Mechanisms and consequences of bacterial resistance to antimicrobial peptides

ContentslistsavailableatScienceDirect

Drug

Resistance

Updates

j ou rn a l h o m e p a g e :w w w . e l s e v i e r . c o m / l o c a t e / d r u p

Invited

review

Mechanisms

and

consequences

of

bacterial

resistance

to

antimicrobial

peptides

D.I.

Andersson

,

D.

Hughes,

J.Z.

Kubicek-Sutherland

UppsalaUniversity,DepartmentofMedicalBiochemistryandMicrobiology,Box582,SE-75123Uppsala,Sweden

a

r

t

i

c

l

e

i

n

f

o

Articlehistory: Received14March2016

Receivedinrevisedform7April2016 Accepted11April2016 Keywords: Antimicrobialpeptides Anti-bacterialdrugs Resistance Innateimmunity Selection Bacterialinfections

a

b

s

t

r

a

c

t

Cationicantimicrobialpeptides(AMPs)areanintrinsicpartofthehumaninnateimmunesystem.Over 100differenthumanAMPsareknowntoexhibitbroad-spectrumantibacterialactivity.Becauseofthe increasedfrequencyofresistancetoconventionalantibioticsthereisaninterestindevelopingAMPsasan alternativeantibacterialtherapy.SeveralcationicpeptidesthatarederivativesofAMPsfromthehuman innateimmunesystemarecurrentlyinclinicaldevelopment.Therearealsoongoingclinicalstudies aimedatmodulatingtheexpressionofAMPstoboostthehumaninnateimmuneresponse.Inthisreview wediscussthepotentialproblemsassociatedwiththesetherapeuticapproaches.Thereisconsiderable experimentaldatadescribingmechanismsbywhichbacteriacandevelopresistancetoAMPs.Asforany typeofdrugresistance,theratebywhichAMPresistancewouldemergeandspreadinapopulation ofbacteriainanaturalsettingwillbedeterminedbyacomplexinterplayofseveraldifferentfactors, includingthemutationsupplyrate,thefitnessoftheresistantmutantatdifferentAMPconcentrations, andthestrengthoftheselectivepressure.SeveralstudieshavealreadyshownthatAMP-resistant bac-terialmutantsdisplaybroadcross-resistancetoavarietyofAMPswithdifferentstructuresandmodes ofaction.Therefore,routineclinicaladministrationofAMPstotreatbacterialinfectionsmayselectfor resistantbacterialpathogenscapableofbetterevadingtheinnateimmunesystem.Theramificationsof therapeuticlevelsofexposureonthedevelopmentofAMPresistanceandbacterialpathogenesisarenot yetunderstood.ThisissomethingthatneedstobecarefullystudiedandmonitoredifAMPsareusedin clinicalsettings.

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

Contents

1. IntroductiontoAMPs...44

1.1. DefinitionofAMPsandearlyhistory...44

2. StructuresofAMPs...44

3. AMPstructureinfluencesactivity...45

4. BiologicalrolesofAMPs...45

4.1. ExpressionofAMPs...45

4.2. Antibacterialactivity(direct)...45

4.3. Immunemodulation(mayindirectlybeantibacterial)...45

5. ModeofactionofAMPs...45

5.1. Membraneinteractionspecificity...45

5.2. MembraneandcellwallactivityofAMPs...46

5.3. AMPsactivitymayalsoinvolveintracellulartargets...46

6. IntrinsicresistancetoAMPs...46

6.1. MembranemodificationsinGram-negativebacteria...46

6.2. MembranemodificationsinGram-positivebacteria...47

6.3. OtherintrinsicAMPresistancemechanisms...47

6.4. MethodstoidentifyintrinsicAMPresistancemechanisms...48

∗ Correspondingauthor.Tel.:+46184714175.

E-mailaddress:Dan.Andersson@imbim.uu.se(D.I.Andersson).

http://dx.doi.org/10.1016/j.drup.2016.04.002

7. Acquiredresistance...48

7.1. MethodstoselectformutantswithacquiredAMPresistance...48

7.2. Mechanismsofacquiredresistance...49

7.2.1. Naturallyoccurringacquiredresistance...49

7.2.2. Serialpassage...50

7.2.3. Directplating...50

7.3. WhatdeterminestheriskofacquiringresistancetoAMPs?...51

7.3.1. Mutationsupplyrate...51

7.3.2. Mutantfitness...51

7.3.3. Selectionstrength...52

7.3.4. Compensatoryevolution...52

7.3.5. Epistaticinteractions...52

8. TherapeuticuseofAMPs ... 52

9. Concludingremarks ... 54

Acknowledgement...54

References...54

1. IntroductiontoAMPs

1.1. DefinitionofAMPsandearlyhistory

Cationicantimicrobialpeptides(AMPs)arerelativelysmall pep-tidesthat, bydefinition,have anetpositivecharge,and exhibit some,oftenbroad-spectrum,antimicrobialactivity.Thisrather tau-tologicaldefinition encompassesa great diversityof molecules, atboth thesequenceandthestructurallevel.Anonline antimi-crobialpeptidedatabase,APD3(Wangetal.,2016),accessibleat http://aps.unmc.edu/AP/,listsover2600examplesofAMPsfrom allkingdomsoflife:bacteria,archaea,andeukaryotes(including plants,animals,fungiandprotists).AMPscanactdirectlytoprotect animalsagainstawidevarietyofbacterial,viral,fungaland proto-zoaninfectionsandbecauseofthisareoftenreferredtoashost defensepeptides.Inanimals,bothvertebratesandinvertebrates, AMPsalsofunctionasauniversalandimportantpartoftheinnate immune system,modulating activities that promote protection against infections by microorganisms. The APD3 database cur-rentlylists112differenthumanhostdefensepeptides ofwhich 100haveexperimentallydeterminedantibacterialactivities.The focusofthisreviewisontheimportanceofhumanAMPsin pro-tectingagainstbacterialinfections,onthemechanismsofintrinsic andacquiredbacterialresistancetoAMPs,andonthe opportuni-tiesandpotentialconsequencesassociatedwiththerapeuticuseof AMPs.

AMPsarenotuniqueaspeptide-basedsmallmoleculeswith antibacterialactivity.Manyoftheantibioticsproducedby microor-ganisms are peptide-based molecules. These are produced by non-ribosomalpeptidesynthesis(NRPS),a processthatinvolves theexpressionoflargearraysofgenesencodingmultipleenzymes that work in succession to catalyze the sequence of chemi-calreactions required tosynthetizethe antibiotic(Baltz, 2006; Hughes,2003).Peptide-basedantibioticsinclude␤-lactamssuch aspenicillin,cyclicpeptideantibioticslikethepolymixinsand bac-itracin,glycopeptideslikevancomycin(Yimetal.,2014),andthe lipopeptidedaptomycin,oneofthemostrecentlyintroducednovel antibioticclasses(RobbelandMarahiel,2010).TheAMPsdiscussed inthisreviewdifferfrompeptide-basedantibioticsinthe mech-anismoftheirsynthesis and alsointheiramino acidmake-up. AMPs, includingthose made in humancells, are in contrastto NRPS-antibiotics,produced bythenormal process of ribosomal translationonan mRNAtemplate.The primaryproduct is usu-allya pre-protein that is then processed tothe final length of theactiveAMP.Forexample,thehumancathelicidinAMP,LL-37, isproducedasaprecursorprotein,humancationicantimicrobial protein-18kDa(hCAP18),andthematureLL-37peptideistheresult ofitsprocessingbyproteinase3(Gudmundssonetal.,1996).This

differenceinthegeneticoriginsofAMPsandNRPS-antibioticshas consequencesforthecompositionsofthefinalproducts. Ribosoma-llyproducedAMPscontainonlythenormalcomplementofamino acidsfoundinproteins,sometimeswithpost-translational modifi-cations.NRPS-producedantibiotics,incontrast,areunconstrained bythelimitations of ribosomaltranslation, and usuallycontain a mixture of normal amino acids together with non-canonical amino acids not foundin any proteins(Walsh et al., 2013). In addition,genes forsomeof thehuman␣-defensins (HNP1 and HNP3)arepresentinmultiplecopiesandareinheritedunequally by different individuals, giving rise toindividuals with two or threecopiesof eitherorbothofthesegenes,potentially affect-ingthelevelsofAMPsproducedamongindividuals(Marsetal., 1995).

Human AMPs vary in length from 5 to 149 amino acids (Bangaloreetal.,1990;Cashetal.,2006).Thegreatmajorityare cationic peptides,but the netcharge of different humanAMPs ranges from ₋₃ to +20. AMPs have been identified in differ-ent human excretions, tissues and cell types, including saliva, tears, sweat and milk, on the skin and tongue, in bone mar-row,plasma, kidneys, liver,heart, brain,eyes, intestine, sperm, urinarytract,amnioticfluid,andrespiratorytract,andinmany dif-ferentcelltypesincludingepithelial/mucosalcells,macrophages, neutrophils, natural killer cells, monocytes, eosinophilic leuko-cytes, Paneth cells, T-cellsand B-cells (http://aps.unmc.edu/AP/ ).Research intothefunctionsofAMPsgoesbackatleasttothe timeofAlexanderFlemingwhodiscoveredlysozyme(Flemingand Allison,1922).However,itisinthepastfewdecades,especially afterthediscoveryofcationicbacteriocidalpeptidesin polymor-phonuclearleukocytes(ZeyaandSpitznagel,1963),thatinterest andresearchhasexpandedexplosively.Duringthe1990s impor-tantnewclassesofAMPswerediscoveredinhumans,including the␤-defensins(Benschetal.,1995)andthecathelicidins(Gallo etal.,1997)andthisstimulatedresearchintounderstandingthe numberandvarietyofhumanAMPsand,morerecently,intotheir potentialapplicationsinantimicrobialtherapytocontrolbacterial infections.

2. StructuresofAMPs

Defensins and cathelicidin were recognized early on as important components of the antimicrobial mechanisms of polymorphonuclear leukocytes (PMNs), and were classified as belongingtosignificantlydifferentstructuralfamilies(Lehrerand Ganz,2002;Zanettietal.,1995).AclassificationofAMPsintofour majorfamilies,basedontheir3Dstructures,hasbeenadoptedby theAPD3database(Wangetal.,2016).

(i)ThealphafamilyiscomposedofAMPswithhelicalstructures, forexample,thehumancathelicidinLL-37(Agerberthetal., 1995).

(ii)ThebetafamilyiscomposedofAMPswith␤-strands,for exam-ple,thehuman␣-defensins(Selstedetal.,1985).

(iii)ThealphabetafamilycomprisesAMPswithboth␣-helicaland ␤-strands structures in thesame3D fold, for example,the human␤-defensins(Harderetal.,1997).Thesequencesofboth the_␣-and_{␤-defensins}includesixcysteineresiduesthat sta-bilizethestructurebyformingthreeintramoleculardisulfide bonds.␤-SheetAMPsarefrequentlycyclicmoleculeswiththe structureconstrainedbytheintramoleculardisulfidebridges. The␣-and␤-defensinsdifferinthebondingpatternoftheir cysteines,probablyreflectingtheirdifferentphylogenetic ori-gins.Intotal,humansencodesixdifferent␣-defensins,over 30different␤-defensins,butonlyone␣-helicalcathelicidin, LL-37(BauerandShafer,2015;Wang,2014).

(iv)The non-alphabeta family contains AMPs with neither ␣-helicalnor␤-strands,forexample,theantimicrobialpeptide indolicidin,isolatedfromcattle(Selstedetal.,1992).Because 3Dstructuresareknownforlessthan400oftheover2600 identifiedAMPs,thisclassificationsystemshouldberegarded asprovisionalandotherclassificationsystemsareconcurrently inuse,basedonpropertiessuchascovalentbondingpatterns, hydrophobicity,netcharge,ormoleculartargets(Wang,2015).

3. AMPstructureinfluencesactivity

TheantibacterialpropertiesofAMPsareassociatedwithtwo interrelatedfeaturesofthepeptides:theirnetcharge,whichfor thegreatmajorityofAMPsispositive,andtheirpropensitytobe amphipathic,meaningthattheycanfoldintostructureswithboth ahydrophobicandahydrophilicsurface.Thesefeaturestogether facilitatetheinteractionofAMPswiththenegativelycharged com-ponentsofthebacterialenvelope(e.g.,lipopolysaccharides(LPS) andtechoicacids(TA))onthesurfacesofGram-negativeand Gram-positive bacteria,respectively, and withthenegatively charged phospholipidsofthebacterialmembrane(Morgeraetal.,2008; Wang,2008).Forexample,AMPslikeLL-37areinitiallyattracted tothebacterialsurfacebyelectrostaticinteractions.LL-37interacts withlipidbilayerssuchthattheamphipathichelixisoriented par-alleltothesurfaceofthebilayer(HenzlerWildmanetal.,2003). Itiswidelybelievedthataftertheinitialelectrostaticinteraction withthebacterialsurfacethat theamphipathic natureofAMPs allowsthemtoinsertintothebacterialcellmembrane,forming aporethatdisruptsmembraneintegrity,leadingtoosmoticlysis ofthebacterialcell.However,asdiscussedbelow,thereisevidence thatthisviewoftheAMPantibacterialmodeofactionmaybean oversimplification.

4. BiologicalrolesofAMPs

4.1. ExpressionofAMPs

HumanAMPs areexpressedina widevarietyof cellsofthe immunesystem,includingneutrophils,NK,andTcells,aswellasin theepithelialcellsliningtissuesfacingtheexternalenvironment (Yangetal.,2004).TheexpressionoftheseAMPscanbe consti-tutive(e.g.␣-and␤-defensinsinneutrophils,and␣-defensinsin panethcells),inducible(e.g.␣-defensinsinCD8Tcells,␤-defensins inepithelialcells)orbothconstitutiveandinducible(e.g. catheli-cidinsand_{␤-defensins}inavarietyofcells,includingneutrophils, epithelialcells,andmacrophages).Expressionofmostofthe ␤-defensinsis alsoinducedbyproinflammatorystimuli,including

bacteria,LPS,TNF␣,andIL-1(Diamondetal.,1996;Fangetal.,2003; Singhetal.,1998).

4.2. Antibacterialactivity(direct)

Human cathelicidin and humandefensinshave the capacity tokillawidevarietyofGram-positiveandGram-negative bacte-riainvitro(Ganzetal.,1985;Garciaetal.,2001b;Harderetal., 2001,1997;OuelletteandSelsted,1996;Zaiouetal.,2003).Like smallmoleculeantibiotics,this antibacterialactivityof AMPsis dose-dependentandbacteriocidalactivitiesareusuallyobservedat ␮Mconcentrations.Evidencefortheirantibacterialefficacyinvivo comesfromstudiesusinginfectionsinanimalmodels(Huangetal., 2002),andalsofrominfectionstudiesusingtransgenicor knock-outmicedefectiveintheproductionofvariousAMPs(Moseretal., 2002;Salzmanetal.,2003;Wilsonetal.,1999).

4.3. Immunemodulation(mayindirectlybeantibacterial)

AMPsareproducedbyepithelialcellsandbycirculatingimmune cells including neutrophils and macrophages, and are among thefirstimmuneeffectorsencounteredbyaninvadingmicrobe (Jenssenet al., 2006).The humancathelicidinLL-37 modulates theinnateimmuneresponsebyactingasachemoattractantfor neutrophils,monocytes,andmastcells(Niyonsabaetal.,2002b; Yang et al., 2000)and human ␤-defensins chemoattract leuko-cytes(Garciaetal.,2001a;Niyonsabaetal.,2002a;Territoetal., 1989).␤-Defensinsalsoattractdendriticcellsthatcanactdirectly byphagocytizingandkillingpathogens(Liu,2001)orby produc-ingcytokines,chemokinesandotherAMPsthatthenparticipatein innateantimicrobialactivity(Duitsetal.,2002).Thuscathelicidin, defensinsandchemokinesoverlapintheirchemotacticactivityand affectsoninnateimmunity.

5. ModeofactionofAMPs

5.1. Membraneinteractionspecificity

Selectivetoxicity,killingbacterialpathogenswithoutdamaging hosttissue,isafeaturethatiscrucialforAMPs.However,while it hasbeendemonstrated thatAMPs canpermeabilizebacterial membranesandcellwalls,andthatthisabilitycorrelateswiththeir antibacterialeffects,theactualmodeofbacterialkillingbyAMPs remainsanareawhere thereisstillapoorlevelof understand-ing.Onecluetothepotentialspecificityofkillingbymembrane disruptionisthattherearesignificantdifferencesinthelipid com-positionofbacterialandeukaryoticcellmembranes(Sohlenkamp andGeiger,2016;Teixeiraetal.,2012).Bacterialmembranesare rich in negatively chargedphospholipidssuch as phosphatidyl-glycerol(PG),cardiolipin(CL)andphosphatidylserine(PS)thatare stabilizedbydivalentcationssuchasMg2+_andCa2+_.Inaddition,

Gram-negativebacteriahaveanoutermembranecontainingLPS that actsasa permeability barrierreducing access tothe cyto-plasmicmembrane.Incontrasttobacteria,humancellmembranes arerichinzwitterionicphospholipidswithaneutralnetcharge, suchasphosphatidylethanolamine(PE),phosphatidylcholine(PC), andsphingomyelin(SM).ThehumanAMPLL-37candiscriminate betweendifferentmembranesbased ontheirlipidconstituents, preferentiallyinsertingintolipidscontainingPG,typicalof bacte-rialmembranes,butnotintomembranescontainingPCorPE,as foundinhumanredbloodcells(Nevilleetal.,2006).These impor-tant differences between bacterial and eukaryotic membranes, wherebacteriahaveanioniclipidsexposedonthesurfacewhereas eukaryoticmembranessequesteranioniclipidsinthemonolayer facingtheinteriorofthecell,mayexplaintheselectivityofAMPs for bacterial membranes. Human cell membranes also contain

significantamountsofcholesterolthataffectthefluidityof phos-pholipidsinthemembrane,andincreasethestabilityofthelipid bilayers(Tytleretal.,1995).Accordingly,anumberof distinguish-ingfeaturesincludingthepresenceofcholesterol,themembrane potential,andtheasymmetricdistributionofphospholipidsinthe membrane,maybeimportantforreducingthebindingofAMPsto humancells(LaiandGallo,2009).

5.2. MembraneandcellwallactivityofAMPs

ThebacteriocidaleffectofAMPsiswidelybelievedtobedue totheformationofporesinthebacterialcytoplasmicmembrane, resultinginalossofcontroloverionflowsacrossthemembrane andcelldeath.ThereisconsiderableinvitrodatashowingthatAMPs candisruptlipidbilayers,supportingthisbasicmodel.Exactlyhow AMPsformporesinbacterialmembranesislesscertain.Theinitial stepisbelievedtoinvolveelectrostaticinteractionsbetweenthe positivelychargedaminoacidsinAMPsandthenegativelycharged LPS,orphospholipidheadgroupof themembrane,resulting in AMPsaccumulatingonthesurfaceofthemembrane(Epandetal., 2015).Afterreachingsomethreshold concentration,AMPs may self-assembleonthemembraneandincorporateintoitcreating aporeintheprocess.SeveralmodelsforAMP-directedmembrane poreformationhavebeensuggested:(i)thebarrel-stavemodel, proposesthatpeptidesinsertperpendicularlyinthebilayer, asso-ciatingtogethertoformapore;(ii)thecarpetmechanism,proposes thatpeptidesadsorbparalleltothebilayerand,afterreaching suf-ficientdensity ofcoverage, producea detergent-likeeffect that disintegratesthemembrane;and(iii)thetoroidalporemechanism, proposesthatpeptidesinsertperpendicularlyinthebilayerand inducealocalmembranecurvaturewiththeporelinedpartlyby peptidesandpartlybyphospholipidheadgroups(Meloetal.,2009). Recentresearchsuggeststhattheinteractionsbetweendifferent AMPsandbacterialmembranesmayinvolvemuchmorespecific interactionsthansuggestedbythesegenericmodelsofpore for-mation.Forexample,phosphatidylethanolamine(PE),presentat ahighconcentrationonthesurfaceofbacterialmembranesbut sequesteredonthecytoplasmicleafletofmammalianmembranes, actsasahighaffinitylipidreceptorforseveralAMPs(Phoenixetal., 2015).CurrentevidencesuggeststhatmanyAMPs,includingthe plantcyclotides,usespecificreceptorssuchasPE,intheir inter-actionswithmembranes(Stromstedtetal.,2016).Otherspecific componentsofbacterialmembranesthat aretargetedby AMPs includeLPS,lipoteichoicacid(LTA)andthepeptidoglycan precur-sorlipidII(Schmittetal.,2015).LipidIIseemstobeafavoritetarget. Theantibioticactivityofthehuman␤-defensin3(Sassetal.,2010) andhuman␣-defensin1(deLeeuwetal.,2010;Varneyetal.,2013), reliesontheirselectivebindingoflipidIItoblockbacterialcell wallbiosynthesis.Accordingly,thebacteriocidalactivitiesofmany humanAMPsmayhavelittleornothingtodowithgenericeffects ontheintegrityofbacterialcytoplasmicmembranesandmoreto dowithspecifictargetinteractionsthathappentobelocalizedto thebacterialcellwalland/ormembrane.

5.3. AMPsactivitymayalsoinvolveintracellulartargets

AMPs have alsobeen shown totraversethe cell membrane andtoblockessentialcellularprocesseswithoutcausingextensive membranedamage(Brogden,2005;Patrzykatetal.,2002)raising thequestionofwhethersomeoftheirantibacterialactivitiesare dependentonintracellulartargeting.Thus,analogsofthehuman ␤-defensin4wereshowntotranslocateacrosstheouterandinner membranesofEscherichiacoliwithoutcausingcelllysis,suggesting thattheobservedbacteriocidalactivitymightdependon cytoplas-micinteractions(SharmaandNagaraj,2015).Human␣-defensin5 wasalsoshowntotranslocateintothecytoplasmofE.coliandto

accumulateatthesiteofcelldivisionandatthecellpoles, suppor-tingthenotionthatatleastpartofitsantimicrobialactivitymight beexertedinthecytoplasm(Chileveruetal.,2015).Indolicidin, abovinecathelicidinwithbroad-spectrumbacteriocidalactivity, hasbeenshownnottolysebacterialcellsbuttoinhibitDNA syn-thesis(Subbalakshmi and Sitaram, 1998)and shown invitroto interactwithduplexDNA(Ghoshetal.,2014).Analogsofbuforin antimicrobialpeptideswerealsoshowntoexerttheir bacterioci-dalactivityonE.colibybindingtoDNAandRNAafterpenetrating thecellmembrane(Haoetal.,2013).OnepossibilityisthatAMP targetingcanbepromiscuousanddependentonlocal concentra-tion,eithercausinglethalmembranedisruptionortranslocation intothecytoplasmtointerferewithcellularmetabolism(Sugiarto andYu,2007).Oneoftheapproaches totestthesignificanceof intracellularinteractionsinthebacteriocidaleffectsofAMPsisto selectformutantswithreducedsusceptibilitytoAMPsandthen geneticallymaptheselectedmutations.Inthefollowingsections thisapproachisdescribed,withthetwinaimsofdeterminingthe probabilityofresistancedevelopinginthecasethatAMPsareused morewidelyintherapy,andtogainabetterunderstandingofthe moleculartargetsofAMPs,whetheratthecellwall,cellmembrane, orintracellularly.

6. IntrinsicresistancetoAMPs

BacteriaencounterAMPsfrequentlyintheirnatural environ-mentsandhaveevolvedmechanismstoresisttheiraction.Intrinsic resistancetoAMPscanoccurviapassiveorinduciblemechanisms. PassiveAMPresistanceoccursinsomebacterialspecies includ-ingProteus,Morganella,Providencia,Serratia,andBurkholderiaasa resultofaninherentlymorepositivelychargedlipidAthatreduces AMPinteraction(Mannielloetal.,1978;ViljanenandVaara,1984). TheinductionofAMP-resistanceinotherbacteriaistightly reg-ulatedin responsetoenvironmentalconditionsandserves asa mechanismforbacterialsurvivalinvariousnaturalenvironments inwhichtheywouldexperiencethethreatofAMPs.Inducible(also referredtoasadaptive)resistanceresultsintransientmolecular modificationsinbothGram-negativeandGram-positivebacteria (summarizedinFig.1).Reversibilityoffsetstheenergeticburden imposedbymanyofthesemodifications,whichlargelyaffectthe compositionof thebacterialmembrane. Themolecularbasis of inducibleAMPresistancehasbeenreviewedextensivelyforboth Gram-negative and Gram-positive bacteria (Anaya-Lopez et al.,

2013; Gunn, 2001; Kraus and Peschel, 2006; Matamouros and

Miller,2015;Nawrockietal.,2014;Nizet,2006;YeamanandYount, 2003).AcommonmechanismofAMPresistanceinbacteriaisthe incorporationofpositivelychargedmoleculesintotheircell sur-faceinordertoreducetheinteractionandbindingofcationicAMPs. Inactivationofthesemechanismsresultsinenhancedsusceptibility toAMPs,includingcomponentsofthehostinnateimmunesystem andofteninreducedvirulence.

6.1. MembranemodificationsinGram-negativebacteria

IntrinsicAMPresistancehasbeenmostextensivelystudiedin Salmonella enterica serovarTyphimurium in which a variety of LPSmodificationsaretriggeredbyenvironmentalstimuli includ-ingnutrientstarvation(McLeodandSpector,1996),lowpH,low magnesium,andhighiron(GunnandMiller,1996)aswellasin varioushosttissues(Kubicek-Sutherlandetal.,2015).ThePmrAB, PhoPQandRcsregulatorysystemsmediatemanyofthese modifi-cationsinordertosupportbacterialsurvivalinvivo.Theaddition of4-aminoarabinose(Ara4N)tolipidA,whichcreatesaless nega-tivelychargedLPSandreducesbindingaffinityofAMPs,isregulated bythepmrCABandpmrHFIJKLMoperonsinS.Typhimurium(Gunn

Fig.1.SummaryofintrinsicantimicrobialpeptideresistancemechanisminGram-negativeandGram-positivebacteria.Themainpathwaysresultingintransienthigh-level AMPresistanceinbacteriaaremembranemodifications,increasedeffluxandproteolyticdegradation.

etal.,2000).InactivationofpmrAorpmrBresultsinattenuated viru-lencewhenadministeredorallybutnotintraperitoneallyindicating theimportanceofAra4Nmodificationin overcomingtheinnate immunityintheintestine.PmrAcanalsoberegulatedbythePhoPQ systemviatheproteinPmrD,whichpreventsdephosphorylationof PmrAresultingincontinuedactivation(Rolandetal.,1994).Other PhoPQregulatedgenesthatmediateAMPresistanceandvirulence inS.TyphimuriumincludeslyA,encodingavirulence regulatory protein; ugtL, encoding an inner membrane protein that forms monophosphorylatedlipidA;pagP,encodinganoutermembrane proteininvolvedinpalmitoylationoflipidA;andyqjA,aninner membraneproteinofunknownfunction(Guoetal.,1998;Shietal., 2004a,b).TheRcsphosphorelaysystemregulatesAMPresistance inS.Typhimurium byactivatingexpressionofthe uncharacter-izedperiplasmicproteinYdeI,whichisrequiredforAMPresistance andoralvirulenceinmice(EricksonandDetweiler,2006).Other AMP resistanceand virulence determinants in S. Typhimurium includesurA,tolB,andpgm.SurAandTolBareinvolvedin mem-branestabilizationwhilePgm(phosphoglucomutase)isinvolved inLPSbiosynthesis(Patersonetal.,2009;Tamayoetal.,2002).

ThePhoPQsystemalsoplaysaroleinAMPresistanceand vir-ulenceinPseudomonasaeruginosa(Gooderhametal.,2009),and PmrABregulatedLPSmodificationsconferAMPresistancein Acine-tobacter baumannii (Adams et al., 2009).In Vibriocholerae, the almEFGoperonencodestheglycylationofLPS,whichconfers resis-tancetopolymyxins(Hendersonetal.,2014).InP.aeruginosathe arnBCADTEFoperonmediatestheadditionofAra4NtolipidA con-ferringAMPresistance(McPheeetal.,2003).Theanioniccapsular polysaccharideofKlebsiellapneumoniae,P.aeruginosa,and Strep-tococcuspneumoniae,hasalsobeenshowntoconferresistanceto polymyxinBandhuman␣-defensin1bybindingcationicAMPs (Llobetetal.,2008).

6.2. MembranemodificationsinGram-positivebacteria

In Gram-positive bacteria, AMP-resistance occurs following theincorporationof positivelychargedmoleculesintocell wall teichoicacids(TA),which playanessentialrole incelldivision,

morphology, adhesion, virulence and antimicrobial resistance (Brown et al., 2013). Although critical to cellular function, the anionicpropertiesofTAsmakethematargetforthebindingof positivelychargedAMPs.ResistancetoAMPsisoftenconferredina transientmannerviaTAmodificationsthatreducetheoverall neg-ativechargeinthebacterialmembrane.TheGraRS(alsoknownas aps)sensor/regulatorsystemsensesAMPsintheenvironmentand inducesAMPresistanceinStaphylococcusaureusbyactivatingthe d-alanylationofTA,theincorporationoflysylphosphatidylglycerol inthebacterialmembrane,andactivationofthevraFGAMP trans-porter(Lietal.,2007).d-Alanineestermodificationsaremadevia componentsofthedltoperoninresponsetolowdivalentcation concentrations(Mg2+_andCa2+_{)andthepresenceofcationicAMPs}

(Koprivnjaketal.,2006).Inactivationofthedltoperonresultsin increasedsusceptibilitytoAMPsinS.aureus(Pescheletal.,1999), Bacillus subtilis(Caoand Helmann, 2004), GroupA streptococci (GAS)(Kristianetal.,2005), GroupBstreptococci(GBS)(Poyart et al., 2003), Listeria monocytogenes (Abachin et al., 2002), and Lactobacillusreuteri(Walteretal.,2007).Similarly,theMprF pro-tein(multiplepeptideresistancefactor)mediatestheadditionof l-lysinetomembranephospholipidstherebyreducingtheiroverall negativechargeandconferringAMPresistanceinS.aureus(Ernst etal.,2009),L.monocytogenes(Thediecketal.,2006),andBacillus anthracis(Samantetal.,2009).InactivationofgraRormprFresults inenhancedsusceptibilitytoAMPsandattenuationofvirulencein mice(Krausetal.,2008;Lietal.,2007;Yangetal.,2012).

6.3. OtherintrinsicAMPresistancemechanisms

Besidesmembrane modification,othereffective methodsfor dampeningtheantimicrobialeffectsofAMPsareeffluxand pro-teolyticdegradation.InS.TyphimuriumAMPsareexportedbythe sap(sensitivitytoantimicrobialpeptides)effluxsystemand the putativeABCtransporterencodedbytheyejABEF,bothofwhichare requiredforvirulenceinmice(Eswarappaetal.,2008;Groisman etal.,1992b).InS.aureus,thechromosomalvraFGgeneencodesthe ABCtransporter-dependenteffluxpump,alsoregulatedbyGraRS (Lietal.,2007).Also,theplasmid-encodedqacA geneencodesa

Fig.2. MethodsusedforisolationofAMP-resistantmutants.(A)Serialpassageinvolvesthegrowthofbacteriainliquidmediacontainingprogressivelyincreasing concen-trationsofAMPs.(B)SelectiononplatesinvolvesdirectplatingofbacteriaonagarplatescontainingAMPconcentrationsabovetheMIC.

protonmotiveforce-dependenteffluxpump(Kupferwasseretal., 1999).EffluxpumpsthatactonAMPshave alsobeenidentified in Neisseria gonorrhoeae (Shafer et al., 1998) and Yersinia spp. (Bengoechea and Skurnik, 2000). Proteases that degrade AMPs includePgtEinS.Typhimurium(Guinaetal.,2000)andOpmTin E.coli(Stumpeetal.,1998).

6.4. MethodstoidentifyintrinsicAMPresistancemechanisms TodatethemolecularbasisofintrinsicAMPresistancehasbeen characterizedlargelyusinggeneinactivationstudies.These stud-ieshave beenconducted usingboth large-scale(genome-wide) andtargeted(single gene)approaches.Screening approachesto identifygenesassociated withAMPresistancetypically involve transposonorchemicalmutagenesisofawildtypeAMP-resistant bacterialisolatetocreateacollectionofmutants,whicharethen screenedphenotypically for resultingAMP hyper-susceptibility. Suchstudieshaveidentifiedgenesconferring AMPresistancein S.Typhimurium(Fieldsetal.,1989;Groismanetal.,1992b),E.coli (Groismanetal.,1992a),GroupAStreptococcus(GAS)(Nizetetal., 2001);S.aureus(Pescheletal.,1999,2001);andNeisseria menin-gitidis(Tzengetal., 2005).Transposon mutagenesiscanalsobe performedonabacterialstrainwithamutantbackgroundinorder toassessthecontributionofparticulargenesinaresistance path-way.Forinstance,mutagenesiswasperformedinS.Typhimurium onaconstitutivelyactivepmrAmutantallowingfortheinitial iden-tificationofnon-regulatorygenesconferringAMPresistance,pmrE (alsoknownaspagAandugd)andpmrF,bothofwhichareinvolved inlipidAmodification(Gunnetal.,1998).

Theselarge-scalescreenshaveprovidedtremendousinsights intothemolecularbasisofintrinsicresistance,howeverthereare disadvantagestothismethodology.Drawbacksfromalarge-scale knockout approach include the difficulty in observing small changesinAMPsusceptibilitythatoftenrequiretime-killassays todetect;AMP-resistancemechanismsinvolvingmultiplegenes arenotdirectlyassayed;andtheinvolvementofessentialgenes cannotbecharacterized.Thismethodcanalsobelimitedbythe highcostofcertainAMPs,whichcanmakelarge-scalephenotypic screening very expensive and the approach hasthus far been largelylimitedtolessexpensivepeptides.Transposon mutagene-siscanalsoinvolveharshconditionsoftenresultinginadditional

unintendedbackgroundmutationsthataretypicallynotaccounted forinlarge-scalescreens.

Once a gene or pathway has been identified, a subsequent moretargetedapproachcanprovidedirectinformationaboutthe relationshipbetweenindividual genesand AMPresistance. The inactivationofindividualgenesofinteresthashelpedtoidentify geneswithinaregulonthat maybeoverlooked inalarge-scale approach.Narrowingtheanalysisdowntoseveralgenesofinterest alsosupportsthetestingofresistancetoavarietyofAMPs, includ-ingthosethataremoreexpensive.However,thistargetedapproach requirespriorelucidationofcandidategenesandtheirregulatory pathwaysand still doesnot include essentialgenes. Lastly,the largestproblemwithgeneinactivationstudiesistheinabilityto assessthemolecularbasisofacquired(stable)AMPresistance.

7. Acquiredresistance

7.1. MethodstoselectformutantswithacquiredAMPresistance ConsideringthecontinuedhighinterestinAMPsaspotential therapeuticdrugs forbacterialinfections surprisinglyfew stud-ieshave tried toassesstherisk of resistancedevelopmentand explorethemechanismsforacquiredresistance.Partlythismight beexplainedbythelong-held,anderroneous,beliefthatresistance toAMPsisverydifficulttoacquireandthatitthereforeisnota bigconcern(Batonietal.,2011;GhoshandHaldar,2015;Nizet, 2006;YeamanandYount,2003).However,asseveralrecent stud-ieshaveshown,resistancetoAMPscaninfactevolveathighrates (atleastinvitro),generatingmutantswith,sometimeshigh,levels ofresistance.

Typically two different methods have been used to select mutants with acquired resistance to AMPs (Fig. 2). The first, and most used, method involves serial passage of bacteria in mediacontainingAMPsatconcentrationsneartheMIC,whichare progressivelyincreasedduringthe courseoftheexperiment as resistantmutantsareenriched.Afterdifferenttimepoints bacte-riaarerecoveredandtestedforresistancebybrothmicro-dilution (toobtainMICsoftheAMPs)and/ortime-killassays(Loftonetal., 2013).Generallythelattermethodismoresensitiveindetecting smalldifferencesinsusceptibility(Loftonetal.,2013).The advan-tagewithserialpassageisthatitisrelativelyeasytoperformand

Table1

Summaryofknownmechanismsofacquiredresistancetoantimicrobialpeptides.

Organism Methodforisolation AMPresistance Genesinvolvedin AMPresistance

Proposedmechanism Reference Staphylococcusaureus Clinicalisolates LL-37,human

␤-defensin2,human ␤-defensin3, lactoferricinB

hemB Inactivationresultsinsmall colonyvariant(SCV) phenotypewithreducedAMP binding/uptake

Glaseretal.(2014)

Escherichiacoli,Klebsiella pneumoniae

Clinicalisolates Colistin,polymyxinB mcr-1 EncodesaPEtNtransferase modifieslipidAtoreduce anioniccharge

Huetal.(2016)

Listeriamonocytogenes Directplatingwith leucocinA

LeucocinA mptACD Unknown Gravesenetal.(2002)

Acinetobacterbaumannii Directplatingwith colistin

Colistin,polymyxinB lpxA,lpxD,orlpxC Inactivationresultsin completelossofLPS production,reducedAMP binding

Moffattetal.(2010)

SalmonellaTyphimurium DESmutagenesis Polymyxin,CAP37, CAP57,protamine, polylysine

phoQ Constitutiveactivationinlow Mg2+_{ofphoP-regulatedLPS} modificationsreduceanionic charge

GunnandMiller(1996)

Directplatingwith colistin

Colistin,polymyxinB pmrA,pmrB Constitutiveactivationof pmrAB-regulatedAra4Nand PEtNLPSmodificationreduce anioniccharge

Rolandetal.(1993) andSunetal.(2009)

Directplatingwith PR-39

PR-39 sbmA InactivationreducesAMP uptake

Prantingetal.(2008)

Directplatingwith protamine

Protamine,colistin, lactoferricin,human ␣-defensin1

hemA,hemB,hemC, hemL

Inactivationresultsinsmall colonyvariant(SCV) phenotypewithreducedAMP binding/uptake

Prantingand Andersson(2010)

Serialpassagewith LL-37orCNY100HL

LL-37,CNY100HL, wheatgermhistones

pmrB,phoP Constitutiveactivationof variousLPSmodifications reducinganioniccharge

Loftonetal.(2013)

Ara4N,4-aminoarabinose;PEtN,phosphoethanolamine.

thatitrequireslowamountsofAMPssincegrowthvolumesare small.However,disadvantagesarethatitonlyselectsforhigh fit-nessmutants(i.e.themutantspectrumwillbebiased)andbecause selection isstepwisethe mutantsoftencontain morethan one mutation,necessitatingsubsequentgeneticreconstitution experi-mentstoconfirmtheroleofeachindividualmutations.Likewise, becausetheselectionisnotlethalandgrowthoftenoccurs dur-ingseveralhundredgenerations,mediaadaptationmutationswill almostalwaysbeacquired(Loftonetal.,2013).Thesemutations conferfitnessincreasesthatareunrelatedtotheAMP-selectionand areidentifiablebyperformingaparallelexperimentintheabsence ofAMP.

Analternativeapproach,andwhichifpossibleispreferable,is toperformadirectselectiononagarplatescontainingAMP lev-elsabovetheMIC.Theadvantageswiththisapproacharethatthe mutantsgenerallyonlycontainonemutation(alleviatingtheneed forgeneticreconstruction,abigadvantageinorganismsthatare geneticallyintractable)andthatalsoresistantmutantswithsevere fitness reductions are recoverable. Furthermore, if this experi-mentisperformedasaclassicalLuria–Delbruckfluctuationtest withmanyindependentcultures(PrantingandAndersson,2010; Prantingetal.,2008;Sunetal.,2009),thisapproachallows deter-minationofmutationratestoresistance.Themainproblemwith theuseofagarplateselectionsisthatitrequireshighamountsof AMPsandthatmanyAMPsarelessactive/inactiveinthecomplex environmentofanagarplate(Dhawanetal.,1997).Furthermore, ifresistanceevolutionrequirestheformationofseveralmutations thedirectselectionmethodisunlikelytoyieldresults,makingthe serialpassagemethodpreferable.

Forbothoftheabovemethods,itispossibletoexploreadditional mutational pathwaysand resistancemechanisms by artificially increasingthemutationrate.Forexample,chemicalmutagenesisor varioustypesofmutatorstrainshavebeenusedtothisend(Hoang etal.,2012;Silvermanetal.,2001).

7.2. Mechanismsofacquiredresistance

The molecularbasisof acquired resistancehasbeen charac-terized ina handful ofbacterialpathogens throughisolation of naturallyoccurringresistantisolates,serialpassageinthepresence ofAMPsanddirectplatingwithAMPs(summarizedinTable1).

7.2.1. Naturallyoccurringacquiredresistance

Recently,forthefirsttimeAMPresistancehasbeenshownto behorizontallytransferablein E.coli. Aplasmidcontaining the mcr-1genewasshowntomediatecolistinresistancebyencoding aphosphoethanolaminemodificationtolipidA(Liuetal.,2016). Themcr-1containingplasmidwasinitiallyisolatedinChinese live-stockanimals,andsinceitsinitialcharacterizationthemcr-1gene hasbeenidentifiedretroactivelyin3of1267humanfecal micro-biomesamplestakenfromChinapriorto2011indicatinganimal tohumangenetransfer(Huetal.,2016).

InS.aureusthesmallcolonyvariant(SCV)phenotype, character-izedbyslowgrowthrate,diminishedtransmembranepotentialand alteredmetabolism,hasbeenassociatedwithresistancetoseveral AMPsincludingLL-37,human␤-defensins2and3,RNase7and lactoferricinB(Glaseretal.,2014;Samuelsenetal.,2005b).The SCVphenotypecanbeinducedbytheintracellularenvironmentat arateof10−3(spontaneousrateislessthan10−7)inS.aureusin ordertoevadethehostimmuneresponse(Vesgaetal.,1996).The lactoferricinBresistantphenotypeofSCVsinS.aureuswasshownto bearesultofreducedmetabolicactivityandnotofreduced trans-membranepotentialoruptake ofthepeptide(Samuelsenetal., 2005b).Glaseretal.(2014)alsoshowedthatclinicallyderivedS. aureusSCVsaswellasahemBauxotrophdisplayedreduced suscep-tibilitytoLL-37andhumandefensins.S.aureushasalsodisplayed constitutiveresistancetoplateletmicrobicidalproteins(PMPs)by increasingmembranefluidityviaexpressionofunsaturatedlipids

(Bayeretal.,2000)orreducingmembranepotential(Yeamanetal., 1998).

7.2.2. Serialpassage

Samuelsenet al.(2005a) described therapid inductionof S. aureusresistancetotheAMPlactoferricinB(activeformofbovine lactoferrin)withover30-foldincreaseinresistancefollowingonly four serial passages in medium containing increasing concen-trationsofthepeptide.High-levellactoferricinBresistancewas maintainedwhenpassagedinmediumwithoutpeptide. Follow-ing30passagesintheabsenceofpeptide,allstrainsdisplayedat leastapartiallystableresistancephenotypewithelevatedMICsup to10-foldhigherthanwildtypealthoughtheexactmechanismof resistancewasnotdefined.LactoferricinBresistantmutantsalso displayedlow-levelcross-resistancetoindolicidinandmagainin. TheseexperimentssuggestthatinS.aureustheacquisitionof low-levelstableresistancecanbeachievedrapidly,withinaslittleas fourpassagesinthepresenceoflactoferricinB,whilehigh-level sta-bleresistancerequirestheacquisitionofmultiplemutationsover time.

Perronetal.(2006)showedthatstableAMPresistancecanarise inE.coliandPseudomonasfluorescensfollowingcontinuous expo-sure(600–700generations)tograduallyincreasingconcentrations oftheAMPpexiganan,amagaininanalogcurrentlyinclinical devel-opment.Pexigananresistancewasobservedwith32-to512-fold and 2- to64-fold increase in resistancetothe parentalstrains forP.fluorescensandE.coli,respectivelywithoutaninvitro fit-nesscost.However,theexact mechanismofresistancewasnot determined.Similarly,itwasshownthatS.aureuscoulddevelop stableAMPresistancefollowingserialpassageinmedium contain-ingincreasingconcentrationsofpexiganan(HabetsandBrockhurst, 2012). These mutants oftendisplayed 10- to50-fold increased resistancetopexigananandcross-resistancetohuman␣-defensin HNP1.Thelargerthelevelofpexigananresistance,thehigherthe invitrofitnesscost.Furtherserialpassageintheabsenceof pexi-gananresultedincompensatorymutationsamelioratingthefitness costwhilemaintainingAMPresistance,butthemolecularbasisof AMPresistancewasnotdefined.Thisstudysuggeststhatthecosts associatedwithpexigananresistanceinS.aureuscanbeeasily com-pensatedresultinginthepersistenceofmutantscapableofevading criticalcomponentsofthehumaninnateimmunesystem.

AcquiredAMPresistanceinS.Typhimuriumhasbeenexamined byfollowingcontinuousexposure(∼500generations)to increas-ing concentrations of LL-37 (a human cathelicidin), CNY100HL (derivativeofhumanC3complement peptideCNY21),orwheat germhistones(WGH) (Loftonetal.,2013).Thisresulted inthe rapid isolation of stable resistant mutants that also displayed 2- to 4-fold increases in resistance to the AMPs not used for selection.AMPresistantisolateswhereshowntocontain muta-tionsintwo-componentregulatorysystems(pmrBorphoP)orin LPSbiosynthesispathways(waaY,alsoreferredtoasrfaY).phoP mutantscontainingasingleaminoacidsubstitutionshowed sim-ilarlevelsofresistancetoallthreeAMPstestedandhadafitness costinvitro,howevernosignificantvirulencedefectwasdetected inamousemodelofsepsis(Loftonetal.,2015).ThephoPmutant alsodisplayedareducedO-antigenchainlength,enhanced toler-anceoflowpHandbile,andincreasedpagPexpressionregardlessof environmentalMg2+_{ionconcentrations,indicatingaconstitutively}

activephenotype.MutationsinpmrB,singleaminoacid substitut-ions,occurredwithoutanimpactonbacterialfitnessinvitroor invivo in micevia intraperitoneal infection, yielding low-level AMPresistance.ActivationofthePmrABtwo-component regula-torysystemsupportstheAra4NmodificationoflipidA,resulting inreducednegativechargeanddiminishedAMPbinding.An inac-tivatingframeshiftmutationinwaaYalsoresultedindividuallyin low-levelAMPresistancewithoutanimpactonbacterialfitness

invitro,howeverinamousemodelofsepsisthismutationresulted inminorattenuation.DisruptionofwaaYresultsinreduced hep-tose phosphorylation of the LPS core, thereby diminishingthe interactionbetweenAMPsandthebacterialmembrane.Although theseindividualmutationseach conferredonlyminor enhance-mentsinAMPresistance,combiningallthreemutationsinasingle mutantfurtherincreasedAMPresistance.pmrBandwaaY muta-tionsappeared togetherin two offour isolatesfollowing serial passagewithLL-37 orCNY100HLindicating a possibleepistatic effectbetweenthesemutationsbothofwhichleadtoareduction inAMPinteractionwithLPS.

7.2.3. Directplating

Roland et al. isolated spontaneous AMP resistant S. Typhimurium by directly plating on agar plates containing various concentrations of polymyxin E (colistin). Characteriza-tion of the resulting AMP-resistant mutants yielded a mutant containing a pmrA mutant allele (pmrA505) that displays con-stitutive activation of PmrAB leading to increased Ara4N and phosphoethanolamine covalent modification of LPS. pmrA505 displaysa1000-foldincreaseinresistancetopolymyxinBanda 2-to4-foldincreaseinresistancetoneutrophilproteins(Roland etal.,1993).Colistin-resistantmutantsinS.Typhimuriumwere alsoobtained bydirect plating onagar containingcolistin at a concentration10or50timestheMICofwildtype(Sunetal.,2009). Spontaneousresistantmutantsdisplayed2-to35-foldincreased colistinresistancewithlittletonofitnesscostinamousemodel of sepsis. 44 independent mutants were characterized and 27 differentmissensemutationswereidentifiedinpmrAorpmrB,the two-componentregulatorysystem.Thecolistin-resistantmutants displayed2-to15-foldincreasesinexpressionofpmrHindicating a dysregulationof pmrAB-regulated genes.Stable colistin resis-tancefollowingdirectplatingwasalsoobservedinA.baumannii followingthecompletelossofLPSduetotheinactivationofone ofthreegenesinvolvedinlipidAbiosynthesis(lpxA,lpxD,orlpxC) (Moffattetal.,2010).

Similarly,aPhoQconditionallyconstitutiveallele(phoQ24or pho-24)displayed∼10-foldincreaseinphoP-activatedgenes(pags) (MillerandMekalanos,1990).WhengrowninlowlevelsofMg2+_,

the phoQ24 allele confers polymyxin resistance as a result of increasedexpressionofPmrAB(GunnandMiller,1996).A simi-larphenotypewasobservedfollowingover-expressionofdltABCD in wild-typeS.aureus, whichincreases d-alanylationofTA and reducescellsurfacenegativecharge(Pescheletal.,1999).

StableresistancetothepigcathelicidinPR-39occurredinS. Typhimuriumfollowing directplatingonPR-39containingagar ataconcentration1.5timestheMICofwildtype(Prantingetal., 2008).TheresultingspontaneousPR-39mutantswerestableand displayed2.5to3.5foldincreasesinPR-39resistanceoverwild type.AlloftheAMPresistantmutantsisolatedcontainedmutations insbmA,aputativeABCtransporter.Theisolatedmutantsdisplayed similarphenotypestothatofasbmAstrainindicatingthataloss offunctiongenotyperesultinginreducedPR-39uptakeintothecell isresponsiblefortheincreasedAMPresistance.Theinactivationof sbmAandtheresultingPR-39resistancedidnotconferafitnesscost invitroorinamousemodelofsepsis.

S.Typhimuriummutantsresistanttoprotamine,an arginine-richAMPisolatedfromsalmonspermcells,wereisolatedbydirect platingonagarcontainingprotamineataconcentrationofaround 3timesthewildtypeMIC(PrantingandAndersson,2010). Spon-taneousmutantswere2-to20-foldmoreresistanttoprotamine thanwildtype,anddisplayedcross-resistancetootherAMPs (col-istin,lactoferricin,human␣-defensinHNP1).Resistancemutations wereidentifiedasdeletions,nonsenseormissenseinhaem biosyn-thesis(hemA,hemB,hemC,hemL)with14differentmutationsin fourstepsofhaembiosynthesiswereobservedindicatingthatthe

Table2

RatesandfitnesseffectsofmutationscausingresistancetoantibioticsorAMPs.

Drugclass Species Mutationratea _Fitnessb _MICincreaseb _{Mutation Reference} AMP

PR-39 S.enterica 0.4×10−6 _{Nochange 4-fold sbmA Prantingetal.(2008)} LL-37 S.enterica ND Nochange/reduced 6-fold waaY,pmrB,phoP Loftonetal.(2013) CNY-100HL S.enterica ND Nochange/reduced 2-to3-fold waaY,pmrB,phoP Loftonetal.(2013) WGHc _{S.enterica ND Nochange/reduced >15-fold waaY,pmrB,phoP Loftonetal.(2013)} Pexiganan P.fluorescens ND Weaklyreduced 2-to32-fold ND Perronetal.(2006)

Pexiganan E.coli ND ND 2-to-32-fold ND Perronetal.(2006)

Colistin S.enterica 0.6×10−6 _{Weaklyreduced 2-to35-fold pmrAB Sunetal.(2009)}

Protamine S.enterica 2.3×10−7 _{Stronglyreduced 2-to20-fold hemACL/cydC PrantingandAndersson(2010)} Antibiotics

Rifampicin E.coli,S.enterica 5×10−9 _{Reduced >200-fold rpoB Brandisetal.(2015)} Ciprofloxacin E.coli <1×10−9 _{Nochange 16-fold gyrA ChinandNeu(1987)} Fusidicacid S.enterica ND Reduced 8-to500-fold fusA Nagaevetal.(2001) Mecillinam E.coli 8×10−8_to2×10−5 _{Nochange/reduced 2-to768-fold >38genes Thulinetal.(2015)} Mupirocin S.enterica 1×10−9 _{Reduced >64-fold ileS Paulanderetal.(2007)} a_{Percellpergeneration.}

b_{Ascomparedtosusceptibleparentalstrain.} c _{Wheatgermhistones.}

lackofanendproductinthispathwayisresponsibleforprotamine resistance.Mutationsinthehemgenesoftenresultindefectsin electrontransportandrespiration,whichresultsinslowgrowth, reducedmembranepotentialandthusloweredaffinityfor bind-inganduptakeofAMPs.Allprotamine-resistantmutantsdisplayed largedefectsinbacterialfitnesswithrelativegrowthratesranging from0.25to0.85ascomparedtothewildtype.Thetremendous fit-nesscostresultedininstabilityoftheprotamineresistanceinsome mutants,particularlyhemCandhemAmutants,andcompensatory evolutionbyserialpassageintheabsenceofprotamineresulted inthelossofprotamineresistance.Threemutantsdisplayed non-sensemutationsincydC,acysteine/glutathioneABCtransporter thoughttobeinvolved inmaintainingcellularredoxbalance.A cydCmutantwouldthereforedisplayreducedmembranepotential aswell,resultinginAMPresistance.

Spontaneoushigh-levelbacteriocinresistance(over2000-fold) occurred in L. monocytogenes following direct plating due to increasedexpressionoftwo putative␤-glucoside-specific phos-photransferasesystemsleadingtoreducedinteractionwithAMPs (Gravesenetal.,2002).

7.3. WhatdeterminestheriskofacquiringresistancetoAMPs? As for any type of drug resistance, the rate by which AMP resistancewould emergeand spread in a populationof bacte-riaina naturalsettingis determinedbyacomplexinterplayof severaldifferentfactors,includingthemutationsupplyrate;the fitnessoftheresistantmutantatdifferentAMPconcentrationsand thestrengthoftheselectivepressure(AMPlevels)(Hughesand Andersson,2015).Also,theoccurrenceofcompensatoryevolution wherefitnesscostsmightbereducedbyadditionalmutations,and epistaticinteractionsinvolvingresistancegenesanddrugs(Hughes andAndersson,2015)thatmaygeneratecross-resistanceand/or collateralsensitivity,complicatesthepictureevenfurther.In addi-tion,numerousotherepidemiologicalfactors(e.g.hostpopulation structure,density,immunity,etc.)willaffecthowaputativeAMP resistantclonewouldspreadinapopulation.Severaloftheabove parametershave beendefinedfor antibioticandantiviral resis-tancesbutwithregardtoAMPresistancefewhavebeenexplored inanydetail(summarizedinTable2).

7.3.1. Mutationsupplyrate

Thegenerationofgeneticheterogeneity,inthiscasemutants thatareresistant/lesssusceptibletoAMPs,inabacterial popula-tionislargelydeterminedbythemutationsupplyrate,theproduct

ofthepopulationsizeandratesofmutation(orHGT)toAMP resis-tance.MutationratestoAMPresistancehavebeenmeasuredin severalcasesandtheyaretypicallyintherangeof10−7 to10−6 pergenerationpercell(PrantingandAndersson,2010;Pranting etal.,2008;Sunetal.,2009),makingthemcomparableto antibi-otics(10−5to10−10pergenerationpercell).Populationsizeswithin aninfectedhumanaregenerallypoorlyknownbutinafewcases wehavenumbersforthetotalbacterialpopulationsize.For exam-ple,forurinarytractinfectionsthetotalpopulationsizesinafull bladderof300mlisintherangeof106_–1010_{perbladderandfor}

pulmonarytuberculosisinfectionsthenumberofbacteriaperml ofsputumcanbe105_–106_{andevenhigherinthelungs(Hughes}

andAndersson,2015).Similarly,formeningitisandbronchitisthe bacterialcountscanbeashighas109_{perml,andoverallthesedata}

suggestthatthetotalbacterialpopulationsizesinaninfected indi-vidualtypicallyissohighthatpre-existingAMPresistantmutants areexpectedtobepresent(i.e.populationsizexmutationrate>1) whentreatmentisinitiated.

7.3.2. Mutantfitness

Thefitnessofa drug-resistantpathogen(intheabsenceand presenceof drug)is akeyparameterin determining evolution-arysuccess—i.e.theprobabilitythatitwillemerge,fix,transmit andremain—withinahostpopulation,ashasbeendemonstrated indifferentstudiesforantibioticresistantbacteria(Anderssonand Hughes,2011).WithregardtofitnessmeasurementsofAMP resis-tantmutantsdataisscarceandonlyafewstudieshaveexploredthis insomedetail(Loftonetal.,2013,2015;PrantingandAndersson, 2010;Prantingetal.,2008).Asisgenerallyobservedfor antibi-oticresistantmutants,AMPresistantvariantsshowsometypeof reductioninfitness,eventhoughcasesareknownwhenwiththe assaysused,theresistantmutantsappearedasfitasthecongenic susceptiblestrain(Prantingetal.,2008).Forexample,PR-39 resis-tantS.Typhimurium mutantsshowednoapparentreductionin fitnessasmeasuredbygrowthrate,starvationsurvivalandgrowth inmice.Inothercases,S.Typhimuriummutantsresistanttocolistin, LL-37,CNY100HLand wheatgermhistones,fitnesswas signifi-cantlyreducedbutthereductionswereforsomemutantsrelatively minor,suggestingthatthesemutantsmightpersistquitewellina hostpopulation,evenwithoutaselectivepressure(Loftonetal., 2013,2015).YetforprotamineresistantS.Typhimuriummutants thereductionsweresoseverethatitisunlikelythattheycould remaininthehostpopulationwithoutastrongselection(Pranting andAndersson,2010).

7.3.3. Selectionstrength

Bacterialpathogensare,asaresultoftreatments,exposedtoa widerangeofantibioticconcentrations,generatingavariablerange ofselectivepressures.Asshownbyrecentstudies,notonly concen-trationswellaboveMIC(lethalselection)canselectforresistance butmutantsmayalsoemergeduringnon-lethalselectionfarbelow theMIC,wheretheirrateofenrichmentisdeterminedbythefitness differencebetweensusceptibleandresistantcells(Gullbergetal., 2011).Sinceonlyafewpeptideantibiotics(e.g.colistin)havehad anywidespreadclinicaluse,andasofyetnoinvitrostudieshave beenperformedtostudyselectionatdifferentAMPconcentrations, itisdifficulttoassesstheimpactofselectionstrengthonwhich mutanttypesemerge.ItisnotablethatforAMPs,selectionfor resis-tantmutantscouldoccurattwolevels:(i)duetotherapeuticuse ofAMPs(whichmightgenerateverystrongselectivepressures), and(ii)duetocontinuousselectionprovidedbyAMPsproducedin thehumanbodyandtowhichbacterialpathogenscanbeexposed. Atpresent, we know that the pressure exerted by therapeutic useisweak (simplybecause fewAMPs are usedclinically) but thepressureexertedbyhuman-specificAMPsislargelyunknown. However,onerecentstudyshowedthatforawaaYmutant resis-tanttoseveralAMPsthedosageofLL-37(0.1–1.5mg/L)requiredto selectforandmaintainthemutantinthepopulationissimilarto theconcentrationsofAMPsfoundin,forexample,secretionsnear hostepithelialcells(Loftonetal.,2013).Thisfindingimpliesthat naturallyproducedAMPswillexertsignificantselectivepressures. However,thedeleteriousnatureofthisparticularwaaYmutation (reducedtoleranceofserum,acidicconditions,andsurvival ina mousemodelofsepsis)likelyexplainsitsabsenceinpathogenic isolates(Loftonetal.,2015).

7.3.4. Compensatoryevolution

Mostmutationsthatoccurinanyorganismaredeleteriousbut compensatoryevolution, wherethecosts arereduced or elimi-natedbyadditionalmutations,mightincreasetheprobabilityof maintenance.Anumberofexperimentalandclinicalstudieshave shownthatthefitnesscostofantibioticresistancecanbeefficiently andrapidlyreducedbycompensatorymutationsandthat compen-satedmutantscanrapidlysweeparesistantpopulationofbacteria (AnderssonandHughes,2010).Towhatextentthisprocesswill influenceevolutionwilldependonthemutationratesof compen-sation,thefitnessofcompensatedmutantsandthepopulationsize oftheresistantmutant.Oneofthefewstudiesofcompensatory evolutioninvolvingAMPswasmadeusingpexigananandS.aureus buttheactualmutationsselectedwerenotdetermined(Habetsand Brockhurst,2012).

7.3.5. Epistaticinteractions

Recentdatafrombacteriashowthatepistasisbetween resis-tancemutationsispervasiveandthatitmighthaveanimpacton resistanceevolution(HughesandAndersson,2015).Forexample, combinationsoffitness-reducingchromosomalresistance muta-tionsand/orresistanceplasmidscaneithergeneratenoepistasis, positive epistasis (a double mutant has a higher fitness than expectedfromthesumofcostsofindividualmutations),or neg-ativeepistasis(adoublemutanthasalowerfitnessthanexpected fromthesumofcostsofindividualmutations).Allofthesetypesof epistasishavebeenobservedinexperimentalstudies.Atpresent, thistypeofepistasishasnotbeensystematicallystudiedforAMP resistancemutations.

Another type of epistasis involves the phenomenon where resistancetoonedrugalterssusceptibilitytootherdrugs.Thus, resistancetoonedrugclass mighteither increaseresistanceto another (cross-resistance), or result in increased susceptibility (Macvanin and Hughes, 2005), so-called collateral sensitivity. Recent large-scale studies have shown that both are common

phenomenaforantibioticresistanceandtheyimplythatcollateral sensitivitycanpotentiallybeappliedtoclinicalsituationstoreduce therate of resistanceevolution (Imamovicand Sommer, 2013; Muncket al.,2014).Thus, treatmentwould combinetwo drugs whereresistanceevolutiontoonedrugincreasesthesusceptibility to the other. Again, for AMP resistant mutants the extent of collateralsensitivityfordifferentAMPsand/orantibioticshasnot beenextensively explored,even thougha few cases havebeen reported(Bertietal.,2015;Garcia-Quintanillaetal.,2015).

8. TherapeuticuseofAMPs

Widespreadantibiotic resistancehasledtothe urgentneed fordiscoveryofnovelantimicrobialstotreatbacterialinfections. Thebroad-spectrumbactericidalactivityofAMPsmakesthema promisingcandidatefortherapeuticuse.Thereareonlyahandful ofAMPscurrentlyinclinicaluse,andofthese,polymyxinBand colistin(polymyxinE)arethemostcharacterizedinaclinical set-ting(FalagasandKasiakou,2005;Landmanetal.,2008;Zavascki et al., 2007). The polymyxins are produced by the bacterium Bacillus polymyxa and were introducedinto clinical practice in the1950s.Thepolymyxinshavebeenadministeredintravenously totreatmulti-drugresistantinfectionscausedbyGram-negative pathogens, asan aerosol totreat bronchopulmonaryinfections, andintraventricularlytotreatmeningitis(Landmanetal.,2008).In additiontoitsuseforsystemictreatmentofcomplicatedinfections, polymyxinBisusedintopicalformulationstotreatandprevent skin,eye and earinfections. The highincidence of nephrotoxi-cityandneurotoxicityassociatedwithintravenousadministration hasresultedinlimitedusageandstudyofthesedrugsoverthe past50years(FalagasandKasiakou,2006).Regardless,the emer-genceofmulti-drugresistantpathogenshasledtoanincreasein thetherapeuticuseofpolymyxinsdespitelimited pharmacologi-calandsusceptibilitystudies.Conflictingclinicalbreakpointshave emerged duetothesensitivity ofpolymyxins, likeotherAMPs, toenvironmentalconditions,resultingindiscrepancies between invitroandinvivoefficacy.Thesediscrepancies havehampered theclinicaldevelopmentofAMPstodate(FalagasandKasiakou, 2005;Landmanetal.,2008;Zavasckietal.,2007).

Currently, there aremany AMPs in clinical developmentfor the treatment of various bacterial pathogens, but the major-ity of these are intended for topical use only (Table 3). This is likely a direct result of the toxicity observed following sys-temic administration of polymyxinB and colistin (Falagas and Kasiakou,2006).The onlyAMPin clinicaltrialsfor intravenous administrationis human-derived Lactoferrin 1–11(hLF1–11)to treat life-threatening infections that occur in stem cell trans-plantationpatients.AlthoughastudyhasshownthathLF1–11is well-toleratedinsingleandmultipledosesashighas5mg follow-ingintravenousadministration(Veldenetal.,2009),itisunclear whetherthisdosewillbesufficienttotreataninfectionsince col-istinhasbeenadministereduptoatotalof720mgperdayin3 divideddoses(Michalopoulosetal.,2005).

AsidefromhLF1–11,thereisanAMP-derivativefromthehuman innateimmunesysteminphaseIIclinicaltrials.OP-145,derived fromthehumancathelicidinLL-37,andformulatedineardrops, isbeingtestedtotreatchronicbacterialearinfections(Malanovic etal.,2015).OtherAMPsderivedfromnaturallyoccurringanimal immunecomponentsincludePexiganan,IsegananandOmiganan. Pexigananisananalogofmagainin,anAMPoriginallyisolatedfrom theskinoftheAfricanclawedfrog,thathasgonethroughclinical trialstwicetotreatdiabeticfootulcersandiscurrentlyinphase III(LambandWiseman,1998;Lipskyetal.,2008).Another animal-derivedAMPisasyntheticprotegrinoriginallyisolatedfrompig leucocytestermedIseganandevelopedasarinseforthepotential

Table3

Antimicrobialpeptidesinclinicaldevelopment.

Peptide AMPsource(host) Status Administration Indication Company OP-145 LL-37(human) PhaseI/II Eardrops Chronicbacterialearinfection OctoPlusInc. hLF1–11(Lactoferrin) Lactoferrin1–11(human) Notspecified Intravenous Neutropenicstemcell

transplantationpatients

AM-PharmaB.V.

Pexiganan(MSI-78) Magainin(frog) PhaseIII Topicalcream Diabeticfootinfection DipexiumPharmaceuticals,Inc. PhaseIII Topicalcream Diabeticfootulcers MacroChemCorporation Iseganan(IB-367) Protegrin-1(porcineleukocytes) PhaseIII Mouthwash Preventionof

chemotherapy-induced mucositis

NationalCancerInstitute(NCI)

PhaseII/III Mouthwash Preventionof ventilator-associated pneumonia

IntraBioticsPharmaceuticals

Omiganan(MBI226, CLS001)

Indolicidin(bovineneutrophils) PhaseIII Topicalcream Topicalskinantisepsis, preventionofcatheter infections

Mallinckrodt

PhaseIII Topicalcream Rosacea CutaneaLifeSciences,Inc. PhaseII Topicalcream Usualtypevulvar

intraepithelialneoplasia(uVIN)

CutaneaLifeSciences,Inc. PhaseII Topicalcream Moderatetosevere

inflammatoryacnevulgaris

CutaneaLifeSciences,Inc. PhaseII Topicalcream Mildtomoderateatopic

dermatitis

CutaneaLifeSciences,Inc. Lytixar(LTX-109) Syntheticantimicrobial

peptidomimetic

PhaseII Topicalcream UncomplicatedGram-positive skininfections

LytixBiopharmaAS PhaseI/IIa Nasal NasalcarriersofStaphylococcus

aureus

LytixBiopharmaAS C16G2 Syntheticspecificallytargeted

antimicrobialpeptide

PhaseII Mouthwash Preventtoothdecaycausedby Streptococcusmutans

C3Jian,Inc.

treatmentofmucositis, asanaerosol totreatrespiratory

infec-tions,andasageltotreatpneumonia(Toney,2002).Omiganan

isaderivativeofindolicidin,whichwasisolatedfrombovine neu-trophilsandisinphaseII/IIIclinicaltrialstotreatskininfections. AsidefromnaturallyoccurringAMPs,therearealsoseveralpurely syntheticAMPsinclinicaltrials(Meloetal.,2006).Thesynthetic AMPLTX-109isinphaseIItrialsforthetreatmentofuncomplicated Gram-positiveskininfectionsaswellastopreventnasalcarriage ofS.aureus(Nilssonetal.,2015).C16G2isanothersyntheticAMP inphaseIItrialsindicatedforthepreventionoftoothdecaycaused byStreptococcusmutans(Kaplanetal.,2011).

Asidefromthedirectadministrationofantimicrobialpeptides, there are several ongoing clinical studiesaimed at modulating theexpressionofAMPsbythebodytoboosttheinnateimmune response(Table4).Manyrecentfindingshaveelucidatedtheeffects ofdietarynutrientsonmodulatingtheinnateimmuneresponse, includingtheexpressionof AMPs,inordertoassessthe poten-tialofutilizing suchdietarysupplementsastherapeutics. Since many microbial infectionsare associated withinhibition ofthe innateimmuneresponse(DiacovichandGorvel,2010),itisthought thatenhancingtheimmunesystemwillbeaneffectivestrategy tocombatinfection.VitaminD3hasbeenshowntodirectly reg-ulateexpressionofthehumanantimicrobialpeptidesLL-37and ␤-defensin2(Wanget al.,2004;Weber etal.,2005).Thereare

manyongoingclinicaltrialsaimedatelucidatingtheutilityof vita-minDsupplementsincombatingbacterialdiseasesrangingfrom uncomplicated diarrheal diseasesto life-threateningpulmonary tuberculosis.Othercompoundsthathavebeenshowntoinduce expressionof hostantimicrobialpeptidesinclude pimecrolimus appliedtopicallyforthetreatmentofdermatitis(eczema)(Buchau etal.,2008)andphenylbutyrateadministeredorallytotreatvarious bacterialinfections(Steinmannetal.,2009).

TherearealsoAMPsusedintheclinicwithindicationsother than as antimicrobials. For example, protamine hasbeen used clinicallyasa heparinantagonist topreventbloodclotting dur-ingcardiacsurgery,vascularsurgery,andinterventionalradiology. Recently,protaminehasbeenproposedasadrugtotreatobesity duetoitsabilitytoinhibitlipaseactivityanddecreasethe absorp-tionofdietaryfat(Duarte-Vazquezetal.,2009).Routineclinical useofprotaminewillinevitablyresultinbacterialexposureand subsequentdevelopmentofresistance,regardlessofwhetherits intendeduseisasanantimicrobialornot.PrantingandAndersson

(2010)have already shown in S.Typhimurium that

protamine-resistant mutantsdisplay cross-resistance toa variety of other AMPs(includingcomponentsofthehumaninnateimmune sys-tem)aswellastoclinicallyprescribedantibiotics(gentamicinand streptomycin),whichwouldmakeinfectionsmorepersistentand difficulttotreat.

Table4

Selectedantimicrobialpeptidemodulatingcompoundsinclinicaldevelopment.

Intervention Administration LocationoftargetedAMPs Status Indication Studysponsor VitaminD3 Oral Skinandsaliva PhaseII Atopicdermatitisand

psoriasis

NationalInstituteofAllergy andInfectiousDiseases(NIAID) VitaminD3 Oral Lung Notspecified Activetuberculosis AtlantaVAMedicalCenter Phenylbutyrateand

vitaminD3

Oral Lung PhaseII Pulmonarytuberculosis InternationalCentrefor DiarrhoealDiseaseResearch, Bangladesh

Lung,blood PhaseII Pulmonarytuberculosis KarolinskaInstitutet

Blood PhaseII HIV KarolinskaInstitutet

Sodiumbutyrate Enema Colon PhaseII Shigellosis InternationalCentrefor DiarrhoealDiseaseResearch, Bangladesh

Fig.3.Potentialconsequenceoftherapeuticadministrationofantimicrobial pep-tides.RoutineclinicaladministrationofAMPstotreatbacterialinfectionswillselect forresistantmutantsmorecapableofevadingtheinnateimmunesystemand causingpersistentinfections.Thusalternateantimicrobialtherapieswillbe imple-mentedresultinginadditionalantibioticuseandselectionformulti-drugresistant pathogens.

TheindiscriminatemodulationofhostAMPproductionisnot

withoutitsownconsequences.Inimmunocompromised

individ-ualsorthosewithinherentdeficienciesinAMPproduction,the

inductionofendogenousAMPsmaybebeneficialincircumventing

bacterialinfections.However,artificiallystimulatingthe

produc-tionofAMPswillincreasetheexposureofbacterialpathogensto

theseagents.TheexactamountofAMPexpressionwillbedifficult

tocontrol,andsinceexposuretohigh-levelsofAMPsoften

corre-spondwithtoxicsideeffectstherewilllikelybeprolongedbacterial

exposuretosub-inhibitoryconcentrationsofAMPs.The

ramifica-tionsofsuchexposureonthedevelopmentofAMPresistanceand

bacterialpathogenesisarenotyetunderstood.

TheclinicalutilityofAMPsisalsoplaguedbyissueswiththe

inherentinstabilityof peptides,theirsusceptibilityto

proteoly-ticdegradation,thecostofpeptideproductionandpurification,

and thehighassociation of toxic side effects which have been

reviewedpreviously (Bradshaw,2003; Marret al.,2006;Vaara,

2009).AnotherhindrancetothedevelopmentofAMPs as ther-apeuticshasbeen a general lackof mechanistic understanding manifestedinthefactthatdespitehighinvitroactivity,someAMPs displaylittletonoactivitywhentestedunderphysiological rele-vantconditions(Marretal.,2006).

Despitetheseobstacles,clinicaldevelopmentofAMPstotreat bacterialinfectionsisontherise.Althoughsomeofthesepeptides arenotnaturallyderived,severalstudieshavealreadyshownthat AMP-resistantbacterialmutantsdisplaybroadcross-resistanceto avarietyofAMPswithdifferentstructuresandmodesofaction. Therefore,routineclinicaladministrationofAMPstotreatbacterial infectionsmayselectforresistantmutantsmorecapableofevading theinnateimmunesystem.Immuneevasionwilllikelyenhance bacterialpersistenceandthustheneedforalternateantimicrobial therapiesultimatelyresultinginfurtherantibioticuseandselection formulti-drugresistantpathogens(Fig.3).

9. Concludingremarks

Itisclearthatatpresentwehavearelativelylimited under-standingof how bacteria would acquireresistance todifferent AMPs,weretheytobeusedtherapeutically.Thus,fewstudieshave examinedhowrapidlyresistanceappearsandhowtheresistance

mechanismswouldimpactfitnessandvirulenceofthemutants. However,fromthelimitednumberofstudiesdonewecanalso concludethat(i)itisnotdifficultforbacteriatoacquireAMP resis-tancebymutation,(ii)anumberofmechanismsexistbywhichthis canoccur,(iii)someofthesemechanismsarenotassociatedwith anyextensivelossoffitness,and(iv)oftentheseresistance muta-tionsconfercross-resistancetohumanAMPsthatarepartofinnate immunity.

Thelatterisapointofspecialconcernsinceitmightimplythat therapeuticuseofAMPscouldselectformutantsthatbecome resis-tanttokeycomponentsofourownimmunesystem.Thispotential problemhasbeenpointedoutbyseveralresearchersandwithmore databeingavailableonacquiredresistancemechanismsand cross-resistanceitseemsclearthatthis issomethingthatneed tobe carefullystudiedandmonitoredwhenAMPsareincreasinglyused inclinicalsettings.

Acknowledgement

ThisworkwassupportedbygrantsfromtheSwedishResearch CounciltoDIAandDH.

References

Abachin,E.,Poyart,C.,Pellegrini,E.,Milohanic,E.,Fiedler,F.,Berche,P.,Trieu-Cuot, P.,2002.Formationofd-alanyl-lipoteichoicacidisrequiredforadhesionand virulenceofListeriamonocytogenes.Mol.Microbiol.43,1–14.

Adams,M.D.,Nickel,G.C.,Bajaksouzian,S.,Lavender,H.,Murthy,A.R.,Jacobs,M.R., Bonomo,R.A.,2009.ResistancetocolistininAcinetobacterbaumannii associ-atedwithmutationsinthePmrABtwo-componentsystem.Antimicrob.Agents Chemother.53,3628–3634.

Agerberth, B.,Gunne,H.,Odeberg, J.,Kogner, P., Boman,H.G., Gudmundsson, G.H.,1995.FALL-39,aputativehumanpeptideantibiotic,iscysteine-freeand expressedinbonemarrowandtestis.Proc.Natl.Acad.Sci.U.S.A.92,195–199. Anaya-Lopez,J.L.,Lopez-Meza,J.E.,Ochoa-Zarzosa,A.,2013.Bacterialresistanceto

cationicantimicrobialpeptides.Crit.Rev.Microbiol.39,180–195.

Andersson,D.I.,Hughes,D.,2010.Antibioticresistanceanditscost:isitpossibleto reverseresistance?Nat.Rev.Microbiol.8,260–271.

Andersson,D.I.,Hughes,D.,2011.Persistenceofantibioticresistanceinbacterial populations.FEMSMicrobiol.Rev.

Baltz,R.H.,2006.Molecularengineeringapproachestopeptide,polyketideandother antibiotics.Nat.Biotechnol.24,1533–1540.

Bangalore,N.,Travis,J.,Onunka,V.C.,Pohl,J.,Shafer,W.M.,1990.Identificationofthe primaryantimicrobialdomainsinhumanneutrophilcathepsinG.J.Biol.Chem. 265,13584–13588.

Batoni,G.,Maisetta,G.,Brancatisano,F.L.,Esin,S.,Campa,M.,2011.Useof antimi-crobialpeptidesagainstmicrobialbiofilms:advantagesandlimits.Curr.Med. Chem.18,256–279.

Bauer,M.E.,Shafer,W.M.,2015.Ontheinvivosignificanceofbacterialresistanceto antimicrobialpeptides.Biochim.Biophys.Acta1848,3101–3111.

Bayer,A.S.,Prasad,R.,Chandra,J.,Koul,A.,Smriti,M.,Varma,A.,Skurray,R.A.,Firth, N.,Brown,M.H.,Koo,S.P.,Yeaman,M.R.,2000.Invitroresistanceof Staphylococ-cusaureustothrombin-inducedplateletmicrobicidalproteinisassociatedwith alterationsincytoplasmicmembranefluidity.Infect.Immun.68,3548–3553. Bengoechea,J.A.,Skurnik,M.,2000.Temperature-regulatedeffluxpump/potassium

antiportersystemmediatesresistancetocationicantimicrobialpeptidesin Yersinia.Mol.Microbiol.37,67–80.

Bensch,K.W.,Raida,M.,Magert,H.J.,Schulz-Knappe,P.,Forssmann,W.G.,1995. hBD-1:anovelbeta-defensinfromhumanplasma.FEBSLett.368,331–335. Berti,A.D.,Baines,S.L.,Howden,B.P.,Sakoulas,G.,Nizet,V.,Proctor,R.A.,Rose,W.E.,

2015.Heterogeneityofgeneticpathwaystowarddaptomycin nonsusceptibil-ityinStaphylococcusaureusdeterminedbyadjunctiveantibiotics.Antimicrob. AgentsChemother.59,2799–2806.

Bradshaw,J.,2003.Cationicantimicrobialpeptides:issuesforpotentialclinicaluse. BioDrugs17,233–240.

Brandis,G.,Pietsch,F.,Alemayehu,R.,Hughes,D.,2015.Comprehensive pheno-typiccharacterizationofrifampicinresistancemutationsinSalmonellaprovides insightintotheevolutionofresistanceinMycobacteriumtuberculosis.J. Antimi-crob.Chemother.70,680–685.

Brogden,K.A.,2005.Antimicrobialpeptides:poreformersormetabolicinhibitorsin bacteria?Nat.Rev.Microbiol.3,238–250.

Brown,S.,SantaMariaJr.,J.P.,Walker,S.,2013.Wallteichoicacidsofgram-positive bacteria.Annu.Rev.Microbiol.67,313–336.

Buchau,A.S.,Schauber,J.,Hultsch,T.,Stuetz,A.,Gallo,R.L.,2008.Pimecrolimus enhancesTLR2/6-inducedexpressionofantimicrobialpeptidesinkeratinocytes. J.Invest.Dermatol.128,2646–2654.

Cao,M.,Helmann,J.D.,2004.TheBacillussubtilisextracytoplasmic-functionsigmaX factorregulatesmodificationofthecellenvelopeandresistancetocationic antimicrobialpeptides.J.Bacteriol.186,1136–1146.