Immunohistochemical identification of Propionibacterium acnes in granuloma and inflammatory

15 Pages·2017·11.07 MB·English

RESEARCHARTICLE
Immunohistochemical identification of
Propionibacterium acnes in granuloma and
inflammatory cells of myocardial tissues
obtained from cardiac sarcoidosis patients
NaoyaAsakawa1☯,KeisukeUchida2☯,MamoruSakakibara1*,KazunoriOmote1,
KeijiNoguchi1,YusukeTokuda1,KiwamuKamiya1,KanakoC.Hatanaka3,
YoshihiroMatsuno3,ShiroYamada4,KyokoAsakawa5,YuichiroFukasawa6,
ToshiyukiNagai7,ToshihisaAnzai7,YoshihikoIkeda8,HatsueIshibashi-Ueda8,
a1111111111 MasanoriHirota9,MakotoOrii10,TakashiAkasaka10,KentaUto11,YasushigeShingu12,
a1111111111 YoshiroMatsui12,Shin-ichiroMorimoto13,HiroyukiTsutsui14,YoshinobuEishi15
a1111111111
1 DepartmentofCardiovascularMedicine,HokkaidoUniversityGraduateSchoolofMedicine,Hokkaido,
a1111111111
Japan,2 DivisionofSurgicalPathology,TokyoMedicalandDentalUniversityHospital,Tokyo,Japan,
a1111111111
3 DepartmentofSurgicalPathology,HokkaidoUniversityHospital,Hokkaido,Japan,4 Departmentof
CardiovascularMedicine,Otaru-kyokaiHospital,Hokkaido,Japan,5 DepartmentofCardiovascular
Medicine,SapporoCityGeneralHospital,Hokkaido,Japan,6 DepartmentofPathology,SapporoCity
GeneralHospital,Hokkaido,Japan,7 DepartmentofCardiovascularMedicine,NationalCerebraland
CardiovascularCenter,Osaka,Japan,8 DepartmentofPathology,NationalCerebralandCardiovascular
Center,Osaka,Japan,9 DepartmentofCardiovascularSurgery,MachidaMunicipalHospital,Tokyo,
OPENACCESS
Japan,10 DepartmentofCardiovascularMedicine,WakayamaMedicalUniversity,Wakayama,Japan,
Citation:AsakawaN,UchidaK,SakakibaraM, 11 DepartmentofPathology,TokyoWomen’sMedicalUniversity,Tokyo,Japan,12 Departmentof
OmoteK,NoguchiK,TokudaY,etal.(2017) CardiovascularandThoracicSurgery,HokkaidoUniversityGraduateSchoolofMedicine,Hokkaido,
Japan,13 DepartmentofCardiology,FujitaHealthUniversitySchoolofMedicine,Aichi,Japan,
Immunohistochemicalidentificationof
14 DepartmentofCardiovascularMedicine,KyusyuUniversity,Fukuoka,Japan,15 Departmentof
Propionibacteriumacnesingranulomaand
HumanPathology,GraduateSchoolandFacultyofMedicine,TokyoMedicalandDentalUniversity,Tokyo,
inflammatorycellsofmyocardialtissuesobtained
Japan
fromcardiacsarcoidosispatients.PLoSONE12
(7):e0179980.https://doi.org/10.1371/journal. ☯Theseauthorscontributedequallytothiswork.
pone.0179980 *[email protected]
Editor:AndrewMcDowell,UniversityofUlster,
UNITEDKINGDOM
Abstract
Received:March17,2017
Accepted:June7,2017
Background
Published:July7,2017
Copyright:©2017Asakawaetal.Thisisanopen Althoughrare,cardiacsarcoidosis(CS)ispotentiallyfatal.Earlydiagnosisandintervention
accessarticledistributedunderthetermsofthe areessential,buthistopathologicdiagnosisislimited.WeaimedtodetectPropionibacterium
CreativeCommonsAttributionLicense,which
acnes,acommonlyimplicatedetiologicagentofsarcoidosis,inmyocardialtissuesobtained
permitsunrestricteduse,distribution,and
fromCSpatients.
reproductioninanymedium,providedtheoriginal
authorandsourcearecredited.
Methodsandresults
DataAvailabilityStatement:Allrelevantdataare
withinthepaperanditsSupportingInformation Weexaminedformalin-fixedparaffin-embeddedmyocardialtissuesobtainedbysurgeryor
files.
autopsyandendomyocardialbiopsyfrompatientswithCS(n=26;CS-group),myocarditis
Funding:ThisworkwassupportedbyaGrant-in- (n=15;M-group),orothercardiomyopathies(n=39;CM-group)usingimmunohistochem-
AidforScientificResearchfromtheJapanMinistry
istry(IHC)withaP.acnes-specificmonoclonalantibody.Wefoundgranulomasin16
ofEducation,Culture,Sports,Scienceand
(62%)CS-groupsamples.Massive((cid:21)14inflammatorycells)andminimal(<14inflamma-
Technology(23591024,http://www.mext.go.jp/).
M.S.receivedthefunding. torycells)inflammatoryfoci,respectively,weredetectedin16(62%)and11(42%)ofthe
PLOSONE|https://doi.org/10.1371/journal.pone.0179980 July7,2017 1/15
CardiacsarcoidosisandPropionibacteriumacnes
Competinginterests:Theauthorshavedeclared CS-groupsamples,10(67%)and10(67%)oftheM-groupsamples,and1(3%)and18
thatnocompetinginterestsexist. (46%)oftheCM-groupsamples.P.acnes-positivereactivityingranulomas,massive
inflammatoryfoci,andminimalinflammatoryfociweredetectedin10(63%),10(63%),
and8(73%)oftheCS-groupsamples,respectively,andinnoneoftheM-groupandCM-
groupsamples.
Conclusions
FrequentidentificationofP.acnesinsarcoidgranulomasoforiginallyasepticmyocardialtis-
suessuggeststhatthisindigenousbacteriumcausesgranulomainmanyCSpatients.IHC
detectionofP.acnesinmassiveorminimalinflammatoryfociofmyocardialbiopsysamples
withoutgranulomasmaybeusefulfordifferentiatingsarcoidosisfrommyocarditisorother
cardiomyopathies.
Introduction
Sarcoidosis,asystemicdisease,ischaracterizedbythepresenceofnoncaseatingepithelioidcell
granulomasinmultipleorgans,includingtheheart,lung,skin,eyes,andcentralnervoussys-
tem.Sarcoidosisisgenerallyabenigndisease,butcardiacinvolvementleadingtofunctional
abnormalitiesisanindependentpredictorforapoorprognosis[1].Cardiacinvolvementisa
majorcauseofsarcoid-relateddeath[2],andaffects25%ofpatientswithsystemicsarcoidosis
[3].Corticosteroidtherapy,astandardtreatmentforcardiacsarcoidosis(CS),improvesthe
survivalrateofCSpatientsbypreventingleftventricular(LV)remodellinginducedby
repeatedgranulomatousinflammationandpost-inflammatoryfibrosis[4].Thispowerfulstan-
dardtreatmentforsarcoidosis,however,isonlyeffectiveforCSpatientswithapreservedLV
ejectionfraction,andnotthosewithadvancedLVdysfunction[5].Therefore,adefiniteCS
diagnosisandrapidinitiationofcorticosteroidtherapyareessentialforimprovingtheprogno-
sisofCSpatients.Atpresent,however,thediagnosisofCSischallengingduetothevariability
oftheclinicalmanifestationsaswellasthepoorsensitivityofendomyocardialbiopsy(EMB)
causedbyfrequentsamplingerrors[6,7].
Althoughtheetiologyofsarcoidosisremainsundetermined,mycobacterialandpropioni-
bacterialorganismsarethemostcommonlyimplicatedetiologicagents[8].Todate,Propioni-
bacteriumacnes(P.acnes)istheonlymicroorganismtobeisolatedfromsarcoidlesions[9].
Thisindigenousbacteriumwasisolatedfromlymphnodebiopsysamplesin78%ofsarcoidosis
patientsand20%ofnon-sarcoidosispatientsinJapan[10].Recently,Negietal.developeda
novelmonoclonalantibodyspecifictoP.acnes(anti-P.acnesantibody)thatreactswithcell
membrane-boundlipoteichoicacid[11].Immunohistochemistry(IHC)usingthenovelanti-
P.acnesantibodyrevealedP.acnes-positivereactivityinsarcoidgranulomatissuesamples
from74%oflungsand88%oflymphnodesfrompatientswithsarcoidosis,whereasnoP.
acnes-positivereactivitywasdetectedinnon-sarcoidgranulomasoftheseorgansfrompatients
withtuberculosisorsarcoidreaction.
Inthepresentstudy,weperformedIHCwiththeP.acnes-specificmonoclonalantibodyto
detectP.acnesinmyocardialtissuesfromCSpatientstoevaluatethepossibleetiologiclinkof
thisbacteriumwithCS,andtoestimatethefeasibilityofIHCanalysisofP.acnesinmyocardial
biopsysamplestodistinguishbetweenCSandothercardiomyopathies.
PLOSONE|https://doi.org/10.1371/journal.pone.0179980 July7,2017 2/15
CardiacsarcoidosisandPropionibacteriumacnes
Methods
Studysamples
Thisstudywasdesignedasaretrospectivemulti-centreobservationalstudy.Wecollected107
samples(61EMB,27autopsy,and19surgicalsamples)frompatientsclinicallydiagnosedwith
CS(n=39;CS-group),myocarditis(n=17;M-group),orothercardiomyopathies(n=51;
CM-group)betweenJanuary1994andApril2015insevenJapaneseMedicalhospitals(Hok-
kaidoUniversityHospital,NationalCerebralandCardiovascularCentre,WakayamaMedical
UniversityHospital,FujitaHealthUniversityHospital,TokyoWomen’sMedicalUniversity
Hospital,HayamaHeartCentre,andSapporoCityGeneralHospital).WecollectedtheEMB
samplesfromtherightventricularseptum.Forthesurgicalsamples,wepunchedaholeinthe
LVmyocardialtissuetoimplantaleftventricularassistdeviceorresectedduringleftventricu-
loplasty.TheCS-groupsampleswerecollectedfrompatientswithhistologically-provenmyo-
cardialsarcoidosis.AllCS-groupsamplesincludedoneormoregranulomasintheoriginal
histologicsectionspreparedforpathologicdiagnosis.AdefinitediagnosisofCSwasbasedon
thepresenceofbothmyocardialgranulomasandcardiacmanifestationwithorwithoutextra-
cardiacsarcoidosisinvolvementofatleastoneorgan,accordingtothe2006revisedversionof
theJapaneseMinistryofHealthandWelfareguidelines[12].CSpatientswithaclinicaldiag-
nosisbutnohistologicevidencewerenotincludedinthestudy.Myocarditiswasdiagnosed
basedonclinicalmanifestation,time-course,andhistologicfindingsaccordingtothe‘Guide-
linesforDiagnosisandTreatmentofMyocarditis’inJapaneseCirculationSocietyGuidelines
2009[13].Histologicfindingsofmyocarditisincludedinflammatorycellinfiltrationinthe
myocardiumandthepresenceofadjacentnecroticand/ordegeneratedmyocytesthatwerenot
typicaloftheischaemicdamageassociatedwithcoronaryarterydisease.Othercardiomyopa-
thies,includingdilatedcardiomyopathy(DCM)andhypertrophiccardiomyopathy,werediag-
nosedbyexcludingCSandmyocarditis,accordingto2016ESCGuidelinesfortheDiagnosis
andTreatmentofAcuteandChronicHeartFailure[14].Weexcluded27sampleswithunex-
pectedpresenceofnon-specificbackgroundstainingduringIHC,andsuccessfulIHCresults
wereobtainedintissuesamplesfrom26patientsintheCS-group,15intheM-group,and39
intheCM-group.Demographicandclinicaldata,includingunderlyingheartdisease,were
obtainedforallpatients.ForpatientswithEMBsamples,wecollecteddetailedinformationon
theCSdiagnosis.TheethicscommitteeofHokkaidoUniversityHospitalapprovedthestudy
protocol,whichwasperformedinaccordancewiththetenetsoftheDeclarationofHelsinki.In
thepresentstudy,individualpatientconsentwasnotrequired,butpatientswereinformedof
entryintotheregistryandallowedtooptout.
Immunohistochemistry
Formalin-fixedparaffin-embeddedmyocardialtissueswereobtainedbyEMB,oratcardiac
surgeryandautopsy.Serialparaffinsectionsonsilane-coatedslides(NewSilaneII,MutoPure
Chemicals,Tokyo,Japan)werestainedwithhaematoxylin-eosin(HE)andimmunostained
withtheanti-P.acnesantibody[11]andtheanti-mycobacterialantibody[15]asacontrol.The
IHCproceduresweresimilartothosedescribedpreviously[11].Briefly,followingdeparaffini-
sationandrehydration,sectionsweremicrowavedat97˚Cfor40min(MicrowaveProcessor
H2850;EnergyBeamSciences,EastGranby,CT,USA)in10mmol/lcitratebuffer(pH6.0)
andthenimmersedfor10minin3%hydrogenperoxideinmethanol.Initialincubationofthe
sectionswithnormalhorseserum(VectastainUniversalEliteABCKit;VectorLaboratories,
Burlingame,CA,USA)wasfollowedbyasecondovernightincubationatroomtemperature
withtheappropriatelydilutedantibodyinahumidifiedchamber.Thesectionswerethen
PLOSONE|https://doi.org/10.1371/journal.pone.0179980 July7,2017 3/15
CardiacsarcoidosisandPropionibacteriumacnes
incubatedatroomtemperaturefor30minwithbiotinylatedsecondaryantibody,followedby
another30-minincubationatroomtemperaturewithstreptavidin–peroxidasecomplex(Vec-
tastainUniversalEliteABCKit).Betweeneachstep,thesectionswerewashedwithphosphate-
bufferedsalinecontaining0.5%Tween-20.Theperoxidasesubstratediaminobenzidine(His-
tofineSimplestainDABSolution;NichireiBioscience,Tokyo,Japan)wasusedtodevelopthe
signalasabrownreactionproduct,andallsectionswerecounterstainedwithMayer’shaema-
toxylin.IHCresultsforeachsampleweredefinedaspositivewhenoneormoresmallround
bodiesweredetectedwithinanyofthegranulomasorinanyinflammatorycellinfiltration
foci,classifiedaseithermassiveorminimalinflammatoryfoci.Massiveinflammatoryfoci
indicateclearinflammatorycellinfiltration(diffuse,focal,orconfluent)comprising(cid:21)14lym-
phocytesandmacrophages/mm2inthemyocardium,consistentwiththecriteriathatdefine
myocarditis[13].Minimalinflammatoryfoci,consideredasnon-specificfindings,indicateill-
definedinflammatorycellinfiltration(mostlyfocal)otherthanmassiveinflammatoryfoci
[16].SevensamplesofCS-group,5samplesofM-group,and1sampleofCM-groupcontained
bothmassiveandminimuminflammatoryfoci.Therefore,weanalyzedeachoftheselesions
individually.ThefrequencyofsamplesorlesionswithP.acnes-positivereactivitybyIHCin
granulomasandmassiveorminimalinflammatoryfociwascomparedbetweentheCS-,M-,
andCM-groupsamplesandbetweenthesurgical/autopsyandEMBsamples.
Statisticalanalysis
Continuousvariablesareexpressedasthemean±standarddeviation,andcomparedwitha
one-factorrepeated-measuresanalysisofvariance.Post-hoccomparisonswereperformed
usingTukey’sHonestlySignificantDifferencetest.Categoricalvariablesareexpressedasfre-
quencieswithpercentages,andwerecomparedusingFisher’sexacttestfollowedbytheHolm
correction.APvalueoflessthan0.05wasconsideredstatisticallysignificant.Statisticalanaly-
seswereperformedusingJMP12(SASInstitute,Cary,NC,USA).
Results
Patientcharacteristics
ThepatientdemographicsandclinicalcharacteristicsaresummarizedinTable1.Weexam-
ined19surgical,27autopsy,and34EMBsamples.Meanpatientagewassignificantlyyounger
intheM-groupthanintheCS-andCM-groups(P’s<0.001).Thepercentageoffemaleswas
significantlyhigherintheCS-groupthanintheM-andCM-groups(P’s<0.001).Detailed
clinicalprofilesofpatientsfromwhomEMBsampleswereobtainedareshowninTable2.Sim-
ilarly,theM-grouppatientswereyoungerthanthoseintheCS-andCM-groups(P=0.006
andP=0.002,respectively).HeartratewaslowerintheCS-grouppatientsthanintheM-
grouppatients(P=0.003).Plasmaangiotensin-1-convertingenzymelevelswerehigherinthe
CS-grouppatientsthanintheM-andCM-grouppatients(P=0.017andP<0.001,respec-
tively).SerumtroponinTwashigherintheM-grouppatientsthanintheCM-grouppatients
(P=0.031).Extra-cardiacsarcoidosis,includingthatinthelung,eye,andskin,wasdetectedin
six(75%)oftheCS-grouppatients.Theclinicalmanifestationswereisolatedtotheheartin
two(25%)patients.
Histologicfindings
Thenumberandfrequency(%)ofsampleswithgranulomasorinflammatorycells(massive
orminimalinflammatoryfoci)detectedbyHE,andthenumberandfrequency(%)ofsuch
sampleswithP.acnesdetectedbyIHCineachcorrespondinglesionaresummarisedforthe
PLOSONE|https://doi.org/10.1371/journal.pone.0179980 July7,2017 4/15
CardiacsarcoidosisandPropionibacteriumacnes
Table1. Patientcharacteristics.
CS-group M-group CM-group
(N=26) (N=15) (N=39)
Typeofsample
Surgical,n(%) 4(15) 5(33) 10(26)
Autopsy,n(%) 14(54) 3(20) 10(26)
EMB,n(%) 8(31) 7(47) 19(49)
Age(years),range 27–92 15–79 18–85
Age(years),mean±SD 61.5±8.3 40.3±19.5*‡ 58.6±13.8
Female,n(%) 20(77)†‡ 2(13) 10(26)
Underlyingheartdisease
Cardiacsarcoidosis,n(%) 26(100) - -
Myocarditis - 15(100) -
DCM,n(%) - - 25(64)
HCM,n(%) - - 8(21)
Others,n(%) - - 7(18)
Continuousvariablesareexpressedasmean±SDandcategoricalvariablesareexpressedasnumber(%)
ofpatients.
*P<0.001vs.CS-group;
†P<0.001vs.M-group;
‡P<0.001vs.CM-group;
EMB,endomyocardialbiopsy;DCM,dilatedheartdisease;HCM,hypertrophicheartdisease.
https://doi.org/10.1371/journal.pone.0179980.t001
surgicalorautopsysamplesandtheEMBsamplesfromeachpatientgroup(Table3).Nonca-
seatingepithelioidcellgranulomaswithorwithoutmultinucleatedgiantcellsareshowninFig
1.Granulomasweredetectedin16(62%)samplesfromtheCS-groupandinnoneofthesam-
plesfromtheM-andCM-groups.Massiveinflammatoryfoci,aspresentedinFig2,were
observedin16(62%)lesionsfromtheCS-groupandin10(67%)lesionsfromtheM-group
sampleswithnosignificantdifferencebetweenthem.Massiveinflammatoryfociwere
observedinonly1(3%)lesionfromtheCM-group,significantlylessfrequentlythaninthe
CS-andM-groupsamples(P’s<0.001).Minimalinflammatoryfoci,asshowninFig3,were
alsoobservedfrequentlyin11(42%)samplesfromtheCS-group,10(67%)samplesfromthe
M-group,and18(46%)samplesfromtheCM-group,withnosignificantdifferenceamong
groups.Nogranulomawasobservedinthesectionsadditionallypreparedforthestudyin10
(38%)oftheCS-groupsamples,despitethefactthatgranulomaswereidentifiedintheinitially
preparedsectionsusedforpathologicdiagnosis.IntheCS-groupsampleswithnogranuloma
observed,5(50%)sampleslackedanyinflammatorylesions,includingmassiveorminimal
inflammatoryfoci.Thefrequencyofgranulomaandmassiveorminimalinflammatoryfoci
wasnotsignificantlydifferentbetweenthesurgicalorautopsyandEMBsamples.
Immunohistochemicalfindings. IHCwiththeanti-P.acnesantibodyrevealedoneora
fewsmallroundbodiesinsomeofsarcoidgranulomacells,includingepithelioidcellsand
multinucleatedgiantcells(Fig1).SuchP.acnes-positivereactivityingranulomaswasfoundin
10(63%)of16CS-groupsampleswithgranuloma(Table3).Afewormanysmallroundbod-
ieswerealsodetectedinsarcoidinflammatorycellsincludingmassive(Fig2)andminimal
(Fig3)inflammatoryfoci.P.acnes-positivereactivityinmassiveinflammatoryfociwas
detectedin10(63%)ofthe16CS-groupsamplesandinnoneofthe10M-groupsamplesor1
exceptionalCM-groupsamplewithmassiveinflammatoryfoci.Thedifferenceinthedetection
frequencybetweentheCS-groupandM-groupsampleswasstatisticallysignificant(P=0.003).
PLOSONE|https://doi.org/10.1371/journal.pone.0179980 July7,2017 5/15
CardiacsarcoidosisandPropionibacteriumacnes
Table2. ClinicalprofilesofthepatientsprovidingEMBsamples.
CS-group M-group CM-group
(N=8) (N=7) (N=19)
Age(years),range 48–85 15–79 40–85
Age(years),mean±SD 63.8±10.6 38.9±22.1*† 62.2±11.9
Female,n(%) 5(63) 0(0) 5(26)
HR(beats/min) 63.7±16.7‡ 91.4±20.3 82.9±20.0
SBP(mmHg) 108.0±17.1 104.4±21.3 116.3±24.7
LVEF(%) 37.5±14.5 28.6±20.2 38.9±19.5
BasalthinningofIVS,n(%) 3(38)‡ 0(0) 0(0)
UptakeGascintigraphyorFDG-PET,n(%) 7/8(88) 1/1(100) 0/2(0)
LGE-CMR,n(%) 3/3 1/3 13/14
Biochemistry
Calcium(mEq/l) 9.3±0.4 8.7±0.3 9.1±0.2
BNP(pg/ml) 468.3±564.6 710.8±721.2 601.0±566.7
ACE(IU/L) 24.2±8.3§k 12.0±5.3 9.3±6.5
Lysozyme(μg/mL) 7.8±1.3 8.0±2.4 9.0±2.4
sIL2-R(μg/mL) 394.0±130.0 - -
TroponinT(ng/ml) 0.03±0.01 6.3±9.1¶ 0.02±0.03
AVblock,n(%) 4(50) 2(29) 1(5)
VTorVF 2(25) 0(0) 0(0)
PPM/ICD/CRT,n(%) 5(63) 0(0) 4(21)
Extra-cardiacsarcoidosis 6(75) - -
Medications
Corticosteroid,n(%) 6(75)k 2(29) 0(0)
ACE-Is/ARBs,n(%) 5(63) 6(86) 16(84)
β-blockers,n(%) 6(75) 3(43) 17(90)
Diuretics,n(%) 2(25) 4(57) 13(68)
Allvaluesareexpressedasthemean±SDornumber(%)ofpatients.P<0.05wasconsideredstatisticallysignificant.
*P=0.006vs.CS-group;
†P=0.002vs.CM-group;
‡P=0.003vs.M-group;
§P=0.017vs.M-group;
kP<0.001vs.CM-group;
¶P=0.031vs.CM-group;
EMB,endomyocardialbiopsy;HR,heartrate;SBP,systolicbloodpressure;LVEF,leftventricularejectionfraction;IVS,intraventricularseptum;Ga,
gadolinium;FDG-PET,18F-fluorodeoxyglucose-positronemissiontomography;LGE,lategadoliniumenhancement;CMR,cardiacmagneticresonance;
BNP,brain-typenatriureticpeptide;ACE,angiotensinconvertingenzyme;IL2-R,interleukin-2receptor;AV,atrio-ventricular;VT,ventriculartachycardia;
VF,ventricularfibrillation;PPM,permanentpacemaker;ICD,implantablecardioverterdefibrillator;CRT,cardiacresynchronizationtherapy;ACE-I,
angiotensinconvertingenzymeinhibitor;ARB,angiotensinIIreceptorblocker.
https://doi.org/10.1371/journal.pone.0179980.t002
P.acnes-positivereactivityinminimalinflammatoryfociwasdetectedin8(73%)ofthe11CS-
groupsamplesandinnoneofthe10M-groupand18CM-groupsampleswithminimal
inflammatoryfoci.ThedifferenceinthedetectionfrequencybetweentheCS-groupandM-
grouporCM-groupsampleswasstatisticallysignificant(P=0.002andP<0.001),respec-
tively.ThefrequencyofP.acnes-positivereactivityingranulomasandmassiveorminimal
inflammatoryfociwasnotsignificantlydifferentbetweenthesurgicalorautopsysamplesand
theEMBsamples,orbetweentheCS-grouppatientswithandwithoutclinicalcharacteristics,
includingextra-cardiacmanifestations.Granulomaswerenotdetectedinthe10(38%)of26
PLOSONE|https://doi.org/10.1371/journal.pone.0179980 July7,2017 6/15
CardiacsarcoidosisandPropionibacteriumacnes
Table3. Summaryofhistologicandimmunohistochemicalfindings.
Numberof Number(%)of Number(%)of Number(%)of
samples sampleswith sampleswith sampleswith
granuloma massive minimal
(HE)and inflammatory inflammatory
number(%)of foci(HE)and foci(HE)and
theabove number(%)of number(%)of
sampleswith theabove theabove
P.acnes lesionswithP. lesionswithP.
detectedin acnesdetected acnesdetected
granulomas therein(IHC) therein(IHC)
(IHC)
HE IHC HE IHC HE IHC
CS-group
Allsamples 26 16 10 16(62)* 10 11 8(73)¶*
(62) (63) (63)§ (42)
Surgical/autopsysamples 18 10 5(50) 10(56)† 5(50) 9(50) 6(67)**
(56)
EMBsamples 8 6(75) 4(67) 6(75)* 5(83)k 2(25) 2
(100)††
M-group
Allsamples 15 0 - 10(67)* 0 10 0
(67)
Surgical/autopsysamples 8 0 - 6(75)† 0 5(63) 0
EMBsamples 7 0 - 4(57)‡ 0 5(71) 0
CM-group
Allsamples 39 0 - 1(3) 0 18 0
(46)
Surgical/autopsysamples 20 0 - 1(5) 0 7(35) 0
EMBsamples 19 0 - 0 - 11 0
(58)
Allvaluesareexpressedasthenumber(%)ofpatients.
*P<0.001vs.CM-group;
†P=0.002vs.CM-group;
‡P=0.007vs.CM-group;
§P=0.003vs.M-group;
kP=0.048vs.M-group;
t¶P=0.002vs.M-group;
**P=0.034vs.CM-group;
††P=0.038vs.CM-group;
HE,haematoxylin-eosin;EMB,endomyocardialbiopsy;IHC,immunohistochemistry.
https://doi.org/10.1371/journal.pone.0179980.t003
CS-groupsamples.P.acnes-positivereactivitywasdetectedwithininflammatoryfocievenin3
of10CS-groupsampleswithoutgranulomas.Nopositivereactivitywasdetectedbythecontrol
IHCwiththeanti-mycobacterialantibodyinanysamplesfromtheCS-group,M-group,and
CM-grouppatients.
Discussion
Inthisstudy,weidentifiedP.acnesbyIHC,forthefirsttime,inmyocardialtissuesfromCS
patients.P.acnes-positivereactivityingranulomacellswasfoundin63%oftheCS-groupsam-
pleswithgranulomas.P.acnes-positivereactivitywasalsofoundinmassiveandminimal
PLOSONE|https://doi.org/10.1371/journal.pone.0179980 July7,2017 7/15
CardiacsarcoidosisandPropionibacteriumacnes
Fig1.Representativesampleswithsarcoidgranulomas.Anautopsysample(A–C)andanEMBsample(D–F)from
patientswithsarcoidosiswerestainedwithhaematoxylinandeosinandimmunostainedwithanti-P.acnesantibody.Many
sarcoidgranulomaswereobservedatthelowermagnification(A,D).Smallroundbodiesindicatedbytheblackarrows(C,F)
werefoundinsomeofepithelioidcellsandmultinucleatedgiantcellsofthesesarcoidgranulomasbyimmunohistochemistry
withanti-P.acnesantibody.Originalmagnification;×200(left),×1000(middleandright).
https://doi.org/10.1371/journal.pone.0179980.g001
Fig2.Representativesampleswithmassiveinflammatoryfoci.SpecimensobtainedfromautopsysamplesinpatientswithCS(A–
C),myocarditis(D–F),orothercardiomyopathy(G-I)werestainedbyhaematoxylinandeosinandimmunostainedwithanti-P.acnes
antibody.MassiveinflammatorycellinfiltrationwasobservedinsamplesfrompatientswithCS(A,B),myocarditis(D,E),orother
cardiomyopathies(G,H).PositiveP.acnesimmunostaininginmacrophagesoftheseinflammatoryfociwasdetectedonlyinsamples
frompatientswithCS(C;blackarrows),andnotinsamplesfromthosewithmyocarditis(F)orothercardiomyopathies(I).Original
magnification;×200(left),×1000(middleandright).
https://doi.org/10.1371/journal.pone.0179980.g002
PLOSONE|https://doi.org/10.1371/journal.pone.0179980 July7,2017 8/15
CardiacsarcoidosisandPropionibacteriumacnes
Fig3.Representativesampleswithminimalinflammatoryfoci.SpecimensobtainedfromautopsysamplesandEMBsamplesinpatients
withCS(A–D),M(E–H),andCM(I–L)werestainedbyhaematoxylinandeosin(A,C,E,G,I,andK)andimmunostainedwithanti-P.acnes
antibody(B,D,F,H,J,andL).BlackarrowsindicatepositiveP.acnes-immunostaining.Minimalinflammatorycellinfiltrationwasobservedin
samplesfromallsixpatients.Evenatthelowestinflammatorycellinfiltration,positiveP.acnes-immunostaininginmacrophagesofthese
inflammatoryfociwasdetectedinsamplesfrompatientswithCS(B,D),butnotinsamplesfromthosewithM(F,H)andCM(J,L),regardlessof
thesampletype.Originalmagnification;×1000.
https://doi.org/10.1371/journal.pone.0179980.g003
inflammatoryfoci,in63%and73%oftheCS-groupsamples,respectively,andinnoneofthe
M-groupandCM-groupsampleswitheachcorrespondinglesion.Thehighfrequencyand
specificityofthedetectionoftheindigenousbacteriuminsarcoidgranulomaandinflamma-
torycellsfromoriginallyasepticmyocardialtissuesmorestronglyimplicatesitsinvolvement
insarcoidosisetiologythanearliersimilardetectionofP.acnesinsarcoidgranulomasofthe
lungsandlymphnodesbecausethebacteriumisthoughttobecommensaltotheseorgans
[17].TheresultsalsosuggestthatIHCdetectionofP.acnesinmassiveorminimalinflamma-
toryfociofmyocardialbiopsysampleswithoutgranulomascouldbeusefulfordifferentiating
sarcoidosisfrommyocarditisorothercardiomyopathies.
AdefinitiveCSdiagnosisusuallyrequiresahistologicexaminationoftissuesamples
obtainedbyEMB.ApreviousstudydemonstratedthatahistologicCSdiagnosisislimitedto
lessthan30%ofsmallEMBsamplesobtainedfromCSpatients[6];thisisalsothecaseforthe
heartsofCSpatientsobtainedatautopsy[3].WecollectedCS-groupsamplesfromhistologi-
cally-provenCSpatientswithgranulomasidentifiedinthemyocardialtissuesectionsorigi-
nallypreparedforpathologicdiagnosis.Sarcoidgranulomaswerenotfoundin10(38%)ofthe
CS-groupsamples,butratherinthehistologicsectionsthatwereadditionallypreparedforthe
presentstudy.Similarly,routinepathologicdiagnosticprocesseswithnogranulomadetected
intheoriginalsectionsbutlaterfoundinthedeep-cutsectionsofEMBsamplessometimes
occurred.Thistypeoftissue-preparationerrorseemstobeoneoftheproblems,inadditionto
themore-commonsamplingerrors,inCSdiagnosiswithEMBsamples.
PLOSONE|https://doi.org/10.1371/journal.pone.0179980 July7,2017 9/15
CardiacsarcoidosisandPropionibacteriumacnes
Granulomasinmyocardialtissues,evenwhendetectedinsurgicalorautopsysamplesand
EMBsamples,requireadifferentialdiagnosisbetweenCSandothergranulomatousdiseases
[18].Aminimumamountoffocalandcentraleosinophilicnecrosisinsarcoidgranulomas
mayneedtobedifferentiatedfromtuberculosisorotherinfectiousdiseases[19].Similarly,
giantcellsthatappearingiantcellmyocarditisaresometimesdifficulttodistinguishfrom
giantcellsintermingledwithsarcoidinflammatorycells[20–22].IntheCS-groupsamples
withgranulomaspresent,P.acnes-positivereactivityinthegranulomaswasfoundin56%of
thesurgicalorautopsysamplesand75%oftheEMBsampleswithnosignificantdifference
betweenthelargeandsmallmyocardialtissuesamples.Negietal.reportedthatnoP.acnes-
positivereactivityisfoundinnon-sarcoidgranulomasoflungsandlymphnodesfrompatients
withtuberculosisorsarcoidreaction[11].Therefore,IHCdetectionofP.acnesmaybeuseful
fordistinguishingsarcoidgranulomasfromothernon-sarcoidosisgranulomaswithasimilar
histologicappearance.
Additionally,minimalinflammatorycellinfiltrationinthemyocardiumisobservedin
somepatientswithDCMclassifiedasinflammatorydilatedcardiomyopathy(DCMI)[16,23].
Asmallnumberofinflammatorycellsarefoundinbiopsysamplesfrompatientswithclini-
callysuspectedCS,butnogranulomas.Insuchcases,apathologicCSdiagnosiscannotbe
madewithonlynon-specificfindings,andconventionalheartfailuretherapyforidiopathic
DCMiscommonlyinitiated.Itisthereforeimportanttodeterminewhethertheinflammatory
cellinfiltrationwascausedbyDCMIorCS.Inthepresentstudy,P.acnes-positivereactivity
wasdetectedin63%oftheCS-groupsampleswithmassiveinflammatoryfociandin73%of
thosewithminimalinflammatoryfoci.ThreeautopsysamplesfromtheCS-groupwithP.
acnesdetectedininflammatoryfocicontainedonlyinflammatorylesionswithoutgranuloma
inthesehistologicsections.ThesefindingssuggestthatIHCwithanti-P.acnesantibodymay
beusefulfordifferentiatingsarcoidosisfromnon-sarcoidosismyocarditisandothercardiomy-
opathies,especiallywhennogranulomaisfoundbutsomeinflammatorycellsareobservedin
themyocardialbiopsysamples.
P.acnes-positivereactivitywasalsoobservedinmanysampleswithsarcoidgranulomasand
accompanyinginflammatoryfoci(presumablyaprecursorlesionofgranulomaformation
causedbyanidenticalpathogen).Ingeneral,granulomasareformedtosequesteranddegrade
aninvadingagent.Anepithelioidcellgranulomaisthepathologichallmarkofsarcoidosis;
thus,thesarcoidgranulomamustcontain,orhavecontained,anetiologicagentofsarcoidosis.
Accordingtothesebasicprinciplesofgranulomaformation,thehighfrequencyandspecificity
ofP.acnesdetectedinsarcoidgranulomaandinflammatorycellsinoriginallyasepticmyocar-
dialtissuessuggestthatthisindigenousbacteriumcausesgranulomaformationinmanyCS
patients.
AhypersensitiveTh1immuneresponsetothiscommensalbacteriumisthoughttoleadto
granulomaformationinthesusceptiblesubjects[24–26].Basedonthehypothesisthatsarcoid-
osisisanallergicendogenousinfectioncausedbyP.acnes[27],P.acnescancauselatentinfec-
tionprimarilyinthelungsandlunghilarlymphnodesandthesecondarilyinothersystemic
organs,suchastheheart,eyes,andskin.ThedormantformofP.acnesisactivatedendoge-
nouslyundercertainenvironmentalconditionsandproliferatesatthesiteoflatentinfection.
InpatientswithP.acneshypersensitivity,granulomasareformedaroundmacrophageswith
intracellularP.acnessupportedbyanactiveTh1immuneresponsetothebacteriumduring
theprocessofgranulomatousinflammation,asshowninFig4.
ThefactthatP.acnesreactivitywasnotdetectedinsomepatientsamples,evenwithinsar-
coidgranulomasoftheCS-group,requiresfurtherinvestigation.Onepossibleexplanationis
thatotherbacteria,suchasPropionibacteriumgranulosumandMycobacteriumtuberculosis,
causegranulomaformationinsomesarcoidsamples.Theanti-P.acnesantibodyusedinthe
PLOSONE|https://doi.org/10.1371/journal.pone.0179980 July7,2017 10/15