Immunosuppressive activity of 13-cis retinoic acid and prevention of experimental autoimmune encephalomyelitis in rats.

Immunosuppressive Activity of 13-cis-Retinoic Acid and Prevention

of

Experimental

Autoimmune

Encephalomyelitis in Rats

LucaMassacesi, EmanuelaCastigli, Marco Vergelli, Jacopo Olivotto,Anna LauraAbbamondi,*

FrancescoSarlo, andLuigi Amaducci

Dipartimentodi ScienzeNeurologicheePsichiatricheUniversitddegliStudi diFirenze, 50134Florence, Italy; *Istituto diNeurologia, UniversitdCattolica, PoliclinicoGemelli, 00100 Rome,Italy

Abstract

Someactivitiesof retinoids on cellular and humoralimmunity have been described, but the available data areconflictingor obtainedatconcentrationsthat aretoxicinvivo. In this study, wedemonstrate that13-cis-retinoic acid(13-cRA),aretinoid welltolerated in human therapy, can suppress Tcell-mediated immunity inrats. Treatment withpharmacological concentra-tions of 13-cRA prevented active as well as passive transfer experimental autoimmune encephalomyelitis (EAE) and sup-pressedlymphocyte responsivenessto T cellmitogens, suggest-ing that the drugactivity includedsuppression ofaneffectorT cellresponse. Inaddition,mitogen- andantigen-induced lym-phocyteproliferationwasinhibitedinvitro in thepresence of concentrations of 13-cRA equivalent to or less than those achievedinvivo, further suggestingthatthepreventionofEAE was due to asuppressive activityon Tcell-mediatedimmunity. Theimmunosuppressive activity of 13-cRA included suppres-sion ofinterleukin 2,whoseproductionwasinhibitedin spleno-cytes.These dataindicate that,in anin vivo mammalian sys-tem,13-cRAexerts a

suppressive

activityonTcell-mediated immunityintensive enoughtosuppress anongoingimmune

re-sponse,andthatthiseffectcanbe achievedatnontoxic concen-trations thatmayalsobeattained in humantherapy.(J.Clin. Invest.1991.88:1331-1337.) Key words:

experimental

autoim-muneencephalomyelitis*centralnervoussystem *isotretinoino Tcell-mediated immunity*immunotherapy

Introduction

Retinoids

arethe classof natural and

synthetic

derivativesof vitamin A. The best known molecule ofthisclass,

all-trans-retinoic

acid

(RA),V

plays

a

physiological

rolein

morphogenesis

Apreliminary reportofoneexperimentcovered in this articlewas

publishedas anextendedabstract in theproceedingsofthe1989 meet-ingofEuropeanCommitteefor TreatmentinMultiple Sclerosis.

AddressreprintrequeststoDr.Massacesi,Dipartimentodi Scienze

NeurologicheePsichiatriche,VialeMorgagni 85,50134Firenze,Italy. Received forpublication5February1991and inrevisedform30

April1991.

1. Abbreviations usedinthispaper:CFA, completeFreund'sadjuvant; EAE,experimentalautoimmuneencephalomyelitis;

ILk2sup,

interleu-kin 2supernatant(s);MBP,myelinbasicprotein;p.i., postimmuniza-tion;RA, all-trans-retinoicacid; 13-cRA, 13-cis-retinoic acid; RAR,

RA receptor;SC, spleen cell;SI,stimulationindex.

and celldifferentiation.Its 13-cis-isomer

(13-cis-retinoic acid,

13-cRA),which is better toleratedforsystemic

administration,

is also used in the therapyofvarious dermatologic and neoplas-tic diseases (1-3).

RA(tretinoin), 13-cRA(isotretinoin), andotherretinoids havebeendescribedtoexert anumber

of

suppressive activities

onTcellfunctions,

including

adecreasein

production

of

inter-leukins (IL, 4-15). However,

immunostimulatory effects

of 13-cRA and otherretinoids oncellular and humoral

immunity

havealsobeenreported(16-24). These

observations

indicate thattheinteractions

of

thesemolecules

with

theimmune sys-tem arecomplex andnot

completely

clarified.

Nonetheless,

the established and

potential

applications of retinoids

in human therapy demand deeper

investigation

of their effects onthe immune system, particularly at

pharmacological

concentra-tions.

Inaprevious articlewereported thatexperimental autoim-mune encephalomyelitis (EAE) was

suppressed

by RA (25). EAEisanautoimmune diseaseofthecentral nervous system (CNS) inducible in

susceptible

animals by immunization with myelin antigens orbypassive transfer of sensitizedTcells to

syngenic

recipients (26-34). EAE induced in Lewisrats has been usedas a

generic

invivo model ofautoimmune disorders, becauseofthe easeof detecting andquantitating disease activ-ity and the consistency with which thisratstraindevelops a T cell-mediated delayed-type

hypersensitivity-like autoimmune

response(26,27). In rodents, EAEismediatedby

CDV4,

CD8-, andIL-2 receptor-positiveTlymphocyteswhich induce peri-vascularinflammatory infiltrates inCNS by the end ofthe 2nd wkpostimmunization (p.i.), inaccordancewith the onsetof neurological signs, whereas encephalitogenic T lymphocytes canberecoveredfrom

peripheral

lymphoid

organs or

periph-eral blood,

beginning

1 wk

p.i.

(30-34). In

passive

transfer

EAE,CNSinflammatory infiltratesandneurologicalsignsare present 5 or6 dpost-transfer of

encephalitogenic

T lympho-cytes(30-34).

The

suppression of

EAEbyRA indicatesa

possible

sup-pressive activity of this drug on

immune-mediated diseases,

butits

usefulness

waslimited by its

toxicity,

and

therefore

its

possible immunosuppressive activity

in the

suppression

of EAE was notstudied further (25). However,13-cRA, which has been demonstratedtoretainmostofthe

pharmacological

activ-itiesofRA,is better tolerated forinvivoadministration

(1-3).

Thisfindingsuggeststhat 13-cRAcouldrepresentan

optimal

retinoid for

experimentation

inautoimmune

diseases,

but the

activity

of

pharmacological

dosesof this

drug

onT cell-me-diated

immunity

hasnotbeen

directly

investigated (4).

Forthesereasons, we

studied

the invivo effects

of

nontoxic pharmacological

concentrations

of13-cRA inEAE, while the

activity

of the drugonT

cell-mediated

immunity

was

studied

insplenocytes in vitro.Wereportthat13-cRA

prevented

EAE

J.Clin.Invest.

©TheAmericanSocietyfor ClinicalInvestigation,Inc.

0021-9738/91/10/1331/07 $2.00

andsuppressedTcellresponse duringthe efferent phase of the immune response. The immunosuppression was functional

andreversible, and was obtained atconcentrations that canbe

achieved in human therapy, without inducing major side

ef-fects.

Methods

Preparation

of

the

antigens

Myelinandmyelinbasicprotein(MBP) werepurified from guinea pig

spinal cords (Pel Freeze Biologicals, Rogers, AR) according to the methodof Norton and Poduslo (35) and Deibler et al. (36),

respec-tively.Thepreparationswerelyophilized, and stored at -20C.

Induction

of

EAE

FemaleLewisrats(175-225 g; Charles River Breeding Laboratories, Inc.,Wilmington,MA) wereimmunized with an emulsion of either (a) 1 mgofguineapig myelin, dispersed in 0.5 ml of complete Freund's

adjuvant (CFA) containing 3 mg/ml of Mycobacterium tuberculosis H37RAor(b) 50

Asg

ofguinea pig MBP in0.1ml of CFA. Theanimals

wereinjected intradermally in both hind foot pads. As "normal"

con-trol, six additional rats were sham-immunized with 0.5 ml of CFA

only.

Assessment

of

EAE

Twoexperimentsweredesigned in order to follow the clinical signs, andtodescribe laboratoryparametersatday 12p.i. In experiment 1, animalswereassesseddailyforweightandclinical signs for 3 wk.

Clini-caldiseaseseveritywasgradedon ascaleof 1-5asfollows: 0, no signs; 1,tail paralysis; 2, tail paralysis and hind leg weakness; 3, hind leg

paralysis; 4, tetraplegia; 5, death. Inexperiment 2, the animals were killedatday 12p.i. in ordertocomparelaboratory parameters and clinicalsigns.Spinalcord andbrainwereexamined for the presence of

perivascularcuffs. Theintensityof theperivascularinfiltrates was

arbi-trarily gradedfrom 1to3bytwoindependentobservers. TotalDNA

concentrationwasassayedin thespinalcord,accordingtoSmith et al.

(37), as aparameterofinflammatory cellinfiltrates. At the time of

sacrifice,peripheralbloodsampleswerecollected. Absolute and rela-tivelymphocytecounts werecarriedoutusinganhemocytometer and Giemsa stained bloodsmears.

Spleen

cell

(SC)

response

to

Con

A

and

antigens

Spleenswerecollectedatday 10p.i.andpassedthroughastainless steel screenmesh.Thesinglecellsuspensionswereseparatedfrom erythro-cytesinaFicoll-Hypaque density gradient centrifugationand washed in HBSS(Flow),resuspendedinafinalconcentration of 2X 106/mlof

RPMI1640(Flow) supplementedwith 5%heat-inactivatedFCS(Flow

Laboratories, Inc., McLean, VA), 1% L-glutamine, 1% Na-pyruvate

(Flow Laboratories, Inc.), 1.25% Hepes buffer (Flow Laboratories, Inc.), 1%nonessential aminoacids,1%penicillin-streptomycin(Flow

Laboratories,Inc.), and 5XI0-M2-mercapoethanol (2-ME). Prolifer-ative assays in responseto6.25gg/mlmyelinand50

tg/ml

MBPandto 4;tg/mlConA,werecarriedoutin96-microwellplatesusing2X I0O

cells per well.[3H]thymidineincorporationwasmeasuredby scintilla-tioncountingand themagnitudeof theproliferativeresposewas

ex-pressedasmeancountsperminute(cpm)or asstimulation indexes (SI; stimulated culturecpm/unstimulatedculturecpm).

Preparation

of

IL-2-containing

supernatant

(IL-2sup)

SCwereremovedfromLewisratsandculturedat5X106/mlin pres-enceof4,ug/mlConA aspreviouslydescribed (38).After 72 h the

IL-2supwereremoved, treated with 20 mg/mla-methyl-mannoside,

andfrozen.

IL-2assay

TheIL-2activityoftheIL-2supwasassayedusingthe murine IL-2--de-pendent cell line,CTL-2(American Type Culture Collection,

Rock-ville,MD)aspreviously described(39). Briefly 104 cells per well were addedto200,AI of culture medium containing increasing dilutions of IL-2sup. 48 h laterthe cultureswerepulsed with[3H]thymidinefor an additional 18 h. Data fromtwoseparateexperimentswerefitted using a programfor analysis of dose-responsecurves(40).Apossible CTLL-2 response toIL-4 wascontrolledinaseparate assayand excluded: up to 50 U/mlofmouseIL-4didnotelicit CTLL-2 proliferation.

Passive

transfer of

EAE

SensitizedSC, harvestedatday 10 from untreated immunized rats, wereculturedin 24-well platesat aconcentration of 5X 106/ml with 4

,ug/ml

ConAfor 3 d aspreviouslydescribed (41).After 72 h ofculture, 4 x 107viable SC peranimal wereinjected i.p.to syngeneic naive recipients.

13-cRA

13-cRAwaskindlyprovided by Dr. W. Bollag(Hoffmann-LaRoche,

Basel). Thepowderwasaliquotedinoxygen-freevials and storedat

4VC in the dark. Any procedureutilizing 13-cRAwascarriedoutin dimlight.

In vivostudies. On each dayoftreatment, the 13-cRA was sus-pended incornoil and administeredbygastricintubation,intwodaily

doses; matchedcontrols received the vehicleonly.In active EAE the animalsweretreatedfromdays6to11 p.i., whereas in passive EAE, from daysIto6after the cell transfer. 1 3-cRAwassuspendedincorn

oilat aconcentration of 7.5 mg/ml. Eachratreceived 1 ml of suspen-sion per 200 g ofweighttwiceaday,inordertoadministeratotal of 75 mg/kg perd.Atday 5oftreatment(day 10p.i.),serafrom 37 immu-nized rats were collected at hours 1, 2, 4, 8, and 12 after the morning administration of the drug. To study the pharmacokinetics, 13-cRA concentrationswereassayedin the serum samples by HPLC as previ-ouslydescribed (42, 43), and the average serum concentration over 24 hwasdetermined.

In vitrostudies. 13-cRA was dissolved at different concentrations in dimethylsulfoxide (DMSO); the stock solutions were then diluted in the culturemedium at0.01% DMSO final concentration. Control

cul-tures wereexposedto0.01%DMSO only.

Results

In vivo studies. In control rats, the onset of EAE signs was observed atday1 1±1p.i. (Fig. 1, Table I), whereasproliferative response to the Ag(SI=4.5), wasobserved in SC beginning at day 6 (datanotshown). Inorder toinvestigatethe activity of 13-cRAontheeffectormechanisms ofthe immune response in EAE,theanimals immunizedat day 0withmyelin were there-fore treated from days 6 to 11 (Fig. 1, Table I), whereas the animalsreceiving cell transferatday 0 were treated from days 1 to6(Table II).Atday5 oftreatment, 75 mg/kg per d 13-cRA induced an average serum concentration during 24 h of 5.4±5.2 x

10-6

M.

No major side effectswere associated with 13-cRA treat-ment. Insometreatedanimalsamodest degreeof weightloss (always< 10%)wasobserved(datanotshown).Nodiarrhea or othersignsofgastrointestinalirritation were noted.

AsshowninFig. 1 adelayed andless severe courseofthe diseasewasobserved inanimalsreceiving13-cRA.

Analysis

at day 12of the

frequency

ofneurological signs and of the pres-enceofinflammatory infiltrates in spinal cords showed thatthe diseasewas

significantly

suppressed

(X2

test: P<0.001; Table I). Theonsetofneurological signs,observed afew daysafter the

suspension

of thetreatment,indicatedthatdiscontinuationof thetherapy resulted in

partial

lossof the

protective

effects(FHg. 1).Thehighly significantsuppression ofDNAconcentrationin the

spinal

cords ofthe treated animals

quantitatively

con-SCORE 4.0 -3.5 -3.0 -2.5 -2.0 - 1.5-1.0 -0.5

0-0 cornoil(n =9) 0-0 RA75mg/kg/day(n=9)

8 10 12 14 16 18 20

I---I DAY AFTERIMMUNIZATION treatment

Figure1.Effect of 13-cRA on the course of neurological signs in EAE.

13-cRA and corn oil were administered from days 6 to 11 after im-munization for induction of EAE. Each point represents the mean of nine observations±SE. A delayed and less severe course of EAE is shownby the group receiving13-cRA. The symptom scoreis reported in the text.

firmed the suppression of the inflammatory response in the CNS (Table I). The characteristic lymphopenia described in ratsduringthe courseofEAE and during endogenous cortico-steroidrelease(44-46)wasobservedinthevehicle-,but not in the 13-cRA-treated animals (Table I). In order to establish if thesuppression ofEAEoccurredin presenceof in vivo effects of the treatment on Tcell functions, lymphocyte response to ConAwasstudiedatday 10,immediately beforethe onset of the disease. At 8 h after the oraladministration, when the 13-cRA serumconcentrationwas2.8±2.4X10-6M, a lower prolif-erativeresponse oflymphocytesfrom the 13-cRA-treated ani-mals was observed (two-tailed Mann-Whitney U test, 2P

=0.02; Fig. 2).

TableLEffects of13-cRA in Active EAEatDay12 Inflammatoryinfiltrates

Perivascular Lymphocyte Treatment* Neurologicalsignst cuffst DNA counts1l

sg/mg

13-cRA 8/26(0.8±0.1) 12/26 (0.7) 6.0±0.5 103±17

P<0.001 P<0.001 P<0.001 P<0.001 Oil 25/26 (2.2±0.1) 26/26 (1.7) 9.6±0.1 53±7

*75mg/kg per d13-cRA incornoilor cornoilonlyfromdays6to

1 1 p.i.

$The numbers represent thefrequencyineach group,of animalswith

neurological signsorperivascularcuffs in thespinalcord.In paren-thesisarereportedthemeanscoresof theneurological signsorof the

intensityof theinflammatoryinfiltrates

(X2

test).

SpinalcordmeanDNAconcentrationsexpressedas,ugDNA/mgof

protein±2 SE. ThemeanDNAconcentration in six sham-immu-nizedanimalswas5.6±0.2.(Student'sttest).

11Percentage±2SEof the values observedinsham-immunized animals(mean±2 SEof six observations =4,416+385). (Mann-Whitney test).

Table IL

Effects

of 13-cRA on EAE Passive Transfer*

Treatment

Invitro Invivoll Neurological signst PerivascularcuflV

DMSO Oil 16/18 18/18

13-cRA Oil 0/12 0/12

DMSO 13-cRA 0/12 0/12

*40

X

106

SC wereinjected i.p. insyngenicrecipients.

$Thenumbers represent thefrequency in each group, of animals with neurological signsorperivascular cuffs in the spinal cord.

IBeforetransfer, the SC were cultured for 72 hwith Con A, in presenceof 0.01% DMSOor1o06M 13-cRAdissolved in0.011% DMSO.

11From days1 to 6after celltransfer, 75 mg/kg per d of 13-cRAwas

administered in 1 ml ofcornoil. Controlanimals receivedcornoil only.

The in vivo activity of 1 3-cRA on the efferent limb of the immune response was also studied in passive transfer EAE. Animals treated withthe vehicle developed neurological signs from days 6±1 to 9±1 after transfer, and perivascular cuffs were observed inspinalcords at day 11 (TableII).By contrast, animalsreceiving 13-cRA, from days 1to6aftercelltransfer, developed neither neurological signs nor perivascular

infil-tratesinthespinal cord (Table II).

In vitro

studies.

Inorderto assesswhetherthe

suppression

ofEAE wasdueto adirectactivity of the drugonT cell-me-diatedimmunity, we studiedinvitrothe effects of 1 3-cRA on

SC collected from untreated animals immunizedwithmyelin or MBP. As shown in Table III, exposure of SC to 13-cRA

concentrations equivalentto or less than those observed in vivo suppressed lymphocyte proliferation to antigens and to Con A in a dose-dependent manner. It is noteworthy that the back-groundproliferationswere notaffectedby the same concentra-tions of13-cRA,indicatingthat up to 96 hin vitro exposure to the drugdidnot exert direct toxicity in lymphocytes. More-over, exposure of theSC to 10-6 M 13-cRAduring in vitro activation with ConA,before passive transfer, abrogated the

W) 0 x E 0. 0 350 + 300

t

1*

0 250

f

200+ 150 OIL 0 0 0o cRA

Figure 2. Invivoimmunosuppressive activity of 13-cRA during the development of EAE. Eachpointrepresentsquadruplicate

prolifera-tionsof SC collectedatday 5 oftreatment(day 10p.i.)and cultured for 48hin presenceof ConA(SE of<15% cpmof themean). When theserumconcentrationof thedrugwas2.8±2.4x10-6M, animals

receiving13-cRA(o) showed lower lymphocyteproliferationthan the oil treated controls (-); mean±SE: 13-cRA=252±47;OIL=355+20; Mann-WhitneyUtest:P=0.02).

Activityof

TableIII. In VitroEffects of 13-cRAon Ag-and Mitogen-induced SC Proliferation DMSOt RA10-8M RA10-7M RA106M cpm Medium*

4,420±5671

5,140±943 2,920+348 6,740±1,130 Exp.1 Con A 319,000±13,400 (72.2) 114,000±9,140 (22.3) 95,500±15,000 (32.7) 70,600±5,870 (10.5) Myelin (6.25 usg/ml) 21,000±3,940 (4.8) 11,400±2,530 (2.2) 10,200±1,790 (3.5) 11,500±712 (1.7) Medium 11,600±3,010 9,630±565 14,700±1,290 5,490±1,030 Exp. 2 Con A 308,000±28,400(26.4) 228,000±14,700 (23.7) 191,000±19,500(12.9) ND Myelin(6.25

,g/ml)

41,400±7,720 (3.5) 22,400±4,170 (2.3) 19,400±3,020 (1.3) 15,000±2,180(2.73) Medium 8,450±2,030 6,390±1,520 6,650±1,460 7,660±1,390 Exp.3 ConA 169,000±6,270(20.0) 93,000±7,110 (14.5) 86,800±6,870(13.1) 66,500±9,520 (8.7) MBP(50

,g/ml)

37,800±4,520(4.5) 22,400±3,070(3.5) 23,100±4,790 (3.5) 15,000+2,310 (1.9) Medium 2,540±422 3,200±545 2,610±278 2,650±4,170 Exp. 4 ConA 139,000±10,200(54.6) 68,700±3,700 (21.4) 59,000±10,000(22.6) 50,300±3,460 (18.9) MBP(50

,g/ml)

14,800±2,000 (5.8) 2,830±411(0.9) 3,410±483 (1.3) 2,810±292 (1.1)

*Myelin,MBP, and ConAweredissolved in the culture mediumat6.25, 50,and 4

jg/ml,

respectively. Eachculturecontained0.01%

DMSO. IThevalues expresscpm±SE ofquadruplicatecultures.Stimulationindices(SI)arereportedinparenthesis.The cellswerecollected from ratsimmunizedwithmyelininexperiments1 and2,and with MBP inexperiments3 and 4.

capacityofencephalitogenic lymphocytestotransferEAE (Ta-bleII).

The previous experiments, indicatingthat T cell-mediated immunity isatleastone ofthe targetsofthe immunological activity of 13-cRA, ledus to study the effect of13-cRA on lymphokine productionby T cells. Asexpected, IL-2sup pre-paredinthepresenceof 10-6M 13-cRAshowed a lower IL-2 activity, expressed as the proliferation ofthe IL-2-dependent T celllineCTLL-2, comparedtocontrolIL-2sup (P=0.048,Fig. 3). However, the proliferation of theCTLL-2 induced by opti-maldilutions of IL-2supwas notinhibitedby theaddition of 10-6M13-cRA (datanotshown).Suppression of IL-2

produc-x E KL 2 0 1 10 100 %IL2-SUP

Figure3.Suppression ofIL-2productioninpresenceof10'- M

13-cRA. Thecurvesrepresent theresponseof theIL-2-dependentT cell line CTLL-2topercentdilutionsof IL-2supproduced by SCin

presenceof10-6 M 13-cRA (RA IL-2sup, o), andtocontrolIL-2sup (.).The dilutionofRAIIH2sup exerting50%ofthe maximalactivity

(ED50)wassignificantly higher (mean±SE= 1.65%±0.21) thanthe

control

ED,0

(0.88%±0.l 1;P=0.048).

tion,but not of IL-2activity,is therefore oneeffectof 13-cRA onTcell-mediatedimmunity.

Finally, the effects of different times ofexposure to10-6M 13-cRA wereexamined in proliferatingT lymphocytesduring astandard 96-h stimulationassay in presenceof Con A. 13-cRAreducedaSCproliferative response to themitogenin a time-dependentmanner, if it was addedduringthe first 48 h of theassay,butdidnotshowanyeffectifthe ConA-stimulated SC wereexposedto 13-cRAduringthe second 48 h(Fig. 4). Thisexperiment demonstrates that exposure to 10-6 M 13-cRAforup to48hdoesnotaffectviability of stimulated proli-ferating lymphocytes.,Becauseunder theseexperimental con-ditionsatleastonemitotic cycle is completedin 48h, thymi-dine

incorporation

equivalent to 100% of the untreated cultures,indicate absence of antimitoticactivityof thedrug.

100__-

-90 C 8-0

700

70 )

60-

50I-*- 50 - _ L 40-020 0

1i0

20 30 40 50 60 70 80

hours of culture

Figure 4.Effectsof different times of exposure of SC to 106M 13-cRAduringastandardproliferationassay with ConA.Triplicate

96-h cultureswereexposedto 13-cRAstartingfromthe hour indi-catedonabscissa:eachpointrepresentsthe percentageproliferation

ofthe 13-cRAexposedcultures in respecttocontrol culturesnot

ex-posed to 13-cRA.(A-cpm of13-cRAexposedcultures/A-cpmof

DMSO-exposedcultures x100). 13-cRA suppresses the responseto

themitogeninatime-dependentmannerif addedduringthe first partof the assay, but hasnoeffect if addedduringthe last 48 h.

Discussion

Wehave studied the effectsof 13-cRAonin vivo and in vitro models of cellularimmunity, at concentrations thatarewell

toleratedinhumantherapy(1-3, 47-50). 13-cRA administra-tionduring the development of the efferent limb of the im-muneresponsesignificantly suppressed the clinical and histo-pathological signs of active EAE. Suppressionwasalmost

com-plete as long as the animals received 13-cRA, whereas

discontinuation of thetreatmentresulted inaloss ofthe protec-tive effect.However,passive transfer EAEwascompletely

pre-vented anddiscontinuation of 1 3-cRA 6 d after passive transfer resulted in nolate disease.

Intheseexperiments, suppression of EAEwasachieved

us-ing atreatment protocol that wasnot associated with major side effectsand thatinducedserumconcentrations equivalent tothehighest reportedtobetolerated in human therapy. Ad-ministration ofdrug boluses through gastric intubation

permit-tedaprecisecontrol ofthedosages, but didnotallowan opti-malabsorption of the drug. For thisreasonandbecauseof the highturnoverrateof13-cRAdescribed in rodents(1-3), high

oraldosageswererequiredtoachieve the desiredserum

con-centrations. Preliminary experiments indicated that diarrhea

andothersigns ofgastrointestinal intolerance (e.g., accelerated weight loss) developed insome animalsby day 7 of 13-cRA therapy. For thisreason a6-dtreatmentregimenwasselected. Inhumans, equivalentserumconcentrations can bereached using much lower dosages of 13-cRA and such regimensare

welltoleratedevenduring chronic administration(47-50). Theresponseof SCtothe T cellmitogen ConAwasalso

inhibitedduring thetreatment,indicating that the suppression

of the disease occurred inpresence ofinvivo inhibition of T

cell functions. Thepossibilitythatthis inhibitionwasnot

func-tional but resulted froma directlymphocytotoxic activity of

thedrugwasexcludedbytheobservationthat,atday 7from

thebeginningof the treatment,animalsreceiving 13-cRA had thesamenumber ofcirculating lymphocytesasthe"normal" controlswhichweresham-immunizedwithCFAonly. The pos-sibilitythatsubclinicaltoxicity of13-cRAmayhaveinduced

release ofendogenous corticosteroids was investigated using peripheral blood lymphocytecountas amarker:indeed,inrats exposedtostress,highlevels of corticosteroidsaresecreted and acorrespondinglymphopeniaisusuallyobserved(44-46, 51).

In the present experiments, the rats treated with 13-cRA

showedlymphocytecountsequivalenttothesham-immunized

rats injected only with CFA, whereas the animals receiving cornoil showed theusuallymphopeniadescribed in EAE

(Ta-bleI,reference46).Thisobservation indicates that the 1 3-cRA administrationdidnotinduce relevantendogenous corticoste-roidrelease, andtherefore that thesuppressionofthe disease

wasdirectly duetothetreatment.Theimmunosuppressive ac-tivityof 1 3-cRAwasalsoobserved in vitro. Continuous expo-sureto 13-cRAconcentrationsequivalenttoorlessthan those observed in vivoduringEAE, suppressedCon A-and

Ag-in-ducedlymphocyteproliferationinadose- andtime-dependent

manner, with maximalactivityat 10-6 M(Table III), which decreasedas afunctionof time when added after the lympho-cyteactivation(Fig. 4).This concentrationprevented passive

transfer of EAEby encephalytogenic lymphocytes (Table II)

and reduced IL-2production (Fig. 3)withoutlymphocytotoxic

effect in vivo(Table I)aswellasinvitro(Fig. 4).Taken

alto-gether, the datareportedinthispaperindicatethat

pharmaco-logicalconcentrations of 1 3-cRA are immunosuppressive and can prevent anongoing T cell-mediatedautoimmune disease evenwhen administered exclusively during the development of theefferent limbofthe immune response. An immunosuppres-sive activity intenimmunosuppres-sive enough to control an ongoing immune responseisshown also bycyclosporineAbut not by cytotoxic drugs (52, 53). These drugscansuppress EAE ifgiven prophy-lactically, butnotifadministered when the immunesystemhas alreadymounted aneffector response (53). It is likely that 13-cRA treatmentdoesnotspecifically deleteclonesof Ag-respon-sive Tcells,and theimmunosuppressive activity isprobably due to anonspecific effecton one or moresubpopulationsof immunocompetent cells, which include suppression of IL-2 production. Indeed, the describedactivities of 13-cRA on T cell-mediatedimmunity and IL-2 productionwereobserved stimulatinganunselectedSCpopulation which included acces-sorycellssuch asmacrophagesand dendritic cells. Inaddition, suppressive activities of RA on accessory cell functions has beenpreviouslydescribed (14, 54), suggestingthat13-cRA ac-tivity on IL-2 production as well as on T cell proliferation, couldinpartbeindirect and mediatedbyaneffectonaccessory cells. The suppression of the T cell proliferative response as

well asofIL-2production, could bethereforedue to adirect effect of 13-cRAonT-lymphocytes orcould be indirect and mediated by other solublefactors produced by a number of immunocompetent cells. Itis noteworthy, for example, that RAstimulatestheproductionoftransforming growth factor-3, oneofthe mostpowerful immunosuppressivecytokines, which is produced by many immunocompetent and nonimmuno-competent cells(55,56).Nonetheless, thepossibility ofa direct effect of13-cRA on IL-2production byTlymphocytes is indi-rectly suggested by the reported inhibition by RA of diacylgly-cerolinducedornithine-decarboxylase activity,anearly event thatplaysakey rolein Tlymphocyte activation(5-7). In

addi-tion,

thedescribedsupression ofIL-2production is consistent withthe recentobservationthat anotherexperimental retinoid, Ro 15-0778 (temarotene,Hoffmann-LaRoche,Basel), inhibits phytohemagglutinin-induced production of IL-2 and IFN-gamma inhumanT-lymphocytes (W. Bollag, personal commu-nication).Temarotene has also beenreportedto exert a marked

therapeutic

activity on lichen planus, a human T cell-me-diateddermatological disease (57).Thepossibility that13-cRA acts onmultipletargetsof the immune activitymusttherefore be taken into accountwhen resultsobtained under different

experimental

conditionsarecomparedand may atleastin part explainthelargenumberof

confficting

resultsreported inthe literature(for review,seereference 4).

Thesimilar pharmacological activities ofRA and 1 3-cRA andtheexistence ofanintracellularretinoic acid

binding

pro-teinthatbinds both theseretinoids(58)suggest that thetwo

isomerscould actthroughacommonreceptor. Five different RAreceptors (RARs), each one encompassing

multiple

iso-forms, have been so far described(59-67). At least three of thesereceptorsbindRA aswell as other natural and

synthetic

retinoids (59-62),including 13-cRA (63), although withless

affinity.

In

addition,

somedegree ofcis-transisomerizationto

RA may in part contribute to the activity of 13-cRA. The RARs aredifferentially expressed in various tissuesandexhibit

different

affinities forRAand otherretinoids

(59,

64),

suggest-ing

thattheymay havedistinct rolesin

mediating

anumberof different activities ofthesemolecules.Veryfewgenesactivated by the RARs are known so far (61, 64,65), but it is noteworthy

that these receptors arehighly conserved from rodentsto hu-mans and areexpressedinboth human and rat spleen, and that theoptimalRAconcentration inducing maximal transcription by the RARsis 10-6 M (59, 60, 63-66), the same nontoxic concentration that we have shown tobeactive on T cell-me-diatedimmunity. Inaddition,RARs are partofasuperfamily ofnuclear receptors that alsomediatetheactivity ofsteroidand thyroid hormones (67), supporting the hypothesis that RA functions inafashion analogoustothese hormones.Moreover, itmust beemphasizedthat steroid and thyroid hormones can repress geneexpressionas well asactivate it(67). It is possible thereforetohypothesize that, like glucocorticoids, 13-cRA ex-ertssuppressive activityon Tcell-mediated immunitythrough areceptor whoseexpression varies in differentcellpopulations andunder differentexperimental conditions.Nonetheless, the observations presented in this article establish that, inan in vivomammaliansystem, the net resultof13-cRAactivity isa suppressive effect on Tcell-mediatedimmuneresponse,and thatthisactivity is intensive enough tosuppress anongoing immuneresponse atnontoxic concentrations whichmay also be reached in humantherapy.Considering that theserum con-centrations of13-cRAinduced in this studycanbechronically maintained in human therapy with minor side effects (1-3, 47-50), theexperimentaluseof13-cRAalsoin humanchronic Tcell-mediated autoimmune diseasescan nowbeproposed. Inaddition, the prevention ofEAE, amodel ofhuman T cell-mediated disease oftheCNS(27), and the excellent bioavail-ability in the CNS of retinoids (68, 69), suggest that

multiple

sclerosis

patients

could constitutean appropriate population foraphase 1 clinical trial.

Acknowledaments

The authorswishtothankDr.LauraRaimondi for the HPLC assayof

13-cRA and Dr. DaleMcFarlin and Dr. Richard McCarron for helpful discussion. The authorsarealsogratefultoDr.MichaelSporn for help intherevisionof themanuscript.

This studywassupported byagrant of the Italian National Re-search Council. Dr.Massacesi was arecipient ofafellowship of the Italian National ResearchCouncil.

References

1.Bollag,W.1983. VitaminAandretinoids: from nutritionto pharmacother-apyindermatology and oncology. Lancet. i:860-863.

2.Goodman,D.S.1984. VitaminAandretinoids in health and disease.N.

Engl.J.Med.310:1023-1031.

3.Sporn,M.B.,A.B.Roberts,and D. S. Goodman. 1984. TheRetinoids.

AcademicPress, Inc., New York.

4.Shapiro,P.E., and R. L. Edelson. 1985. Effect ofretinoidsonthe immune system.InRetinoids: NewTrendsin Research andTherapy.R.Saurat,editor. S. Karger, Basel. 225-235.

5.Kensler, T. W., and G. C. Mueller. 1978.Retinoic acid inhibition of the

comitogenic action of mezerein andphorbol esters inbovine lymphocytes. Cancer Res. 38:771-775.

6.Kensler,T.W.,A.K.Verma,R. K.BoutwellandG.C. Mueller. 1978. Effectsof retinoic acid andjuvenilehormoneontheinduction ofornithine decar-boxylase activity by 12-O-tetradecanoylphorbol-13-acetate. Cancer Res.

38:2896-2899.

7.Wertz, P.W., T. W.Kensler,G. C.Mueller, A. K. Verma, andR. K

Boutwell. 1979.5,6-Epoxyretinoicacid opposes the effect of 12-O-tetradecanoyl-phorbol-13-acetatein bovinelymphocytes.Nature(Lond.).277:227-29.

8.Nordlind,K., and J. Thyberg. 1983. In vitroeffectsof 13-cis-retinoic acid (Ro4-3780)and etretinate (Ro10-9359)onDNAsynthesis andfine structure of

guinea pigandhumanlymphoidcells. Int.Arch.AllergyAppl.Immunol. 71:363-367.

9.Eremin,O.,J.Ashby,and J. Rhodes. 1984.Inhibitionof

antibody-depen-dentcellularcytotoxicitybyretinoic acid. Int. Arch.AllergyAppl.Immunol. 75:2-7.

10.Blalock,J.E.,andG. E. Gifford. 1976.Comparisonof thesuppressionof

interferonproductionandinhibition ofits actionbyvitaminAand related com-pounds(39532).Proc.Soc.Exp.Biol.Med.153:298-300.

1 1.Hughes,T.K.,J. K.Russell,and J. E.Blalock. 1986.Inductionof

inter-feron transformed cells: inhibitionbyretinoic acid.Biochem.Biophys.Res.

Com-mun.138:47-53.

12.Abb,J.,H.Abb,and F.Deinhardt. 1982.Effect ofretinoicacidonthe spontaneous andinterferon-inducedactivityof human naturalkillercells. Int. J. Cancer. 30:307-3 10.

13.Abb, J.,H.Abb,and F.Deinhart. 1982. Retinoic acidsuppressionof humanleukocyteinterferonproduction.Immunopharmacology.4:303-310.

14.Rhodes, J.,and P.Stokes. 1982.Interferonproducedchangesin the mono-cytemembrane: inhibitionbyretinol and retinoic acid.Immunology.45:53

1-536.

15.Uhr,J.W.,G.Weissmann,and L. Thomas.1963. Acutehypervitaminosis Ainguinea pigs.II.Effectsondelayed-typehypersensitivity.Proc.Exp.Biol.

Med.112:287-291.

16. Dresser, D. W. 1968. Adjuvanticity of vitamin A. Nature (Lond.).

217:527-528.

17.Spitznagel,J.K.,and A.C. Allison. 1970. Modeofactionofadjuvants:

retinoid and otherlysosome-labilizingagentsasadjuvants.J.Immunol. 104:119-127.

18.Dennert,G.,and R. Lotan. 1978.Effectsof retinoicacidonthe immune system:stimulation ofTkiller cell induction. Eur. J.Immunol.8:23-29.

19.Barnett,J.B.1983.Immunomodulatingeffectsof13-cis-retinoicacidon theIgGandIgMresponseofBALB/cmice. Int.Arch.AllergyAppl.Immunol.

72:227-233.

20.Sidell, N.,P.Rieber,andS.H.Golub. 1984.Immunologicalaspectsof retinoids in humans. I.Analysisof retinoic acid enhancement ofthymocyte re-sponsestoPHA.Cell.Immunol.87:118-125.

21.Otani, S.,I.Matsui-Yuasa,Y.Mimura,and S. Morisawa. 1986. Potentia-tionbyretinoicacid ofornithinedecarboxylaseinductionby

phytohemagglu-tininorphorbol12-myristate-13-acetateinguineapig lymphocytes.J.Biochem. 99:1789-1797.

22.Sidell, N.,and F.Ramsdell. 1988.Retinoic acidupregulates interleukin-2-receptoronactivated humanthymocytes.Cell. Immunol. 115:299-309.

23.Brinckerhoff,C.E.,J. W.Coffey,andA.C. Sullivan.1983.Inflammation andcollagenaseproductioninratswithadjuvantarthritisreduced with 13-cis-re-tinoic acid. Science(Wash. DC).221:756-758.

24.Trentham,D.E.,andC. E.Brinckerhoff.1982.Augmentationofcollagen arthritisbysyntheticanaloguesofretinoic acid.J.Immunol. 129:2668-2672.

25. Massacesi,L.,A.L.Abbamondi,C.Giorgi,F. Sarlo,F. Lolli,andL. Amaducci. 1987.Suppressionofexperimentalallergic encephalomyelitis by reti-noic acid.J.Neurol.Sci.80:55-64.

26.Sedgwick,J.,S.Brostoff,andD.W. Mason.1987.Experimentalallergic encephalomyelitisin the absenceofaclassicaldelayed-typehypersensitivity

reac-tion:severeparalyticdisease correlates with thepresenceof IL-2receptorpositive

cellsinfiltratingtheCNS. J.Exp.Med. 165:1058-1075.

27.Alvord,E.C.,M.W.Kies,andA.J.Suckling.1984.ExperimentalAllergic Encephalomyelitis:aUseful Model forMultipleSclerosis. Alan R.Liss,Inc.,New York.

28.Paterson, P. Y. 1960. Transfer ofallergicencephalomyelitisinratsby meansoflymphnode cells. J.Exp.Med.111:119-135.

29.Paterson,P. Y. 1976.Experimentalautoimmune(allergic)

encephalomy-elitis:induction,pathogenesis,andsuppression.InTextbookof

Immunopathol-ogy.2ndEdition. P.A.Miescher and H. J.Mueller-Eberhard,editors.Grune & Stratton,Inc. New York.179.

30.Ortiz-Ortiz,L.,and W.0.Weigle.1976. Cellulareventsin theinduction

ofexperimental allergicencephalomyelitisinrats.J.Exp.Med. 144:604-616.

31.Pettinelli,C.B.,andD. E.McFarlin. 1981.Adoptivetransfer of

experi-mentalallergicencephalomyelitisinSJL/Jmice after invitroactivationoflymph

nodecellsbymyelinbasicprotein: requirementfor

Lyt'+2-

lymphocytes.J.

Im-munol. 127:1420-1423.

32.Traugott,U. 1989.Detailedanalysisofearlyimmunopathologicevents

duringlesion formation inacuteexperimentalautoimmuneancephalomyelitis.

Cell. Immunol. 119:114-129.

33.Richert,J.R.,B.F.Driscoll,M. W.Kies,andE.C. Alvord. 1981.

Experi-mentalallergicencephalomyelitis:activationofmyelinbasicprotein-sensitized

spleencellsbyspecific antigenin culture.Cell.Immunol. 59:42-53.

34. BenNun, A.,and Z. Lando. 1983.Detection ofautoimmunecells proli-feratingtomyelinbasicproteinandselection of T-celllinesthatmediate

experi-mentalallergic encefalomyelitis (EAE)inmice. J.Immunol. 130:1205-1213.

35.Norton,W.T.,and S.E.Poduslo. 1973.Myelinationinratbrain: method ofmyelinisolation.J.Neurochem.21:749-758.

preparation ofmyelin basic protein, from central nervous tissue of several mam-malian species. Prep.Biochem. 2:139-165.

37.Smith, M. E., F. P. Somera, R. Saldivar, L. Massacesi, and J. Trotter. 1984. DNA changes in spinal cord ofrats with experimental allergic encephalomy-elitis. J.Neurochem. 43:1635-1641.

38.Kappler, J., and P. Marrack. 1986. Lymphokines. In Handbook ofExperi-mental Immunology. VolumeII. D. M. Weir, editor. Blackwell Scientific

Publi-cations,Ltd.,Oxford,UK.

39.Gillis, S.,W.Ferm,W.Ou,andK.A.Smith. 1978. T cell growthfactor:

parametersofproduction and a quantitative microassay for activity. J. Immunol. 120:2027-2032.

40. DeLean, A., P. J. Munson, and B. Rodbard. 1978. Simultaneous analysis

of families ofsigmoidalcurves:applicationstobioassay, radioligand assay, and physiological dose response curves. Am. J. Physiol. 235:E97-E102.

41.Panitch, H. S., and D. McFarlin. 1977. Experimental allergic

encephalo-myelitis:enhancementofcell-mediated transfer by concanavalin A. J. Immunol. 119:1134-1137.

42. Vane, F. M., J. K.Stoltemborg, and C. Y. L.Bugg6.1982.Determination

of 13-cis-retinoicacid in human blood by reversed phase high performance liquid chromathography. J.Chromatogr.227:471-484.

43.McCormick, A. M., J. L. Napoli, and H. F. DeLuca. 1978. High pressure liquid chromatographic resolution of vitamin A compounds. Anal. Biochem. 86:25-33.

44.Dougherty, T. F., and A.White. 1945. Functional alterations inlymphoid tissueinducedby adrenalcorticalsecretion. Am. J. Anat. 77:81-115.

45.Dougherty, T. F., and A.White. 1947. Evaluations ofalterations produced inlymphoidtissue bypituitaryadrenalcortical secretion.J. Lab. Clin.Med.

32:584-605.

46.MacPhee, I.A.M., F.A.Antoni,and D. W. Mason. 1989.Spontaneous recoveryof rats fromexperimentalallergicencephalomyelitis is dependent on regulationoftheimmune system by endogenous adrenal corticosteroids. J. Exp.

Med.169:431-442.

47.Kessler, J. F., F. L. Meyskens, N. Levine,P.J.Lynch, and S. E. Jones. 1983. Treatmentof cutaneous T-celllymphoma(mycosis fungoides) with

13-cis-retinoic acid.Lancet.i:1345-1347.

48.Peck, G. L., T. G.Olsen,F. W.Yode,J.S.Strauss,D. T.Downing,M. Pandya, D. Butkus, and J.A.Buttandier.1979.Prolongedremissionofcysticand

conglobateacnewith13-cis-retinoicacid. N.Engl.J.Med.300:329-333. 49.Kerr,G.I., M. E.Lippmann,J.Jenkins,andC. Y.Myers.1982. Pharma-cologyof13-cis-retinoicacidin humans.Cancer Res. 42:2069-2073.

*50.Clamon,G., G.G.Chabot,F.Valeriote,E.Davilla,C.Vogel,E.

Gor-owski,and R.Birch. 1985. Phase Istudyandpharmacokineticsofweekly high-doseof13-cis-retinoic acid.Cancer Res. 45:1874-1878.

51.Levine,S., R. Strebel, E. J.Wenk,and P. J. Harman.1962.Suppression of

experimentalallergicencephalomyelitis bystress.Proc. Soc.Exp. Biol.Med.

109:294-298.

52.Bolton, B.,J. F.Borel,M. L.Cuzner,A.N.Davison,and A. M. Turner.

1982.Immunosuppression by cyclosponneAofexperimentalallergic encephalo-myelitys. J.Neurol. Sci. 56:147-153.

53.Levine, S., and R. Sowinski. 1980. Therapy of allergic encephalomyelitis in ratsafter onset of paralysis. In The Suppression of Experimental Allergic

En-cephalomyelitisandMultiple Sclerosis,A.N.Davison andM.L.Cuzner, editors. Academic Press, Ltd., London. 199-209.

54.Bedford, P. A., and S. C. Knight. 1989. The effect ofretinoids on dendritic cellfunction. Clin. Exp. Immunol. 75:481-486.

55.Sporn,M.B., A. B. Roberts, L.M.Wakefield,and B.deCrombrugghe. 1987. Some recent advances in the chemistry and biology oftransforming growth factor-B. J. Cell Biol. 105:1039-1045.

56.Glick,A.B., K. C.Flanders, D.Danielpour,S. H. Yuspa, and M. B. Sporn. 1989. Retinoic acid induces transforming growth factor#2in cultured keratinocytes and mouse epidermis.CellRegul.1:87-97.

57. Bollag, W., and F. Ott. 1989. Treatmentoflichen planus with temarotene. Lancet.ii:974.

58.Chytil,F., and D. E. Yong. 1979. Cellularretinol-andretinoic acid bind-ingproteinsin vitamin A action. Fed. Proc. 38:2510-2514.

59.Giguerre,V.,E.S. Ong,P.Segui, andR. M.Evans.1987. Identification of

areceptorfor the morphogen retinoic acid.Nature(Lond.).330:624-629. 60. Kastner, P. H.,A.Krust, C. Mendelsohn, J. M.Garnier, A. Zelent, P. Leroy, A.Staub, and P. Chambon. 1990. Murineisoforms ofretinoic acid

recep-torgammawithspecific pattern of expression. Proc. Natl. Acad. Sci. USA. 87:2700-2704.

61.Mangelsdorf,D.J., E.S.Ong, J.A.Dick,and R.A.Evans. 1990. Nuclear receptorthatidentifies a novel retinoic acid response pathway. Nature(Lond.)

345:224-229.

62.Hashimoto,Y., H.Kagechika,and K. Shudo. 1990.Expressionofretinoic acid receptor genes and theligand binding selectivity of retinoicacid receptors (RAR's).Biochem. Biophys. Res. Commun. 166:1300-1307.

63.Blomhoff,R.,M.H.Green,T.Berg, and K. R. Norum. 1990.Transport andstorageof vitaminA.Science(Wash. DC).250:339-404.

64.deThe, H.,A.Marchio, P.Tiollais,andA.Dejean. 1989.Differential

expressionandligand regulationofthe retinoic acid receptor alfa and beta genes. EMBO (Eur. Mol.Biol.Organ.) J. 8:429-433.

65.Vasios,G.V., J. D. Gold, M.Petkovitch,P.Chambon, and L. J. Gudas. 1989.Aretinoicacidresponsive element is present in the 5'flanking regionofthe

lamininB1 gene. Proc. Natl.Acad.Sci.USA. 86:9099-9103.

66.Rees, J. L.,A.K.Daly, and C.P. F.Redfern.1989.Differentialexpression

of thealfaand betaretinoic acid receptors in tissuesoftherat.Biochem. J. 259:917-919.

67.Evans, R.M.1988. The steroid andthyroidhormone receptor

superfa-mily. Science(Wash.DC). 240:889-895.

68.Larsen,P.D.,andL. J.DeLallo.1987.Cerebrospinal fluid transthyretinin

multiple sclerosis.Neurology. 37:1262-1263.

69. Kalin,J.R.,M.G.Wells,and D.L.Hill. 1982.Dispositionof 13-cis

retinoic acidandN4hydroxyethyl)retinamideinmice after oral doses. Drug