Antibiotic Resistance and Its Cost

8/3/2019 Antibiotic Resistance and Its Cost

1/12

The intrctin f ntibitics is ne f the mst imr-tnt meic interventins with regr t recinghmn mrbiity n mrtity. Hwever, the intensivese f ntibitics (which ws estimte in 2002 t be100,000200,000 tnnes er nnm wrwie1 nwhich is, in tt, we ver 1 miin tnnes since the1940s) hs rmticy increse the freqency fresistnce mng hmn thgens n thretens ssf theretic tins n st-ntibitic er in whichthe meic vnces t te re negte24. Resistncermticy reces the ssibiity f treting infectinseffectivey n increses the risk f cmictins n f ft tcme.

Resistnce is ften sscite with rece bcte-ri fitness, n it hs been rse tht rectin inntibitic se (n, therefre, in the seective ressre tcqire resistnce) w benefit the fitter sscetibebcteri, enbing them t tcmete resistnt strins

ver time5,6. Exeriment sties n theretic m-eing srt this bsic cncet5,7, bt ther rcesses,sch s cmenstry evtin n genetic c-seectin,cmicte the ictre n mke reversibiity ess rbbein re-ife settings. This Review fcses n exerimen-t sties f the fctrs (in rticr, the fitness csts fresistnce) tht infence the evement f resistncen the tenti fr reversibiity. We escribe the t-cmes f brtry sties n cinic interventins trece the freqency f resistnt bcteri in inivitients s we s t the cmmnity eve n exin whyreversibiity, if it ccrs t , rcees s swy tht, inmst cses, it is nikey t be f rctic imrtnce.

The development of resistance and its reversibility

The rate of appearance of resistant bacteria. The rte fernce f ntibitic-resistnt bcteri is etermineby the cmbine rtes fde novo mttin n hri-znt gene trnsfer (HGT) f resistnce eterminnts(FIG. 1). The mst cmmn mttins e t tertinsin the ntibitic trget n increses in rg effx8, btresistnce is s sscite with gene mifictin9,10,rece exressin f the trget11 n tertin f rgmifictin enzymes12. Mechnisms sscite withHGT ince rg mifictin, trget rtectin,bypass resistance, recement f sscetibe rg trgetsn cqisitin f nve effx ms13. Rtes f mt-tin cn be mesre fr mst cinic thgens, btthere is very itte infrmtin fr HGT n which t bsereictins regring the srces f resistnce genes14 rthe rtes f trnsfer f these genes thrgh the gene .This mkes it iffict t reict hw HGT wi infence

the emergence f resistnce t nve rgs15.

Selection of resistant bacteria. Resistnce becmes cinic rbem when the freqency f the resistnt vri-nt thretens the effectiveness f emiric rg thery.The eve f exsre f the thgen tin t therg infences the seectin f resistnt vrints 16,17.The exsre eve is sscite with pharmacokinetic npharmacodynamic properties f the rg, which ffect thecernce f the thgen r the seectin f resistnt

vrints in the tient. It is s sscite with hygienicmesres n trnsmissin cntr in cinics n hsi-ts n with the vme n istribtin f ntibitics

*Department of Medical

Biochemistry and

Microbiology, Uppsala

University, BOX 582, SE751

23, Uppsala, Sweden.Department of Cell and

Molecular Biology, Uppsala

University, BOX 596, SE751

24, Uppsala, Sweden.

Correspondence to D.I.A.

email: dan.andersson@

imbim.uu.se

doi:10.1038/nrmicro2319

Published online

8 March 2010

FitnessThe capability of a genotype or

individual to survive and

reproduce.

Bypass resistanceThe replacement (bypass) of a

metabolic step that is normally

inhibited by an antibiotic with a

new, drug-resistant metabolic

enzyme.

Antibiotic resistance and its cost:is it possible to reverse resistance?Dan I. Andersson* and Diarmaid Hughes

Abstract | Most antibiotic resistance mechanisms are associated with a fitness cost that

is typically observed as a reduced bacterial growth rate. The magnitude of this cost is

the main biological parameter that influences the rate of development of resistance, the

stability of the resistance and the rate at which the resistance might decrease if antibiotic

use were reduced. These findings suggest that the fitness costs of resistance will allowsusceptible bacteria to outcompete resistant bacteria if the selective pressure from

antibiotics is reduced. Unfortunately, the available data suggest that the rate of reversibility

will be slow at the community level. Here, we review the factors that influence the fitness

costs of antibiotic resistance, the ways by which bacteria can reduce these costs and the

possibility of exploiting them.

RE V I E W S

260 | apRI 2010 | VouME 8 www.au.m/ws/m

RE V I E W S

20 Macmillan Publishers Limited. All rights reserved10
mailto:[email protected]:[email protected]:[email protected]:[email protected]


2/12

Transformation

Conjugation

Transduction

Mutation

Pharmacokinetic propertiesCharacteristics of a drug that

include: its mechanisms of

absorption and distribution;

the rate at which its action

begins and the duration of the

effect; the chemical changes of

the agent in the body; and the

effects and routes of ecretion

of drug metabolites. Oftensummarized as what the body

does to a drug.

PharmacodynamicpropertiesCharacteristics of a drug that

include: the physiological

effects of a drug on the body,

on microorganisms or on

parasites in or on the body; the

mechanisms of drug action;

and the relationship between

drug concentration and effect.

Often summarized as what a

drug does to the body.

tht re reese int the envirnment, where they myseect fr resistnce n the trnsfer f resistnce byHGT. The thret frm ntibitics tht re reese intthe envirnment is istrte by the existence f rgereservir f resistnce genes in the hmn micrfrtht c tentiy serve s nrs fr the trnsfer fresistnce t hmn thgens18.

Biological costs of resistance and compensatory evolu-tion. antibitics trget imrtnt fnctins sch s cew synthesis, regtin f chrmsme serciing,RNa trnscritin n rtein synthesis, s it is nt sr-rising tht resistnt mtnts sy sffer ecrese in

bigic fitness (TABLE 1). The cqisitin f resistncegenes by HGT s crries fitness cst19,20. The ss-citin f resistnce with ecrese fitness hs feethe he tht rectin in the se f ntibitics we t rectin in the freqency f resistnt bcte-ri, thrgh ntr seectin. Hwever, resistnt bc-teri cn meirte the csts f resistnce by cqiringitin fitness-cmenstry mttins (TABLE 2).

Co-selection phenomena.C-seectin f resistnce tmre thn ne ntibitic, wing t genetic inkge fthe resistnce genes, is cmmn fetre f resistncecqire by HGT. The freqency f resistnce t n

ntibitic my rise, even if tht ntibitic is nt crrentyin se. Fr exme, recing the se f shnmiesh n effect n the freqency f sfnmie resistncein Escherichia coli21, ssiby becse the smi-brneshnmie resistnce gene is geneticy inke tther resistnce genes tht cntine t be seecte r-ing the sty eri. a recent exme in which trimeth-rim se ws rece in Sweish cnty is iscssebew. In rincie, c-seectin c s rive thefreqency f inke rg-sscetibe ee, eing tcn recement with this ee thrght the -tin. C-seectin s ccrs when resistnce t nerg cnfers increse resistnce t rgs with simirstrctre. This is eseciy rbemtic if ifferent vri-nts f rg re se in grictre n in meicine,with ifferent regtry cntrs. The evement fresistnce in ne envirnment my rive the freqencyf resistnce in nther envirnment.

The magnitude of fitness costs that is required for rever-sion of resistance. The time reqire t rece the bn-

nce f resistnt bcteri is inversey rete t the cstf resistnce22,23(BOx 1). In ne exeriment sty, thess f smi-ence tetrcycine resistnce frmE. coli ws mesre ver 500 genertins24. The best-fit mthemtic me ince high mttin fre-qency (~104 mttins er genertin) fr the ss ftetrcycine resistnce by eetin cmbine with smbt sttisticy significnt selection coefficient (s = 0.007)sscite with crrige f tetrcycine resistnce. In thiscse, s = 0.007 crresns t 0.7% rectin in grwthrte. With these rmeters, recement f 99.9% f thetin by sscetibe bcteri w tke rxi-mtey 1.5 yers. T rech the sme recement eve in5 weeks, the seectin cefficient sscite with resist-nce w nee t be t est 0.06, which is reistic fr cnstittivey exresse tetrcycine resistnce ernn high-cy-nmber smi bt nt fr n inceresistnce gene n w-cy-nmber smi with br hst rnge24. This exeriment srts thereti-c rgments tht in cmmnity settings the cst fresistnce is the key fctr t sccessf iscementf ntibitic-resistnt tins with ntibitic-sensitivenes23. Hwever, if even tiny frctin f resistnt bc-teri remin in the tin, the reintrctin f thentibitic is reicte t cse the resistnt bcteri trech high freqencies gin, t mch fster ce thnthe rigin ecine24.

Fitness costs and compensatory evolution

Estbishing the cnnectins between resistncen fitness reqires exeriments n isgenic strins(FIG. 2). This genertes n exeriment bsis n whicht interret the hentyes f cinic istes thtre sy f ncertin rigin. In rre cses, isgeniccinic istes tht iffer ny in their resistncen fitness hentyes hve been iste frm thesme tient25 r shwn by mecr nysis t beirecty rete26,27. anysing sch istes cn testthe rbstness f the mes tht re bit singbrtry strins.

Figure 1 | Masms ssa aqus. DNA from the biosphere containing

an antibiotic resistance gene (pink) can be transferred by horizontal gene transfer into

a recipient by several paths: cell-to-cell conjugation; transformation by naked DNA

(on plasmids or as linear DNA) that is released by dead cells; or phage-mediated

transduction. Resistance can also arise by de novo mutation (indicated by a red cross).

RE V I E W S

NaTuRE REVIEWS |Microbiology VouME 8 | apRI 2010 |261

f o c u S o n A n tI m I c Ro bI A l RE S I S tA n c E

http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12319http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12319


3/12

Selection coefficientA measure of the fitness of a

phenotype relative to wild type

(often denoted s), having a

value between 0 and 1. When

s=0, there is no fitness

reduction, and when s=1,

the mutation is lethal.

EpistasisAn interaction between genes

such that the effect of one

gene is modified by one or

several other genes.

Fitness is retive n sh be mesre in iffer-ent envirnments28,29 t estbish its reictive ve inetermining the cinic tcme f infectin. In sme

cses, resistnt mtnts er t be s fit s the wi-tye bcteri29,30, rising the qestin f hw t interretnegtive rests31. Hwever, if fitness csts re bservein brtry exeriments, it is resnbe t ssme thtthere wi s be cinic cnitins ner which theresistnce w imse fitness bren.

Fitness costs of resistance measuredin vitro and in ani-mals. The fitness csts f resistnce cn be cnsierbe(TABLE 1). There is n bvis crretin between the mg-nite f the cst n the tye f trget mece inhibiteby the ntibitic, thgh n inverse retinshi betweenthe resistnce eve f n iste n its grwth rte

is bserve in sme excetin cses32,33. Bew, wetine five essns tht hve emerge frm sties ffitness csts.

The first essn is thtepistasis cn ffect fitness csts.Thefitness csts f stretmycin resistnce cse bymttins in rpsL (which ences 30S ribsmrtein S12) een n envirnment n eistticeffects. In Salmonella enterica sbs. enterica servrTyhimrim, stretmycin-resistnt (SmR) mtntswith K42N r p90S sbstittins hve imire grwthn rich meim bt, srrisingy, grw fster thn wi-tye bcteri in mei with rer crbn srces31.The increse grwth rte refects fire f thesemtnts t ince the stress-incibe RNa ymerse-fctr (S; s knwn sRS), which is key reg-tr f mny sttinry-hse n stress-incibe genes;

Table 1 |The biological cost of antibiotic resistance conferred by chromosomal mutations

baa rssa cs* Assa ssm rs

Salmonella entericasubsp. enterica serovarTyphimurium

Streptomycin Variable Mice and in vitro 29,31,88,91

Rifampicin Variable Mice and in vitro 29,49

Nalidixic acid Yes Mice and in vitro 29

Ciprofloxacin Yes Chickens and in vitro 124

Fusidic acid Variable Mice and in vitro 4042,92

Mupirocin Yes Mice, nematodes and in vitro 125

Actinonin Yes Mice, nematodes and in vitro 96

Escherichia coli Streptomycin Variable In vitro 126,127

Norfloxacin Variable Mice and in vitro 58

Rifampicin Variable In vitro 46

Fosfomycin Yes Urine and in vitro 128

Campylobacter jejuni Ciprofloxacin Variable Chickens 34

Mycobacteriumtuberculosis

Isoniazid Yes Mice 73,99

Rifampicin Yes Macrophages and in vitro 26,129,130

Mycobacterium bovis Isoniazid Yes Mice 72

Mycobacterium smegmatis Streptomycin Variable In vitro 30

Staphylococcus aureus Fusidic acid Variable Rats and in vitro 39,131,132

Rifampicin Variable Biofilms and in vitro 45,47,133

Mupiricin No Mice and in vitro 33,134

Methicillin Yes In vitro 32

Vancomycin Variable In vitro 57

Staphylococcusepidermidis

Fusidic acid Yes Humans 135

Ciprofloxacin No Humans 135

Streptococcuspneumoniae

Gemifloxacin Yes Mice and in vitro 136

Helicobacter pylori Clarithromycin Yes Mice and in vitro 25,27

Chlamydia psittaci Spectinomycin Yes In vitro 137

Pseudomonas aeruginosa Fluoroquinolone Variable In vitro 82,138

Pseudomonas fluorescens Rifampicin Yes Soil 139

Listeria monocytogenes Class IIa bacteriocin Yes In vitro 140

Neisseria meningitidis Sulfonamide Yes In vitro 141

*Yes indicates that the increase in the generation time, which is a measure of the fitness cost, ranges from several percent up to asmuch as 400%. Variable means that some mutations have an associated fitness cost, whereas others do not.

RE V I E W S


RE V I E W S

http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12320http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12320http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12320http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12320http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12320http://www.uniprot.org/uniprot/P0A2E5http://www.uniprot.org/uniprot/P0A2E5http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12320http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12320


4/12

the recise mechnism inking the ribsm genemttin with rece S eves hs nt yet been e-cite. on rer crbn srces, wi-tye cesince S, ths swing their grwth. By nt incingS, the SmR mtnts esce this sef-imse grwthinhibitin. The rests sggest tht the inctin f Scntribtes t ng-term ce srviv, even thgh itcty imits the shrt-term grwth rte ner restric-tive grwth cnitins. These rests s highight theimrtnce f mesring fitness csts ner mtieexeriment cnitins, nt ny t cqire mrereevnt estimte f fitness, bt s t reve hysigi-

c weknesses tht my be exitbe fr rg eve-ment31. anther exme f n eisttic effect n fitnesscsts is shwn by mttin in the gene encing DNagyrse sbnit a (gyrA) tht cses resistnce t cir-fxcin ( frqinne) in Campylobacter jejuni.In the bsence f ntibitic seectin, thisgyrA mttinenhnce the fitness f the resistnt strin when it wsin cmetitin in chicken infectin me34. Hwever,the sme mttin imse fitness cst when it wstrnsferre int ifferent genetic vrint fC. jejuni.The srrising imictin is tht the ri emergence ffrqinne-resistnt Campylobacters. my be et the enhnce fitness tht is sscite with resistnce.

psmis crrying rg resistnce genes syimse fitness cst n the bcteri tht hrbr them,resting in rece grwth rte19,20,3537. one exme isthe smi crrying the gene tht ences amC-tye-ctmse, which is rrey fn in cinic istesfSalmonella s.38. When ampC ws exerimentytrnsfrme int S. Tyhimrim, it rece the grwthrte n invsiveness f the bcterim, sggesting thtthe sscite rece fitness is sibe resn fr therrity f the smi38. Hwever, if bth ampC n itsregtr, ampR, were intrce, mking -ctmresistnce incibe rther thn cnstittive, these fitness

csts were eiminte38. Ths, the genetic cntext f theresistnce gene cn etermine the fitness cst sscitewith its cqisitin.

The secn essn is tht envirnment cnitinsffect fitness csts. Sme resistnce mttins tht shwn cst in bcteri grwn in brtry meim hvehigh csts in brtry mice n, cnversey, smemttins tht shw n cst in mice cn hve sb-stnti csts in vitro28,39. a mttin infusA, the geneencing engtin fctr G (EFG), rests in fsiicci-resistnt (FsR) S. Tyhimrim by tering evesf the trnscritin regtr gnsine tetrhs-hte (G)40; this mttin hs eitric effects

Table 2 |Amelioration of the fitness costs that are caused by chromosomal mutations

baa rssamua

rssa cmsa mua rssa msadmua

S ssm rs

Salmonellaenterica subsp.enterica serovarTyphimurium

rpsL Streptomycin Intragenic Maintained Mice 29

rpsL Streptomycin Extragenic (rpsD or rpsE) Maintained Laboratory medium 29,91

gyrA Nalidixic acid Intragenic Maintained Mice 29rpoB Rifampicin Intragenic Maintained Mice 29

fusA Fusidic acid True reversion Lost Mice 28

fusA Fusidic acid Intragenic Often maintained Laboratory medium 92

fmt and folD Actinonin Intragenic Maintained Laboratory medium 96

fmt Actinonin Extragenic (metZand metWamplification)

Maintained Laboratory medium 96

Escherichia coli gyrA and marR Fluoroquinolones Extragenic (parC) Increased Constructed* or laboratorymedium

62

gyrA, parC andmarR

Fluoroquinolones gyrA-2 Increased Constructed* 62

rpsL Streptomycin Extragenic (rpsD or rpsE) Maintained Laboratory medium 126,127

rpoB Rifampicin Intragenic Maintained Laboratory medium 46

Staphylococcusaureus

fus Fusidic acid Intragenic Maintained or lost Laboratory medium 39

rrl Linezolid Gene conversion Lost Laboratory medium 93,94

Streptocccuspneumoniae

gyrA, parC andparE

Fluoroquinolones Intragenic Maintained orincreased

Constructed* 97

Mycobacteriumtuberculosis

katG Isoniazid Extragenic (ahpC) Maintained Humans 101

Neisseriameningitidis

sul2 Sulfonamide Intragenic Maintained Humans 141

ahpC, alkyl hydroperoxide reductase C gene;fmt, methionyl-tRNA formyltransferase gene;fusA, elongation factor G gene; gyrA, DNA gyrase subunit A gene; katG,catalaseperoxidase gene; marR, multiple drug resistance protein R gene; met, methionine biosynthesis; parC, DNA topoisomerase IV subunit A gene; parE, DNAtopoisomerase IV subunit B gene; rpoB, DNA-directed RNA polymerase subunit- gene; rpsD, 30S ribosomal protein S4 gene; rpsE, 30S ribosomal protein S5 gene;rpsL, 30S ribosomal protein S12 gene; rrl, 23S ribosomal RNA gene; sul2, dihydrofolate synthase 2 gene.*Mutants were constructed rather than selected for.

RE V I E W S



http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12299http://www.ncbi.nlm.nih.gov/gene/1254969?ordinalpos=3&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/1254969?ordinalpos=3&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12299


5/12

n gene exressin, ssiby exining the fitness if-ferences tht re bserve in ifferent envirnments41,42.Interestingy, the eitric hentyes f the fusAmtnt s ince hyersensitivity t nrete nti-bitics43, which is tentiy exitbe achies heef resistnce.

Rifmicin resistnce is cse by mttins inDNa-irecte RNa ymerse sbnit- (RB). af the teste rifmicin-resistnt (RifR) mtnts hve rece fitness4446, with the excetin f the S464pmtnt in Staphylococcus aureus, which ws betterte t grwth in mse bifim infectin methn the sscetibe rent strin47. anther intrig-ing bservtin with mttins tht cnfer resistnce trifmicin is tht they ccmte t very high freqen-cies in bcteri cnies tht hve been we t gen si mei. athgh this ws initiy sggeste tbe the rest f stress-ince mtgenesis48 in n ge-ing cny, it seems tht, in fct, mst re-existing RifRmtnts cntine t grw, wheres nrm ces reminqiescent. Ths, RifR mtnts tht hve rece fitnessin exnenti grwth hve ntbe grwth vntge inthe envirnment f the geing cny49. This istrtestht fitness is strngy eenent n the envirnment inwhich it is mesre.

The thir essn is tht sme mttins re cst free.Resistnce t stretmycin tht is cse by the K42R

sbstittin in 30S ribsm rtein S12 is n r-ent n-cst mttin in S. Tyhimrim n E. coli50,51(FIG. 3a). Themttin my even cnfer sight vn-tge ver the wi-tye bcterim in mse29 n ig52infectin mes. This sbstittin is fn t highfreqency in SmR cinic istes fMycobacteriumtuberculosis30,53. This imies tht the K42R mttinin 30S ribsm rtein S12 is genine cst-freeresistnce mttin.

The frth essn is tht the cst f resistnce cnbe rece by regtin f the resistnce mechnism.The Vna-tye resistnce t vncmycin/teichnina-tye gyceties in S. aureus is cqire by HGT

f the resistnce ern vanA54. Resistnce rests frmthe synthesis f terntive ce w recrsrs (eningin d-nyd-ctte) with very w ffinity fr gyc-eties n the eimintin f the nrm sscetiberecrsrs (ening in d-nyd-nine) t which

vncmycin bins55. The resistnce hentye is incein resnse t the resence f gyceties56. Isgenicmethiciin-resistnt S. aureus strins with r withtne f three ifferent vanA resistnce erns wereientifie in cinic istes; irwise cmrisns fthese istes shwe tht when resistnce ws incethe fitness cst ws cnsierbe n grwth rtes wererece by 2038%57. By cntrst, in the bsence finctin the fitness cst f crrying the vanA ernws 0.040.3% rectin in grwth rte. Ths, Vna-tye resistnce is very csty fr the methiciin-resistntS. aureus hst when ince, wheres its bigic cstis minim in the bsence f inctin.

The fifth essn is tht cst cmenstin n resist-nce cn be inke. Cinic resistnce fE. coli t fr-qinnes reqires mtie genetic tertins invving

mttin58,59 n HGT60,61. Cmetitin in vitro n in mse me shwe tht, n verge, fitness ecresewith the nmber f resistnce mttins62. Hwever, frsme trie mtnts the cqisitin f frth resist-nce mttin increse frqinne resistnce r-mticy n s increse the fitness f the resistntstrins62. Ths, in rticr genetic cntext resist-nce mttin c increse bth resistnce n fit-ness, mesre bth in vitro n in vivo. This imiestht Drwinin seectin fr imrve fitness might,t est fr the frqinnes, rive the seectin frincrese rg resistnce62.

Costs of resistance estimated from clinical studies andepidemiology.By exmining the resistnce genes tht cnbe fn in bcteri iste frm hmns, reictinsfrm exeriment sties f the csts f resistnce cnbe teste. The K42R sbstittin in the 30S ribsmrtein S12, which shwe n cst in S. TyhimrimnMycobacterium smegmatis ner exeriment cn-itins29, is s rimriy resnsibe fr resistnce tstretmycin in cinic istes fM. tuberculosis30,53.This sggests tht n-cst (r w-cst) mtnts rereferentiy seecte in hmns. The sectrm fmttins in rpoB tht cse rifmicin resistnce incinic istes fM. tuberculosis n in S. aureus iss bise in fvr f w-cst mttins, sggest-

ing tht fitness mesre s grwth rte in vitro is nimrtnt eterminnt f strin srviv in the cinicenvirnment45,63.

Fsiic ci resistnce in S. aureus hs sever iffer-ent genetic cses, ech f which hs ifferent effectn fitness n is esignte resistnce css. Fsa-cssresistnce is cse by mttins infusA (the gene thtences the trget f fsiic ci, EFG) n syreces fitness in vivo n in vitro39,64. By cntrst, FsB-css n FsC-css resistnce is cnferre by the cqisi-tin ffusB rfusC, resectivey, which cses EFG t bertecte frm the effect f fsiic ci by nther r-tein n hs very w fitness csts39,65. The FsB n FsC

Box 1 | Measuring fitness costs and compensation

Single culture and competition experiments can be carried out in vitro, and exponential

growth rates can be measured for many bacterial cultures in parallel using microtitre

plate machines. It is possible to discriminate differences of ~5% in generation time.

Greater discrimination can be achieved in pairwise competitions29,62. Mixed cultures

are serially passaged through cycles of growth and re-inoculation and assayed for

changes in the ratio of the two strains. Growth cycles measure several components of

competitive fitness, including lag periods, rates of exponential growth and efficienciesof resource utilization. Neutral genetic tags in one or both strains permit detection of

fitness differences that are 1% per generation137. With FACS (fluorescence-activated

cell sorting) analysis using fitness-neutral fluorescent markers138, it may be possible to

detect fitness differences of 0.1% per generation.

Single culture and pairwise competition experiments between susceptible and

resistant bacteria can also be carried out in animal models, in which the complex

growth environment has more relevance to the clinical infection, as rates of clearance

and mortality can be measured in the presence or absence of host immune

defences41. Bioluminescent measurements now permit tracking of tagged bacterial

strains in vivo and provide real-time measurements of the colonization of different

compartments and of competitive fitness139141.

RE V I E W S


RE V I E W S

http://www.uniprot.org/uniprot/P47768http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12304http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12318http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12318http://www.ncbi.nlm.nih.gov/sites/entrez?db=genomeprj&cmd=search&term=txid1772http://www.ncbi.nlm.nih.gov/sites/entrez?db=genomeprj&cmd=search&term=txid1772http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12318http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12318http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12304http://www.uniprot.org/uniprot/P47768


6/12

Susceptible

Susceptible

Resistant

Resistant

HoursRatioofsusceptible/res

istant

bacteria(logscale)

Numberofbacteria(log

scale)

Days

Days

c Competition in vivob Competition in vitroa Growth in vitro

resistnce csses reminte in istes frm skin infec-tins66,67, wheres Fsa, FsB n FsC resistnce cssesre eqy revent in bcteremi istes68. Resistncecse by mttins in rplF(which ences 30S ribs-m rtein 6), esignte FsE-css resistnce, restsin vrint tht frms sm cnies with n extremeysw grwth rte64. Desite their extremey sw grwth,these vrints re fn in cinic istes fS. aureus69,ssiby becse they cn ersist intrcery, shieefrm the hst immne resnse70. Ths, the secific hstenvirnment c etermine the fitness csts f ifferentmechnisms f ntibitic resistnce.

The impact of resistance on bacterial pathogenesis, viru-

lence and disease pathology. Resistnt mtnts syshw ecrese fitness s we s ecrese virence,even thgh there is n riri resn tht they shbe crrete71. Strins fMycobacterium bovis nM. tuberculosis tht re resistnt t isnizi s rest f T275p sbstittin in KtG (ctseerxise) shwseverey rece virence s mesre by hst kiingn by histthgy72,73. By cntrst, strins fM. tuber-culosis crrying the mst cmmn KtG mttin, nS315T sbstittin, re highy resistnt t isnizin virent in mse me f the isese74,75.

other exmes in which isese thgy might

be ifferent between the resistnt mtnts n the wi-tye bcteri re cehsrin-resistnt rtnisticGrm-negtive bcteri sch s Citrobacter freundii nEnterobacter cloacae, which ence n incibe chr-msm -ctmse. Even with this resistnce r-tein, the bcteri remin sscetibe t thir-genertincehsrins. Hwever, ss-f-fnctin mttinsin ampD (which ences N-cety-nhyrmrmy-l-nine-mise n is reqire fr etigycnrecycing) e t cnstittive -ctmse rctinn resistnce t thir-genertin cehsrins76.as the mttin freqency ring infectin is high,ampD mtnts rise ring nging thery, resting

in cinic fires7779. These ampD mtnts ccm-te N-cety-nhyrmrmy-trietie in the cyt-sm76 n, s rest, ince the rctin f nitricxie in eithei n hgcytic ces infecte withS. enterica80, which might ter the hst infmmtryresnse. Mttins in the gene encing mtirgresistnce ern reressr, mexR, (knwn s the nalBmttin) n innfxB tht increse ntibitic resistncein Pseudomonas aeruginosa by regting rg effxms increse the biity f the bcterim t mke bi-fims, which re reevnt fr the cniztin f cthetersn fr ersistence in chrnic infectins, bt bish itsbiity t ki Caenorhabditis elegans, which is se s me fr virence81,82. The cntrsting fitness effects fthese effx mttins sggests tht ne cnnt neces-sriy infer gener fitness efects in vivo by extrtingfrm in vitro cmetitin t.

Transmission rates for susceptible and resistant variants.Fr infectis bcteri t estbish themseves within hst tin, they mst hve bsic rerctivenmber (R

0) f greter thn 1 (REF. 83). R

0is efine

s the verge nmber f secnry infectins tht rerce when n infecte inivi is intrce int sscetibe hst tin. an imrtnt qestin iswhether trnsmissin rtes fr sscetibe n resistnt

bcteri iffer grety. Sties fM. tuberculosis trns-missin hve rce cntrictry rests, with strinsresistnt t isnizi shwing rece trnsmissin84n RifR strins shwing the site effect85. In i-tin, sersreing86 (the henmenn f few ini-

vis being resnsibe fr the infectin f nsyrge nmbers f secnry cses) cn e t heter-geneity in the R

0f the infecting tin n cn

infence the infectin ynmics. Mre mesrementsf the trnsmissin rtes fr resistnt n sscetibe

vrints tht cse imrtnt infectins re neee, n recent sty fM. tuberculosis shws hw sch tcn be btine frm eiemigic sties87.

Figure 2 | Dm ss aa sas. a | Measuring exponential growth rates of susceptible and resistant

strains in vitro can reveal large fitness differences between the strains. | In an in vitro competition experiment in which

cultures are cycled with a bottleneck each day, fitness differences between wild-type and mutant strains are revealed as a

change in the ratio of the two competing strains. | A suitable animal model can be infected with a mixed bacterial culture

and assayed after several days to measure changes in the ratio of two populations.

RE V I E W S



http://www.ncbi.nlm.nih.gov/sites/entrez?db=genomeprj&cmd=search&term=txid1765http://www.uniprot.org/uniprot/Q08129http://www.ncbi.nlm.nih.gov/gene/877857?ordinalpos=1&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/881079?ordinalpos=6&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/881079?ordinalpos=6&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12339http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=9548http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=9548http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12339http://www.ncbi.nlm.nih.gov/gene/881079?ordinalpos=6&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/877857?ordinalpos=1&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.uniprot.org/uniprot/Q08129http://www.ncbi.nlm.nih.gov/sites/entrez?db=genomeprj&cmd=search&term=txid1765


7/12

1 10 100 1,000

WT rpsL K42R

rpsL K42N

rpsL K42NrpsD K205T

0.01 0.1 1 10

marR

WT

marRgyrA

marRgyrAparC

gyrAparCgyrA

MIC (mg l1)

R

elative

fitness

0.7

0.8

0.9

1

R

elative

fitness

0.7

0.8

0.9

1

MIC (mg l1)

a b

Gene conversionA recombination event in which

one strand of DNA is changed

or repaired using information

from another strand.

Compensatory evolution can rapidly reduce the fitnesscost of resistance. The ecrese fitness cse by resist-nce my be riy n efficienty cnterbnce bycmenstry mttins witht ss f resistnce(FIG. 3; TABLE 2). Cmenstry evtin cn stbiizeresistnt bcteri tins in the bsence f nti-bitics by mking them s fit s sscetibe cnes.Sever in vitro n in vivo sties hve shwn tht inmst cses it is ssibe t etect cmenstin whenresistnt bcteri re seriy ssge, bt the egreef restrtin n the tye n nmber f cmens-try mttins fn vry eening n the rtic-

r resistnce mechnism n bcterim s we s theenvirnment cnitins ner which cmenstinccrs (iscsse bew).

The seectin f cmenste mtnts ring serissge eens n the mttin rtes fr the if-ferent tyes f mttins, the fitness f the ifferentmtnts n the bttenecks ring seri trnsfer8890.Dt frm tw sties sing FsRfusA mtnts r SmRrpsL mtnts fS. Tyhimrim sggest tht the trgetsize fr cmenstry mttins is tyicy mre thn20 times the size f tht fr reversin n tht tinbttenecks n grwth cnitins hve strng effectsn the sectr f cmenstry mttins28,29,39,91. Ths,

cmenstin is mre ikey thn reversin (FIG. 3), sw be execte fr mny ntr tins, inwhich bttenecks re resent. Cmenstry mech-nisms hve been etermine fr sever resistnces(TABLE 2). Cmenstry mttins cn restre fitnessby recing the nee fr the ffecte fnctin f theresistnce rtein, sbstitting the ffecte fnctinwith n terntive fnctin r irecty r inirectyrestring the efficiency f the ffecte fnctin. Fr thecmenstin f resistnce mttins, bth sbstit-tin n restrtin f fnctin hve been bserve,n irect restrtin f the fnctin is by fr the mstcmmn mechnism28,29,39,92. Bew, we tine frinteresting exmes f cmenstry mechnisms.

Cmenstin cn ccr by reversin f mttinr bygene conversion. athgh reversin t the wi-tye gentye is rre, there re cinic exmes f thisccrring by gene cnversin. In S. aureus, resistnce tinezi tht ws e t mttin in 23S ribsmRNa gene (rrl) ws st in ne cse93 n cnsierbyrece in nther cse94 fter remv f the ntibitic

seective ressre. In bth cses, resistnce crrie fitness cst tht c be rece by gene cnversinbetween the mtie cies f 23S rrlgenes in bcte-ri with t est ne wi-tye cy. a simir mech-nism exins hetergens mcrie resistnce innemccci95.

Cmenstin cn s ccr by gene mifictin.Gene ictins n mifictins cn reversibyter resistnce hentye r its fitness. By creting rge genetic trget, gene ictins n mifi-ctins s increse the rbbiity f mttins thtcn ermnenty ter the resistnce hentye n itsfitness9,10. Salmonella s. mtnts tht re resistnt tctinnin, etie efrmyse inhibitr, crry mt-tins in either f tw genes tht re reqire fr thefrmytin f methiny inititr tRNa (tRNai):folDnthe methiny-tRNa frmytrnsferse gene (fmt).In the bsence f ntibitic, the mttins rece thefitness f bcteri grwn in brtry meim n inmice. The fitness csts c be meirte by intr-genic mttins infmtrfolD r by extrgenic cm-enstry mttins. one-thir f the extrgenicycmenstefmtmtnts crrie mifictins f theientic, tnemy reete metZn metWgenes,which ence the methiny-tRNai96. The increse inthe cy nmber fmetZn metW(which ws t40-f higher thn the cy nmber in bcteri witht

cmenstry mttins) ws ccmnie by simirincrese in methiny-tRNai eves tht cmenste frthe ck f Fmt ctivity n we trnstin inititint rcee with nn-frmyte methiny-tRNai.

an increse in resistnce cn s be sscite withcmenstin. Stewise brtry seectin f incresefrqinne resistnce in E. coli shwe tht fit-ness ecrese s t five resistnce mttins wereccmte. This crretin ws ccsiny brken,n n increse in resistnce c be sscite withn increse in fitness t te seectin ste58. a simirrevers f the s retinshi ws nte in strinsfStreptococcus pneumoniae int which ne r tw

Figure 3 | ras w a ssa ad aa ss.

Numbers shown on theyaxes are from empirical data28,61,88. a | High-level resistance to

streptomycin in Salmonella enterica subsp. enterica serovar Typhimurium (that is, strains

with a minimal inhibitory concentration (MIC) of > 1,000 mg l1) can be cost free, as is the

case for the resistance given by the mutation in rpsL (which encodes 30S ribosomal

protein S12) that leads to a K42R substitution in the protein, or can reduce fitness, as

seen for resistance imparted by a K42N substitution in the same protein. Reduced fitness

can be compensated by extragenic mutations (for example, the mutation in rpsD, which

encodes 30S ribosomal protein S4, that leads to a K205T substitution in the protein)

without a loss of resistance29,91. | Resistance to fluoroquinolones in Escherichia colidevelops in multiple genetic steps. High-level resistance (that is, strains with an MIC for

ciprofloxacin of > 32 mg l1) can develop without any substantial loss of fitness (as is the

case for strains with mutations in the DNA gyrase subunit A gene, gyrA) or, following a

different genetic trajectory, can involve a major reduction in fitness (as is the case for

strains with mutations in the multiple drug resistance protein R gene, marR, or in both

marR and gyrA). Note that resistance mutations in marR, which are knockout mutations,

are intrinsically much more frequent than resistance mutations in gyrA, which are

specific nucleotide substitutions. In the double mutant, acquisition of an additional

fluoroquinolone resistance mutation in parC (which encodes DNA topoisomerase IV

subunit A) can substantially compensate the fitness costs of resistance and, at the same

time, greatly increase the level of resistance (the MIC in the double mutant is 0.75 mg l1,

whereas in the triple mutant it is 32 mg l1)62. In this figure, the gyrA mutation refers to the

S83L, D87N double substitution.

RE V I E W S


RE V I E W S

http://www.ncbi.nlm.nih.gov/gene/1252062?ordinalpos=1&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/1254930?ordinalpos=1&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/1254512?ordinalpos=1&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/1254513?ordinalpos=1&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12328http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12328http://www.ncbi.nlm.nih.gov/gene/1254513?ordinalpos=1&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/1254512?ordinalpos=1&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/1254930?ordinalpos=1&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/1252062?ordinalpos=1&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSum


8/12

frqinne resistnce mttins were intrce97.T estbish cse n effect, set f isgenic E. colistrins were cnstrcte tht crrie t five fr-qinne resistnce mttins in ifferent cmbin-tins62. w-fitness trie mtnts were ientifie inwhich the cqisitin f frth resistnce mttinsbstntiy increse fitness in vitro n in vivo whiesimtnesy rmticy ecresing rg ssceti-biity. The rgest effect ccrre with the itin f mttin inparC (which ences DNa tismerse IVsbnit a) t w-fitness strin with resistnce mt-tins ingyrA (which ences DNa gyrse sbnit a) nmarR (which ences mtie rg resistnce rtein R, rtein tht regtes rg effx) (FIG. 3b). Incresefitness ws ccmnie by ntbe chnge in the evef ctivity f thegyrA rmter. Exeriments erfrmein the bsence f the ntibitic seecte frparC mtntstht imrve fitness n rece sscetibiity t thentibitic, imying tht ntr seectin fr imrvegrwth in bcteri with w-eve resistnce t fr-qinnes c s seect fr sbstnti rectins in

rg sscetibiity. athgh frther testing is reqire, sibe mechnism is tht marR mttins ter g-b trnscritin tterns n hve negtive effects nfitness, n the cmbintin f mttins inparC ngyrA irecty r inirecty restre trnscritin tterns,s restring fitness, whie s recing frqinnebining t the DNa gyrse n tismerse, therebyincresing resistnce t the rg. Ths, increse resist-nce t frqinnes c be seecte fr even inthe bsence f frther exsre t the rg62. Fitnesscmenstin cn s ccr in cinic istes, bt esit ccr in bcteri ring grwth in hsts n, if s, issch cmenstin imrtnt fr stbiizing w-fitness,resistnt bcteri strins? These qestins re iffictt ress exerimenty, bt cmenstin in bcte-ri iste frm tients hs been sggeste t ccrin few cses25,39,68,98101, inicting tht cmenstryevtin es ccr tsie the brtry.

Physiological reasons for fitness costs. unerstningthe hysigic resns fr fitness csts my e t nbiity t reict these csts n t insights int tentiweknesses f resistnt bcteri.

Tw ntibitic resistnce n fitness cmbintinsin S. Tyhimrim hve been intensivey stie bthin vitro n in nim exeriments; these re the FsRhentye cse by mttins in EFG n SmR hen-

tye cse by mttins in the 30S ribsm rteinS12. These FsR mtnts rece the rte f rtein syn-thesis bt s hve eitric effects n bcteri hysi-gy, becse they ertrb the ccmtin f Gin the ce40. as G is gb regtr f trnscri-tin, FsR mtnts re efective in the inctin f S,hve rece virence in vivo41, hve rece rte fresirtin n re sensitive t xitive stress42. FsRmtnts re s hyersensitive t nrete cssesf ntibitics, incing -ctms, frqinnes,mingycsies, rifmicin n chrmhenic43.The csts f frqinne resistnce in E. coli mys be csey sscite with istrbe tterns f

trnscritin regtin, ssiby meite thrghMrR n tertins in DNa serciing62. There iss cnnectin between fitness csts n trnscri-tin regtin in stretmycin resistnce cseby mtnt 30S ribsm rtein S12. SmR mtntshve rece rtes f rtein synthesis n recefitness in vivo51,91, bt sme SmR mtnts iste inS. Tyhimrim re s efective in S inctin, sg-gesting tht the hysigic effects f resistnce remre cmex n ince tere gb trnscritinregtin31. The etriment effects f these fitness cstsn the ftes f resistnt rgnisms in ntr -tins c tentiy be exite fr esigning rgs,theretic regimes n interventin strtegies.

The high fitness cst sscite with inctin fVna-tye resistnce in S. aureus57 sggests tht thisfetre c be exitbe in the esign f theries.Fr exme, the effectiveness f tretment might beincrese by cmbining chemic inctin fvanAverexressin with the se f n ntibitic t which thebcteri re sscetibe. It is nt knwn hw the cst f

vanA inctin is mnifeste, thgh the bvis s-sibiities ince the csts f exressing the vanA ernrteins. aterntivey, the itin csts sscitewith trnsgycsytin, trnsetitin r trnsrt fd-nyd-ctte rther thn d-nyd-nine. Thecinic imrtnce f vncmycin rges tht ner-stning the hysigic bsis f fitness csts in isteswith Vna-tye resistnce sh be ririty.

Real-life approaches to test reversibility

The cmexity f bcteri tin ynmics menstht we reqire cinic sties t ress the qestinf whether reversibiity is fesibe in reity. Stieshve been erfrme t test the revers f resistncein inivis, in cmmnity settings n in hsi-ts. at the eves f the inivi n cmmnity, thebigic cst is thght t be the min riving frcefr the rectin in the freqency f resistnt bcteriin the bsence f ntibitics7,23,89. By cntrst, in hsitsettings the riving frce is rse t be itineffect cse by the cntins infx f tients whre either ninfecte r infecte with sscetibe bcte-ri. In hsits, bth meing7,23,102104 n nysis fthe crretins between ntibitic resistnce n vri-tin in ntibitic se105107 shw tht tertins in nti-bitic se cn cse ri chnges (in the rer f yst mnths) in the freqency f resistnce. By cntrst,

when the fitness cst f resistnce is the min rivingfrce behin its revers, the rte f chnge is execte tbe mch swer (mnths t yers). Ths, fter remvf the seective ressre in cmmnity settings, evenwitht cmicting fctrs sch s cmenstryevtin n c-seectin, ng time my ss befre rectin in the freqency f resistnce is seen7,23,89.Bew, we iscss exeriments n interventin st-ies tht ttemte t reverse resistnce. Fr reversibiitysties crrie t in inivis, the rincie is t f-w the freqency f resistnt bcteri in n inivibefre n fter n ntibitic tretment tht seectsfr n increse freqency f resistnt bcteri n t

RE V I E W S



http://www.ncbi.nlm.nih.gov/gene/947499?ordinalpos=1&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/947499?ordinalpos=1&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/945825?ordinalpos=21&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/945825?ordinalpos=21&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/947499?ordinalpos=1&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSum


9/12

BaselineValue for the frequency of

resistance before the

intervention.

etermine hw riy tht freqency is rece fterthe ntibitic ressre is reieve. The rtine is tht,ver time, sscetibe bcteri (either srvivrs f thetretment r imrte frm the envirnment), whichre mre fit, wi tcmete the seecte resistnt bcte-ri, which re ess fit, when ntibitic ressre is bsent.In tw sties erfrme n five ysetic iniviswh h been trete fr 1 week with the mcrie cr-ithrmycin fr Helicobacter pyloriinfectin, the seec-tin n reversin f crithrmycin-resistnt (CR)enterccci in the intestine n CRStaphylococcusepidermidis n the skin ws fwe108,109. as execte,in these five trete tients, enterccci tht wereiste befre tretment were sscetibe t the rg,bt thse recvere immeitey fter tretment shwehigh-eve resistnce, which ws e t the intrctinfermB (encing erythrmycin resistnce rtein B).In three f the tients, the seecte resistnt enter-ccci ersiste fr 1 t 3 yers witht ny frtherntibitic seectin (tht is, the tients receive nei-ther crithrmycin nr ny ther ntibitic fter the

initi crithrmycin tretment). Simiry, fr S. epi-dermidis strins iste immeitey fter tretmentwere crithrmycin-resistnt (wing t the intrc-tin fermC), n the sme resistnt strins ersistefr t 4 yers in three tients in the bsence f fr-ther ntimicrbi tretment. These sties shw hwntibitic tretment seects fr resistnce in the inig-ens micrfr n tht these resistnt bcteri cnersist fr yers witht ny frther ntibitic seec-tin. By cntrst, the ver secies cmsitin f themicrfr cn be qite riy restre fter ecgicistrbnces tht re ince by tretment with certincsses f ntibitic, sch s ivmeciinm110. The rtef reernce f sscetibe strins fter remv fthe ntibitic ressre wi een n sever fctrs(incing retive fitness, the trnver rte f bcterin the sy rte f sscetibe strins), n it is there-fre ikey tht reversibiity rtes wi vry eening nthe inivi, the bcteri secies n the ntibiticresistnce mechnism. at the cmmnity eve, threentinwie interventins hve been erfrme in whichthe effects f rece ntibitic se n resistnce werenyse retrsectivey. a rectin in the se f certinmcrie ntibitics in Finn ws fwe by sb-stnti ecrese in erythrmycin-resistnt Streptococcuspyogenes111. Simiry, in sty erfrme in Icenthe ecrese se f ntibitics in chiren ws fwe

within few yers by ecrese in eniciin-resistntS. pneumoniae112. Hwever, sbseqent nysis sg-geste tht fr bth sties the rectin in resistntbcteri might hve been cse by cn recementstht were nrete t the rectin in ntibitic se,istrting the imrtnce f knwing the cn strc-tre f the bcteri tin bth befre n fter theinterventin113,114. Frthermre, fr the erythrmycinresistnce sty, n nysis ws crrie t t eter-mine the mgnite f the fitness csts, thgh scht w be reqire t exin the ri tertinsin the freqency f resistnce n hw they rete tntibitic se. a meing nysis sing the eniciin

resistnce sty sggeste tht resistnce i nt recebcteri fitness in the hst bt ntby increse thebiity f the resistnt bcteri t be trnsmitte betweenhsts retive t the sscetibe strins115. T r knw-ege, these cncsins hve nt been frther exmineexerimenty. In cntrst t the rmtic effects seen inthese tw sties, 97% ecrese in cnsmtin f c-trimxze in the unite Kingm between 1991 n1999 i nt rest in rectin in sfmethxzeresistnce116. a fw- sty in 2004 bse n i-tin t frm the sme re tht s ince tn stretmycin resistnce shwe tht sfmethx-ze n stretmycin resistnce in E. coli h reminestbe esite cntine w-eve se f these rgs117.The ck f effect n resistnce ws ttribte t w rnn-existent fitness cst f resistnce n t c-seectin,sch tht sfmethxze n stretmycin resistncesre mintine by their cse genetic inkge t ntherseecte ntimicrbi resistnce.

T cntr fr the nefine sects f these threesties (fr exme, r escritin f the bcte-

ri cn strctre n the ttern f ntibitic se,s we s the ck f cntr gr n f ng-termbaseline fr the interventin), rge rsective inter-

ventin sty ws erfrme. a 24-mnth vntryrestrictin ws intrce n the se f trimethrim-cntining rgs in Krnberg Cnty, Sween, with bseine f 14 yers f mnthy t n trimethrimresistnce n cnsmtin t etermine if the fre-qency f trimethrim-resistnt E. coli frm rinrytrct infectins w be rece118. Fr rge sbsetsf cecte E. coli, the fitness cst f trimethrimresistnce, the resence f ny sscite resistncesn the cn cmsitin fE. coli befre n fterthe interventin were etermine. as cmre with neighbring cntr cnty, there ws rmt nsstine 85% rectin in the se f trimethrim-cntining rgs ring the interventin, n thisrectin ws sscite with mrgin bt sttis-ticy significnt effect n trimethrim resistnce.N chnges were bserve in the cn cmsitinfE. coli befre, ring r fter the interventin, nthere ws n mesrbe fitness cst sscite withtrimethrim resistnce, s etermine by in vitromesrements f grwth rtes. The freqency f ss-cite ntibitic resistnces in trimethrim-resistntistes ws high, n mst bcteri were resistnt tne r mre f the ther ntibitics se fr tretment

f rinry trct infectins. In cncsin, rstic2-yer rectin in trimethrim se h very smeffect n the freqency f trimethrim resistncein E. coli118. This cn be exine by the w fitnesscst f trimethrim resistnce cmbine with thec-seectin tht ws e t high eves f ssciteresistnce, bt t resent the retive cntribtinsf these tw effects cnnt be ssesse. Whether thiscncsin cn be generize t ther resistnces nbcteri is ebtbe, bt ness the eri f nn-seis mch nger thn 2 yers, the cst f resistnce ismch higher thn tht fr trimethrim resistnce inE. coli, r the tenti fr c-seectin is w, then

RE V I E W S


RE V I E W S

http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12321http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12305http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12305http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12327http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12327http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12327http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12327http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12305http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12305http://www.ncbi.nlm.nih.gov/sites/entrez?Db=genomeprj&cmd=ShowDetailView&TermToSearch=12321


10/12

we w reict tht it is nikey tht ny cinicysignificnt rectin in resistnce wi be bserve frsimir ttemts.

Conclusions and perspectives

Sties f fitness csts n the ynmics f resistnceevement in vris thgenic bcteri hve e tsever cncsins f meic n bigic interest.The min fining is tht reversibiity in cinic settingsis execte t be sw r nn-existent. The min re-sns fr this re tht the intrinsic ynmics f reversre execte t be sw (even if fitness cst is resent),cmenstry evtin n cst-free resistnces cnrece the cst n thereby rece the riving frcefr reversibiity, n c-seectin between the resistncemechnism n ther seecte mrkers cn sw wnny tenti reversibiity tht my be riven by fitness

csts. a secn key cncsin is tht we c se thisknwege t rece the ikeih f resistnce eve-ment, fr exme, by chsing ntibitic trgets frwhich the resistnce mechnism cnfers high fitnesscst n fr which the rte n extent f cmens-tin is w. Simiry, it c be ssibe t exit theetie knwege f the hysigic bsis f fitnesscsts fr the chice n esign f nve theries thttrget the hysigic achies hee tht is sscitewith rticr resistnce mechnism. Finy, betternerstning f fitness csts n cmenstry ev-tin n f their imct n the emergence n sre fresistnt bcteri sh w s t mke better qn-tittive reictins bt the rte n trjectry f theevtin f resistnce twrs new n rgs n,by inference, sh s give s ssibiities t reventthis evtin.

1. Wise, R. Antimicrobial resistance: priorities for action.J. Antimicrob. Chemother.49, 585586 (2002).

2. Woodford, N. & Livermore, D. M. Infections caused byGram-positive bacteria: a review of the globalchallenge.J. Infect.59 (Suppl. 1), S4S16 (2009).

3. Lew, W., Pai, M., Oxlade, O., Martin, D. & Menzies, D.Initial drug resistance and tuberculosis treatmentoutcomes: systematic review and meta-analysis.Ann.Intern. Med.149, 123134 (2008).

4. Guay, D. R. Contemporary management ofuncomplicated urinary tract infections. Drugs68,11691205 (2008).

5. Levin, B. R. et al. The population genetics of antibioticresistance. Clin. Infect. Dis.24 (Suppl. 1), S9S16(1997).

6. Andersson, D. I. & Levin, B. R. The biological cost ofantibiotic resistance. Curr. Opin. Microbiol.2,489493 (1999).

7. Levin, B. R. Models for the spread of resistantpathogens. Neth. J. Med.60, 5866 (2002).

8. Nikaido, H. Multidrug resistance in bacteria.Annu.Rev. Biochem.78, 119146 (2009).

9. Andersson, D. I. & Hughes, D. Gene amplification and

adaptive evolution in bacteria.Annu. Rev. Genet.43,167195 (2009).

10. Sandegren, L. & Andersson, D. I. Bacterial geneamplification: implications for the evolution ofantibiotic resistance. Nature Rev. Microbiol.7,578588 (2009).

11. Ince, D. & Hooper, D. C. Quinolone resistance due toreduced target enzyme expression.J. Bacteriol.185,68836892 (2003).

12. Robicsek, A. et al. Fluoroquinolone-modifying enzyme:a new adaptation of a common aminoglycosideacetyltransferase. Nature Med.12, 8388 (2006).

13. Li, X. Z. & Nikaido, H. Efflux-mediated drug resistancein bacteria: an update. Drugs69, 15551623 (2009).

14. Canton, R. Antibiotic resistance genes from theenvironment: a perspective through newly identifiedantibiotic resistance mechanisms in the clinical setting.Clin. Microbiol. Infect.15 (Suppl. 1), 2025 (2009).

15. Martinez, J. L., Baquero, F. & Andersson, D. I.Predicting antibiotic resistance. Nature Rev. Microbiol.

5, 958965 (2007).16. Bergman, M. et al. Effect of macrolide consumption

on erythromycin resistance in Streptococcus pyogenesin Finland in 19972001 Clin. Infect. Dis.38,12511256 (2004).

17. Bergman, M. , Nyberg, S. T., Huovinen, P., Paakkari, P.& Hakanen, A. J. Association between antimicrobialconsumption and resistance in Escherichia coli.

Antimicrob. Agents Chemother.53, 912917 (2009).18. Sommer, M. O., Dantas, G. & Church, G. M. Functional

characterization of the antibiotic resistance reservoirin the human microflora. Science325, 11281131(2009).

19. Dahlberg, C. & Chao, L. Amelioration of the cost ofconjugative plasmid carriage in Eschericha coliK12.Genetics165, 16411649 (2003).This article shows that plasmids carrying antibiotic

resistance genes impose fitness costs on their

bacterial hosts but that these costs can quickly be

ameliorated by mutations on either the plasmid or

the chromosome.

20. Bouma, J. E. & Lenski, R. E. Evolution of a bacteria/plasmid association. Nature335, 351352 (1988).

21. Enne, V. I., Bennett, P. M., Livermore, D. M. & Hall,L. M. Enhancement of host fitness by the sul2-codingplasmid p9123 in the absence of selective pressure.

J. Antimicrob. Chemother.53, 958963 (2004).22. Spratt, B. G. Antibiotic resistance: counting the cost.

Curr. Biol.6, 12191221 (1996).23. Levin, B. R. Minimizing potential resistance: a

population dynamics view. Clin. Infect. Dis.33,S161S169 (2001).

24. De Gelder, L. et al. Combining mathematical modelsand statistical methods to understand and predict thedynamics of antibiotic-sensitive mutants in apopulation of resistant bacteria during experimentalevolution. Genetics168, 11311144 (2004).

25. Bjorkholm, B. et al. Mutation frequency and biologicalcost of antibiotic resistance in Helicobacter pylori.Proc. Natl Acad. Sci. USA98, 1460714612 (2001).

26. Davies, A. P. et al. Comparison of fitness of twoisolates ofMycobacterium tuberculosis, one of which

had developed multi-drug resistance during the courseof treatment.J. Infect.41, 184187 (2000).

27. Kanai, K., Shibayama, K., Suzuki, S., Wachino, J. &Arakawa, Y. Growth competition of macrolide-resistantand -susceptible Helicobacter pyloristrains.Microbiol. Immunol.48, 977980 (2004).

28. Bjorkman, J., Nagaev, I., Berg, O. G., Hughes, D. &Andersson, D. I. Effects of environment oncompensatory mutations to ameliorate costs ofantibiotic resistance. Science287, 14791482(2000).This work finds that different fitness-compensating

mutations are selected depending on whether the

bacteria evolve in animal models or in laboratory

medium.

29. Bjorkman, J., Hughes, D. & Andersson, D. I. Virulenceof antibiotic-resistant Salmonella typhimurium. Proc.Natl Acad. Sci. USA95, 39493953 (1998).This study shows that resistance mutations to

different classes of antibiotic strongly reduce

bacterial virulence but that second-sitecompensatory mutations can be selected that

restore virulence without the loss of antibiotic

resistance.

30. Sander, P. et al. Fitness cost of chromosomal drugresistance-conferring mutations.Antimicrob. AgentsChemother. 46, 12041211 (2002).

A demonstration that chromosomal drug resistance

mutations with little or no costin vitro are

preferentially found in clinical isolates of

mycobacteria, arguing that decreased levels of

antibiotic consumption will not result in a decline in

the frequency of resistant mutants.

31. Paulander, W., Maisnier-Patin, S. & Andersson, D. I.The fitness cost ofStreptomycin resistance dependson rpsL mutation, carbon source and RpoS (S).Genetics183, 539546 (2009).

32. Ender, M., McCallum, N., Adhikari, R. & Berger-Bachi, B.Fitness cost of SCCmec and methicillin resistance

levels in Staphylococcus aureus.Antimicrob. AgentsChemother.48, 22952297 (2004).

33. Hurdle, J. G., ONeill, A. J., Ingham, E., Fishwick, C. &Chopra, I. Analysis of mupirocin resistance and fitnessin Staphylococcus aureus by molecular genetic andstructural modeling techniques.Antimicrob. AgentsChemother.48, 43664376 (2004).

34. Luo, N. et al. Enhanced in vivo fitness offluoroquinolone-resistantCampylobacter jejuniin theabsence of antibiotic selection pressure.Proc. Natl

Acad. Sci. USA102, 541546 (2005).In this article, the authors make the surprising

observation that fluoroquinolone-resistant C. jejuni

isolates with a mutation ingyrA outcompete

clonally related fluoroquinolone-sensitive C. jejuni

in the absence of antibiotic selection pressurein

vivo (in the chicken). This indicates that the single

point mutation not only confers a high-level

resistance to fluoroquinolones but also modulates

thein vivo fitness ofC. jejuni.35. Johnsen, P. J., Simonsen, G. S., Olsvik, O., Midtvedt, T.

& Sundsfjord, A. Stability, persistence, and evolutionof plasmid-encoded VanA glycopeptide resistance in

enterococci in the absence of antibiotic selectionin vitro and in gnotobiotic mice. Microb. Drug Resist.8, 161170 (2002).

36. Lenski, R. E. & Bouma, J. E. Effects of segregation andselection on instability of plasmid pACYC184 inEscherichia coliB.J. Bacteriol.169, 53145316(1987).

37. Smith, M. A. & Bidochka, M. J. Bacterial fitness andplasmid loss: the importance of culture conditions andplasmid size. Can. J. Microbiol.44, 351355 (1998).

38. Morosini, M. I., Ayala, J. A., Baquero, F., Martinez,J. L. & Blazquez, J. Biological cost of AmpC productionfor Salmonella enterica serotype Typhimurium.

Antimicrob. Agents Chemother.44, 31373143(2000).

39. Nagaev, I., Bjorkman, J., Andersson, D. I. & Hughes, D.Biological cost and compensatory evolution in fusidicacid-resistantStaphylococcus aureus. Mol. Microbiol.40, 433439 (2001).Here, the authors provide evidence that secondary

mutations that compensate the biological fitnesscosts of antibiotic resistance arise in nature and

may contribute to the stabilization of resistance in

a bacterial population.

40. Macvanin, M., Johanson, U., Ehrenberg, M. &Hughes, D. Fusidic acid-resistant EF-G perturbs theaccumulation of ppGpp. Mol. Microbiol.37, 98107(2000).

41. Macvanin, M. et al. Fusidic acid-resistant mutants ofSalmonella enterica serovar Typhimurium with lowfitness in vivo are defective in RpoS induction.


42. Macvanin, M., Ballagi, A. & Hughes, D. Fusidic acid-resistant mutants ofSalmonella enterica serovarTyphimurium have low levels of heme and a reducedrate of respiration and are sensitive to oxidativestress.Antimicrob. Agents Chemother.48,38773883 (2004).

RE V I E W S





11/12

43. Macvanin, M. & Hughes, D. Hyper-susceptibility of afusidic acid-resistant mutant ofSalmonella to differentclasses of antibiotics. FEMS Microbiol. Lett.247,215220 (2005).

44. Enne, V. I., Delsol, A. A., Roe, J. M. & Bennett, P. M.Rifampicin resistance and its fitness cost inEnterococcus faecium.J. Antimicrob. Chemother.53,203207 (2004).

45. ONeill, A. J., Huovinen, T., Fishwick, C. W. & Chopra, I.Molecular genetic and structural modeling studies ofStaphylococcus aureus RNA polymerase and the

fitness of rifampin resistance genotypes in relation toclinical prevalence.Antimicrob. Agents Chemother.50,298309 (2006).

46. Reynolds, M. G. Compensatory evolution in rifampin-resistant Escherichia coli. Genetics156, 14711481(2000).

47. Yu, J., Wu, J., Francis, K. P., Purchio, T. F. &Kadurugamuwa, J. L. Monitoring in vivo fitness ofrifampicin-resistant Staphylococcus aureus mutants ina mouse biofilm infection model.J. Antimicrob.Chemother.55, 528534 (2005).

48. Bjedov, I. et al. Stress-induced mutagenesis inbacteria. Science300, 14041409 (2003).

49. Wrande, M., Roth, J. R. & Hughes, D. Accumulation ofmutants in aging bacterial colonies is due to growthunder selection, not stress-induced mutagenesis. Proc.Natl Acad. Sci. USA105, 1186311868 (2008).This work finds that most rifampicin resistance

mutations increase bacterial fitness in the

environment of an ageing colony, showing that

environment is a critically important determinant

of relative fitness.50. Kurland, C. G., Hughes, D. & Ehrenberg, M. in

Escherichia coli andSalmonella: Cellular and MolecularBiology (eds Neidhardt, F. C. et al.) 9791004(American Society for Microbiology, Washington DC,1996).

51. Tubulekas, I. & Hughes, D. Suppression ofrpsLphenotypes by tufmutations reveals a uniquerelationship between translation elongation andgrowth rate. Mol. Microbiol.7, 275284 (1993).

52. Enne, V. I. et al. Assessment of the fitness impacts onEscherichia coliof acquisition of antibiotic resistancegenes encoded by different types of genetic element.

J. Antimicrob. Chemother.56, 544551 (2005).53. Bottger, E. C., Springer, B., Pletschette, M. & Sander, P.

Fitness of antibiotic-resistant microorganisms andcompensatory mutations. Nature Med.4,13431344 (1998).

54. Arthur, M., Reynolds, P. & Courvalin, P. Glycopeptideresistance in enterococci.Trends Microbiol.4,

401407 (1996).55. Bugg, T. D. et al. Molecular basis for vancomycin

resistance in Enterococcus faecium BM4147:biosynthesis of a depsipeptide peptidoglycanprecursor by vancomycin resistance proteins VanH and

VanA. Biochemistry30, 1040810415 (1991).56. Arthur, M., Molinas, C. & Courvalin, P. The VanS-VanR

two-component regulatory system controls synthesisof depsipeptide peptidoglycan precursors inEnterococcus faecium BM4147.J. Bacteriol.174,25822591 (1992).

57. Foucault, M. L., Courvalin, P. & Grillot-Courvalin, C.Fitness cost of VanA-type vancomycin resistance inmethicillin-resistantStaphylococcus aureus.

Antimicrob. Agents Chemother.53, 23542359(2009).In this important paper, the authors demonstrate

that the fitness cost of the VanA-type glycopeptide

resistance cassettes that are carried by clinical

methicillin-resistant S. aureus isolates is very high

when induced by the presence of the antibiotic butvery low in the absence of induction.

58. Komp Lindgren, P., Marcusson, L. L., Sandvang, D.,Frimodt-Moller, N. & Hughes, D. Biological cost ofsingle and multiple norfloxacin resistance mutations inEscherichia coliimplicated in urinary tract infections.


59. Komp Lindgren, P., Karlsson, . & Hughes, D.Mutation rate and evolution of fluoroquinoloneresistance in Escherichia coliisolates from patientswith urinary tract infections.Antimicrob. AgentsChemother.47, 32223232 (2003).

60. Martinez-Martinez, L., Pascual, A. & Jacoby, G. A.Quinolone resistance from a transferable plasmid.Lancet351, 797799 (1998).

61. Robicsek, A., Jacoby, G. A. & Hooper, D. C. Theworldwide emergence of plasmid-mediated quinoloneresistance. Lancet Infect. Dis.6, 629640 (2006).

62. Marcusson, L. L., Frimodt-Moller, N. & Hughes, D.Interplay in the selection of fluoroquinolone resistanceand bacterial fitness. PLoS Pathog.5, e1000541(2009).These results reveal that the acquisition of an

additional fluoroquinolone resistance mutation can

not only increase drug resistance (as expected) but

also significantly increase bacterial fitness. Thus,

Darwinian selection for improved fitness can select

for increased drug resistance even in the absence

of the drug.

63. OSullivan, D. M., McHugh, T. D. & Gillespie, S. H.Analysis ofrpoB andpncA mutations in the publishedliterature: an insight into the role of oxidative stress inMycobacterium tuberculosis evolution?J. Antimicrob.Chemother.55, 674679 (2005).

64. Norstrm, T., Lannergrd, J. & Hughes, D. Genetic andphenotypic identification of fusidic acid-resistantmutants with the small colony-variant phenotype inStaphylococcusaureus.Antimicrob. AgentsChemother.51, 44384446 (2007).

65. ONeill, A. J. & Chopra, I. Molecular basis offusB-mediated resistance to fusidic acid in Staphylococcusaureus. Mol. Microbiol.59, 664676 (2006).

66. ONeill, A. J., Larsen, A. R., Skov, R., Henriksen, A. S.& Chopra, I. Characterization of the epidemicEuropean fusidic acid-resistant impetigo clone ofStaphylococcus aureus.J. Clin. Microbiol.45,15051510 (2007).

67. ONeill, A. J., McLaws, F., Kahlmeter, G., Henriksen,A. S. & Chopra, I. Genetic basis of resistance to fusidicacid in staphylococci.Antimicrob. Agents Chemother.51, 17371740 (2007).

68. Lannergard, J., Norstrom, T. & Hughes, D. Geneticdeterminants of resistance to fusidic acid amongclinical bacteremia isolates ofStaphylococcus aureus.


69. Lannergard, J. et al. Identification of the genetic basisfor clinical menadione-auxotrophic small-colonyvariant isolates ofStaphylococcus aureus.Antimicrob.

Agents Chemother.52, 40174022 (2008).70. Proctor, R. A. et al. Small colony variants: a

pathogenic form of bacteria that facilitates persistentand recurrent infections. Nature Rev. Microbiol.4,295305 (2006).

71. Cohen, T., Sommers, B. & Murray, M. The effect ofdrug resistance on the fitness ofMycobacteriumtuberculosis. Lancet Infect. Dis.3, 1321 (2003).

72. Wilson, T. M., de Lisle, G. W. & Collins, D. M. Effect ofinhA and katG on isoniazid resistance and virulenceofMycobacterium bovis. Mol. Microbiol.15,

10091015 (1995).73. Li, Z., Kelley, C., Collins, F., Rouse, D. & Morris, S.

Expression ofkatG in Mycobacterium tuberculosis isassociated with its growth and persistence in mice andguinea pigs.J. Infect. Dis.177, 10301035 (1998).

74. Pym, A. S., Saint-Joanis, B. & Cole, S. T. Effect ofkatGmutations on the virulence ofMycobacteriumtuberculosis and the implication for transmission inhumans. Infect. Immun.70, 49554960 (2002).

75. Cohen, T., Becerra, M. C. & Murray, M. B. Isoniazidresistance and the future of drug-resistanttuberculosis. Microb. Drug Resist.10, 280285(2004).

76. Normark, S. -Lactamase induction in Gram-negativebacteria is intimately linked to peptidoglycanrecycling. Microb. Drug Resist.1, 111114 (1995).

77. Jacobs, C., Huang, L. J., Bartowsky, E., Normark, S. &Park, J. T. Bacterial cell wall recycling providescytosolic muropeptides as effectors for -lactamaseinduction. EMBO J.13, 46844694 (1994).

78. Jacobs, C., Frere, J. M. & Normark, S. Cytosolicintermediates for cell wall biosynthesis anddegradation control inducible -lactam resistance ingram-negative bacteria. Cell88, 823832 (1997).

79. Tuomanen, E. et al. Coordinate regulation of beta-lactamase induction and peptidoglycan compositionby the amp operon. Science251, 201204 (1991).

80. Folkesson, A., Eriksson, S., Andersson, M., Park, J. T.& Normark, S. Components of the peptidoglycan-recycling pathway modulate invasion and intracellularsurvival ofSalmonella enterica serovar Typhimurium.Cell. Microbiol.7, 147155 (2005).

81. Tan, M. W. & Ausubel, F. M. Caenorhabditis elegans: amodel genetic host to study Pseudomonas aeruginosapathogenesis. Curr. Opin. Microbiol.3, 2934 (2000).

82. Sanchez, P. et al. Fitness ofin vitro selectedPseudomonas aeruginosanalB and nfxB multidrugresistant mutants.J. Antimicrob. Chemother.50,657664 (2002).

83. May, R. M., Gupta, S. & McLean, A. R. Infectiousdisease dynamics: what characterizes a successfulinvader? Philos. Trans. R. Soc. Lond. B Biol. Sci.356,901910 (2001).

84. Burgos, M., DeRiemer, K., Small, P. M., Hopewell, P. C.& Daley, C. L. Effect of drug resistance on thegeneration of secondary cases of tuberculosis.

J. Infect. Dis.188, 18781884 (2003).85. Bottger, E. C., Pletschette, M. & Andersson, D. Drug

resistance and fitness in Mycobacterium tuberculosisinfection.J. Infect. Dis.191, 823824 (2005).

86. Lloyd-Smith, J. O., Schreiber, S. J., Kopp, P. E. & Getz,W. M. Superspreading and the effect of individualvariation on disease emergence. Nature438,355359 (2005).

87. Luciani, F., Sisson, S. A., Jiang, H., Francis, A. R. &Tanaka, M. M. The epidemiological fitness cost of drugresistance in Mycobacterium tuberculosis. Proc. Natl

Acad. Sci. USA106, 1471114715 (2009).88. Maisnier-Patin, S., Berg, O. G., Liljas, L. & Andersson,

D. I. Compensatory adaptation to the deleteriouseffect of antibiotic resistance in Salmonellatyphimurium. Mol. Microbiol.46, 355366(2002).

89. Levin, B. R., Perrot, V. & Walker, N. Compensatorymutations, antibiotic resistance and the populationgenetics of adaptive evolution in bacteria. Genetics154, 985997 (2000).Here, the authors use modelling to show that the

trajectory of the adaptive evolution of low-fitness

antibiotic-resistant bacteria will be strongly

influenced by the relative rates of different

mutations (reversion and compensatory mutations)

and the population bottlenecks that these bacteria

experience.

90. Handel, A., Regoes, R. R. & Antia, R. The role ofcompensatory mutations in the emergence of drugresistance. PLoS Comput. Biol.2, e137 (2006).

91. Bjorkman, J., Samuelsson, P., Andersson, D. I. &Hughes, D. Novel ribosomal mutations affectingtranslational accuracy, antibiotic resistance andvirulence ofSalmonella typhimurium. Mol. Microbiol.31, 5358 (1999).

92. Johanson, U., Aevarsson, A., Liljas, A. & Hughes, D.The dynamic structure of EF-G studied by fusidic acidresistance and internal revertants.J. Mol. Biol.258,420432 (1996).

93. Meka, V. G. et al. Reversion to susceptibility in alinezolid-resistant clinical isolate ofStaphylococcusaureus.J. Antimicrob. Chemother.54, 818820(2004).

94. Meka, V. G. et al. Linezolid resistance in sequentialStaphylococcus aureus isolates associated with aT2500A mutation in the 23S rRNA gene and loss of asingle copy of rRNA.J. Infect. Dis.190, 311317(2004).

95. Wolter, N., Smith, A. M., Farrell, D. J. & Klugman,K. P. Heterogeneous macrolide resistance and geneconversion in the pneumococcus.Antimicrob. AgentsChemother.50, 359361 (2006).

96. Nilsson, A. I. et al. Reducing the fitness cost ofantibiotic resistance by amplification of initiator tRNAgenes. Proc. Natl Acad. Sci. USA103, 69766981(2006).

An important paper showing that amplification of

an unrelated wild-type gene can restore growth

fitness to an antibiotic-resistant mutant.

97. Rozen, D. E., McGee, L., Levin, B. R. & Klugman, K. P.Fitness costs of fluoroquinolone resistance inStreptococcus pneumoniae.Antimicrob. AgentsChemother.51, 412416 (2007).

98. Gillespie, S. H., Billington, O. J., Breathnach, A. &

McHugh, T. D. Multiple drug-resistant Mycobacteriumtuberculosis: evidence for changing fitness followingpassage through human hosts. Microb. Drug Resist.8,273279 (2002).

99. Heym, B. et al. Effects of overexpression of the alkylhydroperoxide reductase AhpC on the virulence andisoniazid resistance ofMycobacterium tuberculosis.Infect. Immun.65, 13951401 (1997).

100. Gagneux, S. et al. The competitive cost of antibioticresistance in Mycobacterium tuberculosis. Science312, 19441946 (2006).

101. Sherman, D. R. et al. Compensatory ahpCgeneexpression in isoniazid-resistant Mycobacteriumtuberculosis. Science272, 16411643(1996).

102. Lipsitch, M., Bergstrom, C. T. & Levin, B. R. Theepidemiology of antibiotic resistance in hospitals:paradoxes and prescriptions. Proc. Natl Acad. Sci.USA97, 19381943 (2000).

RE V I E W S


RE V I E W S



12/12

This article describes the mathematical modelling

that is used to propose practical measures to

control or reduce the prevalence of antibiotic-

resistant bacteria in hospital settings. It also

discusses why it will be easier to control resistance

in hospitals that in the community.103. Bonten, M. J., Austin, D. J. & Lipsitch, M.

Understanding the spread of antibiotic resistantpathogens in hospitals: mathematical models as toolsfor control. Clin. Infect. Dis.33, 17391746 (2001).

104.Austin, D. J. & Anderson, R. M. Studies of antibiotic

resistance within the patient, hospitals and thecommunity using simple mathematical models. Philos.Trans. R. Soc. Lond. B Biol. Sci.354, 721738 (1999).

105.Aldeyab, M. A. et al. Modelling the impact of antibioticuse and infection control practices on the incidence ofhospital-acquired methicillin-resistant Staphylococcusaureus: a time-series analysis.J. Antimicrob.Chemother.62, 593600 (2008).

106. Mahamat, A. et al. Impact of infection controlinterventions and antibiotic use on hospital MRSA: amultivariate interrupted time-series analysis. Int. J.

Antimicrob. Agents30, 169176 (2007).107. McGowan, J. E. Jr. Minimizing antimicrobial resistance

in hospital bacteria: can switching or cycling drugshelp? Infect. Control7, 573576 (1986).

108. Sjlund, M., Tano, E., Blaser, M. J., Andersson, D. I. &Engstrand, L. Persistence of resistant Staphylococcusepidermidisafter single course of clarithromycin.Emerg. Infect. Dis.11, 13891393 (2005).

109. Sjlund, M., Wreiber, K., Andersson, D. I., Blaser,M. J. & Engstrand, L. Long-term persistence ofresistant Enterococcus species after antibiotics toeradicate Helicobacter pylori.Ann. Intern. Med.139,483487 (2003).This article shows that an antibiotic treatment that

successfully eradicatesH. pyloriin patients also

selects for resistant entercocci that persist in the

patients for up to 3 years without any further

selection.

110. Sullivan, A., Edlund, C. & Nord, C. E. Effect ofantimicrobial agents on the ecological balanceof human microflora. Lancet Infect. Dis.1, 101114(2001).

111. Seppl, H. et al. The effect of changes in theconsumption of macrolide antibiotics on erythromycinresistance in group A streptococci in Finland. N. Engl.

J. Med. 337, 441446 (1997).112. Kristinsson, K. G. Effect of antimicrobial use and other

risk factors on antimicrobial resistance in pneumococci.Microb. Drug Resist.3, 117123 (1997).

113. Kataja, J. et al. Erythromycin resistance genes in

group A streptococci in Finland.Antimicrob. AgentsChemother.43, 4852 (1999).

114.Arason, V. A. et al. Clonal spread of resistantpneumococci despite diminished antimicrobial use.Microb. Drug Resist.8, 187192 (2002).

115.Austin, D. J., Bonten, M. J., Weinstein, R. A.,Slaughter, S. & Anderson, R. M. Vancomycin-resistantenterococci in intensive-care hospital settings:transmission dynamics, persistence, and the impact ofinfection control programs. Proc. Natl Acad. Sci. USA96, 69086913 (1999).

116. Enne, V. I., Livermore, D. M., Stephens, P. & Hall, L. M.Persistence of sulphonamide resistance in Escherichiacoliin the UK despite national prescribing restriction.Lancet357, 13251328 (2001).In this article, the authors show that reduced

prescription of sulphonamides in the United

Kingdom did not reduce resistance within a period

of years, probably because of genetic linkage to

other resistance determinants that continued to be

under selection.

117. Bean, D. C., Livermore, D. M., Papa, I. & Hall, L. M.Resistance among Escherichia colito sulphonamidesand other antimicrobials now little used in man.

J. Antimicrob. Chemother.56, 962964 (2005).118. Sundqvist, M. et al. Little evidence for reversibility

of trimethoprim resistance after a drastic reduction intrimethoprim use.J. Antimicrob. Chemother. 65,350360 (2010).

A key prospective intervention study in which

trimethoprim consumption in a Swedish county

was reduced by 85% for a period of 2 years. The

authors conclude that reduced antibioticconsumption is unlikely to significantly reduce

resistance levels in the community.

119. Lenski, R. E. Quantifying fitness and gene stability inmicroorganisms. Biotechnology15, 173192 (1991).

120. Lambertsen, L., Sternberg, C. & Molin, S. Mini-Tn7transposons for site-specific tagging of bacteria withfluorescent proteins. Environ. Microbiol.6, 726732(2004).

121. Kadurugamuwa, J. L. et al. Noninvasive monitoring ofpneumococcal meningitis and evaluation of treatmentefficacy in an experimental mouse model. Mol.Imaging4, 137142 (2005).

122. Kadurugamuwa, J. L. et al. Noninvasive biophotonicimaging for monitoring of catheter-associated urinarytract infections and therapy in mice. Infect. Immun.73, 38783887 (2005).

123. Xiong, Y. Q. et al. Real-time in vivo bioluminescentimaging for evaluating the efficacy of antibiotics in arat Staphylococcus aureus endocarditis model.

Antimicrob. Agents Chemother.49, 380387 (2005).124. Giraud, E., Cloeckaert, A., Baucheron, S., Mouline, C.

& Chaslus-Dancla, E. Fitness cost of fluoroquinoloneresistance in Salmonella enterica serovar Typhimurium.

J. Med. Microbiol.52, 697703 (2003).125. Paulander, W., Maisnier-Patin, S. & Andersson, D. I.

Multiple mechanisms to ameliorate the fitness burdenof mupirocin resistance in Salmonella typhimurium.Mol. Microbiol.64, 10381048 (2007).

126. Schrag, S. J., Perrot, V. & Levin, B. R. Adaptation tothe fitness costs of antibiotic resistance in Escherichiacoli. Proc. Biol. Sci.264, 12871291 (1997).

127. Schrag, S. J. & Perrot, V. Reducing antibioticresistance. Nature381, 120121 (1996).This is pioneering work demonstrating that the

initially high cost of chromosomal streptomycin

resistance mutations is rapidly compensated for in

the absence of drug selection without a clinically

significant reduction in the level of streptomycin

resistance. The results argue that prudent

antibiotic use alone may not be sufficient to reduce

the prevalence of antibiotic resistance.128. Nilsson, A. I., Berg, O. G., Aspevall, O., Kahlmeter, G.

& Andersson, D. I. Biological costs and mechanisms offosfomycin resistance in Escherichia coli.Antimicrob.

Agents Chemother.47, 28502858 (2003).129. Mariam, D. H., Mengistu, Y., Hoffner, S. E. &

Andersson, D. I. Effect ofrpoB mutations conferringrifampin resistance on fitness ofMycobacteriumtuberculosis.Antimicrob. Agents Chemother.48,12891294 (2004).

130. Billington, O. J., McHugh, T. D. & Gillespie, S. H.Physiological cost of rifampin resistance inducedin vitro in Mycobacterium tuberculosis.Antimicrob.

Agents Chemother.43, 18661869 (1999).131. Besier, S., Ludwig, A., Brade, V. & Wichelhaus, T. A.

Molecular analysis of fusidic acid resistance inStaphylococcus aureus. Mol. Microbiol.47, 463469(2003).

132. Besier, S., Ludwig, A., Brade, V. & Wichelhaus, T. A.Compensatory adaptation to the loss of biological

fitness associated with acquisition of fusidic acid

resistance in Staphylococcus aureus.Antimicrob.Agents Chemother.49, 14261431 (2005).

133. Wichelhaus, T. A. et al. Biological cost of rifampinresistance from the perspective ofStaphylococcusaureus.Antimicrob. Agents Chemother.46,33813385 (2002).

134. Hurdle, J. G., ONeill, A. J. & Chopra, I. The isoleucyl-tRNA synthetase mutation V588F conferringmupirocin resistance in glycopeptide-intermediateStaphylococcus aureus is not associated with asignificant fitness burden.J. Antimicrob. Chemother.

53, 102104 (2004).135. Gustafsson, I., Cars, O. & Andersson, D. I. Fitness ofantibiotic resistant Staphylococcus epidermidisassessed by competition on the skin of humanvolunteers.J. Antimicrob. Chemother.52, 258263(2003).

136. Johnson, C. N., Briles, D. E., Benjamin, W. H. Jr,Hollingshead, S. K. & Waites, K. B. Relative fitness offluoroquinolone-resistantStreptococcus pneumoniae.Emerg. Infect. Dis.11, 814820 (2005).

137. Binet, R. & Maurelli, A. T. Fitness cost due to mutationsin the 16S rRNA associated with spectinomycinresistance in Chlamydia psittaci6BC.Antimicrob. AgentsChemother.49, 44554464 (2005).

138. Kugelberg, E., Lofmark, S., Wretlind, B. & Andersson,D. I. Reduction of the fitness burden of quinoloneresistance in Pseudomonas aeruginosa.J. Antimicrob.Chemother.55, 2230 (2005).

139. Compeau, G., Al-Achi, B. J., Platsouka, E. & Levy, S. B.Survival of rifampin-resistant mutants ofPseudomonasfluorescens and Pseudomonas putida in soil systems.

Appl. Environ. Microbiol.54, 24322438 (1988).140. Dykes, G. A. & Hastings, J. W. Fitness costs associated

with class IIa bacteriocin resistance in Listeriamonocytogenes B73. Lett. Appl. Microbiol.26, 58(1998).

141. Fermer, C. & Swedberg, G. Adaptation to sulfonamideresistance in Neisseria meningitidis may have requiredcompensatory changes to retain enzyme function:kinetic analysis of dihydropteroate synthases fromN. meningitidis expressed in a knockout mutant ofEscherichia coli.J. Bacteriol.179, 831837 (1997).

AcknowledgementsThis work was supported by the Swedish Research Council, theEuropean Union 5th, 6th and 7th Framework Programmes andthe Swedish Agency for Innovations Systems (VINNOVA).

Competing interests statementThe authors declare no competing financial interests.

DATABASESErez Gee:http://www.ncbi.nlm.nih.gov/genefolD|fmt |fusA|marR | metW| metZ|mexR| nfxB|parC

Erez Gee Prje:http://www.ncbi.nlm.nih.gov/genomeprj

Caenorhabditis elegans | Campylobacter jejuni |

Escherichia coli|Helicobacter pylori | Mycobacterium bovis|

Mycobacterium smegmatis | Mycobacterium tuberculosis |

Pseudomonas aeruginosa|Salmonella enterica subsp.

enterica serovar Typhimurium | Staphylococcus aureus |

Staphylococcus epidermidis|Streptococcus pneumoniae |

Streptococcus pyogenes

uiPrKb:http://www.uniprot.org

KatG | RpoB|S

FURTHER INFORMATIONDa Adersss hepage:http://www.imbim.uu.se/

forskning/anderssondresearch.html

All linkS Are Active in the online pDf

RE V I E W S

NaTuRE REVIEWS | Microbiology VouME 8 | apRI 2010 | 271

http://www.ncbi.nlm.nih.gov/genehttp://www.ncbi.nlm.nih.gov/gene/1252062?ordinalpos=1&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/1252062?ordinalpos=1&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/1254930?ordinalpos=1&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/1254930?ordinalpos=1&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/1254969?ordinalpos=3&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/1254969?ordinalpos=3&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/945825?ordinalpos=21&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/945825?ordinalpos=21&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/1254513?ordinalpos=1&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/1254512?ordinalpos=1&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/1254512?ordinalpos=1&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/877857?ordinalpos=1&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/877857?ordinalpos=1&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/877857?ordinalpos=1&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/881079?ordinalpos=6&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/881079?ordinalpos=6&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/947499?ordinalpos=1&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/947499?ordinalpos=1&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSumhttp://www.ncbi.nlm.nih.gov/gene/947499?ordinalpos=1&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPan

Antibiotic Resistance and Its Cost

Documents

Transcript of Antibiotic Resistance and Its Cost