Detection of multidrug-resistant Mycobacterium tuberculosis strains isolated in Brazil using a multimarker genetic assay for katG and rpoB genes

w w w . e l s e v ie r . c o m / l o c a t e / b j i d

The

Brazilian

Journal

of

INFECTIOUS

DISEASES

Original

article

Detection

of

multidrug-resistant

Mycobacterium

tuberculosis

strains

isolated

in

Brazil

using

a

multimarker

genetic

assay

for

katG

and

rpoB

genes

Luita

Nice

Café

Oliveira

_,

_Jairo

_da

_Silva

_{Muniz-Sobrinho}

_,

Luiz

Alexandre

Viana-Magno

,

Sônia

Cristina

Oliveira

Melo

,

Antonio

Macho

_,

_Fabrício

_Rios-Santos

a_{LaboratóriodeFarmacogenômicaeEpidemiologiaMolecular(LAFEM),UniversidadeEstadualdeSantaCruz(UESC),Ilhéus,BA,Brazil} b_{InstitutoNacionaldeCiênciaeTecnologiaemMedicinaMolecular(INCT-MM),FaculdadedeMedicina,}

UniversidadeFederaldeMinasGerais(UFMG),BeloHorizonte,MG,Brazil

c_{FaculdadeInfóriumdeTecnologia,MestradoemTecnologiadaInformac¸ãoAplicadaaBiologiaComputacional,BeloHorizonte,MG,} Brazil

d_{DepartamentodeCiênciasBiológicas,UniversidadeEstadualdeSantaCruz(UESC),Ilhéus,BA,Brazil}

e_{DepartamentodeCiênciasBásicasemSaúde,FaculdadedeMedicina,UniversidadeFederaldeMatoGrosso(UFMT),Cuiabá,MT,Brazil}

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received11September2015 Accepted26December2015 Availableonline10February2016 Keywords: Biomarkers Diagnosis Multidrug-resistance Mycobacteriumtuberculosis

a

b

s

t

r

a

c

t

Multidrug-resistanttuberculosis(MDRTB)isaseriousworldhealthproblemthatlimits pub-licactionstocontroltuberculosis,becausethemostusedanti-tuberculosisfirst-linedrugs failtostopmycobacteriumspread.Consequently,aquickdetectionthroughmolecular diag-nosisisessentialtoreducemorbidityandmedicalcosts.Despitetheavailabilityofseveral molecular-basedcommercial-kitstodiagnosemultidrug-resistanttuberculosis,their diag-nosticvaluemightdivergeworldwidesinceMycobacteriumtuberculosisgeneticvariability differsaccordingtogeographiclocation.

Here,westudiedthepredictivevalueoffourcommonmycobacterialmutationsinstrains isolatedfromendemicareasofBrazil.Mutationswerefoundatthefrequencyof41.9%for katG,25.6%forinhA,and69.8%forrpoBgenesinmultidrug-resistantstrains.Multimarker analysisrevealedthatcombinationofonlytwomutations(“katG/S315T+rpoB/S531L”)wasa bettersurrogateofmultidrug-resistanttuberculosisthansingle-markeranalysis(86% sensi-tivityvs.62.8%).Predictionofmultidrug-resistanttuberculosiswasnotimprovedbyadding athirdorfourthmutationinthemodel.Therefore,ratherthanusingdiagnostickits detec-tingseveralmutations,weproposeasimpledual-markerpaneltodetectmultidrug-resistant tuberculosis,with86%sensitivityand100%specificity.Inconclusion,thisapproach (previ-ousgeneticstudy+analysisofonlyprevalentmarkers)wouldconsiderablydecreasethe processingcostswhileretainingdiagnosticaccuracy.

©2016ElsevierEditoraLtda.ThisisanopenaccessarticleundertheCCBY-NC-ND license(http://creativecommons.org/licenses/by-nc-nd/4.0/).

∗ _{Correspondingauthorat:DepartamentodeCiênciasBásicasemSaúde,FaculdadedeMedicina,UniversidadeFederaldeMatoGrosso}

(UFMT),Av.FernandoCorrêadaCosta,n◦_{2367,Cuiabá,MT78060-900,Brazil.}

E-mailaddress:[email protected](F.Rios-Santos). http://dx.doi.org/10.1016/j.bjid.2015.12.008

1413-8670/©2016 Elsevier Editora Ltda. This is an open access article under the CC BY-NC-ND license (http://creativecommons. org/licenses/by-nc-nd/4.0/).

Introduction

Tuberculosis(TB)remainsanimportantpublichealthconcern asaboutonethirdoftheworldpopulationis asymptomati-callyinfectedwithlatentMycobacteriumtuberculosis.1Multiple healthcampaignsare beingaddressedtodiminishthe inci-denceofthedisease,butTBeliminationremainsachallenge, duetotheemergenceofclinicalformsofmultidrug-resistant strainsofM.tuberculosis,1_{thatareresistanttoatleastisoniazid}

(INH)andrifampicin(RIF),thefirst-lineantitubercular (anti-TB)drugs.BeyondincreasingTBincidence,mortality,andthe rateofinitialtreatmentfailure,multidrug-resistant tubercu-losis(MDRTB) also increase the costof TBtreatment.1 _For

example,whilea6-monthregimenofINH+RIF(plus pyraz-inamide and streptomycin or ethambutol in the first two months) willcuremostcasesofnon-resistant TB, MDRTB-patientswillneedatleast12–24monthsoftherapywithmore toxicsecond-linedrugs.1,2_{Inaddition,MDRTBaccountsforup}

to84%ofthe“retreatment”casesofTB.1_Withthis

perspec-tive,earlydiagnosisappearstobethebeststrategytocontrol MDRTB.1

Conventional culture-based drug-susceptibility testing (DST)forTBcantakeasmuchas90daystocomplete,due totheslowgrowthofM.tuberculosis.3_{Sincediagnosticdelay}

isanarduousobstacletoeffectiveMDRTBcare,development andimplementationofmolecularapproachesforrapid detec-tionofMDRTBareneededforattenuatingtheMDRTBburden. Towardsthatend,manystudieshaveinvestigatedgene muta-tions that confer drugresistance inM. tuberculosis.4–6 _{As a}

result,manycommercialtestshavebeendevelopedtodetect MDRTBand,consequently,theWHOrecommendstoall gov-ernments the gradual implementation of evidence based commercial testsusing moleculartechniques fordiagnosis andcontrolofMDRTB.1

Overall, these tests aim to detect mutations in the mycobacterialgenesinvolvedindrugmetabolism,whichmay conferphenotypic changesassociatedwithantibiotic resis-tance. In this way,INH resistance is frequently associated withmutationsinthecatalase-peroxidaseenzyme(encoded by the katG gene and involved in metabolic activation of thedrug),5_{and/orintheenoyl-ACPreductasegenepromoter}

(inhA),requiredformycobacterialcellwallbiosynthesis.7

Like-wise,resistancetoRIFisstronglyrelatedtomutationsina regionof81bpintherpoBgene(encodingthebetasubunitof theRNApolymerase,thedrugtarget),3,8_{resultingindecreased}

affinityofRIFfortheactivecentreoftheenzyme.9_Mutations

inseveralothergenesmayalsoleadtoINHorRIFresistance butare lessfrequent.10–12 _{Thus,common genemarkersfor}

MDRTBareS315TinkatG,−15C/T(CGT-CAT)intheoperator regionofinhA,andH526DandS531LinrpoB.4,5

Diagnosticparameters,suchassensitivityandspecificity, ofcommerciallyavailablediagnostictestsarepoorly under-stood in a worldwide context.1 _{This is of interest since}

thereisevidencethatgeographiclocationinfluencesonthe strain-to-strainM.tuberculosisgeneticvariability.13–15_Inthis

context,currentgold-standardmolecularbiomarkersto diag-noseMDRTBmightnotbefoundatreasonablefrequencies to guide anti-TB drug choicein neglected regions such as Brazil.Moreover,itisevenpossiblethatnon-trivialmutations

couldturnout tobenovelbiomarkersforMDRTBdiagnosis insuchregions.Inthisregard,wegenotypedfour mycobacte-rialSNPs(singlenucleotidepolymorphisms)inisolatesfrom TBpatientsofanendemicregionofBraziltoinvestigatetheir predictivevaluetodiagnoseMDRTB,andusingthesenew ref-erencestandardsasgold-standard.Ouraimwastotestand validate a rapidand cost-effectivereal-time PCR assay for detectingRIFandINHresistance.

Methods

Isolation,identificationanddrugsensitivitytestsofM.

tuberculosis

M.tuberculosisisolateswereconsideredasMDRTBwhen resis-tant to at least INH and RIF (worldwide used as first-line antituberculardrugs),andassensitivewhensusceptibletoall anti-tuberculardrugs. Therefore,mono-resistant strainsfor INHorRIF,aswellasotherpoly-drug-resistancenotincluding bothINHandRIFwerenotconsideredforfurtherexperiments inourstudy.Theseisolateswererandomlyselectedand col-lectedfromarepositorylocatedattheCentralPublicHealth Laboratory“Prof.Gonc¸aloMuniz”(LACEN-BA),stateofBahia, Brazil.Allisolatesfromthecollectioncomefromsputumof localpatientswhich completedademographicand clinical questionnaireaftersigninganinformedconsentform.All pro-cedureswereapprovedbytheHumanEthicalCommitteeof theUniversidadeEstadualdeSantaCruz(UESC;Ilhéus,Brazil), underprotocolnumber098/07.

The selected samples were cultivated on Löwenstein-Jensenagar(LJ).Biochemicaltests,includingnitratereduction, niacin test and 68◦_{C catalase inhibition were conducted} in order to classify these strains according to the Man-ual ofTuberculosis Bacteriology.16 _{Drug sensitivitytests to}

INH,RIF,ethambutol(EMB),pyrazinamide(PZA),and strepto-mycin(SM)wereperformedusingtheCanetti(1969)multiple proportionaldilutionmethod.17_{ThestandardstrainsM.}

tuber-culosis H37Rv and 636 were used as sensitive and MDRTB controls,respectively,asrecommendedbytheBrazilian Min-istry of Health.16 _{The minimum inhibitory concentrations}

(MIC), defined as the lowest drug concentration showing complete inhibition of bacterial growth, were as follow: INH≤0.2␮g/mL,RIF≤1␮g/mL,EMB≤5␮g/mL,SM≤2␮g/mL, and PZA≤25␮g/mL. The criterion for drug resistance was growthof≥1%ofthebacterialpopulationonmediacontaining thecriticalconcentrationofeachdrug.

Genotypingassayandamplification

MycobacterialgenomicDNAwasextractedfromthecultured strainsaccordingtoVanSoolingenetal.18_{Thequantityand}

puritywasdeterminedmeasuringspectrophotometricsignals at260nmand280nm.

Assayswereperformed,inaccordancetoEspasaetal.,19

in a spectrofluorometric thermal cycler (ABI Prism 7500, AppliedBiosystems®_{,Carlsbad,CA,USA),usingtheprimers} andprobessummarizedinTable1.TheMycobacterium tuber-culosisComplex-SpecificInsertionSequenceIS6110wasalso amplifiedasinternalcontrol.Fiveamplificationreactionswere

Table1–Primersandprobesdesignedtodetectmutationsincludedinthestudy. Gene (codon) Primers Probes IS6110a ₅′_{-GGATAACGTCTTTCAGGTCGAGTAC-3}′ 5′_{-TCGCCTACGTGGCCTTTG-3}′ 5′_{-GGATAACGTCTTTCAGGTCGAGTAC-3}′ katG(315) 5′_{-GGGCTGGAAGAGCTCGTATG-3}′ 5′_{-GGAAACTGTTGTCCCATTTCG-3}′ 5′_{-CGACCTCGATGCCGCTGGTGAT-3}′ 5′_{-CGACCTCGATGCCGGTGGTGAT-3}′ inhA(−15) 5′_{-CACGTTACGCTCGTGGACAT-3}′ 5′_{-CAGGACTGAACGGGATACGAA-3}′ 5′_{-AACCTATCGTCTCGCCGCGGC-3}′ 5′_{-AACCTATCATCTCGCCGCGGC-3}′ rpoB(526) 5′ -ACCGCAGACGTTGATCAACAT-3′ 5′ -GGCACGCTCACGTGACAG-3′ 5′ -CGCTTGTGGGTCAACCCCGA-3′ 5′ -CGGCGCTTGTAGGTCAACCCC-3′ rpoB(531) 5′_{-ACCGCAGACGTTGATCAACAT-3}′ 5′_{-GGCACGCTCACGTGACAG-3}′ 5′_{-AGCGCCGACAGTCGGCG-3}′ 5′_{-CAGCGCCAACAGTCGGCG-3}′

a _{MycobacteriumtuberculosisComplex-SpecificInsertionSequence.}

performedforeachstudiedsample.Briefly,thereaction con-tained:1XMasterMixsolution,500nMofeachprimer,200nM of fluorogenic 5′ _{exonuclease probes (Taqman}®_{, Applied} Biosystems®_{,FosterCity,CA,USA),and4␮L(20ng/␮L)of} sam-plein 25␮Lof reaction volume.The amplificationsettings werecarriedoutaccordingtothemanufacturer’sinstructions. AnalysisoftheresultswasperformedusingtheSequence DetectionSystem software (Applied Biosystems®_{)and was} basedontwoparameters:thecyclethreshold(CT),forwhich thedetectionoffluorescencereachesthethreshold,andthe cumulativefluorescencesignal(CFS)ofeachprobeattheend of40amplificationcycles.

Statisticalanalysis

Pearsonchi-squaretestwasappliedtoevaluatedifferences in genotypic frequencies between sensitive and resistant isolates.Discriminantanalyseswereperformedtobuild pre-dictivemodelsofINHandRIFresistancebasedongenotyping dataof the four geneticmarkers used, alone or in combi-nation.Comparisonoftheobservedgeneticfrequencieswas performed using UNPHASED software (v.3.1.7).20 _Statistical

measuresofsensitivityandspecificitywereperformedusing IBMSPSSStatisticsversion20(SPSSInc.,Armonk,NY,USA) witha95%confidenceinterval.Thep-valuesignificancewas alwayssetat5%(p<0.05).

Results

First,aconfirmatorysensitivitytestwasperformedonatotal of51resistantand50sensitivecataloguedstrains(Table2). From those, 43(84.3% ofresistant-strains)were considered asMDRTBstrainsexhibitingresistancetobothINHandRIF, andwereselectedforfurtherexperiments.Sixofthestrains showed drug-resistance to only one antibiotic (INH), and onlytwoexhibitedmulti-resistanceotherthan INHandRIF together.Theother50strains,cataloguedassensitive,were indeedconfirmedtobesensitivetoallfiveantibioticstested, andwereusedasnegativecontrols.

FrequencyofpolymorphismsinkatG,inhAandrpoB

genesandtheirassociationwithdrugresistance

As expected,4,5 _{none of the studied mutations was found}

in drug-susceptiblestrains (Table 3). Although this finding

suggests that absenceofthe analyzedkatG,inhA, andrpoB mutationsisindicativeofnon-MDRTBstrainswith100% speci-ficity,itisnoteworthythatasmallnumberofMDRTBsamples (14%, n=6)alsolackedthesemutations(Table3).Thus,our datarevealthattheabsenceofthesecommonkatG,inhA,and rpoBmutationsdoesnotnecessarilyassurenon-MDRTB.

Conversely, wefoundthat the occurrenceof any ofthe mutationstranslatesidentificationofMDRTB.Amongthe43 strainsstudiedforMDRTB,18displayedtheS315T(katG) muta-tion, 11 had the −15C/T (inhA) whereas, 3 and 27 strains showedH526DandS531Lpolymorphisms,respectively,inrpoB (Table3).Althoughdisplaying100%specificity,sensitivityto detect MDRTB varies according to the identified mutation. Mutationswithgreatersensitivity(62.8%)wereS531L(rpoB), followedbyS315T(katG)and−15C/T(inhA),with41.9%and 21.6%,respectively(Table3).TheH526mutation(rpoB)showed only7%sensitivitywithoutstatisticallysignificancetodetect drugresistance.

Given that bacteria may reduce effectivenessof a drug throughvariousbiologicalmechanismsatthesametime,we addressedwhetherdiagnosissensitivitycouldbegreaterthan 62.8% (displayedby the single markerS315T katG) in pre-dictionmodelsincludingmultiplemutations.Inthisregard, westructuredfourmodelsoftwoorthreeSNPs,as summa-rizedinTable4(notethatcombinationswithH526fromrpoB were excluded due tothe lackofstatistical significanceto detect drug resistance).Amongthem,onlymodel1(S315T

Table2–Drugresistanceprofileinsamplesrandomly obtainedfrombiobank.

Resistancetoeach drug wtstrains (n=50)a Resistantstrains (n=51)a INHb _{0(0) 100.0%(51)} RIFb _{0(0) 84.3%(43)} EMB 0(0) 5.9%(3) SM 0(0) 54.9%(28) PZA 0(0) 47.1%(24)

Basalmediumw/o

antibiotics

100%(50) 100%(51)

a _{Previouslycataloguedaswtorresistant-strainsbybiochemical}

tests.Valuesare%ofn(absolutevalues).

b _{ResistancetobothINHandRIFwasdefinedasMultiDrug}

Table3–FrequencyofmutationsinkatG,inhAandrpoBgenesofMycobacteriumtuberculosisstrains. Mutation (gene) Frequencyin susceptiblestrains (n=50)a FrequencyinMDRTB strains(n=43)a X2b _p-Valuec S315T (katG) 0(0) 41.9%(18) 25.953 <0.001 −15C/T (inhA) 0(0) 25.6%(11) 14.507 <0.001 H526D (rpoB) 0(0) 7.0%(3) 3.605 0.095 S531L (rpoB) 0(0) 62.8%(27) 44.239 <0.001 Noneof above 100%(50) 14.0%(6) – –

a _{Previouslycharacterizedbybiochemicaltests.SeeTable1.Valuesare%ofn(absolutevalues).}

b _{Pearsonchi-squaretest.}

c _{p-Valuesfromcomparisonbetweensusceptibleandresistantstrains.Boldvalueswereconsideredstatisticallysignificant.}

katG+−15C/T_inhA) modeldid notreach the “basal” sensi-tivity of 62.8%. The gain in sensitivity of model 2 (S531L rpoB+−15C/T inhA) was only about 2.3%, when compared to the best single marker. In contrast, models 3 (S315T katG+S531L rpoB) and 4 (S315T katG+S531L rpoB+−15C/T inhA)showedaconsiderablegainindiagnosissensitivity,by increasingthisparameterby23.2%.Usingthelasttwomodels, weidentified37out43MDRTBstrainscarryingatleastoneof theanalyzedmarkers.Thus,takeninconsiderationthatboth modelshaveidenticaldiagnosticpredictivevalues,this sug-geststhatitisunnecessarytogenotypethreeofthestudied polymorphismstoachievegreatersensitivity.

Discussion

Itisevidentthatstrains, aswell astypeofmutationsand frequency ofthem in mycobacterial genome, vary accord-ing tothegeographical region.21,22 _{Consequently,sincethe}

accuracy of current molecular diagnostic tests to identify MDRTBarebasedonworldwidepolymorphismsfrequencies, performanceindicatorsforthesediagnosticapproachesina country-by-countrybasismaybecontroversial.Forexample, theS315T(katG)polymorphismshowsawidevariationintheir incidence: althoughit was detected ina 100% ofresistant strainsincountriessuchasTurkey,Canada,andFrance,23–25

inourstudywefoundalowerfrequencyof41.9%,more sim-ilartothoseobtainedinanotherstudyinBrazil(60%)andin placessuchasRussia(77%),Syria(40%)orTaiwan(51%).13,26–28

Likewise,the−15C/T(inhA)polymorphism,describedasthe onemostfrequentgloballyfortheinhAgenepromoter,5,29_has

anobservedfrequencyvaryingbetween15%inChina30_to32%

inSyria,13_{and25.6%inthisstudy.Also,thefrequencyof62.8%}

forS531L(rpoB)foundinthisworkwasconsistentwithother studies,consideringthatpublishedratesrangebetween40.1% and82.4%.31–34_{Inthesameway,H526D(rpoB)polymorphism,}

withafoundfrequencyof7%inthiswork amplyoscillates worldwide,withvaluesrangingfrom5.9%to40%.31,34

There-fore,greatfrequencyvariationsareevidentbetweencountries, making it necessary to study locally the prevalenceof the canonical mutations foreach geographical region in order tofigureoutwhichpolymorphismshavethebestpredictive accuracy.

AsearlydiagnosisisthebeststrategytocontrolMDRTB, several molecular methods todetect drug resistancein M. tuberculosisarecurrentlywidelyemployed.35_Forthat,a

num-ber of commercial kits to identify MDRTB are available, suchasXpertMTB/RIF®_{(Cepheid,Sunnyvale,CA,USA)and} INNO-LiPARif.TB(InnogeneticsN.V.,Gent,Belgium),among others, which detect resistance to only rifampicin not to isoniazid.36,37_{Therationalforassessingonlymutations}

asso-ciated RIF-resistance is that some reports have suggested thatRIF-resistanceseemstoserveasasurrogatemarkerfor MDRTBdetection,38,39_{asalmostallRIF-resistantstrainswere}

alsoINH-resistant.Inourstudy,wefoundalsothat100%of RIF-resistantstrains werealsoINH-resistant(Table2). Nev-ertheless, this finding depends of the endemic strains of M.tuberculosisineachgeographicalregion,21,22 _asdiscussed

Table4–PredictiveprofileofpolymorphismsinkatG,inhAandrpoBgenesrelatedtomultidrugresistancein

Mycobacteriumtuberculosisstrains.

Genes Polymorphisms Statisticalmodels Frequencyinresistant

strainsa_(n=43)

katGorinhA S315T&−15C/T Model1 60.5%(26)

inhAorrpoB −15C/T&S531L Model2 65.1%(28)

katGorrpoB S315T&S531L Model3 86.0%(37)

katGorinhAorrpoB S315T&−15C/T&S531L Model4 86.0%(37)

above.Inthis way,somereportsfrom thePunjab regionof India40,41_{haveshownahighincidenceofRIF-monoresistant}

strains(22%)thatwouldbefalselydiagnosedasMDRTBusing thesetypeoftest.TheXpert®_{MTB/RIFassay,muchcheaper} thanINNO-LiPARif.TB(approximatelyUS$17andUS$45per test,respectively),42,43_{hasbeenrecommendedbytheWorld}

HealthOrganizationforthe detectionofMDRTBfrom 2010, and is exhaustively implemented worldwide.1,44 _However,

onlyalimitednumberoftestscanbeperformedperday, mak-ingitnecessarytoacquireseveralmachinesperlaboratory. Furthermore,thecostpertest,45_{evenifnegotiatedforpublic}

institutionsbytheFINDorganization43_{(Foundationfor}

Inno-vativeNewDiagnostics,Geneva,Switzerland)(Table5),makes oursystemaninterestingcost/benefitdiagnostictoolforour localarea.Inaddition,thesekitshavesensitivityand speci-ficitytodetectdrugresistancesimilartothoseshowninthis study36,37,45_(Table5).

Recently,therehasbeenreportedamoreaccurate commer-cialkit,theGenoTypeMTBDRplus(HainLifescience,Nehren, Germany)whichhashighersensitivitythancompetitors(up to94.4%)andcoulddetect evenresistancetobothINHand RIF.46_{Nevertheless,itrequirestheanalysisofupto10markers}

todiagnoseMDRTB,whereasXpert®_{MTB/RIFandINNO-LiPA} Rif.TBuse5and4,respectively,42,43_{increasingevenmorethe}

processingcosts (betweenD_{42 and} D_{86 per test).}47 _{In our}

study,the proposedanalysisofpolymorphismstodiagnose MDRTBintheBahiastateofBrasil,hasahighsensitivity(86%) evenifonlytwogeneticmarkersareanalyzed.

Inourcase(BahiastateofBrazil),afteridentificationofthe prevalentpolymorphisms,theprotocolforMDRTBanalysis, (Fig.1)wouldconsistinextractingtheDNAfromM. tuberculo-sisdirectlyfromsputumsamples.Thenthesampleswouldbe subjectedtogenotypingforthedetectionofpolymorphismsat S531LrpoBandS315TkatGgenes.PresenceofS531L(62.8%of incidenceinendemicmycobacteria)orS315T(41.9%) muta-tionsinthesampleswouldbediagnosedasMDRTB(86%of

Table5–Comparativeparametersofmolecular diagnosisofMultidrugResistantTuberculosisbetween thisstudyandpublisheddataforXpertMTB/RIF.

Analyticalvariables Thiswork XpertMTB/RIFa

Sensitivity 86.0% 89.2–91.4% Specificity 100.0% 98.0–98.7% No.ofpolymorphisms studied 2 5 Chemotherapyresistance studiedb

INH,RIF RIF

Time(DNAextraction+PCR

reaction)

3h 2h

Test/day(consideringa

workdayof8h)

20–30 8–16c

Priceof1test US$8.53 US$9.98–16.86d

a _{Source:ChangK,LuW,WangJ,etal.Rapidandeffectivediagnosis}

of tuberculosis and rifampicin resistance with Xpert mtb/rif

assay:Ameta-analysis.JInfect2012;64(6):580–588.45

http://www.finddiagnostics.org/about/whatwe do/successes/find-negotiated-prices/xpertmtbrif.html.43

b _{INH:isoniazid;RIF:Rifampicin.}

c _{Xpertmachinewithtwoorfourcartridges,respectively.}

d _{Negotiatedpriceforthepublicsectorineligiblecountriesandreal}

price,respectively.

cases,carryinganyoneofthetwomutations).Thespecificity of100%coupledwithhighaccuracy,easyoperation,andlow cost(duetothesmallernumberofmarkersused),makesita greatmethodtobeusedinroutinediagnosisofMDRTB.Our workshowsthatcombiningafewmarkers,withouttheneed ofanextensivepanelofanalysis,issufficienttoimprovethe accuracyforthedetectionofbothRIF-andINH-resistancein ourregion.

Our resultscame asno surprise, consideringthat most RIF-resistance mutationsoccurclose totheS531 residue,48

andalmostallRIF-resistantstrainsarealsoINH-resistant.38,39

Suspected

Multidrug-Resistant

Tuberculosis

(MDRTB)

Genotyping S315T mutation in katG gene and S531L mutation in rpoB gene

Detection of polymorphism in S315T katG gene?

Detection of polymorphism in S531L rpoB gene? NO

YES Classical Biochemical Sensitivity Tests

YES MDRTB (62.8%) MDRTB (41.9%) MDRTB (14.0%)

Combined apparition:

MDRTB (86.0%)

Fig.1–AlgorithmforthediagnosisofMultidrugResistanceTuberculosis(MDRTB)usingtwopredictivemarkersfor isoniazidandrifampicinresistance.

Moreover,theS315Tmutationisthemostfrequentlyfound inINH-resistant strains that are also MDRTB, occurring in approximately40%ofallcases.49

Forthesmallpercentageoffalse-negativesamples(14%), conventionalcultureanddrugsensitivitytechniquescanbe performedas confirmatory tests, as 86% ofthe multidrug-resistant population will be previously diagnosed in both effectiveandearly-timemanners.Inthiscontext,evenif fur-therstudiesshouldbeconductedtoestablishothermarkers toincreaseevenmorethesensitivityofthe proposed algo-rithm(Fig.1),takentogetherourresultshaveshownthatin ordertoprovideafast,sensitiveandlow-costdiagnosis,the proposedPCR-basedprotocolwouldbesufficiently appropri-ateforourgeographicalregion.Asdemonstratedpreviously byMadaniaetal.,13_{inSyria,asmallsetofprobesdetecting}

prevalentmutationscanbeusedinrealtimePCRinorderto detectmoststrainscarryingRIF-andINH-resistance. Inthe sameway,weproposethatouralgorithmcanberevisedand implementedinothercountriesbesidesBrazil,by conduct-ingasimplegeneticbackgroundevaluationoftheirendemic populationsofM.tuberculosis.

Conflicts

of

interest

Theauthorsdeclarenoconflictsofinterest.

Acknowledgments

Wethankthe LaboratórioCentral da Bahia(LACEN-BA) for providing isolates and drug susceptibility testing for M. tuberculosis. Wealsothank Dr.DaniellaOliveira Campo for the help during this research. This study was funded by the Fundac¸ão de Amparo a Pesquisa do Estado da Bahia (FAPESB) and the Conselho Nacional de Desenvolvimento CientíficoeTecnológico(CNPq).L.A.V.MandA.M.have post-doctoral fellowships from the Fundac¸ão Coordenac¸ão de Aperfeic¸oamentodePessoaldeNívelSuperior(CAPES).

r

e

f

e

r

e

n

c

e

s

1. WorldHealthOrganization(WHO).Globaltuberculosisreport 2013.1sted.Geneva,Switzerland:WorldHealthOrganization; 2013.

2. DaltonT,CegielskiP,AkksilpS,AsenciosL,CamposCaoiliJ, ChoSN,etal.Prevalenceofandriskfactorsforresistanceto second-linedrugsinpeoplewithmultidrug-resistant tuberculosisineightcountries:aprospectivecohortstudy. Lancet.2012;380:1406–17.

3. RiederLH,VanDeunA,KamKM,KimSJ,ChondeTM,Trébucq A,etal.Prioritiesfortuberculosisbacteriologyservicesin low-incomecountries.2nded.Paris,France:International UnionAgainstTuberculosisandLungDisease;2007,150pp. 4. ColeST.Rifamycinresistanceinmycobacteria.ResMicrobiol.

1996;147:48–52.

5. RamaswamyS,MusserJM.Moleculargeneticbasisof antimicrobialagentresistanceinMycobacteriumtuberculosis:

1998update.TuberLungDis.1998;79:3–29.

6.WuX,ZhangJ,ZhuangY,ZhangX,LiG,HeX.Molecular mechanismsofdrugresistanceinMycobacteriumtuberculosis

clinicalisolates.ChinMedJ(Engl).1999;112:524–8. 7.KarakousisPC.Mechanismsofactionandresistanceof

antimycobacterialagents.In:MayersDL,editor.Antimicrobial drugresistanceInfectiousDisease:HumanaPress;2009.p. 271–91.

8.MarrakchiH,LaneelleMA,DaffeM.Mycolicacids:structures, biosynthesis,andbeyond.ChemBiol.2014;21:67–85. 9.FeklistovA,MeklerV,JiangQ,WestbladeLF,IrschikH,Jansen

R,etal.Rifamycinsdonotfunctionbyallostericmodulation ofbindingofMg2+_{totheRNApolymeraseactivecenter.Proc}

NatAcadSciUSA.2008;105:14820–5.

10.GuoH,SeetQ,DenkinS,ParsonsL,ZhangY.Molecular characterizationofisoniazid-resistantclinicalisolatesof

MycobacteriumtuberculosisfromtheUSA.JMedMicrobiol. 2006;55Pt11:1527–31.

11.HeepM,BrandstatterB,RiegerU,LehnN,RichterE,

Rusch-GerdesS,etal.FrequencyofrpoBmutationsinsideand outsidetheclusterIregioninrifampin-resistantclinical

Mycobacteriumtuberculosisisolates.JClinMicrobiol. 2001;39:107–10.

12.LavenderC,GlobanM,SieversA,Billman-JacobeH,FyfeJ. Molecularcharacterizationofisoniazid-resistant

MycobacteriumtuberculosisisolatescollectedinAustralia. AntimicrobAgentsChemother.2005;49:4068–74. 13.MadaniaA,HabousM,ZarzourH,GhouryI,HebboB.

Characterizationofmutationscausingrifampicinand isoniazidresistanceofMycobacteriumtuberculosisinSyria.PolJ Microbiol.2012;61:23–32.

14.SalinaTI,MorozovaTI.Moleculargeneticanalysisofthe isoniazid-resistantstrainsofM.tuberculosiscirculatingover theSaratovregion.MolGenMikrobiolVirusol.2013:26–8[in Russian].

15.TelentiA,ImbodenP,MarchesiF,LowrieD,ColeS,ColstonMJ, etal.Detectionofrifampicin-resistancemutationsin

Mycobacteriumtuberculosis.Lancet.1993;341:647–50.

16.Brasil-MinistériodaSaúde,SecretariadeVigilânciaemSaúde, DepartamentodeVigilânciaEpidemiológica.Manualnacional devigilâncialaboratorialdatuberculoseeoutras

micobactérias.1sted.Brasilia,Brazil:MinistériodaSaúde; 2008,436pp.[inBrazilian].

17.CanettiG,FoxW,KhomenkoA,MahlerHT,MenonNK, MitchisonDA,etal.Advancesintechniquesoftesting mycobacterialdrugsensitivity,andtheuseofsensitivitytests intuberculosiscontrolprogrammes.BullWorldHealthOrgan. 1969;41:21–43.

18.vanSoolingenD,deHaasPE,HermansPW,GroenenPM,van EmbdenJD.ComparisonofvariousrepetitiveDNAelements asgeneticmarkersforstraindifferentiationand

epidemiologyofMycobacteriumtuberculosis.JClinMicrobiol. 1993;31:1987–95.

19.EspasaM,Gonzalez-MartinJ,AlcaideF,AragonLM,LoncaJ, ManterolaJM,etal.Directdetectioninclinicalsamplesof multiplegenemutationscausingresistanceofMycobacterium tuberculosistoisoniazidandrifampicinusingfluorogenic probes.JAntimicrobChemother.2005;55:860–5. 20.DudbridgeF.Likelihood-basedassociationanalysisfor

nuclearfamiliesandunrelatedsubjectswithmissing genotypedata.HumHered.2008;66:87–98.

21.FilliolI,MotiwalaAS,CavatoreM,QiW,HazbonMH,Bobadilla delValleM,etal.GlobalphylogenyofMycobacterium

tuberculosisbasedonsinglenucleotidepolymorphism(SNP) analysis:insightsintotuberculosisevolution,phylogenetic accuracyofotherDNAfingerprintingsystems,and

recommendationsforaminimalstandardSNPset.JBacteriol. 2006;188:759–72.

22.BrudeyK,DriscollJR,RigoutsL,ProdingerWM,GoriA, Al-HajojSA,etal.Mycobacteriumtuberculosiscomplexgenetic

diversity:miningthefourthinternationalspoligotyping database(SpolDB4)forclassification,populationgeneticsand epidemiology.BMCMicrobiol.2006;6:23.

23.BlackwoodKS,HeC,GuntonJ,TurenneCY,KabaniJW,Kabani AM.EvaluationofrecAsequencesforidentificationof

Mycobacteriumspecies.JClinMicrobiol.2000;38:2846–52. 24.DurmazR,GunalS,YangZ,OzerolIH,CaveMD.Molecular

epidemiologyoftuberculosisinTurkey.ClinMicrobiolInfect. 2003;9:873–7.

25.Elia-PasquetS,DabisF,Texier-MaugienJ,Dessus-BabusS, MeynardJ,BouigesM,etal.Transmissionoftuberculosisin Gironde:epidemiologicinvestigationbygenomicanalysisof

Mycobacteriumtuberculosis.RevEpidemiolSantePublique. 2000;48:127–36[inFrench].

26.Afanas’evMV,IkryannikovaLN,Il’inaEN,SidorenkoSV, Kuz’minAV,LarionovaEE,etal.Molecularcharacteristicsof rifampicin-andisoniazid-resistantMycobacteriumtuberculosis

isolatesfromtheRussianFederation.JAntimicrob Chemother.2007;59:1057–64.

27.HöflingCC,PavanEM,GiampagliaCM,FerrazoliL,AilyDC,de AlbuquerqueDM,etal.PrevalenceofkatGSer315

substitutionandrpoBmutationsinisoniazid-resistant

MycobacteriumtuberculosisisolatesfromBrazil.IntJTuberc LungDis.2005;9:87–93.

28.HuangWL,HsuZJ,ChangTC,JouR.Rapidandaccurate detectionofrifampinandisoniazid-resistantMycobacterium tuberculosisusinganoligonucleotidearray.ClinMicrobiol Infect.2014.

29.RouseDA,LiZ,BaiGH,MorrisSL.Characterizationofthe katGandinhAgenesofisoniazid-resistantclinicalisolatesof

Mycobacteriumtuberculosis.AntimicrobAgentsChemother. 1995;39:2472–7.

30.ZhangM,YueJ,YangYP,ZhangHM,LeiJQ,JinRL,etal. Detectionofmutationsassociatedwithisoniazidresistance inMycobacteriumtuberculosisisolatesfromChina.JClin Microbiol.2005;43:5477–82.

31.CastanP,dePabloA,Fernandez-RomeroN,RubioJM,Cobb BD,MingoranceJ,etal.Point-of-caresystemfordetectionof

Mycobacteriumtuberculosisandrifampinresistanceinsputum samples.JClinMicrobiol.2014;52:502–7.

32.JamiesonFB,GuthrieJL,NeemuchwalaA,LastovetskaO, MelanoRG,MehaffyC.ProfilingofrpoBMutationsandMICs torifampicinandrifabutininMycobacteriumtuberculosis.JClin Microbiol.2014;52:2157–62.

33.ZakhamF,ChaouiI,EchchaouiAH,ChetiouiF,Elmessaoudi MD,EnnajiMM,etal.Directsequencingforrapiddetectionof multidrugresistantMycobacteriumtuberculosisstrainsin Morocco.InfectDrugResist.2013;6:207–13.

34.ZhaoLL,ChenY,LiuHC,XiaQ,WuXC,SunQ,etal.Molecular characterizationofmultidrug-resistantMycobacterium tuberculosisisolatesfromChina.AntimicrobAgents Chemother.2014;58:1997–2005.

35.WangX,JiaoJ,XuW,ChaiX,LiZ,WangQ.Asimple,rapidand economicmethodfordetectingmultidrug-resistant

tuberculosis.BrazJInfectDis.2013;17:667–71. 36.SteingartKR,SchillerI,HorneDJ,PaiM,BoehmeCC,

DendukuriN.Xpert®_{MTB/RIFassayforpulmonary}

tuberculosisandrifampicinresistanceinadults.Cochrane DatabaseSystRev.2014;1:CD009593.

37.ViveirosM,LeandroC,RodriguesL,AlmeidaJ,BettencourtR, CoutoI,etal.DirectapplicationoftheINNO-LiPARif.TB line-probeassayforrapididentificationofMycobacterium tuberculosiscomplexstrainsanddetectionofrifampin resistancein360smear-positiverespiratoryspecimensfrom anareaofhighincidenceofmultidrug-resistanttuberculosis. JClinMicrobiol.2005;43:4880–4.

38.EwenKR,BahloM,TreloarSA,LevinsonDF,MowryB,Barlow JW,etal.Identificationandanalysisoferrortypesin high-throughputgenotyping.AmJHumGenet. 2000;67:727–36.

39.YaoC,ZhuT,LiY,ZhangL,ZhangB,HuangJ,etal.Detection ofrpoB,katGandinhAgenemutationsinMycobacterium tuberculosisclinicalisolatesfromChongqingasdetermined bymicroarray.ClinMicrobiolInfect.2010;16:

1639–43.

40.RufaiSB,KumarP,SinghA,PrajapatiS,BalooniV,SinghS. ComparisonofXpertMTB/RIFwithlineprobeassayfor detectionofrifampin-monoresistantMycobacterium tuberculosis.JClinMicrobiol.2014;52:1846–52.

41.KumarP,BalooniV,SharmaBK,KapilV,SachdevaKS,SinghS. Highdegreeofmulti-drugresistanceandhetero-resistancein pulmonaryTBpatientsfromPunjabstateofIndia.

Tuberculosis(Edinb).2014;94:73–80.

42.MorganM,KalantriS,FloresL,PaiM.Acommerciallineprobe assayfortherapiddetectionofrifampicinresistancein

Mycobacteriumtuberculosis:asystematicreviewand meta-analysis.BMCInfectDis.2005;5:62.

43.http://www.finddiagnostics.org/about/whatwedo/successes/ find-negotiated-prices/xpertmtbrif.html[06.09.15].

44.WorldHealthOrganization(WHO).Automatedreal-time nucleicacidamplificationtechnologyforrapidand simultaneousdetectionoftuberculosisandrifampicin resistance:XpertMTB/RIFsystem.Geneva:WorldHealth Organization(WHO);2011.p.1–36.

45.ChangK,LuW,WangJ,ZhangK,JiaS,LiF,etal.Rapidand effectivediagnosisoftuberculosisandrifampicinresistance withXpertMTB/RIFassay:ameta-analysis.JInfect. 2012;64:580–8.

46.Anek-VorapongR,SinthuwattanawiboolC,PodewilsLJ, McCarthyK,NgamlertK,PromsarinB,etal.Validationofthe GenoTypeMTBDRplusassayfordetectionofMDR-TBina publichealthlaboratoryinThailand.BMCInfectDis. 2010;10:123.

47.WHOExpertGroupReport.Molecularlineprobeassaysfor rapidscreeningofpatientsatriskofmulti-drugresistant tuberculosis(MDR-TB).Geneva:WorldHealthOrganization; 2008.Availablefrom:http://www.who.int/tb/featuresarchive/ expertgroupreportjune08.pdf

48.EderthJ,MooneyRA,IsakssonLA,LandickR.Functional interplaybetweenthejawdomainofbacterial

RNApolymeraseandallele-specificresiduesinthe productRNA-bindingpocket.JMolBiol.2006;356: 1163–79.

49.HazbonMH,BrimacombeM,BobadilladelValleM,Cavatore M,GuerreroMI,Varma-BasilM,etal.Populationgenetics studyofisoniazidresistancemutationsandevolutionof multidrug-resistantMycobacteriumtuberculosis.Antimicrob AgentsChemother.2006;50:2640–9.