structure-activity relationship studiesof amantadine (Wang et al., 2015) and breakthroughs in M2structural biology and mechanistic studies (Hong andDeGrado, 2012), steady progress has been made toward thedesign of M2-S31N inhibitors (Wang et al., 2013a,b; Wu et al.,2014; Li et al., 2016). They are broadly classified as S31N-specific inhibitors and S31N/wild-type (WT) dual inhibitors.S31N-specific inhibitors target the S31N channel only, not theWT, and one representative example is N-[(5-cyclopropyl-1,2-oxazol-3-yl)methyl]adamantan-1-amine (WJ379) (Wang et al.,2013b). S31N/WT dual inhibitors target both S31N andWT M2 channels, and one representative example is N-[(5-bromothiophen-2-yl)methyl]adamantan-1-amine (BC035) (Wuet al., 2014). Their channel blockage and antiviral effects havebeen confirmed in electrophysiological two-electrode voltageclamp (TEVC) assays and antiviral plaque assays, respec-tively. Their modes of binding were revealed by both solutionand solid-state NMR studies (Cady et al., 2010, 2011; Wanget al., 2013b; Williams et al., 2013; Wu et al., 2014).In this study, we further evaluated the therapeutic potential

of both classes of M2 inhibitors, S31N-specific inhibitors andS31N/WT dual inhibitors. Specifically, from the standpoint ofdeveloping therapeutics, we are interested in testing whether thepotent antiviral effect of these compounds on the model virus,A/WSN/33 (H1N1), can be extended to the currently circulatinginfluenza A strains among humans. Furthermore, with these toolcompounds in hand, we would like to address the issue of drugresistance. For example, will the S31N mutant virus becomeresistant to the newly developed S31N inhibitors under drugselection pressure? If yes, which mutations will be selected? Howmany passages does it take for the S31N mutant virus to evolvedrug resistance? In addition, as we already learned the lessons ofantibiotics andantivirals, resistance is unfortunately inevitable. Itis not a question of yes or no but rather a question of when (ClavelandHance, 2004;HaydenanddeJong, 2011;McKimm-Breschkin,2013; Walsh and Wencewicz, 2014; Blair et al., 2015). With thisinmind, what one can do is to slow down the resistance evolutionrather than eradicate it. Toward this goal, we would like to testthe combination therapy potential of M2-S31N inhibitors withoseltamivir, as this offers a means to delay resistance evolutionagainst both oseltamivir and M2-S31N inhibitors.

Materials and MethodsCell Lines, Viruses, and Viral Infection. Madin-Darby canine

kidney (MDCK) cells were grown at 37°C in 5% CO2 atmosphere inDulbecco’s modified Eagle’s medium (DMEM; high glucose, with L-glu-tamine) supplemented with 10% fetal bovine serum (FBS), 100 IU/mlpenicillin, and 100 mg/ml streptomycin. MDCK cells overexpressingb-galactoside a-2,6-sialytransferase I (ST6Gal I) were obtained from Dr.Yoshihiro Kawaoka at the University of Wisconsin (Madison, WI)through material transfer agreement and were maintained in thepresence of 7.5 mg/ml puromycin, except when they were used for viralinfection. Influenza A virus strains A/California/07/2009 (H1N1) andA/Texas/04/2009 (H1N1) were obtained from Dr. James Noah at theSouthern Research Institute (Birmingham, AL), and influenzaA virusstrains A/Denmark/524/2009 (H1N1) and A/Denmark/528/2009 (H1N1)were obtained fromDr.ElenaGovorkova atSt. JudeChildren’sResearchHospital (Memphis, TN). Virus stocks were amplified in MDCK cells inthe presence of 2 mg/ml N-acetyl trypsin. Two days post infection, theculture media were harvested, and cell debris was removed by centrifu-gation at 3000 rpm for 30 minutes. Virus titers were determined byplaque assay using MDCK cells expressing ST6Gal I.

Compounds. WJ379 and BC035 were synthesized in house aspreviously described (Wang et al., 2013b;Wu et al., 2014). Amantadinehydrochloride was purchased from Sigma-Aldrich (St. Louis, MO).

Plaque Assay. Plaque assays were carried out as previouslydescribed (Jing et al., 2008; Balannik et al., 2009), except MDCK cellsexpressing ST6Gal I were used instead of regular MDCK cells. In brief,a confluent monolayer of ST6Gal I MDCK cells was incubated with∼100-pfu virus samples in DMEMwith 0.5% bovine serum albumin for1 hour at 4°C, then 37°C for 1 hour. The inoculums were removed, andthe cells were washed with phosphate-buffered saline (PBS). The cellswere then overlaidwithDMEMcontaining 1.2%Avicelmicrocrystallinecellulose (FMC BioPolymer, Philadelphia, PA) and N-acetyl trypsin(2.0 mg/ml). To examine the effect of the compounds on plaque formation,the overlay media were supplemented with compounds at testing concen-trations. At day 2 after infection, the monolayers were fixed and stainedwith crystal violet dye solution (0.2% crystal violet, 20% methanol).

Cytotoxicity Assay and Cytopathic Effect Assay. The cyto-toxicity of compounds was determined in MDCK cells. Cells weretreated with different concentrations of WJ379 or BC035 and stainedwith 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide(MTT) for 4 hours. A virus-induced cytopathic effect (CPE) assaywas performed as previously described (Atkins et al., 2012; Beyleveldet al., 2013). In brief,MDCK cells were grown inDMEMsupplementedwith 10% FBS in 96-well plates. When 90% confluence was reached,cells were washed once with 100 ml of PBS buffer and infected withvirus at multiplicity of infection (MOI) of 0.002. Then DMEM (withoutFBS) with indicated concentrations of compounds was added. Afterincubating at 37°C for 72 hours, 20 ml of 5 mg/ml MTT reagent wasadded to each well, and the plates were incubated at 37°C for 4 hours.Then the cells were washed once with PBS, and 100 ml of 0.1 N HCl inisopropanol was added. The absorbance was read at 560 nm with aplate reader. The assays were repeated in quadruplicate.

Serial Passage Experiments and Resistance MutationIdentification. MDCK cells were infected with A/WSN/33 (H1N1)virus or A/WSN/33 N31S (H1N1) at an MOI of 0.001 for 1 hour. Thenthe inoculum was removed, and MDCK cells were incubated withcompounds at concentrations shown in Table 2 or in the absence ofcompounds to evaluate cell culture–adapted mutations. In the firstpassage, the applied compound concentrations were slightly belowtheir corresponding EC50 values. In the following passages, thecompound concentrations were gradually increased, as shown inTable 2. In each passage, the viruses were harvested when significantcytopathic effect was observed, which usually takes 2–3 days aftervirus infection. The titers of harvested viruses were determined byplaque assay. At selected passages, as shown in Table 2, influenza Msegments were subjected to sequencing. Influenza viral RNA wasisolated using QIAamp viral RNA Mini Kit (Qiagen, Hilden, Ger-many). TheM segment cDNAwas generatedwith reverse-transcriptionpolymerase chain reaction (PCR) using M-specific primers(59-AGCAAAAGCAGGTAGATATTGAAAG-39and59-TAGTTTTTTACTC-CAGCTCTATGTTG-39). Reverse-transcription PCR products wereanalyzed with agarose gel electrophoresis and purified with WizardSV Gel and PCR Clean-up System (Promega, Madison, WI). PurifiedPCR products were sequenced by Genewiz, Inc. (South Plainfield, NJ).

Electrophysiological TEVC Assay. The corresponding drug-resistantmutations identified fromserial drug-passage experimentswereintroduced toM2 sequences in pGEM3 vector with the QuikChange Site-Directed Mutagenesis Kit (Agilent, Santa Clara, CA) as describedpreviously (Khurana et al., 2009).mRNApreparation andmicroinjectionand electrophysiological TEVC recordingswere carried out as previouslydescribed (Jing et al., 2008). The inhibition kinetics time constant wasdetermined by fitting the period of recording trace when the compoundsarepresent to aone-phase exponential decayequation [f(t)5A*exp(2t/G)1C] in Clampfit 10.3 (Molecular Devices, Sunnyvale, CA).

Combinational Therapy. Combination evaluations were per-formed in quintuplicate using the CPE assay described earlier exceptthat neutral red was used for staining instead ofMTT. In brief, MDCKcells were seeded in 96-well plates for 24 hours. Seven serial

Pharmacology of M2-S31N Inhibitors 189

at ASPE

T Journals on February 21, 2022

molpharm

.aspetjournals.orgD

ownloaded from

half-logarithmic dilutions of WJ379 and oseltamivir carboxylate wereprepared and added to MDCK cells infected with influenza A virusA/WSN/33 (H1N1) (MOI 0.001) in the 96-well plate. After incubationat 37°C in a humidified incubator with 5% CO2 for 72 hours, themedium was removed, and cells were stained with 100 ml of neutralred in DMEM (40 mg/l) for 4 hours. After washing once with 150 ml ofPBS, the neutral red was extracted from cells with 150 ml of neutralred destain solution [ethanol/water/glacial acetic acid: 50/49/1 (v/v/v)]to form a homogenous solution (Repetto et al., 2008). The absorbanceat 540 nm was measured using a Multiskan FC Microplate Photom-eter (Fisher Scientific, Houston, TX). Drug-drug interaction wasevaluated using the MacSynergy II software program, which auto-matically calculates volumes of synergy or antagonism for each three-dimensional plot of data (Prichard and Shipman, 1990; Tarbet et al.,2012). The value ranges 0–25, 25–50, 50–100, or over 100 mm2% (mm�mm � %) are considered as insignificant, minor but significant,moderate, or strong synergy or antagonism, respectively. Synergyplots were made at the 95% confidence limit.

M2 Sequence Analysis. All M2 protein full-length sequences ofinfluenza virus A isolated from humans were obtained from theInfluenzaResearchDatabase (http://www.fludb.org, accessed onJanuary18, 2016). A total of 18,393M2 sequences were aligned byMEGA6 (www.megasoftware.net) using the ClustalW method. The frequencies of theappearance of substitutions of L26I and I32T were identified.

ResultsWJ379 and BC035 Are Potent M2-S31N Channel

Blockers and Antivirals against S31N Containing theA/WSN/33 (H1N1) Virus. We previously reported on twoclasses of S31N inhibitors, S31N-specific inhibitors (Wanget al., 2013b; Li et al., 2016) and S31N/WT dual inhibitors(Wang et al., 2013a; Wu et al., 2014), which are represented byWJ379 and BC035, respectively. To select noncytotoxic drugconcentrations for the following serial viral passage experi-ments, the cytotoxicity and selectivity indexes of these twocompounds were determined. The CC50 values of WJ379 andBC035 were 125 and 123 mM, respectively (Table 1). Theirantiviral activities were further quantified in both plaque assaysand cytopathic effect assays using the influenza A virus stainA/WSN/33 (H1N1) (Table 1). Both the plaque assay and CPEassay were standard antiviral assays used to determine theefficacy of antiviral drugs. The difference is that the plaque assayquantifies the change of viral titer upondrug treatment,whereasthe CPE assay quantifies the change of cell survival upon viralinfection and drug treatment (Atkins et al., 2012; Beyleveldet al., 2013). In general, there is a positive correlation betweenthese twoassays. In the plaque assays, theEC50 values ofWJ379and BC035 against the A/WSN/33 (H1N1) virus were 0.49 and2.2mM, respectively, whichwere consistent with those ofWJ379and BC035 in the CPE assays (0.36 and 6.20 mM, respectively)(Table 1). The selective indexes of WJ379 and BC035 were255 and 56, respectively. On the basis of these results, thehighest drug concentrations used for the following serial viralpassage experiments were set to 5 and 16 mM for WJ379 andBC035, respectively, which conferred minimal cytotoxicity.M2-S31N Inhibitors Are Potent Antivirals against

Human Clinical Isolates of Multidrug-Resistant In-fluenza A Viruses. To determine whether the potent antivi-ral activity of WJ379 and BC035 on A/WSN/33 (H1N1) can beextended to other S31N-containing influenza A viruses, WJ379and BC035 were tested in plaque assays against humanclinical isolates of multidrug-resistant influenza A viruses,including A/California/07/2009 (H1N1) (amantadine-resistant,

oseltamivir-sensitive), A/Texas/04/2009 (H1N1) (amantadine-resistant, oseltamivir-resistant), A/Denmark/524/2009 (H1N1)(amantadine-resistant, oseltamivir-sensitive), andA/Denmark/528/2009 (H1N1) (amantadine-resistant, oseltamivir-resistant).As these human clinical isolates did not produce countableplaques using regular MDCK cells (data not shown), anengineeredMDCK cell line overexpressing the human ST6GalI was used for this purpose as it expresses N-acetyl sialic acidlinked to galactose by an a-2,6-linkage on its cell surface(Hatakeyama et al., 2005). This cell line mimics humanepithelial cells and is the only cell line that is suitable toperform plaque assays for human influenza viruses(Matrosovich et al., 2003; Hatakeyama et al., 2005). With thiscell line and optimized conditions, we were able to get uniformand reproducible plaques with all four strains. Representa-tive images of plaque assays with A/California/07/2009(H1N1) and A/Texas/04/2009 (H1N1) are shown in Fig. 1.The antiviral EC50 values of compound WJ379 were in thesubmicromolar range, whereas for compoundBC035, the EC50

values were in the low micromolar range (Fig. 1, D and E). Ofnote, oseltamivir carboxylate only reduced plaque size, but notplaque number, whereas M2-S31N inhibitors completelyinhibited plaque formation.M2-S31N Inhibitors Have a Higher In Vitro Genetic

Barrier to Drug Resistance than Amantadine. To eval-uate the in vitro genetic barrier to drug resistance of these twoS31N inhibitors, serial viral passage experiments in MDCKcells were performed with the influenza A/WSN/33 (H1N1)virus. The drug concentrations of BC035 and WJ379 appliedat passage 1 were slightly below their EC50 values and weregradually increased in subsequent passages (Table 2). In thefirst two passages, the progeny viruses maintained fullsensitivity to their respective compounds, as indicated bythe EC50 values (Table 2). At passage 3, the EC50 values ofWJ379 and BC035 were moderately increased by 4.2- and 2.8-fold, respectively. Significant resistance was observed atpassage 4 in which the EC50 values increased more than

TABLE 1Channel blockage and antiviral activity of M2-S31N inhibitors

Structure

M2-S31N channelinhibitiona

85%/65%b 76%/N.T.c

Plaque assayEC50 (mM)d

0.49 2.2

CPE EC50 (mM)e 0.36 6.20Cellular toxicity

CC50 (mM)f125 123c

SIg 255 56c

N.T., not tested; SI, selectivity index.aM2-S31N channel inhibition was measured using two-electrode voltage clamp

assay as described (Wang et al., 2013b; Wu et al., 2014). The data are expressed aspercentage inhibition at 100/30 mM compound concentration.

bReported in Wang et al. (2013b).cReported in Wu et al. (2014).dEC50 in plaque assays was determined with the A/WSN/33 (H1N1) virus.eEC50 in CPE assay was determined with the A/WSN/33 (H1N1) virus using MTT.fCC50 was assayed using MTT after incubating MDCK cells with increasing

concentrations of compounds for 72 hours.gSI was calculated by dividing CC50 values over plaque assay EC50 values.

190 Ma et al.

at ASPE

T Journals on February 21, 2022

molpharm

.aspetjournals.orgD

ownloaded from

10-fold in both cases. To test the persistence of the resultingdrug-resistant mutants, the viruses were continuously pas-saged for two more rounds in the presence of compounds, andthen five rounds in the absence of compounds. We found thatthe viruses at passage 11 were resistant to both WJ379 andBC035.To rule out the possibility that the observed resistance was

due to cell culture–adapted mutations, the A/WSN/33 (H1N1)virus was also passed in the absence of compound, and thedrug sensitivities of the resulting viruses against WJ379 andBC035 were tested in plaque assays. At passage 6, the EC50

values of WJ379 and BC035 were 0.65 and 2.9 mM, respec-tively, which were similar to their corresponding EC50 valuesat passage 0, indicating that the elevated EC50 values in serialpassage experiments in the presence of WJ379 and BC035were not due to the virus adaption in cell culture.As a comparison, we carried out a serial viral pas-

sage experiment with the A/WSN/33 N31S (H1N1) virusagainst amantadine. The A/WSN/33 N31S (H1N1) virus isamantadine-sensitive. The EC50 value of amantadine againstthe A/WSN/33 N31S virus (H1N1) was 0.26 mM. Amantadine

was applied at 0.2 mM in passage 1 (Table 2), which was closeto its EC50 value. This condition was chosen tomake a side-by-side comparison with the passage experiments of WJ379 andBC035. In contrast to the emergence of complete resistance toWJ379 and BC035 by the A/WSN/33 (H1N1) virus at passage4, complete amantadine resistance against the WSN N31S(H1N1) virus was observed as early as passage 1 (30% plaqueremaining even when treated with 30 mMamantadine). Theseresults indicate that M2-S31N inhibitors, such as WJ379 andBC035, have a higher genetic barrier to resistance thanamantadine.Sequencing M2 Genes Reveals Mutations that Confer

Resistance to M2-S31N Inhibitors. To identify how M2-S31N evolved to become resistant to WJ379 and BC035, theM2 genes from viruses at passages 0, 3, 4, 5, 6, and 9 werereverse-transcribed and sequenced. The sequencing traces areshown in Fig. 2, and the resulting M2 mutations at eachpassage are shown in Table 2. For WJ379, no M2 mutationwas identified up to passage 3. At passage 4, a substantialpopulation of L26I mutation emerged, and its percentagecontinued to rise throughout passages 5 and 6 and completely

Fig. 1. Antiviral efficacy of compounds WJ379 and BC035 on clinically isolated influenza A viruses in plaque assay. The assay was carried out withMDCK cells expressing the ST6Gal I gene. (A) Plaque assays of WJ379 on A/California/07/2009 (H1N1). (B) Plaque assays of WJ379 on A/Texas/04/2009(H1N1). (C) Plaque assays of BC035 on A/California/07/2009 (H1N1). (D) Dose-response curves of WJ379 on various influenza A strains. (E) Dose-response curves of BC035 on various influenza A strains.

Pharmacology of M2-S31N Inhibitors 191

at ASPE

T Journals on February 21, 2022

molpharm

.aspetjournals.orgD

ownloaded from

overtook L26 at passage 9. In addition, another mutation,N31S, was also identified at passage 6. However, this mutantreverted back to N31 (WT) at passage 9 when the drugselection pressure was removed. Nevertheless, the viruscontaining the single mutant L26I at passage 9 was stillresistant toWJ379 (Table 2). AsM2-S31N could revert back toM2-N31S (WT) to become resistant to S31N-specific inhibi-tors, such as WJ379, it is expected that S31N/WT dualinhibitors might have a higher genetic barrier to drug re-sistance, as they inhibit both the S31N and the N31S (WT)M2channels. To test this hypothesis, we repeated the serialpassage experiments with BC035, one of the S31N/WT dualinhibitors. Similar to the results of WJ379, no mutation wasidentified up to passage 3. At passage 4, a small but obviouspopulation of N31D mutants emerged, and its percentagecontinued to rise throughout passages 5 and 6. Similarly, theN31D mutation also completely reverted back to N31 atpassage 9 when the drug selection pressure was removed,whereas anothermutation, I32T, which emerged at passage 5,continued to increase throughout passages 6–9 and becamethe predominant mutation at passage 9 by complete replace-ment of I32.M2 Mutants Selected from Serial Passage Experi-

ments Have Reduced Drug Sensitivity to M2-S31NInhibitors in TEVC Assays. In the aforementioned serialpassage experiments, we identified two amino acid substitu-tions, L26I and N31S, for WJ379 and two amino acidsubstitutions, N31D and I32T, for BC035. As we cannotdistinguish whether these two substitutions occur in the sameM2 protein or at two different M2 proteins, we thereforegenerated all possible M2 constructs and tested their drugsensitivity in TEVC assays. Specifically, in the case of WJ379,we recorded the channel blockage of WJ379 against M2-S31N/L26I (single mutation), M2-N31S (single mutation), and M2-N31S/L26I (double mutations) in TEVC assays. In the case ofBC035, we recorded its channel blockage against M2-S31N/I32T (single mutation), M2-N31D (single mutation), and M2-N31D/I32T (double mutations). The TEVC assay is one of thegold-standard assays to study the proton conductance prop-erty of M2 channels as well as their drug inhibition (Pintoet al., 1992; Wang et al., 1993; Pinto and Lamb, 2006).Channel blockage is expressed as percentage of current

inhibition at a 2-minute time point after compound treat-ment. Figure 3 shows the percentage of current inhibitionof the M2 mutants by 100 mM WJ379 or 100 mM BC035.For theM2mutants selected byWJ379, M2-N31Swas nearlycompletely resistant to WJ379, as its conductance wasinhibited less than 10% by 100 mM WJ379 (Dong et al.,2015). Another single mutant, M2-S31N/L26I, remainedsensitive to WJ379, although it was less sensitive than theoriginal M2-S31N. The percentage of current inhibitions forM2-S31N/L26I and M2-S31N were 64% and 82%, respec-tively, in the presence of 100 mM WJ379. Dose-responseexperiments revealed that the IC50 of WJ379 against theM2-S31N/L26I mutant was 72.9 mM, in contrast to 27.9 mMagainst M2-S31N (Fig. 4A). Similar to the results of thesingle mutant M2-N31S, the double mutant M2-N31S/L26Iwas nearly completely resistant to WJ379, showing 11%inhibition by 100 mM WJ379 (Fig. 3A). For the mutantsselected by BC035, 100 mM BC035 had minimal inhibitionagainst the M2-N31D single mutant and the M2-N31D/I32Tdouble mutant (Fig. 3B). However, the M2-S31N/I32T singlemutant remained sensitive to BC035 (Fig. 3B). The IC50 ofBC035 against M2-S31N/I32T was 50.5 mM, only a 2.5-foldincrease compared with the IC50 value of BC035 againstM2-S31N (Fig. 4B).Correlation between Electrophysiological TEVC IC50

Values with Antiviral EC50 Values Reveals the Impor-tance of Drug Inhibition Kinetics. In general, there is apositive correlation between an M2 channel blocker’s electro-physiological IC50 value and its antiviral EC50 value. A morepotent channel blocker normally has higher antiviral activity(Wang et al., 2013b; Li et al., 2016). However, when correlat-ing the IC50 and EC50 values of WJ379 and BC035 against M2mutants, we found different results (Table 3). The singlemutant (S31N/L26I) selected by WJ379 only conferred a 2.7-fold increase in TEVC IC50 values (calculated based on2-minute time points); however, it led to more than a 60-foldincrease in antiviral EC50 value. This discrepancy led us tofurther investigate the kinetics of drug binding in both S31Nand S31N/L26I channels. It is of note that the percentage ofchannel blockage was recorded 2 minutes after compoundtreatment at pH 5.5. An examination of the current-inhibitiontraces revealed that, for S31N channel inhibition, equilibrium

TABLE 2Resistance development under selection pressureEC50 values were measured after each passage.

Passage No. WJ379 EC50 Mutation BC035 EC50 Mutation Ama EC50a

mM mM mM mM mM mM

0 0.49 N31 2.2 N31 0.261 0.30 0.33 N.T. 2.0 1.2 N.T. 0.20 Resb

2 0.60 0.38 N.T. 4.0 2.6 N.T. Not continued after passage 1 due to significant resistance3 1.20 1.6 No mutation 8.0 7.2 No mutation4 2.50 35.9 L26I 16.0 25.3 N31D/I32T5 5.0 28.0 L26I 16.0 27.3 N31D/I32T6 5.0 15.5 N31S/L26I 16.0 N.T. N31D/I32T7 0 N.T. N.T. 0 N.T. N.T.8 0 .30 N.T. 0 20.0 N.T.9 0 .30 L26I (100%) 0 22.7 I32T (100%)10 0 N.T. N.T. 0 N.T. N.T.11 0 .30 N.T. 0 .20 N.T.

Ama, amantadine; Res, rimantadine; N.T., not tested.aA/WSN/33 M2 N31S virus was used.bComplete resistant viruses were selected as shown by the plaque-reduction assays. No complete plaque inhibition was observed even when amantadine was added at

30 mM (30% of plaque remained compared with no drug control).

192 Ma et al.

at ASPE

T Journals on February 21, 2022

molpharm

.aspetjournals.orgD

ownloaded from

Fig. 2. M2gene sequencing traces.M2 genes at passages 0, 3, 4, 5, 6, and 9were sequenced by Sanger sequencing. The regions (residues 18–37) that coverdetected mutations are shown. Mutations are indicated by arrows. Sequencings from WJ379 serial passage experiments are shown in (A–F): WJ379,passage 0 (A); WJ379, passage 3 (B); WJ379, passage 4 (C); WJ379, passage 5 (D); WJ379, passage 6 (E); WJ379, passage 9 (F). Sequencings from BC035serial passage experiments are shown in (G–L): BC035, passage 0 (G); BC035, passage 3 (H); BC035, passage 4 (I); BC035, passage 5 (J); BC035, passage6 (K); BC035, passage 9 (L).

Pharmacology of M2-S31N Inhibitors 193

at ASPE

T Journals on February 21, 2022

molpharm

.aspetjournals.orgD

ownloaded from

was not reached at the 2-minute time point (Fig. 5A) whentested at 100 mM, which means the true IC50 is actually lowerthan the recorded IC50 if enough time is given to achieveequilibrium. In contrast, for M2-S31N/L26I inhibition byWJ379 and M2-S31N/I32T inhibition by BC035, the TEVCIC50 values at 2-minute time points were true IC50 values, asequilibriums were reached before the 2-minute time point(Fig. 5, A and B). Fitting the current-conductance trace ofWJ379 in inhibiting M2-S31N/L26I using a one-phase expo-nential decay equation gave a time constant of 8.5 seconds,which ismuch faster than the 45.3-second time constant in thecase of M2-S31N inhibition by WJ379. Similar results wereobserved for BC035: the time constants for BC035 in inhibit-ing M2-S31N and M2-S31N/I32T were 29.5 and 5.5 seconds,respectively. Another example of slow drug inhibition kineticsis amantadine in inhibiting the WT-M2 channel (Fig. 5C).To get the true IC50 values of amantadine in inhibiting

M2-WT and WJ379 in inhibiting M2-S31N due to their slowbinding kinetics, we chose to record the percentage of currentinhibition at 5- and 10-minute time points instead of theregular 2-minute time point. As shown in Fig. 6A, when theWT channel was treated with amantadine, the IC50 valueswere 12.5, 4.7, and 2.5 mM when the percentage of currentinhibition data were collected at 2-, 5-, and 10-minute timepoints, respectively. A representative TEVC trace of amanta-dine in inhibiting M2-WT at 10 mM was shown in Fig. 6B,which clearly showed different percentage of inhibition at 2-,5-, and 10-minute time points. Similarly, in the case of M2-S31N inhibition byWJ379 (Fig. 6C), the IC50 values were 27.1,8.7, and 4.0 mM when recorded at 2, 5, and 10 minutes,

respectively, and the representative TEVC trace of WJ379 ininhibiting M2-S31N at 10 mM is shown in Fig. 6D. Insummary, for the M2-WT channel inhibition by amantadineand the M2-S31N channel inhibition by WJ379, the IC50

value in TEVC assays at 2-minute time points underesti-mates their true IC50 values. After adjusting the bindingkinetics, there appears to be a good correlation between theIC50 and EC50 fold increase from passage 0 to passage11 (Table 3): the IC50 increased 18.2-fold (calculated basedon 10-minute time points), and the EC50 increasedmore than60-fold. The percentage of current inhibition was typicallyrecorded at 2-minute time points because oocyte cells usedfor the current recording cannot survive under pH 5.5 for anextended period of time. To get data at 5 and 10 minutes,repetitive recordings had to be performed with multipleoocytes.In summary, although M2-S31N/L26I and M2-S31N/I32T

remained sensitive toWJ379 andBC035, respectively, the fastbinding kinetics rendered these two drugs much less effective.The M2-S31N Inhibitor WJ379 Shows a Synergistic

Antiviral Effect with Oseltamivir. One useful strategy toreduce the pace of drug resistance development is combinationtherapy, which has established efficacy in the treatment ofhuman immunodeficiency virus infection (Gulick et al., 1997,1998; De Clercq, 2007). Unlike M2-S31N inhibitors, oseltami-vir is a neuraminidase inhibitor, which interrupts the viralreplication cycle by preventing virus release from infected cellsurfaces. Therefore, oseltamivir is expected to show a syner-gistic effect with M2-S31N inhibitors in combination therapy.To test this hypothesis, we combined WJ379 and oseltamivircarboxylate using the two-drug combination protocol (Hayden,2013; Dunning et al., 2014). The efficacy of oseltamivircarboxylate alone was first determined by CPE assay. TheEC50 of oseltamivir carboxylate against A/WSN/33 (H1N1) inMDCK cell culture was 18.4 nM in the CPE assay. In thecombination therapy, a matrix was created to combine 10–10,000 nM WJ379 with 1–1000 nM oseltamivir carboxylate.Compounds were added to confluent MDCK cells that wereinfected with the A/WSN/33 (H1N1) virus at MOI 0.01. Cellviability was evaluated at 72 hours after infection by neutralred staining. The synergistic effect was calculated withMacSynergy II software (Prichard and Shipman, 1990;Prichard et al., 1993; Smee et al., 2010). As shown in Fig. 7,a region of significant synergy (29.0–42.4 mM2%) was observedfor WJ379 and oseltamivir carboxylate at concentrations of320–1000 and 32–100 nM, respectively. The calculated net

Fig. 4. Dose-response curves ofWJ379 and BC035 onM2 channels in TEVC assay with 2-minute time points. (A)WJ379 dose-response curve. (B) BC035dose-response curve. Each data point represents at least three oocytes.

Fig. 3. Percentage of M2 current inhibition by 100 mM WJ379 or 100 mMBC035 in TEVC assays at 2-minute time points. (A) Percentage of currentconductance inhibition of M2 channels by 100 mM WJ379. (B) Percentageof current conductance inhibition of M2 channels by 100 mM BC035.

194 Ma et al.

at ASPE

T Journals on February 21, 2022

molpharm

.aspetjournals.orgD

ownloaded from

effect across the entire surface was a volume of synergy of 166,indicating a strong synergy.

DiscussionDespite the existence of influenza vaccines and antivirals,

influenza virus infection is still one of the leading causes ofdeath in the United States (Thompson et al., 2003, 2004).Among the two classes of Food and Drug Administration–approved anti-influenza drugs, adamantanes are no longerrecommended, which leaves oseltamivir as the only orally

bioavailable drug on the market. With the continuous pre-scription of oseltamivir, it will only be a matter of time beforepredominant flu strains become resistant to it. The alarmingfact is that oseltamivir-resistant strains have been continu-ously reported, and certain strains appear to adapt to thefitness of transmission among humans, which could lead tothe next influenza pandemic (Hurt et al., 2009; Kelso andHurt, 2012; Hurt, 2014). In addressing this unmet medicalneed, we revisited the M2 proton channel and aim to developthe next generation of antivirals by targeting the M2-S31Nmutant. M2-S31N is the predominant mutant among cur-rently circulating influenza A viruses, including those thatare resistant to oseltamivir (Dong et al., 2015). Thus,targeting M2-S31N is expected to yield broad-spectrumantivirals that are active against both oseltamivir-sensitiveand oseltamivir-resistant strains (Wang, 2015). To test thishypothesis, we chose two M2-S31N inhibitors, WJ379 andBC035, which represent S31N-specific inhibitors and S31N/WT dual inhibitors, respectively, and tested their antiviralefficacy in plaque assays. It was found that WJ379 hadpotent antiviral activity against all four human clinicalisolateswith submicromolar efficacy, including two 2009 pan-demic strains that are resistant to both amantadine andrimantadine (A/Texas/04/2009 and A/Denmark/528/2009).BC035 similarly inhibited both oseltamivir-sensitive andoseltamivir-resistant influenza A strains with low micromo-lar EC50 values. These results highlight the great therapeu-tic potential of the first-in-class M2-S31N inhibitors: theycan serve as the second line of defense should oseltamivir failto confine the next influenza outbreak caused by influenza Astrains.With these tool compounds in hand, we began to investigate

the potential issue of drug resistance. This is critical, asaddressing the efficacy is only the first step in developingantiviral drugs. Ideal antiviral drugs should also bear a highgenetic barrier to drug resistance such that they can be usedfor a longer period of time. The genetic barrier is defined inthis study as the ease of resistance generation, although it isalso defined in other reports as the number of mutationsneeded for resistance (Vingerhoets et al., 2005; Fofana et al.,2013). The first-generation M2 channel blockers amantadineand rimantadine were rendered ineffective due to the emerg-ing M2-S31N mutant. Related to the second generation of M2channel blockers, which are the M2-S31N inhibitors, theconcern is whether the M2-S31N channel will similarlydevelop resistance to S31N inhibitors. To address this, twoS31N inhibitors, WJ379 and BC035, were subjected to serialpassage experiments in an attempt to select resistant viruses.Not surprisingly, resistant strains were selected for bothcompounds at passage 4. In comparison with the genetic

Fig. 5. Representative TEVC recording traces showing different inhibitionkinetics. (A) WJ379 in inhibiting M2-S31N and M2-S31N/L26I channels.WJ379 inhibited the M2-S31N channel much slower than the M2-S31N/L26I channel. (B) BC035 in inhibiting M2-S31N and M2-S31N/I32Tchannels. BC035 inhibited the M2-S31N channel much slower than theM2-S31N/I32T channel. The inhibition kinetics time constant was deter-mined by fitting the period of recording trace when the compounds arepresent to the one-phase exponential decay equation [f(t) = A*exp(2t/G) + C]in Clampfit 10.3. (C) Amantadine (Ama) in inhibiting the M2-WT channel.

TABLE 3Correlations of antiviral EC50 and TEVC IC50 values.

Passage 0 Passage 11 Fold Increase

WJ379TEVC IC50 (mM) 27.1 (2 minutes) (S31N) 72.9 (S31N/L26I) 2.7

4.0 (10 minutes) (S31N) 72.9 (S31N/L26I) 18.2Antiviral EC50 (mM) 0.49 (S31N) .30 (S31N/L26I) .60

BC035TEVC IC50 (mM) 20.5 (S31N) 50.5 (S31N/I32T) 2.5Antiviral EC50 (mM) 2.2 (S31N) .20 (S31N/I32T) .9

Pharmacology of M2-S31N Inhibitors 195

at ASPE

T Journals on February 21, 2022

molpharm

.aspetjournals.orgD

ownloaded from

barriers to drug resistance for the two classes of Food andDrug Administration–approved anti-influenza drugs, M2-S31N inhibitors outperformed amantadine and are onlyslightly less effective than oseltamivir. In general, completeresistance to oseltamivir was generated around passage5 (Triana-Baltzer et al., 2011; Renzette et al., 2014). Sequenc-ing the resulting resistant viruses revealed interesting find-ings. For the S31N-specfic inhibitor WJ379, M2-S31N simplyreverted back to N31S to become resistant to WJ379. Thismutation is a dominant positive mutation, as the M2-N31S/L26I double mutant was also completely resistant to WJ379.The single mutant M2-S31N/L26I alone was only partiallyresistant to WJ379. For the M2-S31N/WT dual inhibitor

BC035, we observed similar phenomena. As BC035 is a potentinhibitor of both the S31N and the S31 (WT) channels, M2-S31N could not revert back to S31 to become resistant to it;instead, it changed to N31D. N31D is also a dominant positivemutation, and the M2-N31D/I32T was completely resistant toBC035. The single mutation M2-S31N/I32T remained sensi-tive to BC035, and this mutation only caused a 2-fold EC50

increase. Interestingly, both of these highly resistant mu-tants, N31S and N31D, appeared to be less fit than the N31mutant in cell culture; both N31S and N31D reverted back toN31 after releasing drug selection pressure.Fast drug-binding kinetics were observed for bothWJ379 in

inhibiting M2-S31N/L26I and BC035 in inhibiting M2-S31N/I32T. It is possible that M2-S31N/L26I and M2-S31N/I32Tmight either adapt different structures or have differentdynamics than M2-S31N, which allows drugs to easily escapefrom the channel cavity. For example, M2-S31N/L26I andM2-S31N/I32T might have an enlarged N-termini channel en-trance such that drug can easily escape from the channel.Alternatively, M2-S31N/L26I and M2-S31N/I32T might bemore dynamic than M2-S31N; thus, drugs have a higherprobability to escape the channels when they adapt theOpenout-Closedin conformation (Khurana et al., 2009).Although resistant mutants (L26I and I32T) were selected

from serial passage experiments in cell culture under the drugselection pressure of WJ379 and BC035, respectively, bothmutants appear to be rare among human influenza A viruses.We retrieved and analyzed all of the human influenza A virusM2 sequences from the Influenza Research Database (http://www.fludb.org, accessed on January 18th, 2016). L26I oc-curred 70 times, and I32T only occurred twice out of 18,393M2sequences. These results indicate that viruses carrying eitherof these two mutants might have a reduced fitness of trans-mission among humans, althoughmore stringent experimentsneed to be followed to test their fitness and transmissibility.As resistance development is, unfortunately, inevitable, one

standard approach to delay resistance evolution under drug

Fig. 6. Dose-response curves in TEVC assays with 2-, 5-, and 10-minute time points. (A) Amantadine in inhibiting WT M2 (S31). (B) TEVC trace ofamantadine in inhibitingM2-WT at 10 mM for 10minutes. (C)WJ379 in inhibitingM2-S31Nmutant. (D) TEVC trace of WJ379 in inhibitingM2-S31N at10 mM for 10 minutes. Each data point represents at least three oocytes.

Fig. 7. Three-dimensional plot of the combination therapy of WJ379 andoseltamivir carboxylate on influenza A A/WSN/33 virus. A series ofcombinations of WJ379 (0, 10, 32, 100, 320, 1000, 3200, and 10,000 nM)and oseltamivir carboxylate (0, 1, 3.2, 10, 32, 100, 320, and 1000 nM) wereadded to MDCK cells that were infected with influenza virus A/WSN/33 ina 96-well plate. The plate was subsequently incubated for 72 hours. Cellviability was determined by neutral red staining. The synergy orantagonism volumes of drug-drug interaction were generated by theMacSynergy II software program. The synergy for the various combina-tions was significant at the 95% confidence limit.

196 Ma et al.

at ASPE

T Journals on February 21, 2022

molpharm

.aspetjournals.orgD

ownloaded from

selection pressure is combination therapy. Combination therapyhas a proven track record of reducing resistance evolution andside effects, and it is highly recommended for people withcompromised immune systems who might need extended anti-viral treatment to clear the viruses (Ison, 2013). As an example,the triple-drug combination therapy, commonly known as highlyactive antiretroviral therapy,was shown to bemore effective oversingle- or double-drug combination therapy in suppressinghuman immunodeficiency virus drug resistance (Gulick et al.,1997, 1998;DeClercq, 2007). Therefore, exploring a double-drugcombination of M2-S31N inhibitors with oseltamivir in theprevention and treatment of influenza infection is a rationalapproach to mitigate drug resistance. The need for combinationtherapy is further justified because monotherapy with oseltami-vir was suboptimal in the treatment of H5N1-infected patients,with a mortality rate as high as 60% (Abdel-Ghafar et al., 2008;Kandun et al., 2008). It has been shown in a number of studiesthat a combination ofM2 inhibitors (amantadine or rimantadine)with neuraminidase inhibitors (oseltamivir carboxylate or per-amivir), ribavirin, or a combination of all three was in generalsynergistic in in vitro cell culture assays and in vivo mousestudies (Govorkova et al., 2004; Ilyushina et al., 2006, 2007;Simeonova et al., 2007, 2008, 2009; Smee et al., 2009; Bantiaet al., 2010). Gratifyingly, our rationally designed M2-S31Ninhibitor WJ379 displayed strong synergy with oseltamivircarboxylate, with a synergy volume of 166, suggesting thatcombination therapy of WJ379 with oseltamivir carboxylatemight be an option for controlling influenza viruses in the future.In summary, the potent, broad-spectrum antiviral activities

of M2-S31N inhibitors against currently circulating influenzaA viruses, coupled with the strong synergistic effect of M2-S31N inhibitors with oseltamivir, provide compelling evidencefor further development of these second-generation M2 chan-nel blockers.

Acknowledgments

The authors thank Dr. Elena Govorkova at St. Jude Children’sResearch Hospital for providing the A/Denmark/524/2009 (H1N1) andA/Denmark/528/2009 (H1N1) strains, and Dr. James Noah at theSouthern Research Institute for providing the A/California/07/2009(H1N1) and A/Texas/04/2009 (H1N1) strains. The authors alsothank Drs. Govorkova and Noah for insightful advice regardingthe plaque assay protocols and serial passage experiments. TheST6Gal I–expressing MDCK cells were kindly provided by Dr.Yoshihiro Kawaoka at the University of Wisconsin, Madison, viamaterial transfer agreement. The authors thank Dr. David Bishop forproof reading and editing the manuscript.

Authorship Contributions

Participated in research design: Ma, Zhang, Wang.Conducted experiments: Ma, Zhang.Performed data analysis: Ma, Zhang, Wang.Wrote or contributed to the writing of the manuscript: Ma, Zhang,

Wang.

References

Abdel-Ghafar AN, Chotpitayasunondh T, Gao Z, Hayden FG, Nguyen DH, de JongMD, Naghdaliyev A, Peiris JS, Shindo N, and Soeroso S et al.; Writing Committeeof the Second World Health Organization Consultation on Clinical Aspects ofHuman Infection with Avian Influenza A (H5N1) Virus (2008) Update on avianinfluenza A (H5N1) virus infection in humans. N Engl J Med 358:261–273.

Atkins C, Evans CW, White EL, and Noah JW (2012) Screening methods for in-fluenza antiviral drug discovery. Expert Opin Drug Discov 7:429–438.

Balannik V, Wang J, Ohigashi Y, Jing X, Magavern E, Lamb RA, Degrado WF,and Pinto LH (2009) Design and pharmacological characterization of inhibitors ofamantadine-resistant mutants of the M2 ion channel of influenza A virus. Bio-chemistry 48:11872–11882.

Bantia S, Kellogg D, Parker CD, and Babu YS (2010) Combination of peramivir andrimantadine demonstrate synergistic antiviral effects in sub-lethal influenza A(H3N2) virus mouse model. Antiviral Res 88:276–280.

Beyleveld G, White KM, Ayllon J, and Shaw ML (2013) New-generation screeningassays for the detection of anti-influenza compounds targeting viral and hostfunctions. Antiviral Res 100:120–132.

Blair JMA, Webber MA, Baylay AJ, Ogbolu DO, and Piddock LJV (2015) Molecularmechanisms of antibiotic resistance. Nat Rev Microbiol 13:42–51.

Bright R, Shay D, Bresee J, Klimov A, Cox N, and Ortiz J (2006) High levels ofadamantane resistance among influenza A (H3N2)viruses and interim guidelinesfor use of antiviral agents - United States, 2005-06 influenza season. MMWR MorbMortal Wkly Rep 55:44–46.

Cady SD, Schmidt-Rohr K, Wang J, Soto CS, Degrado WF, and Hong M (2010)Structure of the amantadine binding site of influenza M2 proton channels in lipidbilayers. Nature 463:689–692.

Cady SD, Wang J, Wu Y, DeGrado WF, and Hong M (2011) Specific binding ofadamantane drugs and direction of their polar amines in the pore of the influenzaM2 transmembrane domain in lipid bilayers and dodecylphosphocholine micellesdetermined by NMR spectroscopy. J Am Chem Soc 133:4274–4284.

Cheng PK, Leung TW, Ho EC, Leung PC, Ng AY, Lai MY, and Lim WW (2009)Oseltamivir- and amantadine-resistant influenza viruses A (H1N1). Emerg InfectDis 15:966–968.

Clavel F and Hance AJ (2004) HIV drug resistance. N Engl J Med 350:1023–1035.De Clercq E (2007) The design of drugs for HIV and HCV. Nat Rev Drug Discov 6:1001–1018.

Dong G, Peng C, Luo J, Wang C, Han L, Wu B, Ji G, and He H (2015) Adamantane-resistant influenza a viruses in the world (1902-2013): frequency and distributionof M2 gene mutations. PLoS One 10:e0119115.

Dunning J, Baillie JK, Cao B, and Hayden FG; International Severe Acute Re-spiratory and Emerging Infection Consortium (ISARIC) (2014) Antiviral combi-nations for severe influenza. Lancet Infect Dis 14:1259–1270.

Fiore AE, Shay DK, Broder K, Iskander JK, Uyeki TM, Mootrey G, Bresee JS,and Cox NS; Centers for Disease Control and Prevention (CDC); ; Advisory Com-mittee on Immunization Practices (ACIP) (2008) Prevention and control of in-fluenza: recommendations of the Advisory Committee on Immunization Practices(ACIP), 2008. MMWR Recomm Rep 57 (RR-7):1–60.

Fofana DB, Soulie C, Maiga AI, Fourati S, Malet I, Wirden M, Tounkara A, TraoreHA, Calvez V, and Marcelin AG et al. (2013) Genetic barrier to the development ofresistance to rilpivirine and etravirine between HIV-1 subtypes CRF02_AG and B.J Antimicrob Chemother 68:2515–2520.

Govorkova EA, Fang HB, Tan M, and Webster RG (2004) Neuraminidase inhibitor-rimantadine combinations exert additive and synergistic anti-influenza virus ef-fects in MDCK cells. Antimicrob Agents Chemother 48:4855–4863.

Gulick RM, Mellors JW, Havlir D, Eron JJ, Gonzalez C, McMahon D, Jonas L,Meibohm A, Holder D, and Schleif WA et al. (1998) Simultaneous vs sequentialinitiation of therapy with indinavir, zidovudine, and lamivudine for HIV-1 in-fection: 100-week follow-up. JAMA 280:35–41.

Gulick RM, Mellors JW, Havlir D, Eron JJ, Gonzalez C, McMahon D, Richman DD,Valentine FT, Jonas L, and Meibohm A et al. (1997) Treatment with indinavir,zidovudine, and lamivudine in adults with human immunodeficiency virus in-fection and prior antiretroviral therapy. N Engl J Med 337:734–739.

Hatakeyama S, Sakai-Tagawa Y, Kiso M, Goto H, Kawakami C, Mitamura K,Sugaya N, Suzuki Y, and Kawaoka Y (2005) Enhanced expression of an a2,6-linked sialic acid on MDCK cells improves isolation of human influenza virusesand evaluation of their sensitivity to a neuraminidase inhibitor. J Clin Microbiol43:4139–4146.

Hayden FG (2013) Newer influenza antivirals, biotherapeutics and combinations.Influenza Other Respi Viruses 7 (Suppl 1):63–75.

Hayden FG and de Jong MD (2011) Emerging influenza antiviral resistance threats.J Infect Dis 203:6–10.

Hong M and DeGrado WF (2012) Structural basis for proton conduction and in-hibition by the influenza M2 protein. Protein Sci 21:1620–1633.

Hurt AC (2014) The epidemiology and spread of drug resistant human influenzaviruses. Curr Opin Virol 8:22–29.

Hurt AC, Ernest J, Deng YM, Iannello P, Besselaar TG, Birch C, Buchy P,Chittaganpitch M, Chiu SC, and Dwyer D et al. (2009) Emergence and spread ofoseltamivir-resistant A(H1N1) influenza viruses in Oceania, South East Asiaand South Africa. Antiviral Res 83:90–93.

Ilyushina NA, Bovin NV, Webster RG, and Govorkova EA (2006) Combination che-motherapy, a potential strategy for reducing the emergence of drug-resistant in-fluenza A variants. Antiviral Res 70:121–131.

Ilyushina NA, Hoffmann E, Salomon R, Webster RG, and Govorkova EA (2007)Amantadine-oseltamivir combination therapy for H5N1 influenza virus infection inmice. Antivir Ther 12:363–370.

Ison MG (2013) Influenza prevention and treatment in transplant recipientsand immunocompromised hosts. Influenza Other Respi Viruses 7 (Suppl 3):60–66.

Jing X, Ma C, Ohigashi Y, Oliveira FA, Jardetzky TS, Pinto LH, and Lamb RA (2008)Functional studies indicate amantadine binds to the pore of the influenza A virusM2 proton-selective ion channel. Proc Natl Acad Sci USA 105:10967–10972.

Kandun IN, Tresnaningsih E, Purba WH, Lee V, Samaan G, Harun S, Soni E,Septiawati C, Setiawati T, and Sariwati E et al. (2008) Factors associated withcase fatality of human H5N1 virus infections in Indonesia: a case series. Lancet372:744–749.

Kelso A and Hurt AC (2012) The ongoing battle against influenza: Drug-resistantinfluenza viruses: why fitness matters. Nat Med 18:1470–1471.

Khurana E, Dal Peraro M, DeVane R, Vemparala S, DeGrado WF, and Klein ML(2009) Molecular dynamics calculations suggest a conduction mechanism for theM2 proton channel from influenza A virus. Proc Natl Acad Sci USA 106:1069–1074.

Pharmacology of M2-S31N Inhibitors 197

at ASPE

T Journals on February 21, 2022

molpharm

.aspetjournals.orgD

ownloaded from

Li F, Ma C, DeGrado WF, and Wang J (2016) Discovery of highly potent inhibitorstargeting the predominant drug-resistant S31N mutant of the influenza A virusM2 proton channel. J Med Chem 59:1207–1216.

Matrosovich M, Matrosovich T, Carr J, Roberts NA, and Klenk H-D (2003) Over-expression of the a-2,6-sialyltransferase in MDCK cells increases influenza virussensitivity to neuraminidase inhibitors. J Virol 77:8418–8425.

McKimm-Breschkin JL (2013) Influenza neuraminidase inhibitors: antiviral actionand mechanisms of resistance. Influenza Other Respi Viruses 7 (Suppl 1):25–36.

Monto AS and Webster RG (2013) Influenza pandemics: history and lessons learned,Textbook of Influenza, pp 20–34, John Wiley & Sons, Ltd., Hoboken, NJ.

Pinto LH, Holsinger LJ, and Lamb RA (1992) Influenza virus M2 protein has ionchannel activity. Cell 69:517–528.

Pinto LH and Lamb RA (2006) The M2 proton channels of influenza A and B viruses.J Biol Chem 281:8997–9000.

Prichard MN, Prichard LE, and Shipman C, Jr (1993) Strategic design and three-dimensional analysis of antiviral drug combinations. Antimicrob Agents Chemother37:540–545.

Prichard MN and Shipman C, Jr (1990) A three-dimensional model to analyze drug-drug interactions. Antiviral Res 14:181–205.

Rappuoli R and Dormitzer PR (2012) Influenza: options to improve pandemic prep-aration. Science 336:1531–1533.

Renzette N, Caffrey DR, Zeldovich KB, Liu P, Gallagher GR, Aiello D, Porter AJ,Kurt-Jones EA, Bolon DN, and Poh YP et al. (2014) Evolution of the influenza Avirus genome during development of oseltamivir resistance in vitro. J Virol 88:272–281.

Repetto G, del Peso A, and Zurita JL (2008) Neutral red uptake assay for the esti-mation of cell viability/cytotoxicity. Nat Protoc 3:1125–1131.

Samson M, Pizzorno A, Abed Y, and Boivin G (2013) Influenza virus resistance toneuraminidase inhibitors. Antiviral Res 98:174–185.

Simeonova L, Galabov A, and Gegova G (2007) Anti-influenza A synergistic combi-nation effect of rimantadine and oseltamivir in mice. Antiviral Res 74:A46–A47.

Simeonova L, Galabov A, and Gegova G (2008) In vivo synergistic combination effectof rimantadine and oseltamivir against influenza A(H3N2) is manifested in severaldose ratios. Antiviral Res 78:A19–A19.

Simeonova L, Gegova G, and Galabov A (2012) Prophylactic and therapeutic com-bination effects of rimantadine and oseltamivir against influenza virus A (H3N2)infection in mice. Antiviral Res 95:172–181.

Smee DF, Hurst BL, Wong MH, Bailey KW, and Morrey JD (2009) Effects of doublecombinations of amantadine, oseltamivir, and ribavirin on influenza A (H5N1)virus infections in cell culture and in mice. Antimicrob Agents Chemother 53:2120–2128.

Smee DF, Hurst BL, Wong M-H, Tarbet EB, Babu YS, Klumpp K, and Morrey JD(2010) Combinations of oseltamivir and peramivir for the treatment of influenza A(H1N1) virus infections in cell culture and in mice. Antiviral Res 88:38–44.

Tarbet EB, Maekawa M, Furuta Y, Babu YS, Morrey JD, and Smee DF (2012)Combinations of favipiravir and peramivir for the treatment of pandemic in-fluenza A/California/04/2009 (H1N1) virus infections in mice. Antiviral Res 94:103–110.

Thompson WW, Shay DK, Weintraub E, Brammer L, Bridges CB, Cox NJ,and Fukuda K (2004) Influenza-associated hospitalizations in the United States.JAMA 292:1333–1340.

Thompson WW, Shay DK, Weintraub E, Brammer L, Cox N, Anderson LJ,and Fukuda K (2003) Mortality associated with influenza and respiratory syncytialvirus in the United States. JAMA 289:179–186.

Triana-Baltzer GB, Sanders RL, Hedlund M, Jensen KA, Aschenbrenner LM, LarsonJL, and Fang F (2011) Phenotypic and genotypic characterization of influenza virusmutants selected with the sialidase fusion protein DAS181. J Antimicrob Chemo-ther 66:15–28.

Vingerhoets J, Azijn H, Fransen E, De Baere I, Smeulders L, Jochmans D, Andries K,Pauwels R, and de Béthune M-P (2005) TMC125 displays a high genetic barrier tothe development of resistance: evidence from in vitro selection experiments. J Virol79:12773–12782.

Walsh CT and Wencewicz TA (2014) Prospects for new antibiotics: a molecule-centered perspective. J Antibiot (Tokyo) 67:7–22.

Wang C, Takeuchi K, Pinto LH, and Lamb RA (1993) Ion channel activity of influenzaA virus M2 protein: characterization of the amantadine block. J Virol 67:5585–5594.

Wang J (2015) M2 as a target to combat influenza drug resistance: what does theevidence say? Future Virol 11:1–4.

Wang J, Li F, and Ma C (2015) Recent progress in designing inhibitors that target thedrug-resistant M2 proton channels from the influenza A viruses. Biopolymers 104:291–309.

Wang J, Ma C, Wang J, Jo H, Canturk B, Fiorin G, Pinto LH, Lamb RA, Klein ML,and DeGrado WF (2013a) Discovery of novel dual inhibitors of the wild-type andthe most prevalent drug-resistant mutant, S31N, of the M2 proton channel frominfluenza A virus. J Med Chem 56:2804–2812.

Wang J, Wu Y, Ma C, Fiorin G, Wang J, Pinto LH, Lamb RA, Klein ML, and DegradoWF (2013b) Structure and inhibition of the drug-resistant S31N mutant of the M2ion channel of influenza A virus. Proc Natl Acad Sci USA 110:1315–1320.

Wanka L, Iqbal K, and Schreiner PR (2013) The lipophilic bullet hits the targets:medicinal chemistry of adamantane derivatives. Chem Rev 113:3516–3604.

Williams JK, Tietze D, Wang J, Wu Y, DeGrado WF, and Hong M (2013) Drug-induced conformational and dynamical changes of the S31N mutant of the in-fluenza M2 proton channel investigated by solid-state NMR. J Am Chem Soc 135:9885–9897.

Wu Y, Canturk B, Jo H, Ma C, Gianti E, Klein ML, Pinto LH, Lamb RA, Fiorin G,and Wang J et al. (2014) Flipping in the pore: discovery of dual inhibitors that bindin different orientations to the wild-type versus the amantadine-resistant S31Nmutant of the influenza A virus M2 proton channel. J Am Chem Soc 136:17987–17995.

Address correspondence to: Dr. Jun Wang, Room 423, 1657 East HelenStreet, BIO5 Institute, Tucson, AZ 85721. E-mail: junwang@pharmacy.arizona.edu

198 Ma et al.

at ASPE

T Journals on February 21, 2022

molpharm

.aspetjournals.orgD

ownloaded from