Effect of IFN-γ on the kynurenine/tryptophan ratio in monolayer-cultured keratinocytes and a 3D reconstructed human epidermis model

(1)

Original

Article

Effect

of

IFN-

g

on

the

kynurenine/tryptophan

ratio

in

monolayer-cultured

keratinocytes

and

a

3D

reconstructed

human

epidermis

model

Anna

Gustafsson

a,b,

*

,

Zdenka

Prgomet

c

,

Skaidre

Jankovskaja

a,b

,

Tautgirdas

Ruzgas

a,b

,

Johan

Engblom

a,b

,

Lars

Ohlsson

a,b

,

Anette

Gjörloff

Wingren

a,b

a_Bioﬁlms_–

ResearchCenterforBiointerfaces,MalmöUniversity,Malmö,Sweden b

DepartmentofBiomedicalScience,FacultyofHealthandSociety,MalmöUniversity,Malmö,Sweden c_Department_of_Oral_Biology,_Faculty_of_Odontology,_Malmö_University,_Malmö,_Sweden

ARTICLE INFO Articlehistory: Received16March2020

Receivedinrevisedform10July2020 Accepted15July2020

Keywords: IDO Kynurenine

Pro-inﬂammatorycytokines Reconstructedhumanepidermis Tryptophan

ABSTRACT

Background:Interferon-gamma(IFN-g)representsapotentinducerforkeratinocyteinflammatoryand immune activation invitro. Since tryptophan (trp) conversion to kynurenine (kyn) is involved in inflammation,thetopicalkyn/trpratiomayserveasabiomarkerofskininflammation.However,thetrp metabolisminkeratinocytesexposedtoIFN-gisnotyetfullyunderstood.

Objective:Theaimofthisstudywastoestablishahumanepidermismodelinordertoquantifycytokine andkyn/trpsecretionfromIFN-gstimulatedcellsandtissues.Moreover,tocomparethecellresponseof 2D-culturedkeratinocytesandthe3Depidermismodel.

Methods:Polycarbonateﬁlterswereusedonwhichprimarykeratinocytescouldattachandstratifyin order toform the typicallayers of reconstructed human epidermis (RHE). AfterIFN-g treatment, secretion ofkyn/trpwasmeasuredbyhighperformance liquidchromatography.Gene andprotein expressionofindoleamine2,3-dioxygenase1(IDO)wasanalyzedwithreal-timePCRand immunohis-tochemistry.ThesecretionofcytokineswasquantiﬁedwithELISA.

Results:Trpcatabolismtokynwassigniﬁcantlyincreased(P<0.01)inthe2DcultureinresponsetoIFN-g treatment. Beforekyn secretion, IDOwas strongly upregulated (P <0.001). IFN-g treatmentalso increasedthesecretionofIL-6andIL-8fromthekeratinocytes.IntheRHE,IDOwasupregulatedbyIFN-g, andkynsecretioncouldbedetected.Interestingly,IDOexpressionwasonlypresentinthebasalcellsof theRHE.

Conclusion: Our results suggestthat IFN-g acts as aninducer of trp degradationpreferentially in undifferentiatedkeratinocytes,indicatedbytheIDOexpressioninthebasallayeroftheRHE.

by/4.0/).

1.Introduction

Theskinprovidesa continuouslyrenewedprotectivebarrier formedbydifferentiatingkeratinocytes.Theoutermostlayerofthe skin,theepidermis,canbedividedintofourlayers:stratumbasale,

stratum spinosum, stratum granulosum and stratum corneum.

These layersare characterized bythe properties and degreeof differentiationof thekeratinocytes,astheundifferentiatedcells movefromthebasallayer totheoutermostcorniﬁedlayer [1].

Keratinocytescanreleaseanti-microbialpeptidesandin

ﬂamma-torycytokineswhenactivatedbyanimmuneresponsecausedby

injury,orstimulatedbyexogenousfactorssuchasUV-radiation[2]. Keratinocytesconstitutivelyproduceandstoretheinﬂammatory cytokine IL-1α in the cytoplasm. Interestingly, this cytokine requiresnofurthercleavageandisreleasedinresponsetoawide rangeofstimuliwhichleadstoarapidlocalactivationofadaptive

and innate immune cells [3]. Furthermore, endogenous IL-1α

interactswiththeIL-1receptorinthecellmembranefollowingits release,resultinginincreasedproductionofIL-1α,togetherwith productionofpro-inﬂammatorycytokinessuchasIL-6,IL-8,and TNF-α[4].

Thecatabolismoftheessentialaminoacidtryptophan(trp)has

of late been recognized as having a fundamental role in the

immunesystem[5].Mostofthefreetrpisdegradedthroughthe

* Correspondingauthorat:MalmöUniversity,FacultyofHealthandSociety, Departmentof BiomedicalScience,Jan Waldenströmsgata25,214 28Malmö, Sweden.

E-mailaddress:anna.gustafsson@mau.se(A.Gustafsson).

https://doi.org/10.1016/j.jdermsci.2020.07.005

ContentslistsavailableatScienceDirect

Journal

of

Dermatological

Science

(2)

kynureninepathway(KP),thatgeneratesanumberofmetabolites

involvedininﬂammation,immuneresponse,and

neurotransmis-sion[6].Theﬁrstandrate-limitingstepofthispathwayiscatalyzed bytheenzymesindoleamine2,3-dioxygenase1and2(IDO1/IDO2) andtrp2,3-dioxygenase(TDO)whichconvertstrptokynurenine

(kyn). IDO1 expression is highly induced by cytokines and

inﬂammatorymoleculessuchastheT-helpercelltype1 (Th1)-derivedcytokineinterferon

g

(IFN-

g

).Therateoftrpdegradation canbeestimatedbythekyn totrpratio(kyn/trp),thevalueof whichcanbeusedasasensitiveindicatorforanactivatedimmune responseoftheskin.

Trpderivativesactaspotentagonistsofthearylhydrocarbon receptor (AhR), a ligand-activated transcription factor that is

expressedinmanyimmune cellsand mediatesawide rangeof

immunomodulatorydevelopments[7,8].AhRisalsoexpressedin alltypesofskincells[9]andplaysacentralroleinskinintegrity

andimmunity,respondingtoexogenousandendogenousstimuli

byinducingorrepressinggeneexpressionsessentialforbasicskin functions[10,11].

Sheipouriet al.haveshown thattheKPisfullyexpressedin humanprimarymonolayer-culturedkeratinocytesinvitroandthat exposuretoUVBradiationand/orIFN-

g

causessigniﬁcantchanges

in the expression pattern of downstream KP metabolites and

enzymes in these keratinocytes [12]. Immersed keratinocyte

monolayersrepresent a commonlyused in vitro model system

to study keratinocyte behavior. However, monolayers lack the

keratinocyte terminal differentiation and the epithelial sheet

formation, which are present in vivo. Studies addressing the

differencesofkeratinocytebehaviorinmorecomplex3D epider-mismodelsareneeded.Furthermore,skindiseasesprovide non-invasiveaccesstotheaffectedtissue,andthetopicalkyn/trpratio isofinterestasitmayserveasabiomarkerofskininﬂammation andcancer.

To gain more insight into the KP in epidermis, we have

successfullyestablishedaninvitromodelofreconstructedhealthy

humanepidermis(RHE)generatedfromprimaryhuman

kerati-nocytes(HEKn).TheIFN-

g

inducedcytokineproﬁleandkyn/trp ratiointhis3Dmodelaswellasinmonolayer-culturedHEKnwere analyzed.HerewedemonstratethatIFN-

g

actsasaregulatorof inﬂammation,preferentiallyinundifferentiatedkeratinocytes.

2.Materialsandmethods

2.1.Cellculture

Primary human epidermal keratinocytes isolatedfrom

neo-natalforeskin(HEKn,ThermoFischerScientiﬁc)weremaintained

in EpiLife growth medium with 60

m

M calcium chloride

supplementedwith1%of humankeratinocytegrowth

supple-ment (HKGS) and 0.2 % gentamycin/amphotericin (Thermo

FischerScienti_ﬁc).1%ofHKGSprovidesbovinepituitaryextract, 0.2%v/v;bovineinsulin,5

m

g/mL;hydrocortisone,0.18

m

g/mL; bovinetransferrin,5

m

g/mL;humanepidermalgrowthfactor,0,2 ng/mL.Cells wereculturedat37_C _in₅_%_CO

2 inahumidiﬁed

atmosphere.Thegrowthmediumwaschangedeveryotherday

untilcellsreached50%conﬂuency,andeverydaythereafter.The

cells were passaged with trypsin-EDTA (0.05 %) at 75–80 %

conﬂuency.

Formonolayercultureexperiments,5104_HEKn_were_seeded

perwellin12-wellplates.After24 h, theculturemediumwas

exchangedfornewmediumsupplementedwithorwithoutIFN-

g

(R&DSystems)indifferentconcentrations(0.5 20ng/mL).After anadditional6,24,48,or72h,theculturemediumwascollected, andaliquotswereimmediatelyfrozenat 80C,awaitingcytokine andtrp/kynquantiﬁcation.Thecellswerelysedandusedfortotal RNA-extraction.

FortheRHE,protocolsdevelopedbyDeVuystetal.[13]and Frankartet al. [14] wereused as the basis for establishing an optimalprotocol.HEKnatpassage1–3wereseededinMillipore

polycarbonate culture inserts, 12 mm diameter and 0.4

m

diameter pore size (2.5 105 cells/insert), and were kept

submergedin EpiLife growthmediumsupplemented with 1 %

ofHKGS,0.2%gentamycin/amphotericin,1.5mMCa2+_,_and₅₀

_m

_g/

mL vitamin C for 48 h. The inserts were then raised and

maintainedattheair–liquidinterfacefor15days,replacingthe

culture medium every other day. After 15 days of air–liquid

interfacecultivation,RHEsamplesconsistingofafully

differenti-ated epidermis were stimulated for 48 72 h with IFN-

g

in

differentconcentrations (0.5 100ng/mL).IFN-

g

wasdilutedin

Epilifeculturemediumandaddedtothemediumpresentunder

theinsert.

2.2.Cellviabilitytest

The cell viability assay was performed using the

3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT)

assay[15](ThermoFischerScientiﬁc).Forthemonolayersamples, thecells wereseeded in 96-wellculture plates ata densityof 1104_cells/well._Cells_were_allowed_to_attach_for₁₂_h_and_then_the

culturemediumwasexchangedfornewmediumwithorwithout

IFN-

g

(0.5 20ng/mL)andthecultivationwascontinuedfor24,48,

and72h.Subsequently,0.5mg/mLMTTwasaddedtotheIFN-

g

treated cells. After 4 h of incubation at 37 C in 5% CO2, a

solubilizationsolutionwas added and thecells wereincubated overnightat37Cin5%CO2.Absorbancewasmeasuredat570nm

usingaplatereader(BioTekInstruments).

FullydifferentiatedRHEwasincubatedwithdifferent

concen-trations of IFN-

g

(1 100 ng/mL) for 72 h then washed with

phosphate-buffered saline (PBS), and further incubated with

300

m

LofMTT(1mg/mL)for3hat37Cina5%CO2incubator.

Theformazanformedwasextractedbyincubationwith2mLof

isopropanol for 2 h. 250

m

L of the extracted solution were

transferred to a 96-well plate, and the optical density was

measuredat570nmusinga platereader(BioTekInstruments).

Treatmentwith5%SDSwasusedasapositivecontrol.

ForeachIFN-

g

concentrationandtimepoint3replicateswere included.Inthepresentationofresults,theviabilityisexpressedas thepercentageofsurvivingcellscomparedwith100%survivalin untreatedcontrolcellsortissues.

2.3.RNAisolation,cDNAsynthesis,andReal-TimePCR

TotalRNAfromHEKnandRHEwasextractedusingtheRNeasy

Mini Kit (Qiagen). The RHE was dissected fromthe insert and

transferredtoa12-wellplatecontaining600

m

lRLTbufferwith1%

β-mercaptoethanol.Theepidermiswashomogenizedbypipetting

up and downin RLT buffer containing 70 % ethanol and then

transferredtoaspincolumn.For HEKnmonolayers,300

m

lRLT bufferwith1%β-mercaptoethanolwasaddedtothecellsin 12-wellplates.ThenRNAwasextractedaccordingtotheinstructions

of themanufacturer.Extracted RNAwassynthesized intocDNA

usingaSensiFastcDNAsynthesiskitfromBioline.PCRwascarried outusingSYBRGreenTechnology(ThermoFisherScientiﬁc).Each 20

m

lsamplecontained5

m

loftemplatecDNA,1

m

lofeachprimer (10

m

M),10

m

lSYBR MIX (2 Sensifast, Bioline) and 3

m

lof nuclease-freewater.EachreactionwasrunintriplicateonaLC480

LightCycler from Roche using the following program: Initial

denaturationat95Cfor10min,followedby45cyclesconsisting of95Cduring10sformelting,65Cfor10sannealingand72C

for 11 s extension. The program ended with a melting

curve analysis measuring ﬂuorescence continuously from

(3)

DD

-CT-method, using glyceraldehyde 3-phosphate dehydroge-nase(GAPDH)asthereferencegene.Primerswereasfollows:

IDO-1: fwd: GAAAGGCAACCCCCAGCTAT. rev:

GGAGGAACT-GAGCAGCATGT

AhR: fwd: ACGGAGGCCAGGATAACTGT. rev:

AATGCCCAGTG-TAGCTGACG

GAPDH: fwd: AACAGCGACACCCACTCCTC. rev:

GGAGGGGA-GATTCAGTGTGGT

2.4.Histologicalanalysisandimmunohistochemistry(IHC)of reconstructedepidermis

Initially,theRHEwasﬁxedinformalinaceticacid(4%formalin and1%glacialaceticacid)for24hatroomtemperature,followed

by dehydration in four baths of methanol. To dissolve the

polycarbonate filter, the insert was stirred in pure toluene for 30sandthenincubatedfourtimesfor10minintoluene.Following removalof thefilter,theRHEwasembeddedinparaffin. Tissue sectionsof3

m

mwerestainedinhematoxylin-eosin.

TheexpressionofIDO-1inRHEwasevaluatedbyIHC.Tissue sections(3

m

m)weredeparafﬁnizedandrehydratedbefore heat-inducedantigenretrievalin10mMcitratebuffer,pH6.0,for40min at95CinaDecloakingChamberTM(NxGen,BiocareMedical).The

sections were incubated with Background Punisher (MACH4,

Universal HRP-Polymer Detection System, BRI4012 L, Biocare

Medical)for10minbeforeovernightincubationwithmonoclonal

mouse anti-human IDO-1 (R&D Systems) diluted 1/2000 with

antibodydiluent(DAKO).Peroxidase–BlockingSolution(DAKO)

was added tothesectionsfor 10 minprior toincubation with

mouse probe (MACH4) for 15 min, followed by a secondary

antibody(goatanti-rabbit-HRP,MACH4)for30min.Thesections

werethenincubatedwithdiaminobenzidine(DAB,MACH4)for5

min,counterstainedwithhematoxylin,dehydrated,andmounted. Aftereachstep,theslideswerewashedwithTris-bufferedsaline, 0.1%Tween20,pH7.6. Asnegativecontrol,N-Universalmouse negativecontrol(DAKO)wasused,andsectionsoftonsilwasused aspositivecontrolforIDO-1expression.

2.5.DetectionofcytokineswithELISA

TheextracellularreleaseofIL-1α,IL-6,andIL-8wasdetermined intheculturemediumusingquantitativesandwichimmunoassays

(R&D systems) according to manufacturer recommendations.

Assays were carried out in ﬂat-bottom 96-well immunoplates

(MaxiSorp, Nunc). The detection limits of the assays were

1.0pg/mL(IL-1α),0.7pg/mL(IL-6),and7.5pg/mL(IL-8). 2.6.DetectionofkynurenineandtryptophanwithHPLC

A High Performance Liquid Chromatography (HPLC) system

equipped with a G1312A model binary pump, G1322A in-line degasser,

G1316Acolumnoven,G1313AautosamplerandG1315diodearray

detector(Agilent1100Series,Germany)wasusedfortrpandkyn quantiﬁcationintheculturemedium.Priortotrpandkynanalysisby HPLC,culturesupernatantswereloadedoncentrifugal_ﬁlterswith

molecularweightcut-off(MWCO)of10kDa(PES modiﬁed;VWR

International)andcentrifugedfor15minat5000xg,at10C.30

m

lof thesupernatantwasseparatedon2504.6mmi.d.KromasilC18 analyticalcolumn,particlesize5

m

m(AkzoNobel), usinggradient elution based on solvent A (10 mM of NaH2PO4 at pH of 2.8, Merck) and solventB (100 % methanolofHPLC gradientgrade, VWRInternational). Aseparation profileataflowrateof0.9ml/minat40Cwasrealizedas follows:20%Bwasheldfor5min,thenincreasedto50%Bin0.5min, heldfor2.5min;thensolventBwasfurtherincreasedto90%Bin0.5 minandheldfor4.5min;finallysolventBwasdecreasedto20%Bin 0.5minandadditionallykeptat20%Bfor2.5min.

Trpwasmonitoredat280nmandkynat360nm.Thepeaksof kynandtrpwereidentiﬁedbycomparingwiththeretentiontimes

ofpreviouslydeterminedstandardcompounds(L-tryptophanand

L-kynurenine,reagentgrade>98%,SigmaAldrich).Amountofthe compoundswascalculatedbyintegrationofthepeakareasusing theOpenLABsoftware(LabAdvisorBasicSoftware).Allintegrated

peaks were inspected manually. Quantiﬁcation was based on

externalcalibrationwherestandardcurvesforbothtrpandkyn, rangingfrom0to50

m

Mwererecorded.Standardsandcellculture mediumsampleswererunintriplicates.Theconcentrationoftrp

in thesampleswas estimatedbysubtractingtheamountof trp

presentinthecellmedium. 2.7.Statistics

All measurement data are expressed as mean standard

deviation(SD),derivedfromtheindicatednumberofindependent experiments.Allexperimentswererepeatedatleastthreetimes, indicatedineach ﬁgure.Thesigniﬁcanceofany differenceswas analyzedwiththeWilcoxonsignedranktest.Statisticalanalyses

Fig.1.A)MTT-basedviabilityofmonolayerHEKnafter24,48,and72htreatmentwithIFN-g.DatarepresentthemeansSDofquintuplicatedeterminationsfromthree individualexperiments.B)Phasecontrastimages(x20magniﬁcation)ofmonolayerHEKn.Imagescapturedafter48htreatmentwithindicatedconcentrationsofIFN-g.

(4)

were performed by IBM SPSS. Differences were considered as statisticallysigniﬁcant when P was less than 0.05 (P < 0.05); *P0.05**P0.01***P0.001.

3.Results

3.1.IFN-

g

treatmentaffectcellmorphologyandcauseadecreaseinthe thicknessofstratumcorneum

IFN-

g

treatmentinconcentrationsrangingfrom5 20ng/mL

inducedaconcentrationandtimedependentmoderatechangein

cell viability of monolayer HEKn (Fig.1A). Additionally, IFN-

g

treatment in concentrations above 5 ng/mL also caused an

alteration of the growth of HEKn cells, resulting in a more

elongatedcellmorphology(Fig.1B).

After 15 days of culture at theair-liquid interface, theRHE

showed normal stratiﬁcation with well-represented epidermal

layers;thebasallayerwascomposed ofcolumnarkeratinocytes

attachedtothepolycarbonatemembraneandcoveredbyspinous

andgranularlayers(Fig.2).Intheoutersurfacethereweredead, flattened, cornified, nucleus-free keratinocytes. Expression of stratificationmarkerCK10wasdeterminedbyIHCandtheresult

showed that RHE prepared with HEKn expressed CK10 in the

suprabasal layers (supplementary data). IFN-

g

treatment in

concentrations above 10 ng/mL caused a decrease in stratum

corneumthicknesscomparedtothecontroltissue(Fig.2).IFN-

g

treatmentwith100ng/mLreducedtheviabilityby26%(P=0.005) compared tothe control tissue. Notably, IFN-

g

treatment with concentrations50ng/mLshowedtissueviabilityvalues compa-rabletothoseofthecontroltissue(Fig.2).

3.2.AlteredcellularreleaseofcytokinesIL-1α,IL-6,andIL-8in stimulatedmonolayercellsandRHE

HEKncellsrespondtoIFN-

g

treatmentbyincreasedreleaseof

IL-6 and IL-8 into the culture medium in a dose- and

time-dependentfashion(Fig.3AandC).However,HEKnbasallevelIL-1α secretionwasnotincreasedinthepresenceofIFN-

g

(Fig.3B).The releaseofIL-1α,IL-6andIL-8wasalsoquantiﬁedinthemedium

from IFN-

g

treated RHE. As seen in Fig. 3D, only the highest concentrationofIFN-

g

(100ng/mL)couldsigniﬁcantlyinduceIL-8, IL-6andIL-1αreleasefromRHE.

3.3.Increasedkyn/trpratiosinstimulatedHEKnandRHE

Changesofkynandtrpconcentrationsintheculturemedium wasquantiﬁedwithHPLC.Trpcouldbequantiﬁedintheculture

medium from both untreated monolayers and RHE. When

stimulatedwithIFN-

g

,dose-dependentconversionoftrptokyn inmonolayerculturedHEKnwastakingplaceasshownbythekyn/ trp ratio (Fig. 4). The increase in kyn amounts was small but

signiﬁcant (P 0.01), levels of kyn were 3.3 0.05

m

M in

monolayersstimulatedwith7.5ng/mLIFN-

g

during72h. Kyncouldnotbedetectedinculturemediumunderneaththe

RHEwhenstimulatedwithIFN-

g

of10ng/mL.However,higher

concentrationsofIFN-

g

wastestedtoinducetrpconsumptionin

the RHE model and when stimulated for 72 h with IFN-

g

in

concentrationsbetween20 100ng/mL,kyncouldbedetectedbut levelswerebelowlimitofquantiﬁcation(datanotshown). 3.4.IFN-

g

increaseIDO-1geneexpression

Using realtime-PCR,wedemonstrated thatIFN-

g

signiﬁcantly

induced IDO-1 mRNA expression in HEKn (Fig. 5A). After 24 h

stimulationwith0.5ng/mLIFN-

g

,transcriptionlevelsofIDO-1was increased129timesinHEKncells(Fig.5A).TheexpressionofAhRwas notsigniﬁcantlyaffected byIFN-

g

administrationafter6 or24h (Fig.5B).

TheIDO-1andAhRexpression inIFN-

g

treatedRHEwas also determined withPCR.After 48hofstimulation with50and100ng/mL IFN-

g

,thelevelofIDO-1wassigniﬁcantlyincreased,3and23times, respectively (Fig. 5C). The gene expression of AhR was also signiﬁcantly increasedafter48hstimulationwith100ng/mLIFN-

g

.

3.5.IFN-

g

upregulateIDO-1expressioninthebasallayerofRHE

The expression of IDO-1 was determined by IHC and was

noticedonlyinbasalcellsofRHEtreatedwith50or100ng/mLof

Fig.2.HistologyofRHEatday15.TheRHEwastreatedwithIFN-gfor48h(day14and15);3mmtissuesectionsstainedwithhematoxylin-eosin,scalebars100mm.The histogramshowsMTT-basedviabilityoftheRHEaftertreatmentwithIFN-gfor72h(day14-16ofcultureatair-liquidinterface).Themeanopticaldensityat570nmofthe untreatedcontroltissuesexposedtoculturemediumwassettorepresent100%viabilityandtheresultsareexpressedaspercentageofthesecontrols.Datarepresentthe meansSDofduplicatedeterminationsfromthreeindividualexperiments.

(5)

IFN-

g

(Fig. 6). The 48 h IFN-

g

treatment in the lower range, 10ng/mL,didnotinduceanyIDO-1expressioninthekeratinocyte layersofRHE.

4.Discussion

Todate,manycellularevents,includingcytokinesignalingand trpmetabolismindifferentiatingkeratinocytesexposedtoIFN-

g

arenotyetfullyunderstoodoridentiﬁed.Inordertogenerateand maintainthevitalepidermal barrier, thekeratinocytes undergo continuousproliferationanddifferentiation.Although2D mono-layer-basedculturesarecommonlyusedasinvitromodelsystems tofullyunderstandkeratinocytebehavior,keratinocyteterminal differentiationandorganizationoftheepidermaltissueareabsent inthemonolayercultures.Therefore,theRHEisarelevantmodel forinvestigatingthephysiologicalkeratinocyteresponseinvitro.

To reconstruct the human in vitro epidermis model, we used

polycarbonate cellcultureinsertsonwhich thenormal primary keratinocytesHEKncouldattachandthenstratifyinordertoform thetypicalepidermallayers.ThepropertiesoftheRHEarevery similartothe3Ddifferentiatedhumanepidermisinvivo[13,14].

We showfor theﬁrst time, that trp conversionto kynwas

increasedinRHE,inresponsetotheimmunomodulatorycytokine IFN-

g

. IFN-

g

also upregulated gene and protein expression of IDO-1inthebasallayeroftheepidermis.Ourresultsindicatethat maturekeratinocytesdonotproduceIDO-1inresponsetoIFN-

g

. Inundifferentiatedmonolayered-culturedHEKn,IFN-

g

regulated thetranscriptionlevelofIDO-1byinducingasubstantialincrease. FollowingIDO-1expression,trpcatabolismtokynwassigni ﬁcant-lyincreased.ToinducearesponseintheRHEmodel,additionofup

to 20 times higher concentrations of IFN-

g

was required. The

differentresponseistentativelyduetoHEKncellsmimicingbasal

Fig.3. ExtracellularreleaseofA)IL-6,B)IL-1α,C)IL-8frommonolayerculturedHEKn.Thetracesindicatecytokineconcentrationsafter6,24,48,and72htreatmentwith indicatedconcentrationsofIFN-g.D)IL-6,IL-8,andIL-1αreleasefromRHEafter48hofIFN-gstimulation.DatarepresentthemeansSDofduplicatedeterminationsfrom threeindividualexperiments.*P0.05**P0.01***P0.001vsuntreatedcells.

Fig.4.Extracellularconcentrationsofkynurenine(blackbars)andtryptophan(greybars)inculturemediumfrommonolayerculturesofHEKn.Thecellsweretreatedwith IFN-gandwereincubatedfor72hat37_C_before_HPLC_analysis._Data_represent_the_means_SD_of_duplicate_{determinations}_from_three_individual_experiment._**P_0.01_***P 0.001vsuntreatedcells.

(6)

Fig.5. RelativeexpressionofIDO-1(A)andAhR(B)inmonolayer-culturedHEKncells.DatarepresentthemeansSDoftriplicatedeterminationsfromthreeindividual experiments.CrepresentstherelativeexpressionofIDO-1andAhRinRHE48hafterIFN-gtreatment.DatarepresentthemeansSDoftriplicatedeterminationsandis representativeforthreeseparateexperiments.HEKnandRHEweretreatedwithIFN-gatindicatedconcentrationsandwereincubatedat37_C_before_RNA_extraction. *P0.05**P0.01***P0.001vsuntreatedcells.

Fig.6.DetectionofIDO-1withimmunohistochemistrystainingofRHE.IDO-1expressionwasobservedinbasalcells(indicatedwitharrows)afteradministrationof50and 100ng/mLIFN-g(48h).Dataarerepresentativeofthreeseparateexperiments.

(7)

cells,whereasRHEmimicsthewholeepidermis.Thus,theRHEcan

be considered as a more reliable physiological model in vitro

comparedtomonolayer-culturedkeratinocytes.

TheKPhasbeendescribedinskin-derivedkeratinocytesand ﬁbroblasts[12],butstudiesofKPinskinmodelsorskinbiopsies

arelimited.TheKPhasbeenassociatedwithimmunetolerance

andtumorescapemechanismsbytheincreasedexpressionofits rate-limitingenzymeIDO[16].IDOinducesimmunosuppressionin twoways.Firstly,IDOenablesimmuneescapebydepletingtrpin

tumormicroenvironments.Secondly,thekynproducedbyIDOis

anactiveimmunesuppressant,e.g.byinducingregulatoryT-cells viakynbindingtoAhR[17].StudiesexaminingIDOexpressionin skinarealsoverylimited.Zhangetal.[18]exploredthedifference

of IFN-

g

induced IDO-1 expression in epidermis and dermis,

respectively,inaskintissueculturemodel.IDO-1wasexpressed mainlyin the basal layer of epidermisand rarely expressed in

dermis,asshownbybothimmunoﬂuorescenceand

immunohis-tochemistry. However, Scheler et al. [19] investigated IDO-1

expression in skin biopsies from different inﬂammatory skin

disorders,andshowedthat IDO-expressingcells couldbefound exclusivelyin thedermalinﬂammatoryinﬁltrateandnotin the epidermis.

Apart from kyn production, we also studied the cytokine

proﬁleresponseinRHEinducedbyIFN-

g

.Importantly,ourRHE

modelcannotbeusedtostudythecompleteimmuneresponse

since the model is solely composed of keratinocytes, and the

absenceofotherepidermalcelltypesreducesthecomplexityof thesystem.Besideskeratinocytes,othercelltypessuchasMerkel cells,melanocytes,andimmunecells,includingLangerhanscells

and resident memory CD8 + T cells, are present in normal

epidermis[20].However,thekeratinocytesaretheprincipalcell

typein humanskin andrepresent ﬁrst-line responders in the

tissue during skin immunological reactions. It has previously

beenshownthatIFN-

g

iscapableofbroadlyregulatinggenomic

expression in normalprimarykeratinocytes, by demonstrating

morethan3500genetranscriptsinducedbyIFN-

g

[21].IFN-

g

is

secreted by inﬁltrated Th1 lymphocytes in epidermis during

inﬂammation.WhenboundtotheIFN-

g

receptor(IFNGR)onthe keratinocyte,signalingthroughtheJAK-STATpathwayisinitiated [22]. Our results indicate that IFN-

g

treatment increase the

secretionofIL-6and IL-8in HEKnmonolayers.However,these

cytokines could not be detected in stimulated RHE, unless

concentrationsofIFN-

g

upto100ng/mLwasused.ThehighIFN-

g

concentrationsalsoaffectedtissueviability.Itisknownthat

IFN-g

playsaroleintheregulationofproliferationanddifferentiation intheepidermisandhasalsobeenreportedtoinhibitgrowthin culturedkeratinocytes[23].Thedecreasedthicknessofstratum

corneumin RHEis most probabledue tothereductionof cell

viability. The increased cytokine secretion in connection with decreased viability canprobably be explainedby the fact that dyingkeratinocytes secrete alarm signals i.edanger-associated

molecularpatterns(DAMPs)that inducecytokineproductionin

theviablekeratinocytes [24]. In addition,IL-1αis storedin the cytoplasmandthecellmembraneinkeratinocytes,andisreleased whencells aredamaged[25]. Also,IL-1αplaysa centralrolein initiating inﬂammatoryresponse inskinand is consideredasa potentinducer of IL-6 and IL-8 [26]. Interestingly,it has been shownthattheIFNGRisexpressedinalllayersofdifferentiated epidermisinhealthyskin,butonlyinthebasallayerofepidermal psoriatic lesions [27]. Therefore, we hypothesize that the low

cytokine-response from IFN-

g

stimulated RHEcould bedue to

limitedIFNGRexpression.

Inconclusion,ourresultsindicatethattheKPisonlyactivein undifferentiatedkeratinocytes andnot inmature,differentiated keratinocytes. In undifferentiated basal cells of the RHE, IFN-

g

upregulates the transcription level of IDO-1. Following IDO-1

expression, trp conversion to kyn was increased. Our study

establishestheimportanceofusingfullydifferentiated keratino-cytesina3Dmodelforinvestigatingthephysiologicalkeratinocyte responseinvitro.

DeclarationofCompetingInterest

Theauthorshavenoconﬂictofinteresttodeclare. Acknowledgments

ThisworkhasbeensupportedbytheKnowledgeFoundation,

the Royal Physiographic Society of Lundand BioﬁlmsResearch

CenterforBiointerfacesMalmöUniversity,Sweden.

AppendixA.Supplementarydata

Supplementary material related to this article can be

found, intheonline version, at doi:https://doi.org/10.1016/j. jdermsci.2020.07.005.

References

[1]S.Agarwal,K.Krishnamurthy,Histology,Skin,StatPearls,StatPearlsPublishing L.L.C.,TreasureIsland(FL),2019.

[2]D.Sheipouri, N.Braidy, G.J.Guillemin,Kynureninepathwayinskincells: implicationsforUV-inducedskindamage,Int.J.Tryptoph.Res.:IJTR5(2012) 15–25.

[3]M.M.Klicznik,A.B.Szenes-Nagy,D.J.Campbell,I.K.Gratz,Takingthelead -howkeratinocytesorchestrateskinTcellimmunity,Immunol.Lett.200 (2018)43–51.

[4]C.A. Dinarello, Interleukin-1, interleukin-1 receptors and interleukin-1 receptorantagonist,Int.Rev.Immunol.16(5–6)(1998)457–499.

[5]J.C.Mbongue,D.A.Nicholas,T.W.Torrez,N.S.Kim,A.F.Firek,W.H.Langridge, Theroleofindoleamine2,3-dioxygenaseinimmunesuppressionand autoimmunity,Vaccines3(3)(2015)703–729.

[6]T.W. Stone, N.Stoy, L.G.Darlington, Anexpanding range of targets for kynureninemetabolitesoftryptophan,TrendsPharmacol.Sci.34(2)(2013) 136–143.

[7]J.E.Cheong,L.Sun,TargetingtheIDO1/TDO2-KYN-AhRpathwayforcancer immunotherapy-challengesandopportunities,TrendsPharmacol.Sci.39(3) (2018)307–325.

[8]R.J.Kewley,M.L.Whitelaw,A.Chapman-Smith,Themammalianbasic helix-loop-helix/PASfamilyoftranscriptionalregulators,Int.J.Biochem.CellBiol.36 (2)(2004)189–204.

[9]M.Napolitano,C.Patruno,Arylhydrocarbonreceptor(AhR)apossibletarget forthetreatmentofskindisease,Med.Hypotheses116(2018)96–100. [10]M.S.Denison,A.A.Soshilov,G.He,D.E.DeGroot,B.Zhao,Exactlythesame

butdifferent:promiscuityanddiversityinthemolecularmechanismsof actionofthearylhydrocarbon(dioxin)receptor,Toxicol.Sci.124(1) (2011)1–22.

[11]T.Haarmann-Stemmann,C.Esser,J.Krutmann,Thejanus-facedroleofaryl hydrocarbonreceptorsignalingintheskin:consequencesforpreventionand treatmentofskindisorders,J.Invest.Dermatol.135(11)(2015)2572–2576. [12]D.Sheipouri,R.Grant,S.Bustamante,D.Lovejoy,G.J.Guillemin,N.Braidy,

Characterisationofthekynureninepathwayinskin-derivedﬁbroblastsand keratinocytes,J.Cell.Biochem.116(6)(2015)903–922.

[13]E.DeVuyst,C.Charlier,S.Giltaire,V.DeGlas,C.L.deRouvroit,Y.Poumay, Reconstructionofnormalandpathologicalhumanepidermison

polycarbonateﬁlter,MethodsMol.Biol.(Clifton,N.J.)1195(2014)191–201. [14]A.Frankart,J.Malaisse,E.DeVuyst,F.Minner,C.L.deRouvroit,Y.Poumay,

Epidermalmorphogenesisduringprogressiveinvitro3Dreconstructionatthe air-liquidinterface,Exp.Dermatol.21(11)(2012)871–875.

[15]K.Berg, M.B. Hansen,S.E.Nielsen,Anewsensitive bioassay forprecise quantiﬁcationofinterferonactivityasmeasuredviathemitochondrial dehydrogenasefunctionincells(MTT-method),APMIS:ActaPathol.Microbiol. Immunol.Scand.98(2)(1990)156–162.

[16]F.A.Boros,L.Vecsei,Immunomodulatoryeffectsofgeneticalterationsaffecting theKynureninePathway,Front.Immunol.10(2019)2570.

[17]A.Curti,S.Pandolﬁ,B.Valzasina,M.Aluigi,A.Isidori,E.Ferri,V.Salvestrini,G. Bonanno,S.Rutella,I.Durelli,A.L.Horenstein,F.Fiore,M.Massaia,M.P. Colombo,M.Baccarani,R.M.Lemoli,Modulationoftryptophancatabolismby humanleukemiccellsresultsintheconversionofCD25-intoCD25+T regulatorycells,Blood109(7)(2007)2871–2877.

[18]Y.Zhang,Y.Wang,G.Wu,W.Zhang,X.Wang,W.Cai,J.Zhang,S.Han,Y.Li,X. Bai,J.Shi,L.Su,D.Hu,ProlongedskingraftssurvivaltimebyIFN-gammain allogeneicskintransplantationmodelduringacuterejectionthrough IFN-gamma/STAT3/IDOpathwayinepidermallayer,Biochem.Biophys.Res. Commun.496(2)(2018)436–442.

(8)

[19]M. Scheler,J. Wenzel, T. Tuting,O. Takikawa, T. Bieber,D. von Bubnoff, Indoleamine2.,3-dioxygenase(IDO):theantagonistoftypeIinterferon-driven skininﬂammation?Am.J.Pathol.171(6)(2007)1936–1943.

[20]T.Gebhardt,U.Palendira,D.C.Tscharke,S.Bedoui,Tissue-residentmemoryT cellsintissuehomeostasis,persistentinfection,andcancersurveillance, Immunol.Rev.283(1)(2018)54–76.

[21]K.E.Nograles,L.C.Zaba,E.Guttman-Yassky,J.Fuentes-Duculan,M. Suarez-Farinas,I.Cardinale,A.Khatcherian,J.Gonzalez,K.C.Pierson,T.R.White,C. Pensabene,I.Coats,I.Novitskaya,M.A.Lowes,J.G.Krueger,Th17cytokines interleukin(IL)-17andIL-22modulatedistinctinﬂammatoryand keratinocyte-responsepathways,Br.J.Dermatol.159(5)(2008)1092–1102. [22]K.Schroder,P.J.Hertzog,T.Ravasi,D.A.Hume,Interferon-gamma:anoverview

ofsignals,mechanismsandfunctions,J.Leukoc.Biol.75(2)(2004)163–189. [23]N.A.Saunders,A.M.Jetten,Controlofgrowthregulatoryand

differentiation-speciﬁcgenesinhumanepidermalkeratinocytesbyinterferongamma.

Antagonismbyretinoicacidandtransforminggrowthfactorbeta1,J.Biol. Chem.269(3)(1994)2016–2022.

[24]T.A.Wilgus,Alertingthebodytotissueinjury:theroleofalarminsandDAMPs incutaneouswoundhealing,Curr.Pathobiol.Rep.6(1)(2018)55–60. [25]A.Coquette,N.Berna,A.Vandenbosch,M.Rosdy,B.DeWever,Y.Poumay,

Analysisofinterleukin-1alpha(IL-1alpha)andinterleukin-8(IL-8) expressionandreleaseininvitroreconstructedhumanepidermisforthe predictionofinvivoskinirritationand/orsensitization,Toxicol.InVitro17 (3)(2003)311–321.

[26]J.M.Schroder,Cytokinenetworksintheskin,J.Invest.Dermatol.105(1Suppl) (1995)20S–24S.

[27]A. Scheynius, J. Fransson,C. Johansson, H.Hammar, B. Baker, L. Fry, H. Valdimarsson,Expressionofinterferon-gammareceptorsinnormaland psoriaticskin,J.Invest.Dermatol.98(2)(1992)255–258.