Wild-type p53-mediated down-modulation of IL-15 and IL-15 receptors in human rhabdomyosarcoma cells.

Wild-type

p53.mediated down-modulation

of interleukin

15 and

interleukin

15 receptors in human

rhabdomyosarcoma cells

C

De

Giovanni',

P

Nannil,

A

Sacchi2,

S

Soddu2,

I

Manni2,

G

D'Orazi23,

S

Bulfone-Paus4,

T

Pohl4,

LLanduzzi

15,

G

Nicoletti1-,

F

Frabetti1,

IRossi' and P-L

LoIlini'

1Carx

Research

Insttute,

University

of

Boklona, Bologna,

italy Mokcular

Oncogenesis

Laboratory,ReginaElenaCancerInstitute,Rome,Italy

3Departbent

ofOncologyandNeuroscience,GD'AnnunzioUniversity,Chieti,Italy, InrttuteofImmunology,Berlin,Germany5ISTBiotechnoogySatelliteUnit,Boogna,ttaly

Summary We recentty reported thatrhabdomyosarcoma cell linesexpress andsecreteinterleukin 15(IL-15), atightty regulatedcytokine

with IL-2-like actvity.To testwhetherthe p53-impairedfunction that isfrequently found inthis tunour type couldplay a role intheIL-15 producbon,wild-type p53gene wastransducedin thehumanrhabdomyosarcoma cell line RD(whichharboursamutatedp53gene),and its

effectonproliferationandexpression of IL-15wasstudied. Arrestofproliferationwasinducedbywild-type p53:increasedproporionsof

G1-arrested cells andofapoptoticcells were observed. A markeddown-modulatin ofIL-15expression,atboththemRNA andproteinlevel,was

foundinp53-transduced cells.Becauseadirecteffect ofIL-15onnorrnalmuscle cellshas beenreported,thepresenceofIL-15membrane receptors was studied by cytofluorometnc

analysis.

Rhabdomyosarcoma cells showed IL-15 membrane receptors, which are

down-modulated bywild-type p53transfectedgene. In conclusion, wild-type p53 transductn in human rhabdomyosarcoma cells induces the

down-modulation ofbothIL-15productionandIL-15receptorexpression.

Keywords:p53; interleukin15; interleukin-15receptors; genetransducton; human rhabdomyosarcoma

Mutationinthe tumour-suppressorgenep53isthemostfrequent

alteration foundinhumantumours (Ko and Prives. 1996). Wild-typep53 is atranscription and translation regulator acting on a

varietyofgenesinvolved inthecontrol of cellcycle and differen-tiation (Ewen andMiller. 1996; Bourdonet al. 1997)thatcanbe either activatedor repressed. These genes includecytokines and growth factors (Fujiwaraetal. 1994;Shinetal. 1995: Peschetal.

1996:

Zhang

etal. 19%).Therefore. impaired

p53

function could determine a deregulated production ofcytokines/growth factors.

which could be of advantage for the malignant phenotype

(Baserga. 1994). Rhabdomyosarcoma. the most frequent soft-tissuesarcomainchildhood(Pappoetal. 1995).frequentlyshows

impaired p53function(Felixetal.

1992:

Dilleretal. 1995: Keleti

etal. 1996)andproducesavarietyofgrowthfactors (Schweigerer

et al. 1987; El-Badry etal. 1990: De Giovanni et al. 1995) that could be subject to transcription control byp53 (Fujiwara et al. 1994: Shinetal. 1995: Zhangetal. 1996). We recently found that

rhabdomyosarcoma cell lines express and secrete interleukin 15

(IL-15)(Lollinietal. 1997).Thiscytokine. controlled at both

tran-scriptionandtranslation levels.hasanIL-2-likeactivity (Tagaya et al. 1996;Onuetal. 1997)and protectsfromapoptosisboth in vitro and in vivo (Bulfone-Paus et al. 1997). IL-15 transcription has beenfoundinavarietyofnormalandneoplastic tissues (Grabstein

etal. 1994:Tagayaetal. 1996). but secretion of IL-15 protein is

restrictedto somenormal cell types andafew neoplastic celllines.

Received5December1997

Revised27May 1998 Accepted28May 1998

Correspondenceto:C DeGiovanni,Istituodi

Cancerologia,

VkaleFilopanti

22,1-40126Bolgna,Italy

includingrhabdomyosarcoma (Lollinietal. 1997).Totestwhether impairedp53function couldplayaroleinIL- 15 production. we

transfectedp53geneinrhabdomyosarcomacells andstudied the effectsonproliferationandonIL- 15transcriptionandproduction.

Because cells of the muscle lineage can respond to exogenous IL-15 (Quinn et al. 1995). expression of IL-15 receptorswasalso studied.

MATERIALS

AND

METHODS

Cells

and

p53

gene

transduction

Transduction ofp53 gene was perforned in the human rhabdo-myosarcomacell lineRD. which is knowntoharbour a mutation in codon 248 of thep53gene (Felix et al. 1992). Unless otherwise

specified, cells were cultured in Dulbecco's modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum

(FBS)at370Cinahumidified 7%carbondioxide atmosphere.All media constituentswerepurchased fromLifeTechnologies. Milan. Italy. Plasmid pN53cG(Val-135). carrying the temperature-sensi-tivep53(ial-135) gene (ts-pS3)under thecontrol ofHaMSV-LTR

andthe selectable markerneounder the control of the RSV-LTR

(Sodduetal. 1994).andplasmidpRSVneo,carrying the selectable markeralone.were usedfortransfections. Approximately 3x 106

exponentially growing RD rhabdomyosarcoma cells were trans-fectedbyelectroporation(0.25 V. 960

jiF)

with a gene pulser (Bio-Rad

Laboratonres.

Hercules.CA. USA). Transfectants were cloned

by limiting dilution in medium containing 0.75 mg mll G418

(Gibco BRL): one clone designated RD-tsp53 and one control

designated RD-Neowere thenstudied. All transduced cells were

routinelyculturedat37°C andshiftedtothepermissive

tempera-tureof32°Cforexperimentalstudies.

Toverifythetransactivatingactionofts-p53geneproductatthe

two culture temperatures, RD-Neo and RD-tsp53 transfectants were retransfected with a chloramphenicol acetyltransferase (CAT)reporter construct.Approximately 1 x105cellswereplated

in60-mm dishes and transfectedbycalcium

phosphate

precipita-tion withPG

3CAT

vector,carryingtheCAT reporter genedriven

by a polyomavirus promoter and 13copiesof thep53

consensus-bindingsequence,or

MG,5CAT

vectorwith 15copiesofp53

non-bindingsites(Kemetal, 1992)(kindly provided byBVogelstein).

Soon aftertransfection, the cellsweretrypsinizedand seeded in two new dishes, one of which wasincubated at the permissive

temperature and the other at thenon-permissivetemperature. After 48h,thecells were lysed by freezeand thaw. CATactivity was

determined usingthe two-phase fluorescence diffusion assay as describedpreviously(Huffetal, 1993);countsper minute(c.p.m.) were normalized to proteinquantity.

Overexpressionofwild-typep53wasalsoobtainedbyinfection

of RD cells with replication-defective, recombinant adenovirus carrying the human wild-type p53 cDNA (AdJL16), kindly

provided by Silvia Bacchetti (Bacchetti and Graham, 1993). AdherentRDcellswereinfectedwith 20p.f.u.ofAd-p53ord1312

control

adenovinus,

as described previously (Bacchetti and

Graham, 1993). Forty-eight hours later, cells were subjected to

reversetranscnptase polymerasechain reaction(RT-PCR)analysis.

Proliferation studies

RD-tsp53and RD-Neo cellswereseeded in DMEM+10%FBSat

thedensity of 10 cells cm-2 and cultured at37°Cor32°C for 4

days,then cultures were switched toamedium withalow content of exogenous growth factors (DMEM + 2% horse serum) and incubation continued for further 7 days with medium renewal every2-3 days. At 3-4dayintervals,cultures wereharvested and cell number and viability were determined by trypan blue dye

exclusion test.

Forcellcycleanalysis,5x10 trypsinizedcells were washedin

PBS, fixedincoldmethanol-acetone (1:4)solutionfor 30mmat

4°C and DNA was stained with 50

ig

ml' propidium iodide

(Sigma,St.Louis,MO, USA)in

phosphate-buffered

saline(PBS)

with 2

jg

ml-' RNAase A (Boehringer-Mannheim Italia,Milan, Italy)for 30minatroomtemperature. Thecontentof DNA was measuredbyanEpics XLanalyser (Coulter

Corporation,

Miami.

FL, USA).

The proportionofapoptotic cells wasstudied on cellscultured

as above,harvested and stained with Hoechst 33342; about 500 cells were scoredforapoptoticbodies at 40xmagnification.

In someexperiments,cellcultures were treated with exogenous recombinant human IL-15 (Immunex, Seattle, WA, USA) at

concentrationsrangingfrom 100to 1000ng ml-, withanti-IL-15 monoclonalantibodyM11 (Immunex),at a finalconcentrationof

3

jig

ml-', or with anti-IL-15 polyclonal antibody (PeproTech, London,UK),atafinal concentration of 5

jig

ml-';cellyield was determined4days later.

IL-15

expression

Cells

transduced

and cultured as above were utilized forRNA

extraction,cDNA synthesisand RT-PCR,asreported previously

(DeGiovanni etal, 1995;Lollinietal, 1997). RNA was extracted fromatleast three independent

experinments

in whichcontrol and

25- 20-0 x E 15-0. 0 10

C:.)5

O-

-zzzn-I<9

%

/

PG13CAT MG15CAT

FIgure 1

Transadivating actvity

of thebtansfectedts-p53ge produ at

the

permss

tenreratureof32°C.RD-Neo- and

RD-tspb3-transfected

cellswererebtansectedwith

PG,3CAT

vector,carrying theCATreportergene

drivenbypolyofafirusprofoterand 13copes ofthep53

consensus-bdgsequenceorwiththecontrolMG,CATvectorwih15copiesofap53

non-biding

site.CATacivity(seeMaterialsandmethods) in cels cuituredat

320C is shown

p53 transfectants were run inparallel; acDNApanel from each expernment was tested in RT-PCR at least twice. Specificprimer

pairs for glyceraldehyde-3-phosphate dehydrogenase (GAPDH)

andtransforming growth factor

0,

(TGF-3,)

wereobtained from Clontech (PaloAlto,CA. USA); prinmersfor IL-15 were synthe-sized as reportedpreviously(Lollini etal, 1997).

Enzyme-inked

immunosorbent

assay

(EUSA)

Cellswereseeded ataconcentrationof4-8x 104 cellscm-2and cultured at37°Cor32°C. After 4 days, cultures were switched to RPMI + 0.5% FBS andincubation continued for afurther4days at

37°C or at 32°C respectively. Cell supematants were then collected and IL-15 and

TGF-01

prductionmeasured by means of ELISA purchased from Genzyme (Milan, Italy) and used

accordingtothemanufacturer's protocol. Supernatantsfor IL-15 determination were concentrated about 100 times as reported

previously

(Lollini

et al. 1997). At the sametime, cell cultures

wereharvested and cell yield wasdetermined. Cytokine produc-tion was expressed in pg per 106 cells.

Cytofluorometic

sdies

Expression of membrane molecules on cells seeded at 106 cells cm-2 andcultured at 37°C or 32°C was detrmined by indirect

immunofluorescence and cytofluorometric analysis (FACScan, BectonDickinson, MountainView. CA, USA) using thefollowing primaryreagents: IL-15-IgG2b fusion protein, which recognizes all membrane IL-15 binding sites, produced by cells

trsduced

with ahuman IL-15-murine IgG2b fusion construct (Bulfone-Paus etal, 1997); monoclonal antibody M160(Immunex) recognizing thea-chain of the trimeric IL-15receptor, polyclonal C-16 anti-serum(Santa Cruz Biotechnology, Santa

Crz,

CA, USA)

recog-nizing

TGF->-Rll;

anti-human class I major histocompatibility

complex W6/32 antibody (SeraLab, Crowley Down,

UK):

anti-CD44 monoclonalantibody.cloneIM7(PharMingen. San Diego.

Britsh

Joumal ofCancer(1998)78(12), 1541-1546

4p

4pie

_I(FIq

A , RD-Neo RD-tsp3 a,.-//

/y

XRD-Neo

//

/i

X

3TC 37°C 32°C Adeno L0 to a 0 %-Z t Z 0 0 in 'cj 0 cc cc cc EC

IT

3: IL-15 -TGF_ GAPDH 3 6 9 12 Daysofculture

Fhgure2 Proliferationofts-p53genetansduced RD human

rhbdmyosarnacellscultured at37°C (mutated p53 conformation)and at

32°C (wiklype p53conotion). Open symbols,RD-Ne transfectnt

cells;closedsytbols,RD-tsp53tasea cels;dashed ine,37°C;

continuousline,32°C

CA, USA). The results shownarefromanexperiment

representa-tive ofatleasttwoindependent experiments. To study the

expres-sion of p53, the monoclonal antibody PAbl22 (Boehringer Mannheim, Germany) recognizing both wild-typeand mutatedp53

was used in indirect immunofluorescence on cells fLxed with

methanol.

RESULTS

Humanrhabdomyosarcoma cells RDwerewansduced with the

ts-p53gene,whoseproductbehaveslikewild-type p53atthe pennis-sivetemperatureof320C(Sodduetal, 1994). p53 expressionlevel

andtransactivating action at both permissive andnon-pennissive

temperatures were determined on transfectant cells. A high p53

expression was observedby cytofluorometry in RD-tsp53

trans-duced cellsatbothtemperaturesbymeansofanantibody that binds

top53 in the wildtypeaswell asin the mutated conformation(data

not shown). To verify the ransactivating action of ts-p53 gene product, RD-Neo and RD-tsp53 transfectants were retransfected

with PG

3CAT

vector, carrying the CATreporter genedriven by polyomavirus promoter and 13 copies of the p53

consensus-bindingsequence orwith

MG,5CAT

vectorwith 15copies ofap53 non-binding site (Kernetal, 1992). Baseline levels of CAT activity

were observed in all transfectants cultured atthe non-permissive

temperature(datanotshown), whereasaveryhigh abilityto

trans-activate

PG13

drivenCAT expression was observed inRD-tspS3

cellsculturedatthepermissivetemperature(Figure 1).

In RD-tsp53 transfectants, the induction of wild-type p53 conformationby cultureat32°C causedanalmostcompletearrest

ofcellgrowth(Figure 2); thesamecells culturedat37°C showeda

growth rate close to that of RD-Neo transfectant control cells. Proportions of

G,

cells and ofapoptotic cells were studied as

potential causesofgrowtharrestinducedby wild-type p53. The proportion of

G,

cells inRD-tsp53 transfectants culturedat32°C showedatwofoldincreaseoverthat ofRD-Neocontrols(72% and 36%, respectively) and a twofold increase in the proportion of apoptotic cells was also observed (3.9 ± 0.2% and 2.2 ± 0.2%

respectively). B 10 0 10. 0. cm eg 5 IL-15 TGF-1 -0 40 10 a Go 0 Z f I * V - A _ 37°C 32WC 37°C 32WC

Figure3 Productionof IL-15and TGF-1 by p53genebansduced RDcels.

(A)ExpressionofmRNAforIL-15,

TGF-P,

andGAPDHevaluatedby RT-PCR. RNAwasextacted fromRD-NeoandRD-tsp53 transfectant cultures incbated4daysattheiricated teperature andfromadenovifected

cultures48 hafterinfection. ForeachRT-PCRprodud thefoling

ampnlficto cycles areshowntocomparecontrolandp53transfeant

samplesintheexponentialampifica phase: IL-15,30;

TGF-O,

25;

GAPDH,20.(B) IL-15and

TGF-P,

releaseinsqermnaantsbyRD-Neoand

RD-tsp53tansducedcellsculturedat37°Corat32WC

Theexpression of mRNA forIL-15wasstudied by RT-PCR in

comparison with the expression of an unrelated growth factor

(TGF-j1)

and ofahousekeepinggeneglyceraldehyde-3-phosphate

dehydrogenase (GAPDH) (Figure 3A). Foreach primerpair, the

comparison between cDNAs derived from RD-Neo- and

RD-tsp53-transfectedcellswasmade in theexponential amplification phase. GAPDH showed that comparable cDNAs were obtained: PCR

products ofsimilar intensity were always observed after 15-25

amplification cycles. Adown-modulation of IL-15 tanscriptwas

observedinRD-tsp53-transfected cellsat370C andat320C(Figure

3A). The mean decrease in the intensity of ethidium

bromide-stainedbands, evaluatedby densitometricanalysis of five indepen-dentexperiments, was: 38.4% ± 8.6 at370C and 71.0%± 7.3at

32°C. Therefore, p53 down-modulation of 1L-15wassignificantly

higher (P=0.016) in the wild-type confonnation.

Down-modula-tion of the IL-15 transcript was constantly observed at different timepoints (e.g. from 2to 12days) after cell seeding.

Theexpression ofwild-type p53 proteinwasalso induced in RD

cells by infection withanadenovirus vector.againadecrease in

the]11-15 RT-PCRproductwasobserved inAd-p53 infected cells

in comparison with cells infected with the control adenoviral

BrtshJoumalofCancer(1998) 78(12), 1541-1546 1000- 10O-E 0 0 .o0 0 0 3. 0 10-Si I 0 e RD-tsp.

IL-15bindKg IL-1SRa

37C

TGF§-R

RD-Neo

RD-tsp53

RD-Neo

RD4sp53

Fgure4 Cytofluorotic analyss of membrane receptors for IL-15and

TGF-1 inRD-Neoand RD-tsp53transducedcellsculturedat370Corat

32°C. Ineachgraph,theabscissarepresentsfluorescenceintensityona

kogarithmicscalerangingfrom10°to10',andtheordinaterepresentsthe numberof cells.Open profiles,cells stained withsecondaryantbodyalone; shadedprofiles,cells stainedwithantbodyorfusionproteinrecognizing the indicatedspecficity

vector(Figure 3A). Meandecrease ± s.e.from three expenrments

was55.0%± 14.0.

Down-modulation of IL- 15 mRNA was not duetoan

unspecifi-cally

decreased transcription activity because the expression of

TGF-01

(Figure 3A)

wasslightlyincreased.

Post-transcriptional

regulationof IL-15 protein production has been

reported (Tagaya

etal. 1996;Onuetal, 1997).Therefore, the

expression

of secretedIL-15protein wasstudiedbyanELISA in supematants from RD-Neo andRD-tsp53transfected cells (Figure

3B).

AstrongdecreaseinIL-15productionwasalreadyobserved in

RD-tsp53-transfected

cells cultured at 370Cand reached90%

reduction in

32°C

cultures. Thisdecrease isunlikelytobedue to a

non-specific

decrease in metabolicactivitybecausetheproduction of

TGF-13,

whose baseline level was approximately 200-fold

higher

than thatof IL-15. was actuallyincreasedunder the same

experimental

conditions

(Figure

3B).

IL-15 binds to a membrane trimeric receptor. composed of a

specific

a-chain

(IL-15Ra)

andof >- andy-chainsshared withother

interleukins. Moreover,anunknown receptor has also beenreported onmastcells

(Tagaya

etal. 1996). ThepresenceofIL-15receptors

on p53-transduced RD cells was studied by means of

cytofluori-metry.

eitherusingamonoclonalantibody recognizingIL-15Raor the

IL-15-IgG2b

fusion protein (Bulfone-Paus et al. 1997) that detects cellmembranebindingsites.RD-Neo cellsexpressedIL-15

membrane receptors.asshownbythebindingof the fusion protein. andatleastpart of thebindingwasdue to thepresenceofIL-l5Ra

(Figure 4).

Down-modulationofbothtotalIL-15membranebinding

and ofIL-15Rawasobserved in

RD-tsp53-transfected

cells when the

wild-type

p53conformationwasinducedbyculture at32°C. No differencewasobserved for

TGF_j-RHg

(Figure 4).

ThetreatmentofRD-NeoandRD-tsp53 cells withexogenous recombinant humanIL- 15atconcentrationsupto 1000ng ml-'. as well as with either monoclonal or polyclonal anti-IL-15

anti-bodies,

didnotdetermine anychangeincellproliferation and in expression of membraneglycoproteinssuch as HLA-A. B. C and CD44

(data

not

shown).

DISCUSSION

Humanrhabdomyosarcomacell line RDwas transduced with

p53

gene using twodifferent vector systems: ats-p53 plasmid

(Soddu

etal. 1994).giving riseto stable transfectants inwhichwild-type

p53 conformation isdependent uponthe culturetemperature. and

an adenoviral vectorcarrying the wild-type p53 gene (Bacchetti

andGraham. 1993). inwhich ahightransientwild-typep53gene

expressioncould be studied withinafew days afterinfection.With

bothvectorsystemsadown-modulation of IL- 15transcriptionwas

observed. Thiswasnotduetoanunspecificdecrease in

transcrip-tion activity because expression of an unrelated growth factor

(TGF-,B1)

wasslightly increased.

In thets-p53system.down-modulationsof IL-15transcriptionand productionwerealso observedatthenon-pennissiveculture

temper-ature. but the effect was significantly higher when the wild-type

conformation was induced The effect of ts-p53 at

37°C

could be related to itsleakiness leading to alowamountof p53 in the

wild-type conformationeven atthis temperature. However. experiments

performed at a

higher

temperature (390C) confirmed that the

decrease in IL-15

production

wasalready observed withts-p53in the

mutated conformation (data not shown). This suggests that both

mutated andwild-typep53wereactive in down-modulating IL-15. A

similar phenomenonhas been reportedfor other cytokinesin

expen-mental models basedonp53mutants. e.g. IL-2 andIL4are

down-modulated by wild-typep53aswellasby deletionmutantp53(Pesch

etal. 1996). TGF-j31

production

wasreportedtobeincreased byboth wild-type and mutated p53 inaglioblastomacell line (Fujiwaraetal.

1994): this finding was extended to rhabdomyosarcoma by the

presentstudy withtemperature-sensitivep53geneproducts.

Although the above-mentioned IL- 15 down-modulation was

induced by bothwild-typeand mutatedp53. other effectsreported herewerespecific for p53inthewild-type conformation. suchas

the growth arrest and the down-modulation ofIL-15 receptors.

These different behaviours could suggest that different

mecha-nismsofp53actionareinvolved.

The presence of twop53 consensus sequences (the decamer

PuPuPuCAf[T/APyPyPy)inthepromoterand/orintronicregions

of transactivatedgenesisamechanismforp53actionas

transcrip-tionfactor(Bourdonetal. 1997).IL-15promoterhasnot yetbeen

studied.butap53consensussequenceispresentinthefirstintron

of the IL-15 gene (GAACTGGCCT. position 18-27 of the

sequenceX91233of GenBank).

An altemative to p53 consensus sequence binding. a

'non-specific' mechanism basedonp53 proteininteraction with general

transcription factors(Ko and Prives. 1996). has been hypothesized

forp53-induceddown-modulationofIL-2 and IL-4 (Peschetal.

1996). IL-15could share this mechanismwith otherinterleukins.

Itmustbe remembered that IL-2 and IL-15 share receptors and

activities (Grabsteinetal. 1994;Tagayaetal. 1996).

Different p53 propertiesarerelatedtodifferentproteindomains

(Koand Prives. 1996). The ts-p53 gene wetransfectedharbours

the temperature-sensitive Val-135 mutation in the DNA-binding

centralcoreregion: alterations inthis domainareresponsiblefor

properties ofp53 leading to increased tumorigenic potential of

cellsresulting fromchanges in the abilitytobindDNA. Val-

135-mutatedp53. however,couldmaintain, atleast inpart theability

to interact with general transcriptionfactors through the

amino-terminal domain.andthis could leadtothe transcription modula-tion by mutated p53found forsomecytokines(IL-2/4 and IL- 15) (Peschetal. 1996.presentstudy).

It is also possible that some of the reported effects may be related to the stateof the cell cyclemachinery. not directly viap53 protein-protein interactions. The study of p21 --FI. a key down-stream effector of p53 inhibition of cell cycle, will helpin clari-fying which mechanism is involved in different effects. P211lF'

transfection (Meng et al. 1998) as well as treatment with

exoge-nous p2l--'F peptides (Ball et al. 1996) could mimicp53 gene effects related to the state of cell cycle machinery. If p21w-F does not mimicp53 effects. thenitis more likely that the mechanismis

related directly top53 non-specific suppression oftranscnrption.

Afew IL- 15transcription modulators [e.g. UV radiations and y-interferon(IFN-y)] which leadtoanup-modulationinnormal cells

areknown so far (Tagayaet al. 1996). Wild-type p53 leadsto a

decreased transcription: this is the first example of a down-modulationof IL-15.

Thestudyof proliferative ability of stableRD-tsp53transfectants showed that whereas at non-permissive temperature

RD-tspS3-transduced cells had a

growth

curve close to RD-Neo control. expression ofp53inthewild-type conformation ledto analmost complete arrestof growth that could be determined by increasein

GI

arrestandin apoptosisrate. IL-15protectsfromapoptosis both invitro and invivo (Bulfone-Paus et al. 1997). Therefore. IL-15

andp53appearto exert

antagonistic

rolesoncell apoptosisrate.

IL- 15 produced by rhabdomyosarcoma cells could act in a

paracrine wayin the tumourmicroenvironment. interacting with cells

expressing

IL-15 receptors. such as lymphocytes and mast

cells. butatpresentit is notknown whether andhow this cytokine modifies the tumour-hostrelationship. Moreover. IL-15 has been reportedtopromoteangiogenesisin vivo(Angiolilloetal.1997).a

phenomenon thatisbecomingmoreandmorerelevantfor malig-nancy. IL-15 release in culture medium by rhabdomyosarcoma cellsreached very lowlevels.of the orderof 5-10pgml-' (present data and Lollini et al. 1997). These levels areapparently lower than thethreshold(about 25

pg

mn-' )requiredforin vitro

lympho-cyte proliferative responses (Grabstein et al. 1994: our

unpub-lished data). butinthelocalinvivo microenvironmentshort-range

acting

cytokinescould have significant effectsevenatlow

concen-trations.Inparticular.forIL- 15ithas beensuggestedthat

receptor-positive cells able to release the cytokine could store a high

concentration of IL-15 molecules on their surface. ready to be

presentedtopassengercells(Meazzaetal. 1997).

Thepossibility that IL- 15 alsoacts in anautocrinewaymustbe considered because RD rhabdomyosarcoma cellsdisplayboth the

specific a-chain of the trimeric receptor and IL-15 membrane

binding,

sites.asshownbytheIL-15-IgG2bfusionprotein. Some data on the effectiveness of IL-15 in muscle cells have been reported(Quinn et al. 1995). We arecurrently studying whether

autocnne effects (suchas induction ofproliferation or modifica-tionsinthe expressionofmembrane molecules) are determined by IL-15inrhabdomyosarcoma cells. However.in vitrotreatmentof rhabdomyosarcoma cells with exogenous recombinant IL-5 or

withmonoclonal and polyclonal anti-IL-15 neutralizingantibodies did not determine any modification in cell proliferation and in

expressionof membraneglycoproteins suchasHLA-A. B.C and CD44. This is in accordance with the absence of proliferating

effects observed in normal muscle cells

(Quinn

etal. 1995). In conclusion. aselective down-modulation of both IL-15 and IL-15 receptorswas found inrhabdomyosarcoma cells transduced withwild-typep53. This indicatesacontrolof IL-15circuitsbyp53

andsuggeststhatfurther efforts should be directedtothestudy of the role of IL-

15S

in theprogression of human rhabdomyosarcoma.

NOTE ADDED

IN

PROOF

The recent availability of the IL-15 promoter sequence (reported

byAzimi et al. 1998. Proc Natl Acad Sci USA 95:2452)prompted us toreanalyse all available IL-15 sequences forp53 consensus binding sites. No consensus sequence was found in the IL- 15 promoter.whereas several clusters of one. two and four decamers (with 0. 2 and 4 mismatchtolerance respectively) were found in introns3. 4 and 5 and in the last exon(GenBankX91233).

ACKNOWLEDGEMENTS

This work is dedicated to thememory ofProfessor

Giorgio

Prodi. MD. PhD. 10 yearsafter his untimely death. The authors would like to thank Gabriella Madriali for her excellent secretarial assistanceand Dr T Troutt (Immunex)forproviding recombinant human IL-15 and monoclonal antibodies Ml 11 and Ml 60. This work was supported by grants from National Research Council. Associazione Italiana per la Ricerca sul Cancro and Ministero

dellFUniversit,aedella RicercaScientificaeTecnologica. Italy. IR is in receiptof a Fellowship from the AIRC.

REFERENCES

AngiolilloAL.KaneganeH.SgadariC. Reaman GH and Tosato G 1997i

Interleukin-15 promotesangiog-enesisin vivo. BiochemBiophYsRes Commun 233: 231-237

Bacchetti S andGraham F 1993ihnhibition of cell proliferationb%adenovirus vectorexpressingthehumanwild type p53 protein IntJ Oncol3: 781-788 Ball KL Lain S. Fahraeus R.Smvthe C and Lane DP(1996 Cellcvclearrest and

inhibitionofCdk4 activitv by small peptides based on thecarbox%--terminal domain ofp2l "> CurrBiol7:71-80

BasergaR 1994( Oncoggenes and the strategN ofgrowthfactors. Cell 79: 927-930 Bourdon J-C.Deg-uin-ChambonV.Lelong J-C. DessenP.Ma P. Debuire Band

Mav E (1997 Furthercharacterization of thep53responsiveelement

-identification of new candidategenes for trans-activation b% p53. Oncogene 14: 85-94

Bulfone-Paus S.Ung-ureanuD.PohlT. LindnerG.PausR.RuckertR.Krause Hand Kunzendorf U(1997( Interieukin-15 protects from lethalapoptosisin -%'o. .Vature Med 3: 1124-1128

DeGiovanni C. Melani C.NanniP. LanduzziL. Nicoletti G. Frabetti F.Gnrffoni C. Colombo MP and Lollini PL(1995 Redundancy ofautocrineloopsinhuman rhabdom%-osarcoma cells. Induction ofdifferentiationb'-suramin. BrJCancer 72:1224-1229

Diller LSexsmithEGottliebA. LiFP andMalkin D(1995 Germlinep53

mutationsarefrequently detectedinyoungchildrenwith rhabdomyosarcoma-JClinInvest95: 1606-1611

El-Badr OM.MtinnitiC. Kohn EC.HoughtonPJ.Daughada% WAHand HelmanLI

(1990 Insulin-likegrowthfactor II actsasanautocrine growth and motlitv factor inhuman rhabdomvosarcoma tumors. Cell Growth Different 1: 325-331 EwenMEandMillerSJ(1996(p53andtranslational control.BiochimBiophvsActa

1242:181-184

Felix CA.Chavez Kappel C.MitsudomjT. Nau MM. TsokosM1.Crouch GD. Nisen PD.WinickNJand HelmanLI 1992' Frequency and diversity of p53 mutationsinchildhood rhabdomr osarcomaCancerRes52: 2243--2247

FujiwaraT.Mukhopadhvax T. CaiDW.Morris DK. Roth JA andGrimm EA(1994(

Retroviral-mediatedtransductionof p53 gene increases TGF-beta expression in

ahumanglioblastomacell fine.IntJ Cancer56:834-839

Grabstein KH. Eisenman J. Shanebeck K_ Rauch C. SrinivasanS.Fung\V Beers C.

RichardsonJ.SchoenbornMA. Ahdieh M. JohnsonLAlderson AR. Watson

JD. Anderson DM andGiri JG(1994)CloningofaTcellgrowth factorthat interacts withthebeta chainoftheinterleukin-2 receptor. Science 264: 965-968

HuffCA. YuspaSH andRosenthalD( 1993( Identification ofcontrol elements3'to the human keratin Igenethatregulatecell tpeanddifferentiation-specific expression.J BiolChem268: 377-384

Keleti J.QuezadoMM. Abaza WM. RaffeldM andTsokosMi(1996(TheMDM2 oncoproteinisoverexpressedinrhabdomrnosarcomacell lines and stabilizes

wild-typep53protein.AmJPathol 149:143-151

Kern SEPietenpol JA. Thiagalingam S. SeymourA. Kinzler KW and Vogelstein B 1992)Oncogenicformsof p53 inhibitp53-regulatedgene expression. Science 256:827-830

KoLU andPrives C (1996) p53: puzzle and paradigm. Genes Dev 10: 1054-1072

Lollini P-L Palhneri G. De Giovanni C.Landuzzi L Nicoletti G. Rossi I.Gnrffoni C. Frabetti F. ScotlandiK.Benini S. Baldini N. SantoniA andNanni P 1997) Expression ofinterleukin-15(IL- 15)inhumanrhabdomyosarcona. osteosarcoma- and Ewing's sarcoma Int J Cancer 71: 732-736

Meazza R.GaggeroA.NegliaF. Basso S. Sforzini S. Pereno R.AzzaroneB and FerriniS 1997) Expressionof tw-ointerleukin- 15 mRNAisoformsin human

tumorsdoesnotcorrelate withsecretion: role of different sigmalpeptides. Eur J Immnunol 27: 1049-1 054

MengRD. Shih H.PrabhuNS. GeorgeDLandEI-Deirn WS (1998)Bypassof abnormalMDM2 inhibition of p53-dependent growthsuppression. Clin Cancer

Res4:251-259

Onu A.Pohl T. Krause H and Bulfone-Paus S (1997)Regulationof IL-15 secretion via the leader peptide of two IL- 15 isoforms. JImmunol 158: 255-262

PappoAS.ShapiroDN. Cn'st WM andMaurerH-M (1995)Biolog- andtherapyof pediatncrhabdomr osarcoma J Clin Oncol 13: 2123-2139

PeschJ. Brehm U. Staib C and Grummt F(1996)Repressionofinterleukin-2 and interleukin-4 promoters b- tumorsuppressorproteinp53. J InterferonCr-tokine

Res 16:595-600

QuinnLS.HaugkKLandGrabsteinKHR 1995)Interleukin-1i:anovelanabolic c,vtokine for skeletal muscle.Endohrinology136: 3669-3672

SchweigererL Neufeld G.MergiaA.Abraham JA.Fiddes JC andGospodarow-icz

D(1987)Basicfibroblast growthfactorin humanrhabdomyosarcoma cells: implications fortheproliferationandneovascularization of myoblast-derived tumors.ProcNail Acad Sci LISA 84:842-846

ShinTH. PatersonAMand KudlowJE (1995)p53 stimulates transcription fromthe

humantransforminggrowthfactor alpha promoterapotential growth-stimulatorn role for p53. Mol Cell Biol 15: 4694-4701

SodduS. Blandino G. Citro G. ScardigliR. Piae.eio G.Ferber A. Calabretta Band SacchiA(1994)Wild-tvpe p53geneexpressioninduces granulocytic differentiation of HL-60 cells.Blood 83: 2230-2237

Tagaya Y.Bamford RN.Defilippis AP andWaldmann TA(1996%) L- 15: a pleiotropic cytokinewithdiversereceptor/signalingpathways whose expressioniscontrolledatmultiple levels.Immunity4: 329-336

Zhang L. Kashanchi F. Zhan Q. Zhan S. Brady IN. Fornace AJ. Seth P and Helman Ui (1996)Regulationof insulin-likegrowthfactor 11 P3 promoterbyp53:a

potential mechanism fortumori2enesis.Cancer Res56: 1367-1373