PAPER 1

(1)

PAPER 1

(2)

ContentslistsavailableatSciVerseScienceDirect

Lung Cancer

jo u rn al h om epa g e :w w w . e l s e v i e r . c o m / l o c a t e / l u n g c a n

CD4/CD8 co-expression shows independent prognostic impact in resected non-small cell lung cancer patients treated with adjuvant radiotherapy

Sigurd M. Hald

^a,∗

, Roy M. Bremnes

^a,b

, Khalid Al-Shibli

^c,d

, Samer Al-Saad

^c,e

, Sigve Andersen

^a,b

, Helge Stenvold

^a,b

, Lill-Tove Busund

^c,e

, Tom Donnem

^a,b

aDepartmentofClinicalMedicine,UniversityofTromsø,Norway

bDepartmentofOncology,UniversityHospitalofNorthNorway,Tromsø,Norway

cDepartmentofMedicalBiology,UniversityofTromsø,Norway

dDepartmentofPathology,NordlandCentralHospital,Bodø,Norway

eDepartmentofPathology,UniversityHospitalofNorthNorway,Tromsø,Norway

a r t i c l e i n f o

Articlehistory:

Received3October2012 Receivedinrevisedform 28November2012 Accepted3December2012

Keywords:

NSCLC Radiotherapy

Adaptiveimmunesystem CD4

CD8

Prognosticimpact

a b s t r a c t

Background:Thoughtraditionallyregardedasimmunosuppressive,radiotherapymayalsostimulate immunecellsandfacilitateananti-tumorimmuneresponse.Wethereforeaimedtoexplorethepro- gnosticsigniﬁcanceofimmunecellmarkersinnon-smallcelllungcancer(NSCLC)patientstreatedwith postoperativeradiotherapy(PORT).

Methods:Inadditiontodemographicandclinicopathologicalinformation,tumortissuesampleswere collectedandtissuemicroarrays(TMAs)wereconstructedfrom55patientswithstageI-IIIANSCLCwho receivedPORT.TumorandstromalexpressionofCD1a+,CD3+,CD4+,CD8+,CD20+,CD56+,CD68+,CD117+

andCD138+cells,aswellasM-CSFandCSF-1R,wasassessedbyimmunohistochemistry.

Results:Inunivariateanalysis,highco-expressionofCD4+andCD8+Tlymphocytesaswellashighexpres- sionofCD1a+dendriticcellsinthetumorstromacorrelatedwithimproveddisease-speciﬁcsurvival(DSS).

Inmultivariateanalysispatientswithstromal↓CD4/↓CD8expressionhadahazardratioof21.1(CI95%

3.9–115.6,P<0.001)whencomparedtothosewith↑CD4/↑CD8expression.

Conclusions:Stromal↓CD4/↓CD8expressionwasanindependentnegativeprognosticfactorforsurvival inNSCLCpatientsreceivingPORT,indicatingahighlydetrimentalprognosis.

1. Introduction

Lungcancerremainstheleadingcauseofcancer-relatedmor- talityinthewesternworld,andisprojectedtoaccountfor28%ofall cancerdeathsintheUnitedStatesin2012[1].Non-smallcelllung cancer(NSCLC)represents80–85%ofalllungcancers,andsurgi- calresectionofearlystagediseasepresentsthebestopportunity forlongtermsurvival[2].Despiteextensiveresearchefforts,the prognosisofNSCLCpatients,evenwithcompletesurgicalresection, remainsdisappointing[3].Immunotherapy hasshown potential impactinthetreatmentNSCLC,andclinicalstudiesonthesigniﬁ- canceofimmunologicalmarkersarewarranted[4].

Theimmunesystemcanbedividedintotwocompartments,the innateandtheadaptiveimmunesystems.Theinnatesystemcon- sistsofdendriticcells(DCs),naturalkiller(NK)cells,NKTcells,

∗Correspondingauthorat:DepartmentofClinicalMedicine,FacultyofHealth Sciences,UniversityofTromsø,9037Tromsø,Norway.Tel.:+4790876256;

fax:+4777626779.

E-mailaddress:sigurd.hald@uit.no(S.M.Hald).

macrophages,neutrophils, basophilsandeosinophils, andis the body’sfirstlineofdefenseagainstpathogens.Bcells,CD4+Thelper cellsandCD8+cytotoxicTcells,expressadiversesetofsomatically generatedantigen-specificreceptors,therebyenablingthehighly specificadaptiveimmuneresponse[5].

Tumor-promotinginﬂammationmediatedbycellsoftheinnate immunesystemisrecognizedasanenablingcharacteristicofcan- cer development, and the tumor’s ability of avoiding immune destructionisrecognizedasanemerginghallmarkofcancer[6].

Innatecellssuchasmacrophages,mastcellsandneutrophilscon- tributetotumorangiogenesis,andtumorinﬁltrationbysuchcells oftencorrelateswithapoorprognosis[6,7].Incontrast,anabun- danceofinﬁltratinglymphocytesoftencorrelateswithafavorable prognosis[7].

WhilecelldeathbydamagetotumorDNAisthoughttobethe mainmodeofactionofradiotherapy,evidencesuggeststhat,itin additionmobilizestumorspeciﬁcimmunityandstimulatesananti- tumorresponse[8,9].Hence,radiotherapycanimprovetheeffect ofimmunotherapyincancertreatment[10].Recentstudieshave alsoshownthattheefﬁcacyofhighdoseradiotherapydepends onthepresenceofCD8+Tcells[11,12].Wepreviouslyreported 0169-5002/$–seefrontmatter© 2013 Elsevier Ireland Ltd. All rights reserved.

http://dx.doi.org/10.1016/j.lungcan.2012.12.026

(3)

210 S.M.Haldetal./LungCancer80 (2013) 209–215

ontheprognosticimpactofbothinnateandadaptiveimmunecell markersinNSCLC[13–15].Inaddition,wehaveshownthatangio- genicmarkershaveprognosticimpactinsurgicallyresectedNSCLC patientsreceivingpostoperativeradiotherapy(PORT)[16].Tothe bestofourknowledge,nostudieshaveexploredtheprognostic signiﬁcanceofimmunecellmarkersinthisgroupofpatients.In lightofthelinkbetweenradiotherapyandtumorspeciﬁcimmune responses,weaimedtoexploreifinsituimmunityhadanimpact onsurvivalinNSCLCpatientstreatedwithPORT.

2. Materialsandmethods 2.1. Patients

PatientssurgicallyresectedforNSCLCstageI-IIIAattheUniver- sityHospitalofNorthernNorwayandNordlandCentralHospital from1990through2004wereidentiﬁedinthisretrospectivestudy.

Intotal,371patientsfromthehospitaldatabaseswereregistered.

Of these, sixty-three patients received radiotherapy within 12 weekspostoperatively,withacumulativeradiationdoseof≥50Gy.

Eightpatientswereexcludeddueto:Preoperativechemotherapy (n=3),othermalignancywithin5yearspriortoNSCLCdiagnosis (n=3)orinadequateparafﬁn-embeddedsurgicalspecimens(n=2).

Atotalof55patientsweretherebyincludedinthestudy.Adju- vantchemotherapyhadnotbeenintroducedinNorwayduringthis period(1990–2004).Clinicopathologicanddemographicdatawere collectedretrospectively.Thisstudyincludesfollowupdataasof January2011.Patientswerestagedaccordingtotherevised7th editionofUICCTNMclassiﬁcationof lungcancer [17], andhis- tologicallygradedand subtypedaccordingtotheWorld Health Organizationguidelines[18].TheNorwegianDataInspectorateand TheRegionalCommitteeforMedicalandHealthResearchEthics approvedthestudy.

2.2. Microarrayconstruction

Allspecimenswereexaminedbytwopathologists(S.Al-Sand K.Al-S). The most representative paraffin blocks were selected andtwo areasofviabletumorcells(neoplasticepithelium) and two from the central tumor-surrounding stroma were chosen and marked onthe donor blocks. The tissue microarrayswere assembled using a tissue-arraying instrument (Beecher Instru- ments,Silver Springs,MD,USA). Thedetailed methodologyhas beenreported previously [19]. Using a 0.6mm-diameter stylet, cores from two separate predefined neoplastic epithelial areas and two stromalareas weretransferred torecipient blocks. To includeallcoresamples,atotalofeighttissuearrayblockswere constructed. Multiple 4-␮m sections were cut with a Micron microtome (HM355S) and stained with specific antibodies for immunohistochemicalanalysis.Bothnormallungtissueslocalized distanttotheprimarytumorandoneslidewithnormallungtissue samplefrom20patientswithoutadiagnosisofcancerwereused ascontrols.

2.3. Immunohistochemistry

ThefollowingantibodiesfromVentanaMedical(Tucson,Ariz, USA)wereusedinthisstudy:CD20(cloneL26),CD8(clone1A5), CD68(cloneKP1),CD138(cloneB-A38),CD1a,CD3(clone PS1), CD117(cloneanti-CKit,9.7)andCD138(cloneB-A38).AllVen- tanaantibodieswerepredilutedfromthemanufacturer.Inaddition CD4(clone1F6,NovocastraLaboratoriesLtd.NewcastleuponTyne, UK,dilution1:5),M-CSF(SantaCruzBiotechnology,SantaCruz,CA, USA,dilution1:5)andCSF-1R(cloneH-300,SantaCruzBiotechnol- ogy,dilution1:25)wereused.Thedetailedimmunohistochemical

procedureshavebeenpublishedpreviously[13–15].Foreachanti- body,includingnegativestainingcontrols,allstainingwasdone inasingleexperiment.Asnegativestainingcontrols,theprimary antibodieswerereplacedwiththeprimaryantibodydiluents.

2.4. ScoringofImmunohistochemistry

Tissuesectionswerescoredbylightmicroscopyfordegreeof inﬁltrationofthespeciﬁedimmunecells.

TheCD8+cellswerescoredaslowif≤5%orashighif>5%ofthe wholesurfaceareaoftheepithelialcompartmentswereinﬁltrated, andwasscoredaslowif≤50%orhighif>50%ofthetotalnucleated surfaceareaofthestromalcompartmentswereinﬁltrated.CD4+

cellswerescoredashighifrepresenting≥5%or≥25%ofthetotal nucleatedcellsintheepithelialandstromalcompartments,respec- tively.FewCD4+andCD8+Tcells(0to<5%ofthetotalnucleated cells)wereobservedintheinterstitialtissueofthenonneoplastic controls.

CD1a+cellswerescoredaslowifabsentorifrepresenting<1%of thenucleatedcellsandhighotherwise,inbothepithelialandstromalcompartments.IntraepithelialCD68+cellswerescoredaslow ifabsentorrepresenting<1%ofthenucleatedcellsandhighoth- erwise,whilethemoreabundantstromalCD68+cellswerescored aslowiftheyrepresented<25%of thetotalnucleatedcells and highotherwise.CD56+cellswerescoredaspresent(highscore)or absent(lowscore)inbothepithelialandstromalcompartments.

Theintensity ofM-CSF and CSF-1R inboth epithelial and stro- malcompartmentswerescoredasfollows:0=negative;1=weak;

2=intermediateand3=strong.Thecelldensityofthestromawas scoredastheratioofpositivecellscomparedtothesurfacearea oftheextracellularmatrixinthefollowingmanner:1=lowden- sity(<25%cell/matrixratio);2=intermediatedensity(25–50%)and 3=highdensity(>50%).Highexpressioninthetumorepithelium wasdeﬁnedasascore≥1.5forbothM-CSFandCSF-1R.Expression in thestromawascalculated byadding densityscore tointen- sityscorepriortocategorizingintolowandhighexpression.High expressionwasdeﬁnedas>3.5forM-CSFand>3forCSF-1R.

CD3+cellswerescoredaslowiftheyrepresented<1%ofthe nucleatedcellsintheepithelialcoresandhighotherwise,andas highif representing>50% of nucleated cells in the stroma and lowotherwise.CD138+cellswerescoredashighifrepresenting

>5%ofthenucleatedcellsintheepithelialcompartmentor>25%

inthestromalcompartment,andaslowotherwise.AsCDD138+

cellsalsostainepithelialcellsthemselvesthestainingintensityin theepithelialcompartmentwasscoredinthefollowingmanner:

0=negative;1=weak;2=intermediateand3=strong.Highexpres- sionwasdeﬁnedasascore>1.CD117+cellswereextremelyrare intheepithelialcompartmentsandsparseinthestromalcompart- ment,theywerethereforescoredaspresent(highscore)orabsent (lowscore)andonlyinthestromalcompartment.

Allsamples were anonymizedand independently scoredby two pathologists(S.A.Sand K.A.S). Incase of disagreement,the slideswerere-examinedandtheobserversreachedaconsensus.

Whenassessingonemarkerinagivencore,bothobserverswere blindedtothescoresoftheothermarkersaswellastothepatient’s outcome.Theinterobserverscoringagreementbetweenthetwo pathologistswastested onthecurrentmaterialpreviously[20], withameancorrelationcoefﬁcientof0.95(range0.93–0.98).

2.5. Statisticalmethods

Allstatistical analyses werecarried out usingthe statistical packageIBMSPSS,version20(SPSS Inc.,Chicago,IL,USA).Uni- variateanalysisoftheassociationbetweenmarkerexpressionand survivalwasdoneusingtheKaplan–Meiermethodandthestatisti- calsigniﬁcanceofdifferencesbetweensurvivalcurveswasassessed

(4)

Fig.1. Disease-speciﬁcsurvivalcurvesaccordingtotheco-expressionofstromal CD4andCD8in54NSCLCpatientsadministeredpost-operativeradiotherapy.

bythelog-ranktest.Thedisease-speciﬁcsurvival(DSS)wasdeter- minedfromthedateofsurgerytothetimeoflungcancerdeath.

Onlyvariablesofsigniﬁcantvaluefromtheunivariateanalysiswere enteredintothemultivariateanalysis,usingtheCoxproportional hazardsmodel.Probabilityforstepwiseentryandremovalwasset at0.05and0.10.ThesigniﬁcancelevelwassetatP<0.05.

3. Results

3.1. Clinicopathologicvariables

Demographic, clinical and histopathology variable are pre- sentedinTable1.Themediansurvivaltimeforall55patientswas 24months(range3–197).The5-yearDSSwas44%andthe10-year DSSwas42%.Medianpatientagewas65years(range39–76)and themajorityofpatientsweremen(69%).TheNSCLCtumorswere comprisedof33squamouscellcarcinomas,16adenocarcinomas and6largecellcarcinomas.

In univariate analysis, weight loss >10% (P=0.029), histology (P=0.048), poor tumor cell differentiation (P=0.026) and nodalmetastasis(P=0.010)weresignificantprognosticvariables (Table1).Theassociationbetweenmolecularmarkerexpression anddisease-specificsurvivaldataispresentedinTable2.Theco- expressionof CD4and CD8wasa strongsignificantprognostic factor(Fig.1andTable2),aswasstromalCD4expression(Table2).

Inaddition,patientswithhighstromalexpressionofCD1ahada signiﬁcantlybetterDSSthanthosewithalowexpression(Fig.2).

3.2. Multivariateanalysis

None of the clinicopathologic variables emerged signiﬁcant inmultivariateanalysis,while thehazardratiowas21.2(CI95%

4.5–120.4, P<0.001) for the ↓CD4/↓CD8 combination and 1.8 (CI95%0.4–8.4,P=0.430)forotherCD4/CD8combinations,when comparedtothereferencegroup↑CD4/↑CD8(Table3).LowCD1a hadahazardrationof2.5(CI95%0.97–6.2,P=0.058)whencom- paredtohighexpression.

Fig.2. Disease-speciﬁcsurvivalcurvesaccordingtotheexpressionofstromalCD1a in53NSCLCpatientsadministeredpost-operativeradiotherapy.

4. Discussion

We present theﬁrststudyexaminingtheprognostic impact of immunecell markerexpression insurgically resectedNSCLC patients treated with adjuvant radiotherapy. Our main ﬁnding is thatthestromalco-expression ofCD4+ andCD8+ Tlympho- cytesisastrongandindependentprognosticfactorinthisgroup.

Patients with ↓CD4+/CD8+↓ expression seem to have remark- ably poorprognosisand willtherefore mostlikely have a very limitedbeneﬁtofadjuvantradiotherapy.The5-yearsurvivalrate for patients with↑CD4+↑CD8+ expression(16%,n=9)was78%, whereas↓CD4+↓CD8+patients(22%,n=12)hadmediansurvival rateofonly9months,withnonesurvivinglongerthan19months from the time of diagnosis. The observed hazard ratio of 21.1 between↓CD4+↓CD8+and↑CD4+↑CD8+indicatesasubstantialand independentimpactonDSS.However,duetothesmallnumberof patientstheresultshavetobeinterpretedcautiously.

Hiraokaetal.havepreviouslyshownthatthereisasynergistic effectofsimultaneoushighCD4+andCD8+T-cellexpressionon survivalinNSCLC[21],whilewepreviouslyshowedthatstromal expressionofCD4andCD8bothareindependentprognosticfac- torsinNSCLC[13].Thehighhazardratioobservedinoursubgroup ofpatientsindicatesthatCD4/CD8expressionhashigherprogno- sticsigniﬁcanceinPORTtreatedpatients,andmaysuggestalink betweenstromalinsituimmunityandradiotherapyresponse.

Resultsfromcelllinesandmurinemodelsrevealcloseinter- playbetweentheimmunesystemandtheeffectsofradiotherapy.

Radiotherapymayenhanceexpressionoftumor-associatedanti- gens,facilitateimmune-mediatedtargetingofthetumorstroma and diminish the activity of regulatory T-cells. [4]. However, ourresultssuggestthatradiotherapyalonedoesnotup-regulate the immune response sufﬁciently to inhibit tumor growth in

↓CD4/↓CD8patients.Inamurinemodelofmelanoma,Leeetal.

observedthatthetherapeuticeffectofradiotherapywasdependent onCD8+T-cells,sincetumorsofCD8depletedmicebecameradio- resistant[11].Guptaet al.recentlydescribed howCD8+ T-cells arecrucialfortheeffectoflocalhigh-doseradiotherapy,whereas CD4+T-cellsandmacrophageswerenot[12].LowstromalCD4/CD8 mayindicatean insufﬁcientlevel ofthesecells for asuccessful

(5)

Table1

Prognosticclinicopathologicvariablesaspredictorsofdisease-speciﬁcsurvivalin55NSCLC-patientsreceivingadjuvantpostoperativeradiotherapy.

Characteristic Patients(n) Patients(%) Mediansurvival(months) 5-Yearsurvival(%) P

Age 0.471

≤65years 31 56 44 42

>65years 24 44 41 48

Sex 0.433

Female 17 31 64 53

Male 38 69 26 41

Smoking 0.491

Never 1 2 NR 100

Current 31 56 41 40

Former 23 42 47 47

Performancestatus 0.159

ECOG0 28 51 47 50

ECOG1 23 42 26 36

ECOG2 4 7 NR 67

Weightloss 0.029

<10% 49 89 47 47

>10% 6 11 8 20

Histology 0.048

Squamouscellcarcinoma 33 60 NR 61

Adenocarcinoma 16 29 21 19

Largecellcarcinoma 6 11 18 17

Differentiation 0.026

Poor 27 49 18 21

Moderate 21 38 127 65

Well 7 13 NR 63

Surgicalprocedure 0.795

Lobectomy 29 53 47 43

Pneumonectomy 26 47 18 45

Pathologicalstage 0.084

I 7 13 NR 83

II 20 36 NR 51

III 28 51 21 30

Tumorstatus 0.923

1 7 13 44 40

2 32 58 26 44

3 16 29 47 45

Nodalstatus 0.010

0 14 25 NR 75

1 19 35 41 50

2 22 40 19 21

Surgicalmargins 0.174

Free 38 69 21 37

Notfree 17 31 NR 60

Vascularinﬁltration 0.146

No 42 76 64 51

Yes 13 24 26 21

ClinicanreasonforPORT 0.063

Insufﬁcientmarginortumorcellsinresectionmargin 18 33 NR 65

N1 14 25.5 16 50

N2 20 36 19 22

Localrecurrence 3 5.5 NR 67

Fractioningregime 0.460

2.8×15=42Gy 29 53 19 41

2×30=60Gy 21 38 47 48

2×25–29=50–59Gy 5 9 24 40

NR:notreached;PORT:postoperativeradiotherapy;NCSLC:non-smallcelllungcancer.

Boldvaluesindicatep<0.05

“boosting” of the radiotherapy effect. Stimulating the immune responseviaimmunotherapycouldthereforepossiblyaugmentthe responsivenesstoradiotherapyinthoseindividualslackingconcur- renthighCD4/CD8levelsinthetumorstroma.

Experimentaldatasuggestthatradiotherapyandimmunother- apymayhaveadditiveandsynergisticeffects.Reitsetal.showed thatradiotherapyprior toadoptivetreatmentwithcytotoxicT- cells greatlyenhanced the efﬁcacyof theimmunotherapy [10].

Takeshimaetal.observedthatlocaltumorirradiationaugmented thetherapeuticeffectofTh1celltherapy,accompaniedbyinduc- tion of cytotoxic T-lymphocytes in the tumor draining lymph nodes and tumormass [22]. In a murine model of LewisLung Carcinoma,Yokouchietal.reportedgreaterefﬁcacywhencom- bining radiotherapy with an agonistic monoclonal antibody to

␣OX40(CD134),which augmentsT-cellexpansionand survival, whencomparedtoeithersingletreatmentgivenseparately[23].

Similar resultswere presented by Gough et al., witha signiﬁ- cantportionof long-termtumor-freesurvivors[24]. Combining CTLA-4 blockade with radiation, Demaria et al. were able to induceanimmune-mediatedinhibitionofmetastasesinamouse model of breast cancer [25]. Similarly, Dewan et al. were able toinduceanabscopaleffectbycombiningfractionatedradiothe- rapywithananti-CTLA-4antibody[26].Adjuvantimmunotherapy has shown encouraging results in NSCLC [27], but few trials havelookedatcombiningimmunotherapywithradiotherapy.As theabovepresentedpre-clinicalstudiesindicate,thistreatment combinationmaybeaninterestingapproachforresectedNSCLC patients.

(6)

Table2

Prognosticmolecularvariablesaspredictorsofdisease-speciﬁcsurvivalin55NSCLC-patientsreceivingadjuvantpostoperativeradiotherapy.

Markerexpression Patients(n) Patients(%) Mediansurvival(months) 5-Yearsurvival(%) P

CD4

Tumor 0.799

Low 40 73 47 40

High 14 25 44 49

Missing 1 2

Stroma <0.001

Low 12 22 9 0

High 42 76 NR 59

Missing 1 2

CD8

Tumor 0.525

Low 41 74.5 41 45

High 13 23.5 47 45

Missing 1 2

Stroma 0.072

Low 45 82 41 39

High 9 16 NR 78

Missing 1 2

CD4/CD8

Stroma <0.001

↑CD4+/↑CD8+ 9 16 NR 78

OtherCD4+/CD8+combination 33 60 127 54

↓CD4+/↓CD8+ 12 22 9 0

Missing 1 2

Tumor 0.476

↑CD4+/↑CD8+ 6 11 NR 63

OtherCD4+/CD8+combination 15 27 26 31

↓CD4+/↓CD8+ 33 60 47 47

Missing 1 2

CD20

Tumor 0.059

Low 40 73 26 40

High 14 25 NR 61

Missing 1 2

Stroma 0.419

Low 10 18 16 34

High 44 80 47 47

Missing 1 2

CD68

Tumor 0.661

Low 23 42 19 45

High 31 56 47 45

Missing 1 2

Stroma 0.414

Low 38 69 47 45

High 16 29 44 48

Missing 1 2

CD56

Tumor 0.316

Low 52 94 47 47

High 2 4 18 0

Missing 1 2

Stroma 0.108

Low 49 89 41 41

High 5 9 NR 80

Missing 1 2

CD1a

Tumor 0.499

Low 32 58 28 39

High 22 40 64 54

Missing 1 2

Stroma 0.025

Low 36 65 24 38

High 17 31 NR 64

Missing 2 4

M-CSF

Tumor 0.939

Low 15 27 16 47

High 38 69 46 44

Missing 2 4

Stroma 0.843

Low 28 51 47 45

High 24 44 41 48

Missing 3 5

(7)

Table2(Continued)

Markerexpression Patients(n) Patients(%) Mediansurvival(months) 5-Yearsurvival(%) P

CSF-1R

Tumor 0.215

Low 21 38 16 34

High 22 40 127 55

Missing 12 22

Stroma 0.701

Low 26 47 47 42

High 27 49 41 48

Missing 2 4

CD3

Tumor 0.619

Low 38 69 41 41

High 16 29 64 55

Missing 1 2

Stroma 0.212

Low 42 76 41 41

High 12 22 NR 57

Missing 1 2

CD138

Tumor 0.292

Low 24 43.5 19 35

High 29 52.5 64 54

Missing 2 4

Stroma 0.165

Low 24 43.5 24 33

High 29 52.5 64 53

Missing 2 4

CD138ofthecancercells

Negative 12 22 13 25 0.058

Positive 41 74 64 51

Missing 2 4

CD117inthestroma

Negative 36 65 47 49 0.305

Positive 17 31 44 35

Missing 2 4

NR:notreached;NCSLC:non-smallcelllungcancer.

Table3

ResultofCoxregressionanalysissummarizingprognosticfactorswithP<0.10.

Variable Hazardratio 95%Conﬁdenceinterval P

StromalCD4/CD8 <0.001^*

↑CD4+/↑CD8+ 1.000

OtherCD4+/CD8+combination 1.842 (0.404–8.390) 0.430

↓CD4+/↓CD8+ 21.123 (3.860–115.584) <0.001

StromalCD1a

Low 2.454 (0.969–6.213) 0.058

High 1.000

NoneoftheclinicopathologicvariablesemergedasstatisticallysigniﬁcantduringCoxregressionanalysis.

*Overallsigniﬁcanceasaprognosticfactor.

Though only shown in univariate analysis, we foundthat a higherexpressionofstromalCD1a+DCsconferanincreasedDSS whencomparedtolowexpressionforpatientstreatedwithPORT.

DCsareprofessionalantigenpresentingcells,whocanprocessand presenttumorassociatedantigensandtherebyactivateadaptive immune cells [28]. Radiation-inducedtumorcell deathmaybe associatedwiththeproductionofmaturationsignalsforDCs[29].

Teitz-Tennenbaum et al. observed that the efﬁcacy of DC immunotherapywasenhancedbyradiotherapy[30].IncreasingDC inﬁltrationthoughimmunotherapycouldthereforebeapotential strategytoimprovesurvivalinPORTtreatedpatients.

Inconclusion,wehaveshownthatlowCD4/CD8expressionis anindependentnegativeprognosticfactorinsurgicallyresected NSCLCtreatedwithPORT. Thoughourresultsarestriking,they shouldbe considered withcaution, asthe number of included patientsislow.Nevertheless,furtherstudiesarepivotalinorder toelucidatethepotentialsigniﬁcanceofCD4/CD8expressionasa predictivemarkerinadjuvantlyirradiatedNSCLC.

Conﬂictsofintereststatement Nonedeclared.

Acknowledgements

ThestudywasfundedbytheNorwegianCancerSocietyandthe NorthernNorwayHealthRegionAuthority(HelseNordRHF),and theauthorswouldliketothankthemfortheirsupport.Thefunders hadnoroleinstudydesign,datacollectionandanalysis,decision topublish,orpreparationofthemanuscript.

References

[1]SiegelR,NaishadhamD,JemalA.Cancerstatistics,2012.CACancerJClin 2012;62:10–29.

[2] SanghaR,PriceJ,ButtsCA.Adjuvanttherapyinnon-smallcelllungcancer:

currentandfuturedirections.Oncologist2010;15:862–72.

(8)

[3]HolmbergL,SandinF,BrayF,RichardsM,SpicerJ,LambeM,etal.Nationalcom- parisonsoflungcancersurvivalinEngland,NorwayandSweden2001–2004:

differencesoccurearlyinfollow-up.Thorax2010;65:436–41.

[4]VandenHeuvelMM,BurgersSA,vanZandwijkN.Immunotherapyinnon- small-celllungcarcinoma:frominﬂammationtovaccination.ClinLungCancer 2009;10:99–105.

[5]BremnesRM,Al-ShibliK,DonnemT,SireraR,Al-SaadS,AndersenS,etal.The roleoftumor-inﬁltratingimmunecellsandchronicinﬂammationatthetumor siteoncancerdevelopment,progression,andprognosis:emphasisonnon- smallcelllungcancer.JThoracOncol2011;6:824–33.

[6]HanahanD,WeinbergRA.Hallmarksofcancer: thenextgeneration.Cell 2011;144:646–74.

[7]deVisserKE,EichtenA,CoussensLM.Paradoxicalrolesoftheimmunesystem duringcancerdevelopment.NatRevCancer2006;6:24–37.

[8]KamravaM,BernsteinMB,CamphausenK,HodgeJW.Combiningradiation, immunotherapy,andantiangiogenesisagentsinthemanagementofcan- cer:theThreeMusketeersorjustanotherquixoticcombination.MolBiosyst 2009;5:1262–70.

[9]DemariaS,BhardwajN,McBride WH,FormentiSC.Combiningradiothe- rapyandimmunotherapy:arevivedpartnership.IntJRadiatOncolBiolPhys 2005;63:655–66.

[10]ReitsEA, HodgeJW,HerbertsCA,GroothuisTA,ChakrabortyM,Wansley EK,etal.Radiationmodulatesthepeptiderepertoire,enhancesMHCclass Iexpression,andinducessuccessfulantitumorimmunotherapy.JExpMed 2006;203:1259–71.

[11]LeeY,AuhSL,WangY,BurnetteB,MengY,BeckettM,etal.Therapeuticeffects ofablativeradiationonlocaltumorrequireCD8+Tcells:changingstrategies forcancertreatment.Blood2009;114:589–95.

[12]GuptaA,ProbstHC,VuongV,LandshammerA,MuthS,YagitaH,etal.Radiothe- rapypromotestumor-speciﬁceffectorCD8+Tcellsviadendriticcellactivation.

JImmunol2012;189:558–66.

[13] Al-ShibliKI,DonnemT,Al-SaadS,PerssonM,BremnesRM,BusundLT.Progno- sticeffectofepithelialandstromallymphocyteinﬁltrationinnon-smallcell lungcancer.ClinCancerRes2008;14:5220–7.

[14]Al-ShibliK,Al-SaadS,DonnemT,PerssonM,BremnesRM,BusundL-T.The prognosticvalueofintraepithelialandstromalinnateimmunesystemcellsin non-smallcelllungcarcinoma.Histopathology2009;55:301–12.

[15]Al-ShibliK,Al-SaadS, AndersenS, DonnemT, BremnesRM, BusundLT.

The prognostic value of intraepithelial and stromal CD3-, CD117- and CD138-positivecells innon-small celllung carcinoma. APMIS2010;118:

371–82.

[16]AndersenS,DonnemT,Al-SaadS,Al-ShibliK,BusundLT,BremnesRM.Angio- genicmarkersshowhighprognosticimpactonsurvivalinmarginallyoperable non-smallcelllungcancerpatientstreatedwithadjuvantradiotherapy.JTho- racOncol2009;4:463–71.

[17]Rami-PortaR,CrowleyJJ,GoldstrawP.TherevisedTNMstagingsystemforlung cancer.AnnThoracCardiovascSurg2009;15:4–9.

[18]GibbsAR,ThunnissenFB.Histologicaltypingoflungandpleuraltumours:third edition.JClinPathol2001;54:498–9.

[19]BremnesRM,VeveR,GabrielsonE,HirschFR,BaronA,BemisL,etal.High- throughputtissuemicroarrayanalysisusedtoevaluatebiologyandprognostic signiﬁcanceoftheE-cadherinpathwayinnon-small-celllungcancer.JClin Oncol2002;20:2417–28.

[20]DonnemT,Al-Saad S,Al-ShibliK,DelghandiMP, PerssonM,NilsenMN, etal.Inverseprognosticimpactofangiogenicmarkerexpressionintumor cellsversusstromal cells innonsmallcelllung cancer. ClinCancerRes 2007;13:6649–57.

[21]HiraokaK,MiyamotoM,ChoY,SuzuokiM,OshikiriT,NakakuboY,etal.Con- currentinﬁltrationbyCD8+TcellsandCD4+Tcellsisafavourableprognostic factorinnon-small-celllungcarcinoma.BrJCancer2006;94:275–80.

[22] TakeshimaT,ChamotoK,WakitaD,OhkuriT,TogashiY,ShiratoH,etal.Local radiationtherapyinhibitstumorgrowththroughthegenerationoftumor- speciﬁcCTL:itspotentiationbycombinationwithTh1celltherapy.Cancer Res2010;70:2697–706.

[23] YokouchiH,YamazakiK,ChamotoK,KikuchiE,ShinagawaN,OizumiS,etal.

Anti-OX40monoclonalantibodytherapyincombinationwithradiotherapy resultsintherapeuticantitumorimmunitytomurinelungcancer.CancerSci 2008;99:361–7.

[24]GoughMJ,CrittendenMR,SarffM,PangP,SeungSK,VettoJT,etal.Adju- vant therapywith agonistic antibodies to CD134 (OX40) increases local controlaftersurgicalorradiationtherapyofcancerinmice.JImmunother 2010;33:798–809.

[25]DemariaS,KawashimaN,YangAM,DevittML,BabbJS,AllisonJP,FormentiSC.

Immune-MediatedInhibitionofMetastasesafterTreatmentwithLocalRadia- tionandCTLA-4BlockadeinaMouseModelofBreastCancer.ClinCancerRes 2005;11:728–34.

[26] DewanMZ,GallowayAE,KawashimaN,DewyngaertJK,BabbJS,Formenti SC,etal.Fractionatedbutnotsingle-doseradiotherapyinducesanimmune- mediatedabscopaleffectwhencombinedwithanti-CTLA-4antibody.Clin CancerRes2009;15:5379–88.

[27] TuckerZCG,LagunaBA,MoonE,SinghalS.Adjuvantimmunotherapyfornon- smallcelllungcancer.CancerTreatRev2012;38:650–61.

[28]DePasT,GiovanniniM,RescignoM,CataniaC,ToffalorioF,SpitaleriG,etal.

Vaccinesinnon-smallcelllungcancer:rationale,combinationstrategiesand updateonclinicaltrials.CritRevOncolHematol2012;83:432–43.

[29]RosesRE,XuM,KoskiGK,CzernieckiBJ.RadiationtherapyandToll-likereceptor signaling:implicationsforthetreatmentofcancer.Oncogene2008;27:200–7.

[30]Teitz-TennenbaumS,LiQ,RynkiewiczS,ItoF,DavisMA,McGinnCJ,ChangAE.

Radiotherapypotentiatesthetherapeuticefﬁcacyofintratumoraldendriticcell administration.CancerRes2003;63:8466–75.