Kinetics of hepatitis C virus RNA decay, quasispecies evolution and risk of virological failure during telaprevir-based triple therapy in clinical practice

ContentslistsavailableatScienceDirect

Digestive

and

Liver

Disease

j o ur n a l ho me p a g e :w w w . e l s e v i e r . c o m / l o c a t e / d l d

Liver,

Pancreas

and

Biliary

Tract

Kinetics

of

hepatitis

C

virus

RNA

decay,

quasispecies

evolution

and

risk

of

virological

failure

during

telaprevir-based

triple

therapy

in

clinical

practice

夽

Valeria

Cento

a,∗

_,

_Monica

_Tontodonati

b,c

_,

_Velia

_Chiara

_Di

_Maio

a

_,

_Maria

_Concetta

_Bellocchi

a

_,

Fabrizio

Valenti

d

_,

_Alessandra

_Manunta

e

_,

_Serena

_Fortuna

f

_,

_Daniele

_Armenia

a

_,

Luca

Carioti

a

_,

_Francesco

_Paolo

_Antonucci

a

_,

_Ada

_Bertoli

g

_,

_Francesca

_Trave

b

_,

Pierluigi

Cacciatore

b

_,

_Mario

_Angelico

h

_,

_Pierluigi

_Navarra

f

_,

_Avidan

_U.

_Neumann

i

_,

Jacopo

Vecchiet

c

_,

_Giustino

_Parruti

b

_,

_Sergio

_Babudieri

e

_,

Carlo

Federico

Perno

a,g

_,

_Francesca

_{Ceccherini-Silberstein}

a a_{DepartmentofExperimentalMedicineandSurgery,UniversityofRome“TorVergata”,Rome,Italy} b_{InfectiousDiseaseUnit,PescaraGeneralHospital,Pescara,Italy}

c_{InfectiousDiseaseClinic,Chieti,Italy}

d_{DepartmentofEconomics,InstitutionsandLaw,UniversityofRome“TorVergata”,Rome,Italy} e_{InfectiousDiseasesUnit,DepartmentofClinicalandExperimentalMedicine,UniversityofSassari,Italy} f_{PharmacologyUnit,“A.Gemelli”Hospital,Rome,Italy}

g_{MolecularVirologyUnit,UniversityHospitalofRome“TorVergata”,Rome,Italy} h_{HepatologyUnit,UniversityHospitalofRome“TorVergata”,Rome,Italy} i_{GoodmanFacultyofLifeSciences,Bar-IlanUniversity,Israel}

a

r

t

i

c

l

e

i

n

f

o

Articlehistory: Received8July2014 Accepted3December2014 Availableonlinexxx Keywords: Drug-resistance Mathematicalmodelling Proteaseinhibitors Viralkinetic

a

b

s

t

r

a

c

t

Background:Theusedfirstgenerationproteaseinhibitorsmaybehamperedbyvirologicalfailurein

partiallyinterferon-sensitivepatients.

Aim:ToinvestigateearlyhepatitisCvirus(HCV)-RNAdecayandquasispeciesmodifications,anddisclose

viraldynamicsunderlyingfailure.

Methods:Viraemiadecayatearlytime-pointsduringtelaprevirtreatmentwasmodelledaccordingto

Neumannetal.(1998).NS3-sequenceswereobtainedbypopulation-sequencingand

ultradeep-454-pyrosequencing.

Results:13treatment-experienced(8non-responders,5 relapsers),andtwocirrhoticnaïvepatients,

receivedtelaprevir+pegylated-interferon-␣+ribavirin.

Viraemiadecaywasbiphasic.Inallpatients,first-phasewasrapidandconsistent,withamedian

[interquartile-range]viraemiadecayof2.8[2.6–3.2]logIU/ml within48h. Second-phasedecaywas

slower,especiallyinfailingpatients:3/3showed<1logIU/mldecaybetween48hand2weeks,and

HCV-RNA>100IU/mlatweek2.Onlyonepatientexperiencingsustainedviralresponseshowedsimilar

kinetics.

Bypyrosequencing,mutationalfreezewasobservedinall15patientswithinthefirst24h,butonly

inpatientswithsustainedresponseafterwards.Indeed,2/2failingpatientsshowedearlyresistance,as

minor(V36A-T54A:prevalence<26%at48h)ormajor(V36M/A-R155K:prevalence,99.8%atweek2)

variants.

Conclusions:Followingtelapreviradministration,first-phaseHCV-RNAdecayisconsistentwith

muta-tionalfreezeandlimited/noviralreplication,whilesecond-phaseissignificantlyslowerinfailingpatients

(withappearanceofresistance),suggestingtheusefulnessofearlyHCV-RNAmonitoring.

©2014EditriceGastroenterologicaItalianaS.r.l.PublishedbyElsevierLtd.Allrightsreserved.

夽 Thesedatahavebeenpresentedinpartatthe48thAnnualMeetingoftheEuropeanAssociationfortheStudyoftheLiver(EASL),heldinAmsterdam,Netherlands,April 24–28,2013andattheTheLiverMeeting®2014AmericanAssociationfortheStudyoftheLiver,November1–52013,WashingtonDC,USA.

∗ Correspondingauthorat:DepartmentofExperimentalMedicineandSurgery,UniversityofRome“TorVergata”,ViaMontpellier1,Rome00133,Italy.Tel.:+390672596560; fax:+390672596039.

E-mailaddress:valeriacento@gmail.com(V.Cento). http://dx.doi.org/10.1016/j.dld.2014.12.004

1. Introduction

InhepatitisCvirus(HCV)infection,thedevelopmentofdirectly actingantiviralagents(DAAs)providesanoutstanding opportu-nitytoimprovecureratesforpatientswithchronichepatitis C, allowinghopesforalarge-scaleeradicationofHCVinfection.The introductionintoclinicalpracticeoffirst-generationHCV-1 pro-teaseinhibitors(PIs)improvedtreatmentoutcomeswhenadded topegylated interferon-_␣ (pegIFN) and ribavirin (RBV)in both treatment-na¨

ι

veandtreatment-experiencedpatients[1–4]. Nev-ertheless,tripletherapy,particularlyfordifficultpatients,isoften limitedbysafetyandtolerabilityissues,whichlimititsuse,along witharapidonsetofdrug-resistanceinvirologicalfailures. Para-doxically, drug-resistance is considered a hallmark of antiviral drugs,anddefactoisbothamajorcauseandconsequenceof viro-logicalfailure[5,6].

Withinaninfectedindividual,thehighinvivoHCVmutation rate(2.5–3.2×10−5pernucleotidepergenomereplicationround) [7],insynergywithashorthalf-life,andanextremelyhighdaily viralproductionrate(1012_{particles/day)leadstothegeneration} ofaswarmofgeneticallydistinctbutcloselyrelatedviralvariants, knownasquasispecies[8].Inthisheterogeneousviralpopulation, theprobabilitytohaveatleastoneviralstrainharbouringoneor twonucleotidemutationsiscloseto100%[9].

Inan“idealscenario”,withcompletelywild-typeandsensitive viralquasispecies,viraleradicationbytreatmentfollowsabiphasic profile:inthefirstphaseviralparticleproductionissteeplyblunted, completelyblockingviralreplicationintheliver,theninthe sec-ondphaseallinfectedcellsarecleared,eitherbycell-deathorloss ofreplicativeintermediated(i.e.“cellcure”).Mathematicalmodels estimatedthat,inthisidealsituation,8weeksoftelaprevir-based therapywouldbeenoughtocure99%ofHCV-infectedpatients[10]. Nevertheless,whenananti-HCVagentisusedagainstan“actual scenario”ofviralquasispecies,drug-effectivenessinblockingviral productionmaynotbe100%andminorityresistantvariantsinthe livercanconsequentlygetthechancetoreplicate.Whenreplication underpositiveselectingpressureisallowed,itincreasestherisk ofgeneratingviralvariantswithbetterresistantprofilesandviral fitness.

In the past, mathematical models of HCV dynamics have contributedin understandingthe mechanismsof action of IFN, pegIFNandpegIFN+RBV,andconsequentlyinbettercharacterising viralbehaviourunderdrugpressure [11–13].Recently, dynamic parametershavebeenestimatedin patientstreated withDAAs, particularlyinmonotherapy[14].However,sofarnodataare avail-ableontheeffectsofPI-basedtripletherapyinrealclinicalpractice, oftenincludingdifficult-to-treatpatientswhomayactuallyhave thegreatestneedoftreatment,especiallyinEurope.

Forallthesereasons,theaimsofthisstudyweretodescribethe dynamicsofgenotype1HCV-RNAdecaywithinthefirsthoursand daysoftelaprevir-basedtripletherapyintreatment-experienced and/orcirrhoticpatients,andtosimultaneouslyevaluatetheearly modificationsinthequasispeciespopulationbymeanof ultrasensi-tivetechniques,correlatingbothfactorswiththeclinicalsituation, plasmatelaprevirconcentration,andtherapeuticoutcome.

2. Materialsandmethods 2.1. Studypopulation

ChronicallyinfectedHCVgenotype1patients,startinga triple-therapywithPegIFN/RBVplustelaprevirbetweenJune2012and November2012,whoconsentedtofrequentbloodsamplingwere consideredforinclusion.Exclusioncriteriawereageunder18years andotherconcomitantchronicliverdiseases.Treatmentschedules

and stopping rules followed telaprevir prescribing information [15].

Theanalysiswasperformedin15patients(Table1).Ofthese,13 hadpreviouslyfailedapegIFN+RBVtreatmentasnon-responders (7null-respondersand1partial-responder),orasrelapsers(N=5). Theother2/15patientswerecirrhoticsnaïvetotreatment. Correla-tionoftreatmentoutcomewithviralkineticswasperformedwith aperprotocolapproach.Thisstudywasconformedtolocalethical considerationsandtheprinciplesoutlinedinthe1975Declaration ofHelsinki(6threvision,2008).Allpatientsallowedthe participa-tiontothestudywithawrittenconsentfortheuseofsamplesfor researchpurposes.

2.2. HCV-RNAquantification

SerumHCV-RNAwasquantifiedbyRocheCOBAS® _TaqMan® HCV Test v2.0 (lower limit of quantification: 25IU/ml; LLOD, lower limit of detection: 15IU/ml) at baseline, at very early time points aftertelaprevir+pegIFN+RBV start (1h–2h–3h–4h– 5h–6h–8h–12h–24h–28h–48h–1w–2w) and during per protocol follow-up.

2.3. MathematicalmodellingofHCVdynamics

Aconstant-effectiveness(CE)modelwasusedtofitobserved data[11].In ordertonormalisefluctuations, baselineHCV-RNA values werecalculated as a meanamong 2–3pre/post-therapy measurements(baseline–1h–2haftertriple-therapystart).

Viralloaddatawerealsotestedforfittingwithamodelwith varying-efficacyorintracellulardynamics[16],butthesampling frequencydidnotallowenoughinformationtomakeasignificant differencerespectingtothestandardbi-phasicmodel.

2.4. Populationsequencing

NS3-protease sequences (aa 1–181) were obtained by an home-made sequencing protocol [17] at baseline and at early time-points(8h–12h–24h–48h–1w–2w–4w),whereplasma sam-pleswereavailableandHCV-RNAwasdetectable,andthenincase ofvirologicalfailure.

2.5. Ultra-Deep454-Pyrosequencing(UDPS)

NS3-protease UDPS was performed in all patients at base-line. Tests at8h–12h–24h–48h wereperformed in allavailable plasmasampleswithadequateHCV-RNA.Detailedinformationis reportedinAppendixA.Briefly,454juniorplatformwasusedwith 2genotypespecificprotocolsbasedon2overlappingamplicons spanning theentireNS3-protease-sequence.For each amplicon, readswerealignedand correctedusinghome-madePerlscripts andShorahpackage[18].Byusingaplasmidcontrol,mutations foundwithfrequency>0.1%,bothinreverseandforward orien-tationswereconsideredreliable.Globalhaplotypereconstruction wasperformedusingShorahpackage.

2.6. Geneticdiversityanalysis

Evolutionary divergence between sequences was calculated on NS3-sequences from population-sequencing and on UDPS-amplicons using theTamura 3-parameter model. The variation rate among sites was modelled with a gamma distribu-tion (shape parameter=1). Codon positions included were 1st+2nd+3rd+Noncoding.Allpositionswith<95%sitecoverage wereeliminated.Phylogenetic-trees werecontextually inferred. AllphylogeneticanalyseswereperformedwithMEGAv5.1[19].

Table1

Characteristicsofthestudypopulation.

Naïvecirrhoticpatients(N=2) Treatmentexperiencedpatients(N=13)

Non-responders(N=8) Relapsers(N=5)

Age(years),Median(IQR) 59(55–62) 52(46–58) 66(65–69)

Males,N(%) 1(50.0) 7(87.5) 3(60.0) IL-28Bgenotype,N(%) CC 0(0.0) 1(12.5) 0(0.0) CT 1(50.0) 5(62.5) 1(20.0) TT 0(0.0) 1(12.5) 0(0.0) Notdetermined 1(50.0) 1(12.5) 4(80.0) HCV-genotype,N(%) 1a 1(50.0) 4(50.0) 0(0.0) 1b 1(50.0) 4(50.0) 5(100)

Patientswithcirrhosis,N(%) 2(100) 3(37.5) 0(0.0)

BaselineLog10HCV-RNA(IU/ml),Median(IQR) 5.5(5.1–5.8) 6.0(5.5–6.7) 5.5(4.5–5.8)

BaselineAST(IU/ml),Median(IQR) 72(68–75) 71(53–96) 89(68–106)

BaselineALT(IU/ml),Median(IQR) 81(65–97) 105(68–149) 118(78–187)

HCV,hepatitisCvirus;IU,internationalunits;IQR,interquartilerange;ALT,alaninetransaminase;AST,aspartatetransaminase.

GeneticcomplexitywasalsoevaluatedbyShannonentropyvalues,

asdescribedin[20]. 2.7. Telaprevirconcentration

Telaprevir was measured in plasma samples by ultra-performanceliquidchromatography–tandemmassspectrometry (UPLCMS/MS).DetailedinformationisreportedinAppendixA. 3. Results

3.1. Treatmentoutcomefollowingtelaprevir+pegIFN+RBV administration

Out of 15 patients included in the study, virological-failure was detected in 3 previous null-responders (1=non-response; 2=viral-breakthrough);1previousrelapserdropped-outatweek-8 for non-virological reasons (HCV-RNA at the time of interrup-tion<LLOD,Table2).Theother11patients(2cirrhotic-naïve;4 relapsers;1partial-responder;4null-responder)reachedEndof Treatment(EOT)withundetectableHCV-RNA(targetnotdetected, TND)andarenow experiencinga SustainedVirologicResponse (SVR)afteratleast12weeksoffollow-up.

Treatmentoutcomewasnotcompromisedbythebaseline pres-enceofT54S(detectedinPT12,SVR24),norbyQ80K(detectedin PT15andPT8,SVR24andSVR12respectively).

Arapidviralresponse(RVR),definedasTNDHCV-RNAat week-4oftriple-therapy,wasobservedin11/15(73.3%)patients.Outof 10RVRpatientswhocompletedtreatment,9(90.0%)obtaineda SVRvs.2/4(50.0%)non-RVRpatients(P=0.176).All4patientswho didnotexperienceaRVRwerepreviousnull-responders.

3.2. EarlyHCV-RNAdecayfollowingtelaprevir+pegIFN+RBV administration

AninitialHCV-RNAdecaywasalreadyvisiblein11patientsat 4hsincefirsttelapreviradministration(median[IQR]HCV-RNA decayBL–4h=0.3[0.2–0.4]logIU/ml)(Fig.1).HCV-RNAdeclinedin allpatientsat6h(median[IQR]decayBL–6h=0.6[0.6–0.8]logIU/ml) and at 8h, before taking the second dose (median [IQR] decayBL–8h=1.2[1.0–1.9]logIU/ml).

In a subgroupof 7 patientsfor whom12hHCV-RNA value wasavailable,HCV-RNAdecayshoweditsmaximumvalueatthis time-point (median [IQR] decay_BL–12h=2.1 [1.5–2.2]logIU/ml), slowing down afterwards (median [IQR] decay12h–24h=0.5 [0.4–0.9]logIU/ml).

At24haftertreatmentstart,median[IQR] HCV-RNAdecline frombaselinevalueswasof2.4[2.2–2.7]logIU/ml(Fig.1.). This decay was not affected by baseline viraemia (P=n.s. by lin-earregressionanalysis), norbyHCV-subtype1aor 1b,previous treatmentexperience,stageofliverdiseaseandfinalvirological outcometotriple-therapy(Fig.2).

Thesameresultwasobtainedalsoanalysingthe baseline-to-48hHCV-RNAdecline(Fig.2),thusindicatingthattheroughvalue ofviraldecayfrombaselineto48hseemsnottobeaffectedby baselineclinicalandvirologicalparameters,nortoaffecttherate ofvirologicalsuccess.Nevertheless,eveniftheroughvalueof HCV-RNAdecayat48hwasconsistentamongallpatients,themedian [IQR]slope ofHCV-RNAdecaybetween24and48hwas signif-icantlylowerinfailing-patients,respectingtoSVRs(−0.4[−0.7; −0.1]vs.−3.3[−3.9;−2.9]logIU/ml/week,respectively;P=0.04by Mann–Whitneytest)(Table2).

Overall,failing-patientsexperiencedaslowingdowndecayof HCV-RNA,startingfrom24haftertreatmentstart.Thisdifferent kineticwasmaintainedalsoduringsecond-phaseviraemiadecline. Indeed,the3failing-patientsweretheonlyoneswith<1logIU/ml decline between 48h and week-2 (Table 2), while all other patientshadamedian(IQR)decay48h–2wof1.9(1.4–2.3)logIU/ml (P=0.030 byMann–Whitney test). Atweek-2 of triple-therapy, all patients experiencing RVR (and afterwards SVR) had HCV-RNA values below LLOD (N=5) or <100IU/ml (N=4). The only RVRpatient(PT9) experiencingvirological-failurehadHCV-RNA atweek-2of430IU/ml(Table2).Interestingly,3/4patientswith HCV-RNA>100atweek-2experiencedvirological-failurevs.0/9 ofthosewithHCV-RNA<100IU/ml(P=0.003byFisherExacttest) (Table2).

At week-2,suboptimal viral-kinetic impactedalso on rough HCV-RNAdeltavaluesfrombaseline,whichwassignificantlylower inallpatientsexperiencingvirological-failurerespectingtoSVRs (median[IQR] HCV-RNAdecayBL–2w=−3.1[−3.4;−0.5]vs. −4.6 [−4.8;−4.2]logIU/ml,respectively;P=0.014byMann–Whitney test)(Fig.2).

3.3. Viralloadchangesanalysisusingthestandardmodelofviral dynamics

To verify whether an alteration of first-phase/second-phase viraldeclineoccursinpatientsexperiencingvirological-failure,a CEmodelforviral-kineticswasused.

In our dataset, the mean (SE) delay time t0 assumed for pegIFN+RBV+telaprevirtriple-therapytobeeffectivewasof10.7 (2.2)h,or 0.444(0.092) days(Table3).Highert0 values among

HC V-RN A de cay (Log IU /ml ) 1.0 0.0 -1.0 -2.0 -3.0 -4.0 HC V-RNA d ecay (Log IU /ml) 1.0 0.0 -1.0 -2.0 -3.0 -4.0 Baseline 4h 8h 12h 24h 28h 48h

Median (IQR) Log HCV-RNA12hdecay: 2.1 (1.5-2.2) IU/ml

Median(IQR) Log HCV-RNA48hdecay: 2.9 (2.6-3.1)

Median(IQR) Log HCV-RNA24hdecay: 2.4 (2.2-2.7) IU/ml

HC V -RN A de cay (Log IU /ml ) 0.0 -1.0 -2.0 -3.0 -4.0 -5.0

Baseline 24h 48h Week-1 Week-2

A

B

C

Fig.1. EarlypatternsofHCVRNAdecayduringtelaprevir-basedtripletherapy.TheHCVRNAdecayfrombaselinevalues(calculatedasameanofpre-therapy,1hand2h HCV-RNAvalues)withinthefirst48h(AandB)and2weeks(C)oftelaprevir-basedtripletherapyisreportedforeachpatient.LineinpanelBrepresentspopulationmean. IU,internationalunits;IQR,interquartilerange.

patientswerenotrelatedtotreatmentoutcomenortothestageof diseaseprogression,sinceitwasnotsignificantlyhigherincirrhotic patients(datanotshown).

Themeanfreevirions clearancerate(c),characterising first-phase HCV-RNAdecline, was 9.4day−1 (Table3).It washigher thanthe5.95day−1estimatedwithstandardIFN,thussuggesting alowerdurationofthefirst-phaseofHCV-RNAdrop(<48h)with telaprevir.

Also the mean value of per capita rate of loss of infected cells (ı) was slightly higher than that reported with standard IFN(0.22day−1vs.0.16day−1).Patientsexperiencing virological-failuretended tohavelower meanı values,and thus aslower second phase of HCV-decline, relative to patients achieving SVR (mean [SE]=0.137 [0.093]day−1 vs. 0.244 [0.131]day−1, respectively),even if thedifferencewasnot statistically signif-icant.Alsomean treatment-effectivenessε waslowerin failing patients relative to patientswith SVR (88% vs. 94%) (Table 3) and linear regression analysis highlighted a correlation among treatment-effectivenessεandı(P=0.0133),sustainedby Spear-man’scorrelationcoefficient(r=0.671).Overall,thesedataconfirm asuboptimalsecond-phasedeclineintherapy-failingpatients.

3.4. Telaprevirpharmacokineticsanalysisinthecontextof cirrhosisandvirologicalfailuretotripletherapy

To evaluatethe factors potentially impacting on treatment-efficacy, telaprevir plasma concentration was calculated in all patientsatdifferenttime-points(6h–8h–12h–24h–48h–1w–4w): in all cases, the concentration curve was always found to be above the recommended therapeutic-ranges. Nevertheless, the peak of plasma concentration at 6–8h since treatment start tended to belower in cirrhoticpatients (median [IQR] plasma telaprevir-concentration=2345 [1476–2692]ng/ml) in compar-ison to non-cirrhotic (2701 [1658–3318]ng/ml), even if the differencewasnotstatisticallysignificant(P=0.4923by Wilcoxon-Test),anddidnotcorrelatetoprevioustreatmentoutcome(data notshown).

Analysing thepeak (6–8h) totrough(24h) decline slope of plasma telaprevir-concentration, no differences were observed among previous relapsers or previous non-responders, if the patients were non-cirrhotic and excluding those experiencing virological-failure(Fig.3).Onthecontrary,naïvecirrhoticpatients tendedtohavelowertelaprevirpharmacokineticsvaluesat24h,

HC V-RN A decay Bas e line -24 h ( logIU /m l) 0 -1.0 -2.0 -3.0 -4.0 -5.0 -6.0 Naïve patients GT-1a GT-1b Previous relapsers Previous no n-responders Cirrhotic patients Non-cirrhotic patients Patient failing triple therapy SVR patients to triple therapy HC V-RN A decay Base line -48 h ( logIU /m l) 0 -1.0 -2.0 -3.0 -4.0 -5.0 -6.0 Naïve patients GT-1a GT-1b Previous relapsers Previous no n-responders Cirrhotic patients Non-cirrhotic patients Patient failing triple therapy SVR patients to triple therapy HC V-RN A decay Bas e line -2w (l ogIU /m l) 0 -1.0 -2.0 -3.0 -4.0 -5.0 -6.0 Naïve patients GT-1a GT-1b Previous relapsers Previous no n-responders Cirrhotic patients Non-cirrhotic patients Patient failing triple therapy SVR patients to triple therapy Median(IQR):

-4.6(-4.8;-4.2) _Median(IQR): -3.1(-3.4;-0.5)

P=0.014

Fig.2.EarlyHCVRNAdecayinrelationtoclinicalandvirologicalpatients’characteristics.BoxplotsrepresentingHCVRNAdecayfrombaselinevalues(calculatedasameanof pre-therapy,1hand2hHCV-RNAvalues)to24h,48handweek-2oftriple-therapyadministration.PatientswerestratifiedaccordingtoHCV-genotype,previoustreatment experience,diagnosisofcirrhosisandvirologicaloutcometotelaprevir-basedtriple-therapy.P-valuewascalculatedthroughMann–Whitneytest.IU,internationalunits; SVR,sustainedvirologicalresponse;IQR,interquartilerange.

suchthose who later experiencedvirological-failures, whoalso showedadecline inconcentrationrespectingtothe6–8hpeak (Fig.3).

Therefore,weconfirmedanappropriatedrug-exposureduring thefirstperiodoftriple-therapyadministration,inall15patients includedinthestudy.

3.5. HCVquasispeciesevolutionanddrugresistancedetectionin theearlyphasesoftelaprevir+PegIFN+RBVadministration

ToassessHCV-quasispeciesrearrangementatearlytimepoints duringtreatment, viralevolution bypopulation-sequencing and UDPSwasperformed.

Bypopulation-sequencing,viralevolutioninthefirsthoursand dayssincetreatmentstart(8h–24h–48h)waslow,withamedian (IQR)intrapatientevolutionarydivergence,betweenbaselineand lastdetectableearlytime-point,of0.001(0.000–0.003)nsubs/site. Thiscorrespondstoamedian(IQR)of1(1–7)nucleotidevariations onatotalof543nucleotidepositionsanalyzed.

Then,UDPSanalysiswasperformedtobettercharacterisethe quasispeciesdynamicsand rearrangement.Atbaseline,patients experiencingvirological-failuretendedtohavehigherintra-patient nucleotide genomiccomplexity in comparison to SVR patients, both in terms of evolutionary-divergence (mean [SD]=0.018 [0.009]nsubs/site vs. 0.013 [0.005], respectively, P=0.35) and Shannon Entropy (mean [SD]=0.029 [0.015]×10−3 vs. 0.022 [0.010]×10−3,respectively).

Withinthefirsthours,viral-quasispeciesdetectedatbaseline, remainedsubstantiallystableatallearlytime-points(8h,12h,24h), inallpatientsstudiedforthispurpose,withnorearrangementsof speciesprevalenceatbothnucleotideandaminoacidlevels(data notshown).Thissuggestsamutational-freezeandalimited num-berof(orevenabsent)cyclesofviralreplicationinthisobservation period.

After48h,stillnoquasispeciesrearrangementsweredetectedin SVRpatientsanalyzed.Onthecontrary,in2/2failing-patientswith availablesamples(PT7andPT13),earlydevelopmentof resistance-associatedvariants(RAVs)wasobserved(SupplementaryTableS1).

Table 2 Early viral kinetics and treatment outcome. Patient ID Prior treatment experience HCV genotype IL28B Fibrosis (Metavir) Baseline HCV-RNA (log 10 IU/ml) a Decay slope (log 10 /week) HCV-RNA 2 w (IU/ml) HCV-RNA 4 w (IU/ml) Treatment length (weeks) Treatment outcome 0–24 h 24 h–48 h 48 h-2 w PT15 Naive 1a n.a. F4 5.8 − 18.7 − 1.9 >− 2.8 TND TND 24 SVR24 PT16 1b CT F4 5.1 − 18.5 − 8.9 >− 1.2 TND TND 48 SVR12 PT10 Relapsers 1b n.a. F2 5.7 − 21.4 − 3.0 − 1.0 14 TND 24 SVR24 PT11 1b n.a. F1 5.4 − 18.3 − 2.9 >− 2.4 TND TND 24 SVR24 PT12 1b n.a. F3 5.9 − 15.4 − 4.0 − 1.6 38 TND 24 SVR24 PT17 1b n.a. F0 5.6 − 16.8 − 3.4 n.a. n.a. TND 24 SVR16 PT2 1b CT F2 5.7 − 15.2 − 2.8 >− 1.9 TND TND 8 Drop out PT3 Partial responder 1b CT F4 5.2 − 15.6 − 3.2 >− 2.5 TND TND 48 SVR12 PT1 Null responders 1b CT F4 6.0 − 16.4 − 3.3 − 1.8 24 <15 48 SVR12 PT4 1a CT F4 7.0 − 18.7 − 3.8 − 2.0 58 TND 48 SVR12 PT5 1b CC F3 6.2 − 12.2 − 5.4 − 2.2 32 TND 24 SVR24 PT8 1a CC F3 6.7 − 17.3 − 0.17 − 1.4 545 16 48 SVR12 PT7 1b TT F3 6.2 − 15.6 0.2 − 0.2 669 19 20 Breakthrough PT9 1a CT F3 5.7 − 13.6 − 0.95 − 0.9 430 TND 20 Breakthrough PT13 1a CT F4 5.6 − 19.3 n.a. n.a. 122,528 200,000 6 Breakthrough HCV, hepatitis C virus; IU, international units; n.a., not available; TND, target not detected (below the limit of detection); SVR, sustained virologic response; AE, adverse events. a Baseline HCV-RNA values were calculated as mean among pre-therapy value, hour 1 and hour 2 values after triple-therapy start, in order to normalise fluctuations. -10 -5 0 5 10 15 20 0 2000 4000 6000 8000 P ea k-to-t rough TV R de c line sl ope

TVR plasma concentration at 24h (ng/ml)

Viral failure and

previous null-responders Previous null-responders Previous relapsers Naive -10 -5 0 5 10 15 20 0 2000 4000 6000 8000 P ea k -to -t rough TV R de c line sl ope

TVR plasma concentration at 24h (ng/ml) Viral failure

and cirrhotic Cirrhotic

Viral failure non

cirrhotic Non cirrhotic

A

B

Fig.3. Peaktotroughdeclineslopeoftelaprevirplasmaconcentration.Thepeak (6–8h)totrough(24h)declineslopeoftelaprevirplasmaconcentrationinrelation tothe24hvalueaftertriple-therapystartarereportedforeachpatient.Patients weregroupedaccordingtotriple-therapyoutcome(blackcontentforviralfailures) and(A)previousresponseor(B)diagnosisofcirrhosis.TVR,telaprevir.

InPT13,theR155Kaccountedfor99.8%ofinfectingviral pop-ulationalreadyat2weeksoftripletherapy,anditwasassociated withanincreaseinHCV-RNA(122,528IU/mlvs.728IU/ml,from 24h).

In PT7, experiencing late viral-breakthrough, UDPS analy-sis instead revealed, already at 48h, resistant strains only as minority-variants: V36A (459/5244 reads, 8.7% prevalence, mutational-load=2.0logIU/ml)andT54A(1885/5248reads,26.1% prevalence, mutational-load=2.5logIU/ml). These RAVs were foundlateratfailure,withamajorprevalenceofT54A(4726/5624 reads, 84.0% prevalence), and a minority contribution of V36A (49/5529reads,1.0%prevalence)andT54S(165/5624reads,2.8% prevalence).

4. Discussion

ThekineticsofearlyHCV-RNAdeclineandquasispecies evolu-tionundertelaprevir-basedtriple-therapywereanalysedinasmall pivotalgroupofHCV-1infectedpatients(failingpreviousIFN-based

Table3

Estimationofviraldynamicparametersusingconstanteffectivenessmodel.

StudyID Treatment experience InitialHCV-RNA (logIU/ml) Delay (t0,days)a Virionclearance (c,day−1)b

Efficacy(ε,%)c _{Infectedcelldeath}

(ı,day−1)d

2

Virologicalfailuretotelaprevir-basedtripletherapy

PT7 Nullresponder 6.2 0.384 15.0 85.3 0.034 2.064

PT9 Nullresponder 5.7 0.317 12.5 83.2 0.162 0.778

PT13 Nullresponder 5.6 0.347 11.4 95.0 0.215 0.587

Sustainedvirologicalresponsetotelaprevir-basedtripletherapy

PT15 Naive 5.8 0.306 14.8 92.5 0.092 1.800 PT16 Naive 5.1 0.561 6.3 99.2 0.304 1.900 PT2 Relapser 5.7 0.532 5.5 97.1 0.445 1.183 PT10 Relapser 5.7 0.533 7.9 97.8 0.156 1.975 PT11 Relapser 5.4 0.372 10.1 96.0 0.226 1.495 PT12 Relapser 5.9 0.471 8.3 88.4 0.177 1.182 PT17 Relapser 5.6 0.542 6.5 97.1 0.250 1.310 PT3 Partial responder 5.2 0.513 5.8 96.2 0.449 0.822 PT1 Nullresponder 6.0 0.507 5.8 98.3 0.400 1.002 PT4 Nullresponder 7.0 0.518 7.0 98.5 0.233 2.160 PT5 Nullresponder 6.2 0.413 10.4 77.1 0.119 2.229 PT8 Nullresponder 6.7 0.347 13.6 88.5 0.079 2.038 Overallmean(SD) 5.9(0.5) 0.444(0.092) 9.4(3.4) 92.7(6.7) 0.223(0.129) 1.502(0.557)

HCV,hepatitisCvirus;IU,internationalunits;SD,standarddeviation. a_t

0isthepharmacologicaldelay,representingthetimeafterwhichtreatmentaffectsviralload.

b_{cistheclearancerateoffreevirionsinplasma,calculatedbynon-linearfittingofHCV-RNAdecayduringthefirst48hoftripletherapytothemathematicalmodel.} c _{εistheeffectivenessoftherapeuticregimeninblockingvirionproduction.Isassumedtohaveavaluebetween0(noeffect)and1(100%effective)anditwascalculated} bymodelfittingofthebiphasicdecline,usingHCV-RNAdecayduringthefirst48hoftripletherapy.

d _{ıistherateatwhichinfectedcellsarelost,calculatedbynon-linearfittingofHCV-RNAdecayduringthefirst48hoftripletherapytothemathematicalmodel.} treatments,orwithadvancedliverdisease),withfrequentsampling

atmultipletime-points,includingtheveryfirsthoursoftreatment. Ourresultsshowthatfirst-phaseHCV-RNAdeclinemaybe actu-allyshorterthanpreviouslyestimatedwithstandardIFN-treatment [11],beingcomprisedwithinthefirst24hsincetreatment-start. Duringthisfirst-phase,all15patientsexperiencedaneffectiveand optimalHCV-RNAdecline,withmutational-freeze,suchasitoccurs inmoreeasilytreatablepatients,andindependentfrom virological-outcome[10].Viceversa,thesecond-phaseofviralkineticsisin generalslowerthanthatfoundineasierpatients[10],andeven morecompromisedinthosethatexperiencedvirological-failure.

Previousresultsonnon-cirrhoticdrug-naïvepatientstreated witheither telaprevir-monotherapyor triple-therapy [10] have shownamuchmorerapidfirstphaseandsecondphaseHCV-RNA declineincomparisontodualIFN-basedtreatments[11,13].

Inourpopulation,theslopeoffirst-phasedeclinewasoptimal andconsistentwithpreviousestimates[11,13,10],thus highlight-ingan independenceofthefirst HCV-RNAdecay frombaseline clinicaland virologicalcharacteristics.Inparticular,theslopeof HCV-RNAdecaywasatitsheightat24haftertreatmentstart.After thistime-point,thedecayratesufferedaslowdown,thus assum-inganearlystartofthesecond-phasekinetic.Thehigherpotency oftelaprevirinblockingviralproduction,respectingtoIFN,can accountfortheshorteningoffirst-phasedecline,assupportedby thehighervalues oftreatmenteffectiveness(ε)and freevirions clearancerate(c)[11,13].

Duringthisshortfirst-phase,HCV-RNAdecaywasthusdeep, rapid,and,notably,consistentinall15patients,includingthose thatlaterfailed.ByUDPS,norearrangementofviralquasispecies (noremergenceofRAVs,atthelowestlimitofUDPSreliability) wasdetected within these first 24h of treatment, indicating a mutational-freezeandalimitednumberof(orevenabsent)cycles ofviralreplication.

Onthecontrary,inall15patients,second-phaseHCV-RNAdecay kineticwascharacterisedbyvaluesofinfectedcellclearancerate (ı)lowerthanthosepreviouslydefinedinnaïveandnon-cirrhotic

patientsundergoingthesame treatment[10]. Inparticular, the mean valueof ı wasmuch lower than thevalueof 1.19day−1 reportedusingthesameCEmodelinnaïvepatientstreatedwith TVR-monotherapy or triple-therapy (whose determination was neverthelessconsideredinaccurate)andmoreclosetothevalue of0.58day−1calculatedusingavarying-effectivenessmodel[10]. Inaddition,whilefirst-phasedecaywasidenticalamongSVRsand failing-patients,after24hsignificantdifferencesstartedtoappear. Indeed, inall telaprevir-failingpatientsthesecond-phase viral-declinewascharacterisedbyapoor(ornull)24h–48hand48h–2w HCV-RNAdeclines,andresultedinHCV-RNAvalues>100IU/mlat week-2.

Independentfrombaseline viraemia,3/4patientswith HCV-RNA>100IU/mlatweek-2failedtriple-therapy.Inagreementwith thisresult,inalargerdatasetoftelaprevir-andboceprevir-treated patients (N=83), we recently reported that week-2 HCV-RNA decayissignificantlylowerin patientsexperiencing virological-failure than in SVRs (median[IQR]=−3.7 [−4.3; −3.2] vs. −4.6 [−5.2;−4.0]logIU/ml,P=0.007)andthatHCV-RNA>100IU/mlat week-2issignificantlyassociatedwithvirological-failure(adjusted OR: 27.955; 95% C.I.: 1.885–414.500; P=0.015), after correc-tionbybaseline/earlyRAVs,HCV-subtype,cirrhosis,unfavourable IL28B and previous null-response [21]. Thus, assessment of HCV-RNA at week-2 may actually represent a good proxy for virological-response,promptlyidentifyingpatientsathigherrisk ofvirological-failureifhigherthan100IU/ml.

The suboptimalsecond-phase HCV-RNA decline observed in failing-patientscanbeassociatedtoalowerı,and/ortoalower ε.Aspreviouslyreported[10],weindeedobservedadirect cor-relation among ε and ı. This indicates that, within biological limits, the second slope of viral decline is improved as drug-effectivenessimproves.Thebasemechanismofthisphenomenon lies in the hypothesis that the second-phase HCV-RNA decline may be driven not only by loss of infected cells, but also by a continuous decline in viral production and “cell cure” via riddance of intracellular replication complexes [16]. Therefore,

suboptimaltreatment-effectivenesscanslowerthesecond-phase declineindependentlyofimmuneresponse,increasingtheriskof virological-failure.

Alowertreatmentefficacyisprobablydependentbymultiple factors,suchas:(a)thesuboptimalactionofpegIFN+RBV back-bone,supportedbythefactthatall3failing-patientswereprevious null-respondersandwithunfavourableIL28Bgenotype;(b)bythe poorertelaprevir-pharmacokinetics(indeedobservedinall3 fail-ingpatients);and(c)selectionordevelopmentofRAVs,further affectingdrug-effectivenessduringthecourseoftreatment[6].

Whateverthemechanism,lowerεandıvaluesinbothpatients experiencinglateviral-breakthroughcorrespondedtothelowest 24h–2wHCV-RNA decayslopes.Assuming a complete therapy-mediatedcontrolofviralreplication(ε=100%),thisdecaypattern couldsimplyrepresentadelayininfectedcellclearance,andthus just underlie the utility of a longer treatment. Since also ε is reduced,however,thechanceoflow-levelviralreplicationunder drug-pressureishigh,possiblyallowingthedevelopment/selection ofminorityRAVs.Indeed,earlyrearrangementofviralquasispecies occurredexclusivelyinfailing-patient,anditwasassociatedwith thedevelopment of RAVsat failure. During second-phaseviral decline,mutational-freezewasthusnolongerobservedinpatients experiencingvirological-failure,sinceRAVsweredetectedin2/2 analyzedpatientsbetween48hand2weeks,thoughwithdifferent prevalence.

InPT7,experiencinglateviralbreakthrough,RAVswerealready detectedat48hjustasminorvariants,andbecamethemost preva-lentviralpopulationonlyatthemomentoffailure(week-20).These 48hminorityRAVswerecompletelyabsentatearlierevaluations (8hand24h),suggestingaprocessofdevelopmentandselection, ratherthanpureselectionofpre-existentRAVs,atleastwhenlate failureoccurs.

Ontheotherhand,inPT13,RAVsdetectedatweek-2accounted for 99.8% of infectingviral population, and indeed determined anincreaseinHCV-RNAfrompreviousvalues.Thispatternmay suggestinthispatientthepre-existenceofminorityRAVs. Unfor-tunately,no 24h–48h samples were available for PT13 and, at baseline,noevidencesofresistancewerefound.Nevertheless,the absenceofRAVsdetectionatbaselinedoesn’texcludetheirminor presence,belowthesensitivity-limitofUDPS(<0.1%prevalence). InPT13,forinstance,0.1%ofbaseline440,465IU/mlwould corre-spondto440IU/ml,thusstill2.6logIU/mlofcirculatingpotentially resistantstrains,evenifthishypothesisneedstobeconfirmedby meansofmoresensitivetechniques(suchasillumina).

Forthepresent study,week-2100IU/mlthresholdwas eval-uatedbyRocheCOBAS®_TaqMan®_{HCVTestv2.0assay.Although} thisassayandAbbottRealTimeHCVassayhavedifferentsensitivity, theyshowhighclinicalconcordance[22].Indeed,inanotherstudy onalargerpopulationweconfirmedthesamethresholdwithboth tests[21].

In conclusion, now more than ever, clinicalmanagement of HCV-infectedpatientstreatedwithfirst-generationPIsmay ben-efitfromacloserandaccurateearlyvirological-monitoring,given therapidkineticsofHCV-RNAdeclineandtheriskof resistance-development.ForthisreasonHCV-RNAmeasurementsperformed duringsecond-phaseHCV-RNAdecline(i.e.atweek-2)under triple-therapy may provide relevant clinical information to monitor treatmentefficacy.SuboptimalearlyHCV-RNAdeclinefrom48h tothefirst2weeksoftreatmentmayindeedbethefirsthintforan ongoingresidualviralreplicationunderpharmacological-pressure, potentiallyleadingtodevelopmentandselectionofresistance,and, eventually,virologicalfailure.

Conflictofinterest

P.Navarra, M.Angelico, C.F.Perno andF. Ceccherini-Silberstein have received funds for attending symposia, speaking,

organizing educational activities, grant research support, con-sultancyand advisoryboardmembership,fromViiV andMerck Sharp&Dohme(PN),fromRoche,Gilead,MerckSharp&Dohme, Janssen Cilag (MA), form Abbott, Boehringer Ingelheim, Bristol MyersSquibb,Gilead,MerckSharp&Dohme,JanssenCilag,Pfizer, Tibotec,Roche,andViiV(CFP),andfromAbbvie,MerckSharp& Dohme, Gilead, Janssen Cilag,Roche, Bristol MyersSquibb, and ViiV(FCS).

Theotherauthorsofthisstudydonothaveanythingtodisclose regardingfundingfromindustryorconflictsofinterest.

Acknowledgments

ThisworkwassupportedbytheItalianMinistryofInstruction, UniversityandResearch(MIUR)[RBAP11YS7K001,andPB05]and byAviraliaFoundation.Thefundershadnoroleinstudydesign, datacollectionandanalysis,decisiontopublish,orpreparationof themanuscript.

We thank Rochefor providing laboratorymaterial for HCV-RNAdetermination,theClinicalHepatologyandMolecularVirology groups of “Tor Vergata” University Hospital, and particularly DanieleDiPaolo,DomenicodiCarloandFrancescaStafoggiafor statistical,clinicalandtechnicalsupport.

AppendixA. Supplementarydata

Supplementarydataassociatedwiththisarticlecanbefound,in theonlineversion,athttp://dx.doi.org/10.1016/j.dld.2014.12.004. References

[1]Bacon BR, Gordon SC, Lawitz E,et al. Boceprevir for previously treated chronic HCV genotype 1 infection. New England Journal of Medicine 2011;364:1207–17.

[2]JacobsonIM,McHutchisonJG,DusheikoG,etal.Telaprevirforpreviously untreatedchronichepatitisCvirusinfection.NewEnglandJournalofMedicine 2011;364:2405–16.

[3]PoordadF,McConeJrJ,BaconBR,etal.Boceprevirforuntreatedchronic HCV genotype 1 infection. New England Journal of Medicine 2011;364: 1195–206.

[4]ZeuzemS,AndreoneP,PolS,etal.TelaprevirforretreatmentofHCVinfection. NewEnglandJournalofMedicine2011;364:2417–28.

[5]PawlotskyJM.Treatmentfailureandresistancewithdirect-actingantiviral drugsagainsthepatitisCvirus.Hepatology2011;53:1742–51.

[6]SorianoV,PerelsonAS,ZoulimF.Whyaretheredifferentdynamicsinthe selectionofdrugresistanceinHIVandhepatitisBandCviruses?Journalof AntimicrobialChemotherapy2008;62:1–4.

[7]RibeiroRM,LiH,WangS,etal.QuantifyingthediversificationofhepatitisC virus(HCV)duringprimaryinfection:estimatesoftheinvivomutationrate. PLoSPathogens2012;8:e1002881.

[8]MartellM,EstebanJI,QuerJ,etal.HepatitisCvirus(HCV)circulatesasa pop-ulationofdifferentbutcloselyrelatedgenomes:quasispeciesnatureofHCV genomedistribution.JournalofVirology1992;66:3225–9.

[9]RongL,DahariH,RibeiroRM,etal.Rapidemergenceofproteaseinhibitor resistanceinhepatitisCvirus.ScienceTranslationalMedicine2010;2:30ra32. [10]Guedj J, PerelsonAS.Second-phasehepatitis CvirusRNAdecline during telaprevir-basedtherapyincreaseswithdrugeffectiveness:implicationsfor treatmentduration.Hepatology2011;53:1801–8.

[11]NeumannAU,LamNP,DahariH,etal.HepatitisCviraldynamicsinvivo and the antiviral efficacy ofinterferon-alphatherapy. Science 1998;282: 103–7.

[12]Rosen HR,RibeiroRR,WeinbergerL, etal.EarlyhepatitisCviralkinetics correlatewithlong-termoutcomeinpatientsreceivinghighdoseinduction fol-lowedbycombinationinterferonandribavirintherapy.JournalofHepatology 2002;37:124–30.

[13]DixitNM,Layden-AlmerJE,LaydenTJ,etal.Modellinghowribavirinimproves interferon responseratesinhepatitis Cvirusinfection.Nature 2004;432: 922–4.

[14]ChatterjeeA,GuedjJ,PerelsonAS.MathematicalmodellingofHCVinfection: whatcanitteachusintheeraofdirect-actingantiviralagents?Antiviral Ther-apy2012;17(PtB):1171–82.

[15]EuropeanMedicinesAgency(EMA).Incivek(telaprevir)FilmCoatedTablets. EUPrescibingInformation;2011May1.

[16]Guedj J, Neumann AU. Understanding hepatitis C viral dynamics with direct-acting antiviral agents due to the interplaybetween intracellular

replication andcellularinfectiondynamics.JournalofTheoreticalBiology 2010;267:330–40.

[17]CentoV,MirabelliC,SalpiniR,etal.HCVgenotypesaredifferentlypronetothe developmentofresistancetolinearandmacrocyclicproteaseinhibitors.PLoS ONE2012;7:e39652.

[18]ZagordiO,BhattacharyaA,ErikssonN,etal.ShoRAH:estimatingthegenetic diversityofamixedsamplefromnext-generationsequencingdata.BMC Bioin-formatics2011;12:119.

[19]TamuraK,PetersonD,PetersonN,etal.MEGA5:molecularevolutionary genet-icsanalysisusingmaximumlikelihood,evolutionarydistance,andmaximum parsimonymethods.MolecularBiologyandEvolution2011;28:2731–9.

[20]NasuA,MarusawaH,UedaY,etal.GeneticheterogeneityofhepatitisCvirusin associationwithantiviraltherapydeterminedbyultra-deepsequencing.PLoS ONE2011;6:e24907.

[21]CentoV,DiMaioVC,DiPaoloD,etal.NS3-resistanceinadvancedHCVpatients treatedwithBOC/TRV-basedtherapy:impactonvirologicalfailureofbaseline andearlytimepoints.TheLiverMeeting2013,AASLDAbstracts.Hepatology 2013;58(Suppl.1):1960.

[22]TaylorN,Haschke-BecherE,GreilR,etal.Performancecharacteristicsofthe COBASAmpliprep/COBASTaqManv2.0andtheAbbottRealTimehepatitisC assays–implicationsforresponse-guidedtherapyingenotype1infections. AntiviralTherapy2014;19:449–54.