Associations
between
properties
linked
with
persistence
in
a
collection
of
Staphylococcus
aureus
isolates
from
bovine
mastitis
Marjorie
Bardiau
a,*
,
Johann
Detilleux
b,
Fre´de´ric
Farnir
c,
Jacques
G.
Mainil
a,
Isabelle
Ote
a,1aBacteriology,DepartmentofInfectiousDiseases,FacultyofVeterinaryMedicine,UniversityofLie`ge,Sart-Tilman,Baˆt.43a,B-4000Lie`ge,
Belgium
b
QuantitativeGeneticsGroup,DepartmentofAnimalProduction,FacultyofVeterinaryMedicine,UniversityofLie`ge,Sart-Tilman,Baˆt. 43a,B-4000Lie`ge,Belgium
c
Biostatistics,BioinformaticsandAnimalSelection,DepartmentofAnimalProduction,FacultyofVeterinaryMedicine,UniversityofLie`ge, Sart-Tilman,Baˆt.43a,B-4000Lie`ge,Belgium
1. Introduction
Staphylococcusaureus(S. aureus) is one of themost commonaetiological agentsof contagious bovine mas-titis(Watts,1988).AlthoughS.aureusstrainscancause
acuteandclinicalmastitiswithmacroscopicalterationof the milk, such bacterial infections can evolvetowards chronic and subclinical mastitis, without macroscopic alteration of the milk but with high cell counts and persistence of the bacteria in the mammary gland. Staphylococcal subclinical mastitis account for up to 30%ofallcasesofbovinemastitis(Halasaetal.,2007). They representa seriouseconomicalproblemfor dairy farmers withreductionofmilk productionandquality, prolonged antibiotictreatments andprematureculling. The identification of the staphylococcal properties
ARTICLE INFO Articlehistory: Received6May2013
Receivedinrevisedform3December2013 Accepted13December2013 Keywords: S.aureus Bovinemastitis Biofilm Capsule Intracellularsurvival Persistence ABSTRACT
Staphylococcus aureus is recognized worldwide asa pathogen causing many serious diseasesinhumansandanimals,andisoneofthemostimportantetiologicalagentsof clinicalandsubclinicalbovinemastitis.Theaimofthepresentstudywastoinvestigate andcorrelateproperties,thatmaybe associatedwithpersistentmastitis,ofS.aureus strainsisolatedfrommilkof cowssufferingfrommastitis: (i)expressionof capsular antigens(CP5orCP8)byspecificELISA;(ii)intracellularsurvivalbyinvasionofMAC-T cells;and(iii)biofilmproductionbyspectrophotometryanalysisaftergrowthinTSBglc.
Theresultsshowedthat(i)theproportionofstrainsexpressingcapsularantigenwas higherincap8-thanincap5-positiveisolates;(ii)acorrelationwasobservedbetweenthe capsularprofileandtheintracellularsurvivalaswellasthebiofilmproduction;and(iii) thecapsularprofile,biofilmproductionandintracellularsurvivalwereassociatedwith onlytwoagr-groups.Statisticalandclusteringanalysisallowedustoestablishdifferent profilesthatcouldbeassociatedwithinvivopersistence.Indeed,isolatesbelongingtoagr groupII,expressingthecapsularantigenCP8andshowinglowintracellularsurvivalare probablybetteradaptedtoanextracellularniche.Conversely,isolatesbelongingtoagr groupIthatdonotexpressanycapsularantigen(CP5orCP8)butshowhighintracellular survivalareprobablybetteradaptedtoanintracellularniche.
ß2013ElsevierB.V.Allrightsreserved.
*Correspondingauthor.Tel.:+3243669522;fax:+3243664261. E-mailaddresses:mbardiau@ulg.ac.be,mbardiau@gmail.com
(M.Bardiau).
1
Currentaddress:CorisBioConcept,ScienceParkCREALYS,rueJean Sonet4a,B-5032Gembloux,Belgium.
ContentslistsavailableatScienceDirect
Veterinary
Microbiology
j o urn a l hom e pa ge : ww w. e l se v i e r. c om / l oca t e / v e t mi c
0378-1135/$–seefrontmatterß2013ElsevierB.V.Allrightsreserved.
associatedwithlong-lastinginfectionsisthereforeofthe upmostimportance.
For instance,biofilm formationimpairs theaction of both the host immune system and the antimicrobial agents,andthereforerepresentsoneofthemostimportant survival mechanisms of bacteria persistently colonizing theextracellular niche (Costerton et al.,1999; Melchior etal.,2006).Nevertheless,severalfactors,environmental aswellasgenetic,influencebiofilmformationbyS.aureus. S.aureusisalsoknowntopenetrate,survive,andeven multiplywithinalargevarietyofeukaryoticcells,suchas theepithelialcellsofthemammaryglandortheimmune cells(Almeidaetal.,1996;KerroDegoetal.,2002).This survivalwithintheintracellularnicheprotectsthebacteria fromtheantibioticscommonlyusedinmastitistreatment, butalsoenablesthemtopersistinthehostforalongtime withoutcausingapparentinflammation(Boulangeretal., 2003;GarzoniandKelley,2009).
Theroleofthestaphylococcalcapsulepolysaccharides (CP) in this internalization has been described. More precisely, the absence of CP expression was shown to enhance adherence (Pohlmann-Dietze et al., 2000) and invasion (Tuchscherr et al., 2005) of endothelial and epithelialcellsbyS.aureus.Furthermore,capsule-negative S.aureusbacteriawereshowntoinducechronic experi-mentalmastitisinmice,suggestingthatthelossofcapsule expressionmayenhancethepersistenceofS.aureusinthe mammaryglands(Tuchscherretal.,2005).
The aimof the present work was thus to determine whetherassociationofspecificinvitrotestscouldestablisha profileofinvivopersistence.Wethereforeinvestigatedthe correlationoffeaturesdescribedtobeassociatedwith long-lastinginfections,respectivelycapsularantigenidentityand expression,biofilmformation,andintracellularsurvival,on acollectionofS.aureusstrainsisolatedfromcasesofbovine mastitisinBelgium.Thesephenotypicfeatureswere also supplementedbyphylogeneticanalyses,suchasagr-typing and PFGE. Statistical and clustering analyses were per-formed to associate the different tests together and to evaluatethehomogeneityandthedistributionofprofiles.
2. Materialsandmethods 2.1. Bacterialisolates
A total of 229 bovine S. aureus isolates, collected between 2005 and2008 frombovine mastitis casesin differentareas of Wallonia(southern partof Belgium), wereassayedinthisstudy.Allisolateshavealreadybeen virulotyped(Oteetal.,2011).Tworeferencestrainsfrom bovinemastitiswereincludedinthiscollection:S.aureus ATCC29740(N305)andS.aureusATCC31885(NL6). 2.2. Capsularserotyping
Theprevalenceofthecapsule-encodinggenes(cap5and cap8)amongtheisolatesofthecollectionhasalreadybeen determinedbyPCRandareshowninTable1(Oteetal., 2011).However,capsularserotypingwasperformedusing ELISAtestthroughthedetectionofcapsular polysacchar-idestypes5and8(CP5andCP8)frombacteriagrownon Columbiaagar.Briefly,strainswereculturedfromfrozen stocks onto Columbia sheep blood agar plates and incubatedat378Covernight.Bacteriawerethencollected fromagarplatesbywashingtheplateswithsterilePBSand adjustedto1010CFU/mlbyopticaldensity(OD)measure.
The capsular serotype of 109CFU of each strain was determined by enzyme-linked immunosorbent assay (ELISA)in96-wellmicroplates,usingspecificmonoclonal and polyclonal antibodies (kindly provided by GSK Biologicals, Belgium) and horseradish peroxidase-based quantification. The OD of each well was measured at 490nmusinga microplatereader(Bio-Rad).Theresults werecollectedfromatleasttwoindependentexperiments inwhichthecapsuleproductionofeachtestedculturewas evaluatedintriplicates.ODvalueswerecomparedtothose obtainedwithS.aureusCPreferencestrains,respectively theCP5-positivestrainLowenstein(ATCC49521)and CP8-positive strain Wright (ATCC 49525), and isolates that testednegativeforCP5-andCP8-ELISAweredefinedas non-typeable(NT).
Table1
Capsularprofile,intracellularsurvival,biofilmformationandagr-groupsfrequencydistribution.
Intracellular
survival Biofilm formation agr-groups
<2% >2% No Weak Moderate Strong I II III IV
Capsular profile cap5/NT 13 32 6 17 12 10 39 4 2 0 cap5/CP5 5 23 4 13 7 4 21 7 0 0 cap8/NT 16 20 1 12 12 11 33 2 1 0 cap8/CP8 95 25 29 63 20 8 72 46 1 1 Intracellular survival <2% 33 64 21 11 80 46 2 1 >2% 7 41 30 22 85 13 2 0 Biofilm formation No 18 20 2 0 Weak 76 28 1 0 Moderate 43 7 0 1 Strong 28 4 1 0
Significantcorrelations(P<0.0005)arehighlightedinbold.cap,capsule-encodinggene;CP,capsularpolysaccharide;NT:non-typeable.Intracellular survivalispresentedasthe%oftheinitialinoculum.agr-typing:I,II,III,IV,agr-groups.
2.3. Invasionassay
Bovinemammaryepithelialcells(MAC-T)wereusedfor in vitro bacterial internalization assays as previously described (Boulanger et al., 2007; Brouillette et al., 2003; Buzzola et al., 2007; Tuchscherr et al., 2005). MAC-T cells weregrownin Dulbecco’s modified Eagle’s medium (DMEM), supplemented with 10% heat-inacti-vated foetal bovine serum (FBS), non-essential amino acids,insulin and antibiotics. Prior toeach experiment, MAC-Tcellswereseededat6104cells/wellin24-well
cultureplatesandincubatedfor3daysat378Cin5%CO2.
Approximately 16h prior to the invasion experiments, cells were washed twice with PBS and incubated with invasion medium (growth medium without antibiotics containingonly1%heat-inactivatedFBS).Concurrently,S. aureus strains were incubated in Brain-Heart Infusion (BHI) broth overnight with shaking at 378C. Bacteria concentration wasthenadjusted to109CFU/ml in fresh
BHIbyopticaldensity(OD)measure,andconfluentMAC-T cell monolayers (2.5105cells/well) were inoculated
with 107CFU of S. aureus (10
m
l/well, multiplicity ofinfection40) and incubated at 378C in 5% CO2. After
1h, supernatants of theco-cultures wereremoved and cellswerewashedtwicewithPBStoremoveunattached bacteria.Onemlofinvasionmediumsupplementedwith lysostaphin (20
m
g/ml) was added to each well to lyse extracellularbacteria,andincubationwasprolongedforan additional2hat378Cin5%CO2.SupernatantswerethencollectedandplatedonColumbiasheepbloodagarplates toverifythelysisofextracellularbacteriabylysostaphin. After two PBS washing steps, the MAC-T cells were detachedbyincubationfor15minat378Cwith100
m
l/ well of trypsinand then lysed to release intracellular staphylococcibyadditionof900m
l/wellofsteriledistilled watercontaining 0.025% TritonX-100.The cell lysates werecarefullysuspended,seriallydiluted,andplatedon ColumbiasheepbloodagarplatestoquantifyCFUnumber afterovernightgrowthat378C.Resultsareexpressedasa percentageofrecoveredbacteriafromtheinitial inocu-lum.2.4. Biofilmproduction
Biofilmformationwasevaluatedbyspectrophotometry inmicroplatesusingsafraninstaining.Briefly,strainswere incubated in Tryptic Soy Broth (TSB) overnight with shaking at 378C. Cultures were then diluted 1:100 in TSB containing 0.25% glucose (TSBglc) (Melchior et al.,
2009) and transferred into wells of sterile 96-well polystyrene tissue culture (TC) plates. TSBglc without
bacteriaserved asnegative control.The TC plates were incubated at 378C for 24h. The supernatant was then discardedandthewellswerecarefullywashedtwicewith sterilePBS.The platesweredried,stainedwithsafranin 0.1%(w/v)for10min,furtherwashedtwicewithdistilled wateranddriedagainat378C.Amixtureof50%ethanol and50%aceticacidwasaddedtoeachwellandplateswere incubatedatroomtemperaturefor15min.Finally,theOD ofeachwellwasmeasuredat490nmusingamicroplate reader(Bio-Rad).Theresultswerecollectedfromatleast
two independent experiments in which the biofilm formation of each culture tested was evaluated in triplicates. The quantitative classification of biofilm productionwasbasedoncut-offvalue(ODc)andaverage OD values (Stepanovic et al., 2007), leading to four categories of strains: ODODc=no biofilm producer; ODc<OD2ODc=weakbiofilmproducer;2ODc! OD4 ODc=moderate biofilm producer; 4 ODc< OD=strongbiofilmproducer.
2.5. agr-typing
agr-groups were determined by multiplex PCR as previouslydescribed(Gilotetal.,2002).Inbrief,multiplex PCRs were performed on 2
m
l of DNA using Taq DNA polymerase(NewEnglandBioLabs) and1m
Mofeachof thefollowingprimers:Pan(50-ATGCACATGGTGCACATGC-30),agr1(50-GTCACAAGTACTATAAGCTGCGAT),agr2
(50-TATTACTAATTGAAAAGTGGCCATAGC-30),agr3(50
-GTAATGTAATAGCTTGTATAATAATACCCAG-30),and
agr4(50-CGATAATGCCGTAATACCCG-30).Amplifications
were performed with the following PCR programme: 1 cycleat948Cfor1min;26cyclesat948Cfor30s,558Cfor 30s,and728Cfor1min;andfinally1cycleat728Cfor 10min.AllPCRproductswereseparatedby electrophor-esis in 1.5% (w/v) agarose gel. Gels were stained with ethidium bromide and photographed under UV light. PatternswerecomparedtothoseobtainedwithS.aureus agrgroupreferencestrains,respectivelytheN305strain (ATCC 29740) for agr group I, the Mu50 strain (ATCC 700699) for agrgroup II,theSeattle 1945 strain(ATCC 25923)foragrgroupIII, andtheA880740strainforagr groupIV.
2.6. Pulsedfieldgelelectrophoresis
Pulsedfield gelelectrophoresis(PFGE) analyseswere performedonallstrains.Inbrief,lysostaphinpre-treated bacterialcellswereembeddedin1.8%certifiedlow-melt agarose (Bio-Rad), lysed in a 0.5M EDTA (pH 8) buffer containing10%N-lauroylsarcosineand2mg/mllysozyme, andtreatedwith1mg/mlproteinaseK.ThegenomicDNA wasdigestedwith20UofSmaI(Sigma)accordingtothe manufacturer’s instructions. The restricted fragments wereseparatedthrougha1%pulsed-fieldcertifiedagarose (Bio-Rad)gelin0.5%TBEbufferbyusingaCHEFMAPPER (Bio-Rad),at6.0V/cmfor21h,withpulsedtimesranging from5to60s,withanangleof 1208anda linearramp factor. The size of each DNA band was estimated by Biogene(VilberLourmat).
2.7. Statisticalanalysis
For calculation of the statistical significance of the observed frequency distributions, contingency tables of the randomexpected values weredeterminedand Chi-square testswere performed. P values lowerthan 0.05 wereconsideredsignificant.K-meansclusteringmethod, adaptedfortheclusteringofcategoricaldata,wasusedto groupsimilarstrainsintohomogeneousgroups(procedure fastclus,sas9.1).
3. Results
3.1. Capsularserotyping
Seventy-threestrains(31.9%)harbouredthecap5gene while 156 (68.1%) tested positive for the cap8 gene (Table 1) (Ote et al., 2011). Here, we further assessed theeffectiveproductionoftheCP5andCP8 polysacchar-idesbyELISAanalysis.Twenty-eightofthe229S.aureus isolates(12.2%)expressedCP5,120(52.4%)expressedCP8, and81(35.4%)weredefinedasnon-typeable(Table1). 3.2. Invasionassay
We observeda highvariability in the percentage of recovered bacteria from the initial inoculum 2h post-internalization(from0to95%)intoMAC-T cells. Never-theless,ourdatahighlightedthreepopulationsofstrains, corresponding tostrainsshowingdistinctsurvivalrates, respectively(i)0.50.5%,(ii)6.73.9%and(iii)50.629% oftheinitialinoculum.However,afterfurtheranalysis(see contingencytablesbelow),weobservedthatpopulations(ii) and(iii)behavedsimilarlyinallotherfeatures.Inorderto simplifytheresultspresentation,wethuschosetokeepthe distribution of the strains of our collection into two populations. We determined that 129 (56.3%) showed an invasionratelowerthan2%oftheinitialinoculum,while100 (43.7%)showedaninvasionrategreaterthan2%(Table1). 3.3. Biofilmproduction
InTSBglc,only40strains(17.5%)didnotproduceany
biofilm, while the majority were biofilm producers, respectively weak (n=105; 45.8%), moderate (n=51; 22.3%)orstrong(n=33;14.4%)producers(Table1). 3.4. agr-typingandPFGE
Inourcollection,allfouragr-groupsweredetected.The majorityof theisolates (n=165; 72.1%)belongedtoagr group I, 59 (25.8%) belonged to agr group II, 4 (1.7%) belonged to agr group III and only one strain (0.4%) belongedtoagrgroupIV(Table1).
ConcerningPFGEresults,the229strainsshowed157 distinct pulsotypes (data not shown). The four main pulsotypes, namely A, B, C and D, grouped 33 (14.4%), 28 (12.2%), 7 (3.1%) and 8 (3.5%) isolates respectively (Table 1). All the other pulsotypes were typically represented by only one strain (n=153; 66.8%). These results confirm the heterogeneity of our collection, information quiteimportant here regarding the lack of anamnesisandfarmoriginofourstrains.
3.5. Statisticalassociations
In order to determine whether the studied features couldbecorrelated,wedisposedtheirfrequency distribu-tionsincontingencytables,andtheobservedvalueswere comparedtotherandomexpectedvalues byChi-square tests.Pvalueslowerthan0.05wereconsideredsignificant. The observed frequencies that were statistically higher
thanrandomlyexpectedwerehighlightedinbold in all contingencytables.
We first observed that more cap8-positive strains express CP8, whereas more cap5-positive strains are definedasnon-typeable(Table1,P<0.0005),suggesting thatCP8strainstendtobecapsulated,whileCP5strains tendtoremainacapsulated.
Wethenobservedthatmorecap8/CP8-strainsshowa lowinvasionrate(<2%),whereasmoreNT-strainsshowa highinvasionrate(>2%)(Table1,P<0.0005).Similarly, morecap8/CP8-strainsdonotproduce,orweakly,biofilm inTSBglc,whereasmoreNT-strainsaremoderateorstrong
producers(Table1,P<0.0005).
We also observed that more strains showing a low invasionratedonotproduce,orweakly,biofilminTSBglc,
and thatmore strainsshowinga highinvasion rateare moderateorstrongbiofilmproducers inTSBglc(Table1,
P<0.0005).
Furthermore,weobservedthatmorecap8/CP8strains belongtoagrgroupII,whilemoreNT-strainsbelongtoagr groupI(Table1,P<0.005).Concerningbiofilmproduction, moreagrgroupIstrainsproducebiofilminTSBglc,while
moreagrgroupIIstrainsdonotproducebiofilminTSBglc
(Table1,P<0.0005).Finally,wefoundthatmoreagrgroup IIstrainsshowalowinvasionrate(<2%),whereasmoreagr groupIstrainsshowahighinvasionrate(>2%)(Table1, P<0.0005).
Wethendividedthestrainsintoeightgroupsaccording totheircapsularprofileandintracellularsurvivalability, andwecomparedtheirPFGEpulsotypes(Table2).Strains frompulsotypes C and D wereshown tobe NT-strains exclusively, either cap8- or cap5-positive respectively. Conversely,mostofthestrainsfrompulsotypesAandB were CP8-strains showing a low rate of internalization (<2%). Interestingly, 55.8% (n=53) of this subgroup of strains(cap8/CP8,<2%)wereonly representedby these twopulsotypesAandB,suggestingthatthesestrainsmay bemoreclonalthanallothergroups.
3.6. Clusteringmethod(Table3)
Thecluster1groupedstrainsshowingalowinvasion rateinMAC-TcellsandbelongingtoagrgroupI.Thecluster
Table2
Distributionofthestrainsaccordingtotheircapsularprofile,intracellular survivalandPFGEpulsotypes.
Capsular profile Intracellular survival PFGEtyping A B C D Other cap5/NT <2% 0 0 0 1 12 13 >2% 0 0 0 7 25 32 cap5/CP5 <2% 0 0 0 0 5 5 >2% 0 0 0 0 23 23 cap8/NT <2% 0 2 3 0 11 16 >2% 2 0 4 0 14 20 cap8/CP8 <2% 27 26 0 0 42 95 >2% 4 0 0 0 21 25 33 28 7 8 153 229 Capsularprofile:cap,capsule-encodinggene;CP,capsular polysacchar-ide;NT,non-typeable.Intracellularsurvival:presentedasthe%ofthe initialinoculums.PFGEtyping:A,B,C,D,thefourmainpulsotypes.
2 grouped strains that were cap5-positive but tested negativeinvitroforCP5ELISA,showedahighinvasionrate inMAC-Tcells,formedweakbiofilminTSBglcandbelonged
toagrgroupI.Thecluster3groupedstrainsshowingahigh invasionrateinMAC-TcellsandbelongingtoagrgroupI. Thecluster4groupedstrainsshowingahighinvasionrate inMAC-TcellsandbelongingtoagrgroupII.Thecluster5 grouped strains that were cap8-positive and tested positiveinvitroforCP8ELISA,showedalowinvasionrate inMAC-Tcells,formedweakornobiofilminTSBglcand
belonged to agr group II. Statistical significances were included in the Table 3 and highlighted in bold (p<0.0001).
4. Discussion
Although different properties associated with the persistence of S. aureus in the host, especially the mammary gland, have already been individually, or by pair, explored, toour knowledge,this is the first study correlatingallfeaturesatthesametimeon suchalarge collectionofstrains.
Althoughherepresentedas%oftheinitialinoculum, theinvasionratevaluesthatweobservedcorrespondto thosedescribedintheliterature(Boulangeretal.,2007; Brouilletteetal.,2003; Buzzolaetal., 2007;Tuchscherr etal.,2005).Thus,inaccordancewiththeobservationthat the presence of a capsule impedes cellular invasion (Buzzola et al., 2007; Tuchscherr et al., 2005), we confirmed that capsulated strains are less able than acapsulated toinvade and survive intracellularly. How-ever, the fact that some capsulated isolates still show internalizationcapacityfurthersuggeststhattheabsence ofcapsuleisnottheonlypropertyfacilitatinginvasionof eukaryoticcells(forreview,Tuchscherretal.,2010).
We also correlated biofilm production to both the capsule profile and theintracellular survival. Thus, NT-strainsorstrainswithanintracellularsurvivalratehigher than 2% were shown tobe biofilm producers in TSBglc.
Inversely,morecap8/CP8-strainsthanexpectedorstrains withanintracellularsurvivalratelowerthan2%do not producebiofilminTSBglc.
Itshouldbenotedthat,althoughourstatisticalanalysis suggestsagoodcorrelationbetweenthestudiedfeatures, evenifperformedindifferentmedia,ourresultscouldbe strengthened by analysis of bacteria grown in similar conditions.Itisbesidesquiteintuitivetoconsiderthatmilk serum used in vitro reflects more closely the situation encounteredinvivoduringmastitis.Furtherinvestigations on capsular profile, intracellular survival and biofilm formationwillthereforebeconductedonbacteriagrown inmilkserum,orinclassicalmediumsupplementedwith different milk compounds, considering the potential influenceofthesecompoundsongenesexpression.Such influenceonbiofilmproductionhasalreadybeen demon-strated for S.aureus(Melchior etal., 2009)but alsofor otherbacterial species,i.e.Streptococcusuberis (Mitchell etal.,2010;Varhimoetal.,2011).Particularly,Melchior etal.haveshownthatthegrowthmediainfluencesbiofilm formation.Indeed,theyobtaineddifferentbiofilm produc-tionresultsinTSBglcandin milkserum.Moreover, they
observedthatbiofilmformationwashigherforagrgroupI strainsthanagrgroupIIstrainsinTSBglc,butlowerforagr
groupIstrainsthanagrgroupIIstrainsinmilkserum.Their resultsinTSBglcareinagreementwithourresults.Thus,
the study of biofilm formation in milk serum for our collectionofstrainswouldbeparticularlyinteresting.
In accordance with the literature (Boulanger et al., 2007; Brouillette et al., 2003; Buzzola et al., 2007; Tuchscherret al.,2005),we furtherobserved thatmore agrgroupIstrainsshowahighinvasionrate(>2%),while moreagrgroupIIstrainsshowalowinvasionrate(<2%).In parallel,wealsohighlightedthatmoreagrgroupIstrains remainacapsulated(NT-strains),whereasmoreagrgroup IIstrainsexpressCP8.However,aswedidnotassessthe functionality of the agr system of all strains of the collection,wecannotdeterminewhethertheabsenceof CPexpressionorthehighinvasionratearedirectlylinked toregulationbytheagrlocus,orwhether otherfactors, suchasthoserelatedtotheculturemediumused,could play a role. In accordance with the literature, strains belongingtoagrgroupIproducebiofilminTSBglc,while
conversely, morestrainsthanexpectedbelongingtoagr groupIIarenon-producersinTSBglc.
Based on all these in vitro studied features, the clusteringofthestrainsallowedustoestablishdifferent profiles that could be associated with specific niches (intracellularorextracellular).Thus,wecanassumethat cap5-positive strainsbelongingto agrgroupI, whichin vitrotestnegativeforCP5ELISAandshowahighinvasion rateinMAC-Tcells(cluster2),couldcorrespondtostrains adapted to the intracellular niche. Conversely, cap8-positivestrainsbelongingtoagrgroupII,whichexpress CP8invitroandshowalowinvasionrateinMAC-Tcells (cluster5),couldcorrespondtostrainsbetteradaptedto theextracellularniche.Wecanhypothesizedthatstrains
Table3
Clusteringofthestrainsaccordingtotheircapsularprofile,intracellular survival,biofilmformationandagr-groups(K-Meansmethod).
Clusters
1 2 3 4 5 Capsularprofile cap5/NT 10 23 5 4 3
cap5/CP5 2 8 6 9 3 cap8/NT 10 7 12 1 6 cap8/CP8 21 9 14 2 74 Intracellularsurvival <2% 43 0 0 0 86 >2% 0 47 37 16 0 Biofilmformation No 14 3 1 1 19 Weak 0 37 4 4 64 Moderate 19 0 5 5 2 Strong 10 7 6 6 1 agr-groups I 40 46 37 2 40 II 0 0 0 13 46 III 2 1 0 1 0 IV 1 0 0 0 0 Total 43 47 37 16 86 Significantcorrelations(P<0.0001)arehighlightedinbold.Capsular profile:cap, capsule-encodinggene; CP,capsularpolysaccharide;NT, non-typeable.Intracellularsurvival:presentedasthe% oftheinitial inoculums.agr-typing:I,II,III,IV,agr-groups.
of thefirst group, probably adapted to an intracellular niche, could lead to persistent infections by surviving insidethehostcells,andthatstrainsofthesecondgroup, probablyadaptedtoanextracellularniche,couldleadto acuteinfections,moreeasilyeliminatedbythehost.Allthe intermediatecases(clusters1,3and4)couldberelatedto acertainprobabilityofchronicity.
To further challenge these correlations of chronic features,experimentswillbenowcarriedoutonstrains whose case histories are known and whose chronic tendencyhasbeenhighlightedinvivo.Suchconfirmation couldleadtothedevelopmentofapredictive algorithm-based series of tests that would enable to predict the chronictendencyofagivenstrainandwouldthushelpthe farmer and/or the veterinarian to choose the best therapeuticapproachsubsequentlytoamastitisdiagnosis. Nevertheless,economical aspectsof thesekinds oftests havetobeevaluatedandcomparedtothecurrentlytaken measuresincaseofbovinemastitis.
Acknowledgements
ThisworkwassupportedbyWallonia(SPW-DGARNE, conventionD31-1231).WethankGSKBiologicals(Belgium) forthekindgiftofthecapsuleantibodies.
References
Almeida,R.A.,Matthews,K.R.,Cifrian,E.,Guidry,A.J.,Oliver,S.P.,1996.
Staphylococcusaureusinvasionofbovinemammaryepithelialcells.J. DairySci.79,1021–1026.
Boulanger,D.,Brouillette,E.,Jaspar,F.,Malouin,F.,Mainil,J.,Bureau,F., Lekeux,P.,2007.HelenalinreducesStaphylococcusaureusinfectionin vitroandinvivo.Vet.Microbiol.119,330–338.
Boulanger,D.,Bureau,F.,Melotte,D.,Mainil,J.,Lekeux,P.,2003.Increased nuclearfactorkappaBactivityinmilkcellsofmastitis-affectedcows. J.DairySci.86,1259–1267.
Brouillette,E.,Grondin,G.,Shkreta,L.,Lacasse,P.,Talbot,B.G.,2003.In vivo andinvitrodemonstrationthatStaphylococcus aureusis an intracellularpathogenin thepresenceorabsenceof fibronectin-bindingproteins.Microb.Pathogenesis35,159–168.
Buzzola,F.R.,Alvarez,L.P.,Tuchscherr,L.P.,Barbagelata,M.S.,Lattar,S.M., Calvinho,L.,Sordelli,D.O.,2007.Differentialabilitiesofcapsulatedand noncapsulatedStaphylococcusaureusisolatesfromdiverseagrgroups toinvademammaryepithelialcells.Infect.Immun.75,886–891.
Costerton,J.W.,Stewart,P.S.,Greenberg,E.P.,1999.Bacterialbiofilms:a commoncauseofpersistentinfections.Science284,1318–1322.
Garzoni,C.,Kelley,W.L.,2009.Staphylococcusaureus:newevidencefor intracellularpersistence.TrendsMicrobiol.17,59–65.
Gilot,P.,Lina,G.,Cochard,T.,Poutrel,B.,2002.Analysisofthegenetic variabilityofgenesencodingtheRNAIII-activatingcomponentsAgr andTRAPinapopulationofStaphylococcusaureusstrainsisolated fromcowswithmastitis.J.Clin.Microbiol.40,4060–4067.
Halasa,T.,Huijps,K.,Osteras,O.,Hogeveen,H.,2007.Economiceffects ofbovinemastitisandmastitismanagement:areview.Vet.Q.29, 18–31.
KerroDego,O.,vanDijk,J.E.,Nederbragt,H.,2002.Factorsinvolvedinthe earlypathogenesisof bovineStaphylococcus aureusmastitiswith emphasisonbacterialadhesionandinvasion.Areview.Vet.Q.24, 181–198.
Melchior,M.B.,Vaarkamp,H.,Fink-Gremmels,J.,2006.Biofilms:arolein recurrentmastitisinfections? Vet.J.171,398–407.
Melchior,M.B.,vanOsch,M.H.,Graat,R.M.,vanDuijkeren,E.,Mevius,D.J., Nielen,M.,Gaastra,W.,Fink-Gremmels,J.,2009.Biofilmformation andgenotypingofStaphylococcus aureusbovinemastitisisolates: evidenceforlackofpenicillin-resistanceinAgr-typeIIstrains.Vet. Microbiol.137,83–89.
Mitchell,G.,Brouillette,E.,Seguin,D.L.,Asselin,A.E.,Jacob,C.L.,Malouin, F.,2010.AroleforsigmafactorBintheemergenceofStaphylococcus aureussmall-colonyvariantsandelevatedbiofilmproduction result-ingfromanexposuretoaminoglycosides.Microb.Pathogenesis48, 18–27.
Ote,I.,Taminiau,B.,Duprez,J.N.,Dizier,I.,Mainil,J.G.,2011.Genotypic characterization by polymerase chain reaction of Staphylococcus aureusisolatesassociatedwithbovinemastitis.Vet.Microbiol.153, 285–292.
Pohlmann-Dietze,P.,Ulrich,M.,Kiser,K.B.,Doring,G.,Lee,J.C.,Fournier, J.M., Botzenhart,K., Wolz,C., 2000. AdherenceofStaphylococcus aureus to endothelialcells: influenceof capsularpolysaccharide, globalregulator agr,andbacterial growthphase. Infect. Immun. 68,4865–4871.
Stepanovic,S.,Vukovic,D.,Hola,V.,DiBonaventura,G.,Djukic,S.,Cirkovic, I.,Ruzicka,F.,2007.Quantificationofbiofilminmicrotiterplates: overview of testing conditions and practical recommendations forassessmentofbiofilmproductionbystaphylococci.APMIS115, 891–899.
Tuchscherr,L.,Loffler,B.,Buzzola,F.R.,Sordelli,D.O.,2010.Staphylococcus aureusadaptationtothehostandpersistence:roleoflossofcapsular polysaccharideexpression.FutureMicrobiol.5,1823–1832.
Tuchscherr,L.P.,Buzzola,F.R.,Alvarez,L.P.,Caccuri,R.L.,Lee,J.C.,Sordelli, D.O.,2005.Capsule-negativeStaphylococcusaureusinduceschronic experimentalmastitisinmice.Infect.Immun.73,7932–7937.
Varhimo, E., Varmanen, P., Fallarero, A.,Skogman, M., Pyorala, S., Iivanainen,A.,Sukura, A.,Vuorela,P., Savijoki,K., 2011. Alpha-andbeta-caseincomponentsofhostmilkinducebiofilmformation inthemastitisbacteriumStreptococcusuberis.Vet.Microbiol.149, 381–389.
Watts,J.L.,1988.Etiologicalagentsofbovinemastitis.Vet.Microbiol.16, 41–66.