High affinity insulin binding to an atypical insulin like growth factor I receptor in human breast cancer cells

High-affinity insulin binding to an atypical

insulin-like growth factor-I receptor in human

breast cancer cells.

G Milazzo, … , R Vigneri, I D Goldfine

J Clin Invest. 1992;89(3):899-908. https://doi.org/10.1172/JCI115670.

We studied the nature of insulin receptor binding in MCF-7 breast cancer cells. In both intact cells and solubilized receptor preparations, high-affinity insulin binding was seen. However, unlabeled insulin-like growth factor-I (IGF-I) was five-fold more potent in inhibiting 125I-insulin binding than 125I-insulin itself. With monoclonal antibodies to the 125I-insulin receptor, 30% of 125I-insulin binding was inhibited. In contrast when alpha-IR3, a monoclonal antibody that recognizes typical IGF-I receptor, was employed over 60% of 125I-insulin binding was inhibited. The B29-MAB-125I-insulin photoprobe was then cross-linked to MCF-7

membranes. Cross-linking was inhibited by both unlabeled insulin and IGF-I. Further, the B29-MAB-125I-insulin photoprobe cross-linked to MCF-7 membranes was strongly immunoprecipitated by alpha-IR3. Employing sequential affinity chromatography with insulin-Affi-gel followed by insulin receptor monoclonal antibody agarose, atypical insulin binding activity was separated from insulin receptor binding activity. This atypical receptor had intrinsic tyrosine kinase activity. Both insulin and IGF-I stimulated the phosphorylation of the receptor's beta subunit. In MCF-7 cells both IGF-I and insulin stimulated

[3H]thymidine incorporation; alpha-IR3 blocked all of the IGF-I effect but only 50-60% of the insulin effect. This study demonstrates in MCF-7 cells that, in addition to typical insulin and IGF-I receptors, there is another receptor that binds both insulin and IGF-I with high affinity.

Research Article

Find the latest version:

High-Affinity

Insulin

Binding

to

an

Atypical

Insulin-like

Growth

Factor-I

Receptor in Human Breast Cancer Cells

Giovanni Milazzo,* CecilC.Yip,t BettyA.Maddux,* RiccardoVigneri,§and Ira D. Goldfine*

*Divisionof Diabetesand EndocrineResearch, MountZionMedicalCenter, University of California, SanFrancisco, SanFrancisco, California 94120; tBantingandBestDepartment ofMedicalResearch, University of Toronto, Toronto, Ontario, M5GIL6,

Canada;andOCattedradiEndocrinologiadell'Universita diCatania, Ospedale Garibaldi, 95123Catania,Italy

Abstract

We studied thenature of insulinreceptorbinding in MCF-7 breastcancercells.Inbothintact cells and solubilizedreceptor

preparations,high-affinityinsulinbindingwas seen.However, unlabeled insulin-like growth factor-I (IGF-I) was five-fold

morepotentininhibiting

"HI-insulin

binding than insulin itself. Withmonoclonalantibodiestotheinsulinreceptor,30%of125I1

insulin bindingwasinhibited.In contrastwhena-IR3,a

mono-clonalantibody thatrecognizes typicalIGF-Ireceptor, was

em-ployedover60% of

1251-insulin

bindingwasinhibited. The

B29-MAB-'251-insulin

photoprobewasthencross-linkedtoMCF-7 membranes.Cross-linkingwasinhibited by both unlabeled in-sulin and IGF-I. Further, the B29-MAB-1251-insulin

photo-probe cross-linkedtoMCF-7 membraneswasstrongly

immu-noprecipitatedbya-IR3.Employing sequential affinity chroma-tography with insulin-Affi-gel followed by insulin receptor

monoclonal antibodyagarose,atypicalinsulinbindingactivity wasseparated from insulinreceptorbinding activity. This

atypi-calreceptorhadintrinsictyrosinekinaseactivity.Bothinsulin and IGF-I stimulated thephosphorylation of thereceptor's,B

subunit. In MCF-7 cells both IGF-I and insulin stimulated

13H1thymidine

incorporation; a-IR3 blocked all of the IGF-I effect but only50-60%of theinsulineffect.This study

demon-stratesinMCF-7 cells that, in additiontotypical insulin and

IGF-Ireceptors,there is anotherreceptorthatbinds both insu-lin andIGF-I with highaffinity. (J. Clin.Invest. 1992.89:899-908.)Key words: insulinreceptor* insulin-like growth

factor-I*

oncogene *

photocross-linking.

tyrosinekinase

Introduction

Receptors that havetyrosinekinase activity in their intracellu-lardomain playakey role in both normal and neoplastic cell growth. Theinsulinreceptor

(IR)'

belongsto the

tyrosine-ki-nasegrowth factorreceptorfamily(1-3), and insulin mediates

proliferative responsesina varietyofboth normal and trans-Addressreprintrequests toDr.Milazzo, Division of Diabetesand En-docrineResearch,MountZion MedicalCenter,University of Califor-nia,SanFrancisco,P.O. Box 7921, San Francisco,CA 94120.

Receivedforpublication10 July 1991 andinrevisedform23 Oc-tober1991.

1.Abbreviationsusedinthispaper:IGF-I,insulin-like growthfactor-I; IR, insulinreceptor;MAB, monoazidobenzoyl;WGA, wheat germ

ag-glutinin.

formed cells (4, 5). However, the role of the IR molecule in human neoplasia hasnot yetbeen established. Recently, we

havereported thatoverexpression oftheIRisacharacteristic feature ofmanyhuman breastcancerspecimens(6). Byusinga specific IRradioimmunoassay, we foundthat theaverage IR contentofhuman breastcancerspecimenswasfive-tosix-fold

higher than thatofnormal breasttissue. With immunohisto-chemical analysiswelocalizedtheincreasedexpressionof the IR to the malignant epithelial cells. The IR content ofthe breastcancerspecimenswaspositivelycorrelated withtumor

grade andtumorsize. Theseobservationssuggested,therefore, apossibleroleforIRsinhumancancerinitiation and/or pro-gression.

Humanbreastcancercells intissuecultureareimportantin vitro modelsforstudyingtheregulation ofbreastcancertissue byhormones andgrowth factors,andestablished estrogen-sen-sitive lines ofbreastcarcinomacells intissue culture,havebeen usedtounderstandtheregulation ofhuman breastcarcinoma

growth bypolypeptide growth factors(7). Accordingly, these cells have beenemployedtostudythe effectofinsulinonthe

regulation of growthandother cellularfunctions (8). However,

whether insulin mediates the growthofbreastcancercells via theIRiscontroversial.Ononehand, Osborne,

Lippman

and colleagues havereportedthatconcentrations of insulinaslow

as 10pM stimulate the growth ofMCF-7 cells,an estrogen-sen-sitiveline thatcontainsrelativelyhighlevelsof insulinbinding (8, 9). Further,theeffects of insulin on mitogenesisonthese cells were not inhibited by the monoclonal antibody a-IR3 (10),aspecific antagonist oftherelatedinsulin-like growth

fac-tor-I(IGF-I)receptor( 11).Thesefindingshavesuggested,

there-fore, that insulin, via its own receptor, regulates growth in MCF-7 cells. On the other hand, two other groupshave re-portedthat muchhigher concentrations of insulinareneeded

tostimulate the growthofthese cells (12, 13). Since the related

ligand, IGF-I,stimulated the growth ofMCF-7 cells at much

lowerconcentrationsthan insulin, and since insulin interacts with IGF-Ireceptors,thesegroupsconcludedthat theeffect of

insulinonthe growthofMCF-7 cellswasduetotheinteraction of insulin with theIGF-I receptor. However, in noneof the abovestudieswasinsulinbindingtoreceptorsineitherMCF-7

orother breastcancercellsinvestigatedindetail.

The study of insulin and IGF-Ireceptorsin target cellsis

complicatedbothbythesimilarity of thetwoligands, insulin

andIGF-I, and thesimilarityof theirtworeceptors(14). Also,

bothatypicaland/or alternativeIGF-Ireceptors andIGF-I/IR hybrids havebeenreportedin several cell types(15-18).Inthe present study, in order to investigate the insulin receptor in MCF-7 cells,weemployed(a)monoclonal andpolyclonal anti-bodiesspecific forinsulinandIGF-Ireceptorsand(b) photo-probes of insulin and IGF-I. We now find evidence that in

MCF-7cells thereisanunusualreceptor that hashigh affinity

J.Clin. Invest.

© The American Society for Clinical Investigation,Inc. 0021-9738/92/03/0899/10 $2.00

for both insulin and IGF-I.This unusual receptor maysignal the effects ofboth insulin and IGF-I.

Methods

Thefollowing materialswerepurchased:Dulbecco'smodified Eagle's medium (DME H-21), nonessential amino acids, fetal calfserum (FCS),glutamine, penicillin,streptomycin, and trypsin from the Cell CultureFacility, UniversityofCalifornia,SanFrancisco;tissueculture flasksanddishes from FalconLabware,LosAngeles, CA; MCDB-170 mediaandbovinepituitary extractfrom Clonetic Corp., SanDiego,

CA;epidermal growth factor from CollaborativeResearch, Bedford,

CA; bovineserumalbumin (BSA,radioimmunoassay grade),Triton X-100,bacitracin,phenylmethylsulfonyl fluoride(PMSF), transferrin,

hydrocortisone, isoproterenol, human y-globulin, sonicated salmon spermDNA, dextransulfate, poly (Glu-Tyr) 4:1,andporcine insulin werepurchasedfrom Sigma Chemical Co., St. Louis,MO; wheat germ agglutininagarose (WGA) from ICNBiomedicals, Costa Mesa, CA; formamide and proteinaseKfrom GibcoBRLLife Technologies, Inc.,

Gaithersburg, MD;Affi-gel 15, dithiothreitol and reagents for SDS PAGE from Bio-Rad Laboratories, Richmond, CA; IGF-I from Eli Lilly & Co., Indianapolis, IN; '25I-labeledinsulin (sp act 2,200Ci/ mmol), [3H]thymidine (82.3 Ci/mmol) and [y-32P]ATP (3,000 Ci/ mmol) fromNewEngland Nuclear, Boston, MA; '25I-labeledIGF-I (sp act2,000 Ci/mmol) and[a-32P]dCTP(3,000Ci/mmol)from Amer-shamCorp.,ArlingtonHeights,IL.

Sepharose-proteinA wasfrom Pharmacia(Canada).The radioac-tive photoprobe ofinsulin,

N'29-monoazidobenzoyl

insulin

(B29-MAB-'25I-insulin)was prepared as previouslydescribed (19). N`-monoazidobenzoylIGF-I(B28-MAB-l251-IGF-I)will bedescribed else-where.

For moststudiesweused a lineofMCF-7kindlydonatedbyDr.I. Perroteau(UniversitadiTorino, Turin, Italy).Other MCF-7 and ZR-75-1cellswerefrom Cell CultureFacility,UniversityofCalifornia,San Francisco; T-47Dcells were from Dr. C. Sonnenscheinand Dr. A. Soto, TuftsUniversity, Boston, MA. These cell lineswere routinely grown in DME H-21 with 10% fetalcalfserum, nonessentialamino

acids, glutamine, penicillin, and streptomycin. The medium was

changedevery2 d. 184nonimmortalizedbreastepithelialcells(from Dr.M.Stampfer,LawrenceBerkeley Laboratory,University ofCalifor-nia, Berkeley, CA) were cultured as previously indicated (20) in MCDB-170media, supplementedwithbovinepituitaryextracts,

insu-lin, epidermalgrowthfactor,hydrocortisone, isoproterenol,and

trans-ferrin.Anti-IR monoclonalantibodieswerepreparedaspreviously de-scribed(21). a-IR3,amousemonoclonalantibodytothe IGF-I recep-tor (11) was a gift from Dr. S. Jacobs (Burroughs Wellcome Co., ResearchTrianglePark, NC).AntibodytopeptideIIcorrespondingto

the insulinreceptor sequence 241-251waspreparedaspreviously de-scribed(22).

RNA

analysis

Poly(A)+RNA wasextracted from cellmonolayers(typicallyI0O cells) usinga new, one-step methodas previouslydescribed _{(23). Briefly,} adherent cellswerereleased withproteinaseK(final0.3mg/ml)and solubilizedin 1%sodiumdodecylsulfate. Oligo(dt)-cellulosewas di-rectlyaddedtothelysateandincubatedovernightat22°C.Poly(A)+ RNAwaselutedfromoligo(dt)-cellulose by adding3 mlof10 mM Tris with0.1 mM EDTAand0.2% SDS.Poly(A)+RNA(8Mg)wasthen electrophoresedon1% agarosegelcontaining2.2 Mformaldehydeand then transferredtonitrocellulosefilters.

Northern blot hybridization for insulin receptor mRNA was

carriedoutusingtwohuman IR cDNAprobes, 18.2 and 13.2(1 and 4.2kb,respectively)akindgiftof Dr. G. I.Bell,UniversityofChicago.

Northern blothybridizationfor IGF-I receptor mRNAwascarriedout

usinganIGF-I receptor cDNApreparedas_previouslydescribed(24).

These werelabeledwith100MACi [32PJCTPby randomprimers (25) to a specific activity of 109cpm/ug. Thenitrocellulose filterswere prehy-bridized andhybridizedaspreviously described(25). Filters were then washedunderhigh stringency conditionstwicefor30minat420Cin lxSSC(20x SSC=3 MNaCI,300 mM Nacitrate, pH 7.0) with0.1% SDSandtwice for 30min at50C in 0.5xSSC with0.1%SDS and autoradiographed.Incertain studies,lowstringencyhybridization was carriedoutbywashing the filter fourtimes for10min at room tempera-tureinlxSSC with0.l%SDS.

'25I-insulin

and

'251-IGF-I

binding to intact

cells

Cells attached to 24-mmtissue disheswererinsed twicewith 1 ml of phosphatebuffered saline(PBS) andincubated for16 h at4VCin 0.5 mlof binding buffer containing 120 mM NaCl, 1.2 mM MgSO4, 15 mMNa acetate, 5 mMKCI, 10 mM glucose, 1 mM EDTA, 10 mg/ml BSA,1 mg/mlbacitracin,50 mM Hepes, pH 7.8, and 40 pM

1251-insu-lin withorwithout increasing concentration ofunlabeled hormones or antibodies. The incubation mediumwasthen removed, the mono-layerswerewashedtwice,the cell lysedwith0.03% SDS and the

radioac-tivity associated with cell lysate was measured in a y-scintillation counter(Beckman Instruments, Inc., Palo Alto, CA). Binding was corrected fornonspecific'251-insulinbinding(< 5%oftotalbinding)as determined in thepresenceof1MMunlabeledinsulin. Specific '25I-in-sulin binding/mg proteinwas7.5%, 5.4%,17.4%and1.2%total

radioac-tivityinMCF-7, T-47D, ZR-75-1, and 184 cells, respectively. Specific '25I-IGF-IbindingtoMCF-7 cellswascarriedoutinthe same manner.Nonspecific binding(< 5%oftotalbinding)was deter-mined inthepresenceof 100 nM unlabeled hormone. Specific

'25I-IGF-Ibinding/mg proteinwas83.9%, 31.6%, 16.7%, and 50% total radioactivityinMCF-7, T-47D, ZR-75-1,and184 cells,respectively.

MCF-7 membrane

preparation

Confluentcellsfrom five flasks (150cm2)wereharvested and centri-fuged. The cell pelletwasresuspended in 50mMTris buffer,pH7.9,

containing10mMbenzamidine,30 ug/mlaprotinin,0.2 mg/ml

baci-tracin, 10mMPMSF,sonicated three times for 20s,andcentrifuged for10 minat2,500rpm.Thepellet obtainedwastreated asaboveone moretime.Thesupernatantsofeachcentrifugationwere pooled and

centrifugedat 11,000rpmfor 10min. Thepelletwasresuspended in 1-2 mlof Trisbuffer,aliquotedinI00-Mlfractions and microfuged for 30min. The supernatantwasfinallyremoved, and thepelletwasstored at-700C.

Photoaflinity

labeling

and immunoprecipitation

Membranesfrom MCF-7 cellswereincubatedovernightat4°Cat a

finalprotein concentration of0.3-0.5 mg/ml with the'251-labeled

pho-toprobein50mMTris-HClbuffer,pH 7.4,containingbacitracin(1 mg/Mml), BSA(0.01%), aprotinin (0.05 mg/ml),benzamidine (10

mM),pepstatin (1.35 Mg/ml),leupeptin (0.95 Mg/ml),and PMSF(1

mM). Afterphotolysis(19),the membranepellet, recovered after

cen-trifugationinaBiofugefor 30min,waseithersolubilized in SDS

sam-ple buffer for SDS-PAGEin 7.2%gel,orsolubilizedinTritonX-100to be usedforimmunoprecipitation.SolubilizationinTritonwascarried out at aproteinconcentration of 1.2mg/mlin1%Triton X-100 in 50 mMHepesbuffer, pH 7.4,containingNaCl(150mM), bacitracin(0.08

mg/ml), aprotinin (0.01 mg/ml), benzamidine (10 mM), pepstatin

(1.35,ug/ml),leupeptin (0.95

Mg/ml),

andPMSF(1 mM). Solubiliza-tionwascarriedoutfor90 min inanice bath followedby centrifuga-tionat29,000g for 23 min.Aliquotsof the supernatant, each contain-ingabout 1.5xI05cpm,wereincubatedovernightat4°Cwith 10Mgof monoclonalantibodytoIR,1

Mg

ofa-IR3,or50 ugofimmunoglobulin

from the human polyclonalantiserumtotheIR or normal human immunoglobulin. Sampleswerethenincubated with

Sepharose-pro-teinA(7mg)for 90 minat4°C.Theimmunocomplexboundtothe

Radioautogramsof the dry gels wereobtainedwith X-OmatAR film (Eastman KodakCo., Rochester, NY)andLightening-Plus

en-hancingscreens.Exposuretime ranged from 3 to 5 d.

Receptor purification by

affinity

chromatography

Step1.Confluent 150-cm2flasks of MCF-7 cells were solubilized in 50 mM Hepes buffer pH 7.6containing1mg/mlbacitracin,ImM PMSF, and 1% Triton X-100 for 60 min at4VC. A clear supernatantwas obtained bycentrifugationat 100,000g for 35minat4VC.The super-natant was thenappliedto a l-ml WGA agarose columnwhichhad beenpreviously equilibratedwith WGA column buffercontaining150 mMNaCl,0.1% TritonX-100, 1 mMPMFS,and 50 mMHepes, pH 7.6.Glycoproteinswereeluted with the same buffercontaining0.3 M N-acetylD-glucosamine (26).

Step 2. The fractions withinsulin-binding activitywerepooledand appliedtoa2.5-mlcolumncomposedof insulincoupledtoAffi-gel15 (27).After an 18-h incubation at4VC,thepass-throughof the insulin-Affi gel column wascollected,the column was washedwith50 mMTris HC1 buffer, pH7.4,containing 1 MNaCl, 0.1% Triton X-100,and 1 mM PMFS, and then elutedwith50 mM Na-acetatebuffer, pH 5.0, containing1 MNaCl, 0.1% Triton X-100,and 1 mM PMSF.Fractions were collectedintubes containing 0.1mlof 1 MTris-ClbufferpH 8. Step 3. Thefractionswithinsulin-bindingactivitywereappliedtoa column of agarosecoupledto monoclonalantibody51directedat the human IR a subunit(28).Thepass-throughwas collected and applied to aWGAcolumn asdescribedin step 1. The columnwaswashed with columnbuffer composed of 50mMHepes, pH 7.6, 1MNaCl, 0.1% Triton X-100, 1 mg/mlbacitracin, and 1 mMPMSF. The IR was elutedwith 1.5 M MgCI2, 0.12 M sodiumtetraborate,0.1% Triton X-100, pH 6.5. Eluted fractions werediluted 10-fold with washing bufferandfinally appliedtoWGA column asdescribedinstep1.

'25I-insulin

and

'25I-IGF-I

binding

to

affinity-purified,

solubilized

receptors

Insulin-bindingactivitywas measured at eachpurificationstepby in-cubatingthe receptor preparation at 4°C for 16 h with '25I-labeled porcine insulin (40 pM)in afinalvolumeof 240 Ml ofbindingbuffer (100 mM Hepes, pH 7.9, 120 mM NaCl, 1.2 mM MgCI2, 2 mM MnC12, 2.5 mMKCl, 0.5 mM Naacetate, 1 mM EDTA, 1 mg/ml bacitracin, 10mg/ml BSA)in the presence orabsenceofexcess unla-beled insulin. Carrier human y-globulin (10 mg/ml)wasadded to each tube. 0.2 ml of 22%polyethyleneglycol(PEG) 6000wasthenaddedand incubated for 15 min at4°C. Aftercentrifugation,precipitateswere washedonce with 1 1% PEG andrecentrifuged, and thesupernatant wasdiscardedand thepelletwascounted. Specific bindingwas ob-tained by subtracting bindingwith1251-insulinplus unlabeled insulin from binding with 251I-insulinalone. Specific125I-IGF-Ibindingto solu-blereceptorswascarriedoutin the same manner.

Tyrosine

kinase studies

Autophosphorylation of atypical IGF-I receptors. Aliquots (20

Ml)

of purified atypical IGF-I receptor preparations obtained asdescribed above, were preincubated with 2mM MnCl2and 100 nM of either insulinor IGF-I in a totalvolumeof 25,lofWGA buffer. After 60min at20°C,thephosphorylation reactionwasinitiatedbyaddingATP (10 AM)and 1MCi['y32-P]ATP.After60min at20°C,samples were added to 10Ml of4XLaemmli buffer containing 100 mMdithiothreitoland boiledfor 5_{min. Phosphorylated receptor preparations were analyzed} on7.5% _{SDS-PAGE under reducing conditions. Radioautogramsof} the dry gels wereobtainedasdescribed above.

Artificalsubstratephosphorylation.For studies of exogenous sub-stratephosphorylation, atypicalIGF-Ireceptor preparationsin 20Mlof WGAbuffer werepreincubated for 60min at200Cwith increasing concentrations of either insulinorIGF-Iinpresenceof 2 mMMnCl2 and 10 mM

MgCl2.

Next, ATP (10

MM),

1

MCi

['y32-P]-ATP

and 1 mg/mlof thesyntheticsubstrate

_{poly(Glu-Tyr)}

4:1were_added,andthe

Kb

Kb

Figure

1.

Northern blot

analysis of IGF-I (A) and

11.0- 1.5-8 _ _{insulin(B) receptors}

5.5-

*.

mRNA.Poly(A)+RNAwas

preparedfrom MCF-7 cells

1GF-I-R

IR and

_(8Mg)

subjectedto

aga-rose

_{gel electrophoresis}

fol-lowed

_by

transferto nitro-cellulose filtersandhybridizationutilizedwithlabeled insulinor

IGF-I receptor cDNA.

incubationwascontinuedforafurther60minat20'C. The reaction

was_stopped

_by

theaddition ofasolution at4VCcontainingafinal concentration of 10 mMATP, 0.1mMEDTA,and0.5% BSA.Finally,

15-,Ml

aliquotswere_spottedin_4-cm2disksof 3 Mpaper(Whatman,

Inc., Clifton,

NJ)

anddried. Diskswerewashedinfour

_changes

of_10% TCA

containing

20 mMsodium

pyrophosphate

andinafinalacetone wash.

_{Radioactivity}

wasdetermined

by

scintillation

counting.

DNA

_synthesis

40 xI03cellswere_platedin24-well tissue cultureplatesintheirregular growthmedium.After48h,themediumwasremoved and_replaced

_by

DME H-21

containing

0.1%BSAand 10Mg/mltransferrin.72 hlater,

variableconcentrations ofeitherinsulin, IGF-I,ora-IR3wereaddedin freshmedium.After 24h,0.5

_MCi

of

_{[3H]thymidine}

wereaddedtoeach wellfora2 h. Cellswerethen

_harvested,

and therateof DNA_synthesis

wasmeasuredasdescribedpreviously( 12).

Other methods

Proteinsweremeasured

by

themethodof

_Lowry

etal._(29).Insulin receptor

radioimmunoassay

was_performedas_previouslydescribed

(30).

Results

Insulin and IGF-I receptor mRNA content in MCF-7 cells.

Poly(A)+

RNAwasextracted from MCF-7 cells and

_subjected

to_agarose

_{gel electrophoresis}

_{and then transferredto}

nitrocel-lulose filters. The mRNAwas

probed

with

specific

32p cDNAs for either insulin or IGF-I _{receptors. Under}

_high

_stringency

conditions, typical

bandsat11 and 8.5 kbwere seenforthe IR

receptorandat 11 and 5.5 kb for the IGF-I receptor

(Fig.

1).

Thesame_patternofbandsforboth receptorswas seenwhen low

stringency

conditionswereused

(data

not

shown).

Characteristics

of

'25I-insulin

binding

toMCF-7andother

cells. We first measured

1251-insulin binding

tohuman MCF-7 breastcancer cells. This

_binding

was inhibited

_by

unlabeled insulinwithahalf-maximal effect_occurringat2 nM

_indicating

that these cells have

high-affinity

insulin

_binding

_(Fig.

2

A).

Surprisingly,

unlabeled

_IGF-I,

which

typically

competes for

'25I-insulin

_binding

tothe

typical

IRwithapotencyof -1%of insulin

_(31),

wasmore _{potent than} insulin in competing for

'251-insulin

binding

toMCF-7 cells thaninsulin itself.The half-maximal effect of IGF-I occurredat0.4nM.2

Inordertodeterminewhether thisunusual

'251-insulin

bind-ing

wasduetothe membraneenvironmentof MCF-7 cells, the

2. Themajorityof studieswerecarriedoutwith MCF-7 cellsobtained fromDr. I._{Perroteau, Turin, Italy.}Similar effectsof insulin and IGF-I

a

B.,

0 0

Up.9

oA

z N

H

-0

H

In

N '.4

100

(A)

80

60-40

-20

I~~GF-l

c

20

-01 At

~

o~10

100

10

IP ,

Up

H

H

Sn

C4a

80

60

40'

20

(B)

INSULIN

IGF-1

SI

10

>

8

o 4

6

0 4

H 2 z

-n 2

z 0 _

Ha

0 ."an Z dp

-.

a ff -10 -9 -8 -7

0 10- 10- 10- 10

(X)

Figure 2. Inhibitionbyinsulinand IGF-I of'251-insulinbinding to intactMCF-7 cells and MCF-7 wheat germ extracts. (A) Cells were platedin culturedishesandspecific '25I-insulinbindingwasthen measured. Binding isexpressed as a percentageof maximalspecific

'25I-insulinbinding.Eachpoint isthemean±SEMof6 separate ex-perimentsintriplicate. (B)MCF-7 cells weresolubilized with1% Tri-tonX-100andthesolubilized materialwasappliedto aWGA col-umn.'251-insulin(20,000 cpm)wasincubated withreceptors in the presenceofdifferent concentrations ofunlabeledinsulinandIGF-I for 16 h at4VC.The percentageof maximalspecific binding of1251_

labeled insulinisplottedagainst the peptide concentrations. Each point isthemean±SEMoftwoseparateexperimentsperformed in duplicate.

cellsweresolubilizedin 1%TritonX-100 and partially purified

byaffinity chromatography withWGA. When

'25I-insulin

bind-ingwascarriedoutwiththisreceptorpreparation,high-affinity

insulinbindingwasagainseen. Aswith intactcells, unlabeled IGF-Iwas more potentthaninsulin ininhibiting

'25I-insulin

binding (Fig.2B).

Inordertodeterminewhetherthisunusual

'25I-insulin

bind-ingobserved inMCF-7cellswas seenin other cells,westudied

both a nonmalignant human breast epithelialcell line (184) and twohuman breastcancercelllines (T-47DandZR-75-1)

(Fig.3).In allthree celllines'25I-insulinbindingwasinhibited by unlabeled insulin withahalf-maximumeffect

occurring

at

300, 500, and <200 pM in 184, T-47D, and ZR-75-1 cells, respectively.In contrast toMCF-7cells,in these three celllines,

unlabeled IGF-Ionlyweaklycompeted for

'25I-insulin

binding;

the potency of unlabeled IGF-Iwas- _{1%thatof insulin.}

Influence ofanti-receptor monoclonal antibodieson

'251-in-sulin

binding

to MCF-7and otherbreast cell lines.Tofurther understandthenatureof

'25I-insulin

bindingto MCF-7 cells,

weemployedthree monoclonal antibodies directed againstthe

a,subunitoftheIR(MA-10,MA-51,MA-20)and oneantibody directed againstthe asubunit of the IGF-I receptor (a-IR3)

104

100

-a

z

.

.94x 0 a

-A

H .49

0-10

!0

INSULIN

0O

101 10

10-9o-8

10-7

184

T-47D

ZR-75-1

I .

.

*-1I

0 '-9 I. * I

0 10- 10-9 108 10-7

(M)

Figure 3. Inhibition byunlabeledinsulinand IGF-Iof'25I-insulin

bindingtothenontransformedbreast cellline (184)and breast cancer celllines (T-47DandZR-75-1).Cells wereplatedin culturedishes andspecific '25I-insulin bindingwasthenmeasured.Binding is ex-pressed as a percentageof maximal specific'25I-insulinbinding.Each point ismean±SEMoftwoseparateexperimentsperformedin tripli-cate.

(Fig. 4). At 100 nM(the highest concentrationtested) a-IR3

inhibited 60% of '251I-insulin binding to MCF-7 cells witha potency that was similar to that of insulin. In contrast,at 100 nM the monoclonal antibodiesagainstthe IRonlyinhibited

125I-insulin binding

to MCF-7 cells by 30%. Thesestudies indicated, therefore, that 251I-insulinbindingto MCF-7 cells was notpredominantly tothe insulin receptor but rather to another receptor whichwas moresimilartothe IGF-I receptor. Other studieswerecarriedoutwith 251I-insulin bindingto receptors in T47D and ZR-75-1 breast cells. In contrast to MCF-7 cells, in these cells two monoclonal antibodiestothe IR (MA-lO, MA-51) completely inhibited 125I-insulin binding whereasa-IR3wasnearlywithout effect(Fig.5).

-o10 -9 _-8

10-7

10 10 10 0

(N()

Figure4.Inhibition of '25I-insulin boundtoMCF-7 cells by five monoclonal antibodies,mouseIgG, and insulin. Cellswereplated in culture dishes and specific '25I-insulin bindingwasthen measured. Binding is expressedas apercentageof maximal specific 1251-insulin binding. Eachpoint is the mean±SEM of threeseparateexperiments performedintriplicate.

IGF-Iand a-IR3 completely inhibited the binding of'25I-IGF-I

toMCF-7 cellswithhalf-maximal effect occurringat1nM and

500 pM, respectively. Unlabeled insulin inhibited '25I-IGF-I binding with half-maximal effect occurringat400 nM. Mono-clonal MA-lO and normal mouse IgG were without effect.

Thesestudies indicated, therefore, that MCF-7 cells havea

nor-malIGF-Ireceptorwhich has only weak affinity for insulin.

1

a

z D M .,

0

z

-H d

H _

H

In N

T-47D

-7

:) 80 M-0 80-.,0g

0 MA-51

H

co 40-ZR-75-1

HL

IH 20- _INSULIN

Un

4 NA-10

010 10-9 10- 10-7

(M)

Figure 5. Inhibition of '25I-insulinboundtoT-47D andZR-75-1cells

bymonoclonalantibodies and insulin. Cellswereplated inculture

dishesandspecific '25I-insulin bindingwasthen measured. Binding is

expressedas apercentageof maximal specific '25I-insulin.A repre-sentativeexperiment is shown.Eachpoint isthemeanof triplicate determinations.

(K)

Figure 6. Inhibition of '251-IGF-ItoMCF-7 cellsbyIGF-I, insulin, monoclonal antibodies, and IgG. Cellswereplatedinculturedishes

andspecific 1251-IGF-I bindingwasthenmeasured.Bindingis

ex-pressedas apercentageof maximal specific '251-IGF-I binding.A representative experiment is shown. Each point ismeanoftriplicate

determinations.

Photoaffinitylabeling.Asanalternativeapproachto

under-stand thenatureof insulinbindingtoMCF-7cells,wecarried outaseriesofexperiments wherebywe photoaffinity-labeled

membranes prepared from MCF-7 cells (Fig. 7). Membranes photolabeled withB29-MAB-'251-insulinandthenanalyzed by SDS-PAGE underreducing conditionsshowedanintensely la-beled band of 120kD. Thelabelingofthis bandwasblockedby both insulin and IGF-I. Asimilar labeled bandwasobserved whenmembraneswerephotoaffinity-labeled with

B28-MAB-'25I-IGF-I. The labeling of this bandwasblockedbyIGF-I but notby insulin. These observations thereforeareconsistent with

theresults obtainedfromligand bindingstudies showninFigs. 2and 6.

We next immunoprecipitated the MCF-7 membrane

re-ceptorsphotoaffinity-labeledwitheitherB29-MAB-'25I-insulin

or B28-MAB-'251-IGF-I (Fig. 8). Theprecipitates were

ana-lyzed by SDS-PAGE after reduction. When membraneswere

photolabeled with B29-MAB'251-insulin,alabeledband of 120

KD

123456

120-

-Figure7.Photoaffinity-labelingofMCF-7 cell membranes with pho-toprobesof insulin(lanes1-3)andIGF-I(lanes4-6)in the presence

ofeither ALMunlabeled insulin(lanes2and5)orunlabeled IGF-I

(lanes3 and6).Thephotolabeledmembranesweresolubilizedfor SDS-PAGE in 7.2%gelunderreducing conditions.

AnAtypical ReceptorinHuman Breast Cancer 903

Q

z

3

0 _ M j

of

z H P

_

In

100

KA-20

iNA-10 MA-51

m

d%

0 r

0a

H I id

v1-I

UN N

80

60

40

20

INSULIN

KD

120-IGF-I

KD

120-

i

MONOCLONAL p (1 i

-ANTIBODIES CD _°

IX

U- ra

-o

Figure8.Immunoprecipitationof receptors from MCF-7 mem-branes, after photocross-linkingwithinsulin photoprobe (left) or IGF-I photoprobe(right).Immunoprecipitation wascarriedout using monoclonalantibodies directedagainsteither the IR (51 =antibody MA-5 1;20 =antibodyMA-20; 10 =antibodyMA-lO)or the IGF-I receptor(a-IR3). Theimmunoprecipitatewas separated by SDS-PAGE on7.2%gel.

kD wasprecipitated by the IGF-I receptor monoclonal anti-body a-IR3.Incontrast,precipitation obtained with IR mono-clonalantibodiesshowedlittle or no labeled 120-kD band (Fig. 8,left panel).

Whenmembranes were photolabeled with the IGF-I photo-probe, a 120-kD band precipitated bythe monoclonal anti-body a-IR3 wasmuch more intensely labeled than with the

insulinphotoprobe. Inaddition,averyweaklylabeled bandof

- 200 kD was also detected.This bandmostlikely isanIGF-I

receptor precursor. Precipitation ofthelabeled 120-kD band

by the otherantibodies was minimal (Fig. 8, right panel). These

observationsareinagreement with the dataonligandbinding

MCF-7

KD

HTC-IR

which demonstrated that binding ofinsulintoMCF-7 cells was

primarily to receptors recognized by antibody a-IR3 to the IGF-Ireceptor,andnotbyantibodiestothe IR.

Wehavepreviously reportedthat residues241-251 of the IR asubunit mayplayan importantrole in insulin binding

(32). Inaprevious study (22)weobserved that antiserum AP-II raised against a synthetic peptide containing this sequence,

while neither reactive towards the nativeIR norinhibitoryto

insulinbinding,wasabletoprecipitatethedenaturedasubunit

ofthe IR photolabeled with B29-MAB-'25I-insulin.

Accord-ingly, MCF-7 membraneswerephotolabeledwith theinsulin

Up

p ,, 0d

H 0

I n V

-dP

H -..

UrI

100'

z

o, 80

m ,,

H A 60

H 4

H

en 20

'-A

B

7

1

07

130-:>1:

_

XV

-o

-i

m

z

0

I-X:

c)

a0 -U

-u

-_i

m

a z

3

0

m^I

'-4

H dP

K

* H

d

_U)

N

C)

Figure 9.Immunoprecipitationof membrane receptorsphotolabeled with insulin photoprobe by antiserum AP-IIafterreduction and al-kylation. Left:Membranesfrom MCF-7 cellsphotolabeledwith the radioactive insulin photoprobeweresolubilized in Triton X-100. The solubilized materialwasreduced andalkylated,immunoprecipitated

withantiserumAP-Il,andanalyzedby SDS-PAGE as described(22). Right:MembranesfromHTC-IR cells(expressinghumanIRs)were

photoaffinity-labeledwith insulin photoprobe. The labeled mem-branesweresolubilized,reduced andalkylated,and immunoprecipi-tatedwith antiserum AP-IIasdescribed(22).

10c _C

1 on7

(M)

Figure 10. Separationof theinsulin, IGF-I,andtypicalreceptor binding activities byaffinity chromatography.MCF-7 cellswere sol-ubilized and the insulinbindingactivity subjectedtoaffinity chro-matographyoninsulin-Affi-gel followedbychromatographyonIR monoclonalantibody-agarose. (A)SpecificIGF-IbindingtoIGF-I receptors in thepass-throughof theinsulin-Affi-gelcolumn.(B)

Spe-cific IGF-Ibindingtoatypicalreceptors in thepass-throughof the IRmonoclonalantibodycolumn.(C)Specificinsulinbindingto in-sulin receptors eluted from theinsulinreceptormonoclonalantibody

column.

inn-10

KD 1 2 3 X

N

0

-W

103-~~ ~ ~ ~ ~

0E4

c)

so z 0

~ ~

I_LUU-nn_

80-

60-

40-

20-01

r

IGF-I

INSUL

-lo

10-(4)

Figure 11. IGF-I and insulin-stimulated

tyro-sinephosphorylation oftheatypicalreceptor. Left:Autophosphorylation oftheatypical re-ceptor. Theatypicalreceptorwasallowedto undergophosphorylationin absence (lane 1) orpresenceof 100nMofeither IGF-I (lane 2) or insulin(lane3),andanalyzed by

SDS-PAGE under reducing conditions. Right: Ty-,IN rosine kinase activity of the atypicalIGF-I

re-ceptor. Theatypicalreceptorwasincubated in the presenceof increasingconcentrationsof either insulinorIGF-I,and 1MCi[y32-P]-ATP, with 10 M unlabeled ATP, and 1 mg/ml of '' the substrate poly (Glu-Tyr) 4:1. Valuesare

themean+SD for a representative of three ex-perimentscarriedoutinquadruplicate.

photoprobe, solubilized,and, after denaturation, immunopre-cipitated with antiserum AP-I. Fig.9 shows the absence of the

labeled 120-kD band in the precipitate. Undersimilar condi-tions the antiserumAP-IIprecipitatedahighly labeledbandof 135 kD from membranes prepared from HTC-IR rat hepa-tomacells that have been transfected with human IR cDNA,

andexpress IR(33). The precipitation of the 135-kD bandwas

inhibited byanexcessof the synthetic peptide (Fig. 9). These data provide further evidence thatthebinding of insulin to

MCF-7 cells predominantly occurred on receptorsthatwere nottypicalIRs.

Twoother polyclonal antiserato the IR werestudied; an

antiserum ARS-2totheIRfromapatient with insulin

resis-tance(22) andanantiserum raisedtothecarboxylterminalof theIR (3subunit (34). Neither antisera reacted with the MCF-7

receptors photolabeledwith B29-MAB-'251-insulin (data not

shown).

Purification oftheatypical IGF-I receptor byaffinity

chro-matography.After MCF-7 cells were solubilized and glycopro-teins purified by WGA affinity chromatography, they were subjected toinsulin affinity chromatography on insulin-Affi-gel column. The pass-throughofthis column contained recep-tors thatbound IGF-Iwithhighaffinity (Fig. 10 A), and this

bindingwasweakly inhibited by insulin (< 1% thepotencyof IGF-I). The material eluted from the insulin-Affi-gelcolumn was thensubjectedtoaffinity chromatographyon a columnof

IR monoclonal antibody MA-51 agarose. The pass-through from this columnboundIGF-I with high affinity,but unlike

typicalIGF-I receptors, thisbindingwasinhibited by insulin withapotencyof about 10%thatofIGF-I(Fig. 10 B).

High-af-finityinsulin bindingwasalsoseenin the fraction, but unla-beled IGF-Iwasmorepotentthaninsulin ininhibiting

'251I-in-sulinbinding(datanotshown). Thereceptoreluted from the MA-51 affinity column bound insulin with high affinity and thisbindingwasveryweakly inhibited by unlabeled IGF-I(Fig.

10 C). SpecificIRradioimmunoassay (30) revealed thatthe

MA-5 1 columnabsorbed > 99% ofthe IRimmunoreactivity. Tyrosine kinase activities of the atypicalIGF-I receptor.

EmployingtheatypicalIGF-Ireceptorpreparation purifiedby affinitychromatography,we nextcarriedoutstudiesto deter-mine whether thisatypical receptor contained

figand-stimu-latedtyrosine kinase activity,acharacteristic of typical insulin andIGF-Ireceptors.

When theatypicalreceptor wasallowedtoundergo auto-phosphorylationand then analyzed by SDS-PAGE under

re-ducingconditions,itcontainedalabeledband of 103kD(Fig. 11 A). The phosphorylation of this 103-kD band increased

when theatypicalreceptor wasstimulated with either 100nM

insulin or IGF-I (Fig. 11, left). This band was

significantly

larger thanthetypical 95-kDIR (3subunit (not shown).

Next,thetyrosine kinaseactivity of theatypical receptor

toward the exogenoussubstrate, poly(Glu-Tyr) 4:1,was stud-ied.Stimulation oftheatypicalreceptorwith both insulin and

IGF-I produced adose-dependent increase of 32P

incorpora-tion intopoly(Glu-Tyr)4:1 (Fig. 11,

right).

Stimulation of

sub-strate phosphorylation occurred at lower concentrations of IGF-I than insulin.

Insulin and IGF-I stimulated DNA synthesis in MCF-7

cells.Inordertoinvestigatethemitogeniceffect ofboth insulin and IGF-IonMCF-7cells,westudied

[3H]thymidine

incorpo-ration into DNA. Both insulin and IGF-I stimulated [3H]-thymidine incorporationintoDNAinadose-dependent man-ner;however, IGF-Iwas 10 foldmorepotentthan insulin

(Fig.

12).We nextstudied theeffectof a-IR3oninsulin and IGF-I-stimulated[3H]thymidine

incorporation (Fig.

13).MCF-7cells

werestimulated with either1 nMinsulinor 1 nMIGF-I.Inthe

o 400

04W 300

00

aU / - IN~~~SULIN

200-@0

0

10 0

-10 _{1 °- 1o0} 8 10-7

(M)

Figure 12.Stimulationof[3H]thymidineincorporationbyinsulin and

IGF-Iin MCF-7 breastcancercells.MCF-7 cellswereplatedin tissue cultureplatesintheirregulargrowthmedium. After 48 h mediumwas

replacedwithserumfree medium. Afterafurther 72 h the cellswere

stimulated with either insulinorIGF-Ifor 24 h.Thymidine incorpo-rationwasperformed duringthe last 2 hof stimulationbyadding0.5 MCiof[3H]thymidine.Cellswerethen harvested and therateof

[3H]-thymidineincorporationwasdeterminedaspreviouslydescribed(12).

AnAtypicalReceptor in HumanBreastCancer 905

I

a

0 .64

a 00r

a

64o *d0

.64

E4V _

rq

A

100

80 60

40

20

n

0 1011 10 10 10 9 10-8 0-7

alpha IR-3 (M)

Figure 13.Effectof monoclonal antibodya-IR3on

_{[3H]thymidine}

incorporation stimulated by insulin and IGF-I in MCF-7 breast cancercells. Cells were cultured in tissue culture plates as described aboveandstimulated for24 h with1nM of either insulin or IGF-I in the absence and presence of different concentrations ofa-IR3.

[3H]-thymidine incorporation wasperformed as above described. The amount of[3H~thymidineincorporatedin absence of either insulin orIGF-Iwassubtracted to the amount of[3H]thymidineincorporated in presence of insulin andIGF-I.Normal mouse IgG at 100 nM pro-ducedslight inhibition (10%) of[3H]thymidine incorporation. Each valueismean±SD of three separate experiments performed in tripli-cate.

presence of a-IR3 there was a complete inhibition of IGF-I-stimulated[3H]thymidineincorporation at 10 nM ofantibody.

Incontrast, at this a-IR3 concentration, the effect of insulin was inhibited only by 50-60%.

Discussion

Insulinand IGF-I exerttheir biological effectsoncellular

me-tabolism and growth by bindingtotheirdistinctplasma mem-brane receptorswhichshow manysimilarities ofstructure and

function.Botharea22tetramerswithtwoextracellular a

sub-unitsthatbindthe hormone and two transmembrane A

sub-unitsthat havetyrosine kinase activity (14). Theinsulin and typeIIGF receptors haveahigh degree ofsequence homology, aresimilarin molecularsize,andsignal via activation of

tyro-sinekinasedomains (14, 35-38). However, thetworeceptors

functionallyaredistinct,asassessedby clearly different binding affinities for variousligands (38)andcertainreceptor anti-bodies(39).Thecloning and sequencing ofcDNAscoding for

thesetworeceptors havedemonstratedthat receptorsfor

insu-lin and IGF-I are productsofseparate genes(40, 41). Insulin

and IGF-I themselves have sequence homology and can weakly interact with each other's receptors (31, 42).

Inthepresent study, several linesofevidence suggested that theinteraction of insulinwith MCF-7 cellswaspredominantly

via a receptor that was not the typical insulin receptor. First,

competitionstudies of

1251-insulin

bindingtoboth intact

MCF-7cells and solubilizedreceptors demonstratedhigh-affinity in-sulin binding. However, thisbinding was inhibited by unla-beled IGF-Iatlowerconcentrationsthanunlabeledinsulin it-self. Second, the binding of insulin to MCF-7 cells was

markedly inhibited by monoclonal antibody a-IR3, an anti-body thatreactswith the IGF-I receptor butnotwith theIR. Third, insulin bindingwasweaklyinhibitedbythemonoclonal

antibodies which react with the IR, but not with the IGF-I receptor. Fourth, similar observations were obtained in experi-ments where membrane receptors of MCF-7 cells were

pho-toaffinity-labeled withphotoprobesof insulinand IGF-I. The presence ofthis atypical receptor in MCF-7 cells was not a

general feature of eitherbreastepithelialorbreast cancer cells.

Typical IR bindingwas seen with 184 breast epithelial cells as well as in two breast cancer cell lines, ZR-75-1 and T-47D.

Thequestionwhich arises, therefore, is the nature of the receptor or receptors that insulin binds to in MCF-7 cells. MCF-7 cells havetypicalIR mRNA and a small fraction of

insulin bindingwas blocked by anti-IR monoclonal antibod-ies. Moreover, typical IR were isolated by affinity chromatogra-phy. Thus a small proportion of125I-insulinbinding was most likely to atypicalIR. SinceMCF-7 cells have IGF-I receptors andinsulin is known to weakly bind to IGF-I receptor, one

possibility isthat theatypical insulin bindingwasto the IGF-I receptor. For two reasons, this possibility was considered highly unlikely. First, since MCF-7 cells have typical IGF-I receptors and insulin only weakly competed with IGF-I bind-ing to the IGF-I receptor, it is highly unlikely that the high-af-finityinsulin binding istothe typical IGF-I receptor. Second,

184breastepithelialcells, ZR-75-1 cells, and T47-D cells also havehighconcentrations ofIGF-I receptors (43), but atypical

insulin bindingwas not observed in these cells.

Invitro, hybridreceptors can beformedbetween the insu-lin receptor and the IGF-I receptor (44, 45). Moreover, it has beenreported thatin both human placenta(16)andNIH-3T3 cellstransfectedwith human IR cDNA andexpressinghuman IR

(17),

IR/IGF-I receptorhybridscan occur in vivo. The possi-bility was considered, therefore, that the receptor studied herein was a hybrid insulin/IGF-I receptor, and thatinsulin

wasbindingtothehybridreceptormoeity.For several reasons,

this

possibility

wasconsidered unlikely. First,the monoclonal

antibodytotheIRemployed herein, which interacts with insu-lin/IGF-I receptorhybrids (44), neitherblockedinsulin bind-ingnorimmunoprecipitated insulin photocross-linkedtothis

receptor. Second,anantipeptide antibody directed against

resi-dues241-251 (22) oftheIRfailedto

immunoprecipitate

thea

subunit ofthe receptor

photoaffinity-labeled

with insulin pho-toprobe. Third,it has beenreported that,when

hybrid

insulin/

IGF-I receptors are autophosphorylated, there is labeling of two

,B-subunit species (45):

asmallerone

(95

kD)

representing

the IR, andalargerone(102 kD)

representing

the IGF-I

recep-tor. In our

atypical

receptoronlyone 103 kDsubunit

species

wasseen.Fourth, this atypical insulin bindingwasnotseen in

184, ZR-75-1,and T47-Dcells,allof whichalso express both

insulinand IGF-I receptors.

For several reasons, themost likely

possibility

is that the

atypicalinsulinbindingseeninMCF-7cellswas to areceptor that is relatedtothe IGF-I receptor.

First,

the

cross-linking

of

the

B29-MAB-1251-insulin

revealedareceptorasubunitof 120 kD,asize which is

typical

for the humanplacentaIGF-I

recep-tor asubunit,butis smaller than that of theIRasubunit

(C.

Yip, unpublished data). Moreover, in

autophosphorylation

studies the

_(l

subunitwas103kD,asize which is

larger

than 95

kD IR,Bsubunit. Second,theatypicalreceptor interacted with thea-IR3,anantibodywhich is directed toward the IGF-I

re-ceptor. Therefore, in addition to the

typical

IGF-I receptor which binds IGF-Ibut not

insulin,

another IGF-I receptor

ex-ists which binds both IGF-I and insulin.Inother

tissues,

two

groups havereported atypical receptors that havehigh-affinity

binding for both insulin and IGF-I. In fetal muscle,

Alexan-dridesand Smith(46) reportedthat both insulin and IGF-Iat

similar concentrations stimulated thephosphorylationof 105-kD receptor (3 subunit whichwas moreclosely relatedto the IGF-I receptor (3 subunitthan the IR. Also, in bovine retina

preparations, Waldbillingand Chadder (47) reportedthat la-beledinsulinbindingwasinhibited byunlabeledIGF-Iwitha

potencyequalto orgreater than that of insulin. These results support theconcept that two IGF-I receptors may exist: one

with highaffinity only for IGF-I, and another withhigh affinity for both IGF-I and insulin. Thesetworeceptors could be pro-duced by thesamegene via eitherprocessingof IGF-I receptor mRNAor byposttranslational modificationsof the IGF-I

re-ceptorprotein. Althoughthehuman genome appearsto con-tainonly one IR andoneIGF-Ireceptor gene,it ispossiblethat theatypicalreceptorexpressed inMCF-7cellscouldbea

prod-uctof a distinct gene that encodes a receptorproteinthat has

high-affinitybinding forbothIGF-I andinsulin. Further stud-ies,therefore,including purification and cloning will be

neces-sary tounderstand thebiochemicalnatureof this

atypical

IGF-Ireceptor inMCF-7 cells and othertissues.

The atypical IGF-I receptor hadintrinsic tyrosinekinase and regulated biological functions. Both receptor

autophos-phorylation and phosphorylation ofan exogenous substrate werestimulatedby bothinsulin andIGF-I.Moreover, in

MCF-7cells, bothinsulinand IGF-Istimulated[3H]thymidine

incor-poration. a-IR3 blocked all theeffects ofIGF-I, but only 50-60% oftheeffects of insulin. These studiessuggesttherefore

thatinsulinexerts aportion ofitseffectthroughitsown recep-tor, and anotherportionthrough theatypicalreceptor. Interest-ingly, Cullen et al.(10)have reported thatca-IR3does not block theeffects of insulinoncellgrowthin MCF-7 cells.

Insummary, in MCF-7 cells, there are threefunctionally

distinct receptors: a typical insulin receptor, a typical IGF-I receptor, and anatypical IGF-I-like receptor which binds both

insulinandIGF-Iwithhigh affinity.Thepresenceofthese re-ceptors in MCF-7 cells mayexplainin part theconflicting re-sults concerning the growth effects ofinsulin onthese cells. Furtherstudies,however, will be necessary to understand the roleofthese receptors in the growthof MCF-7 cells.

Acknowledaments

WethankElaineJackfor technical assistance.

This

_study

wassupported

by

agrantfrom MedicalResearch Coun-cil, Canada;the_JayGershow CancerFund;the JohnA.Kerner Foun-dation;and theAssociazione ItalianaperlaRicerca sul Cancro_(Dr. Milazzo).

References

1._Hanks,S.K.,A.K.Quinn,and T.Hunter.1988.Theproteinkinasefamily:

conserved featuresand deducedphylogenyofthecatalyticdomains. Science

(Wash. DC).241:42-53.

2.Ullrich, A.,J. R.Bell,E. Y.Chen,R.Herrera,L.M.Petruzzelli,T.J.Dull,

A._Gray,L._Coussens,Y.C.Liao,M.Tsubokawa,etal. 1985. Humaninsulin receptoranditsrelationshiptothe_tyrosinekinasefamilyof oncogenes. Nature

(Lond.).313:756-761.

3.Goldfine, I. D. 1987. The insulinreceptormolecularbiologyand

trans-membranesignaling.Endocr.Rev.8:235-255.

4.Strauss,D. S. 1984.Growth-stimulatoryactionsofinsulin invitroand in vivo.Endocr.Rev.5:356-369.

5.Giorgino,F.,A.Belfiore,G.Milazzo,A.Costantino,B.Maddux,J.

Whit-taker,I.D._Goldfine,andR.Vigneri.1991.Overexpressionof insulinreceptorsin

fibroblast and ovary cells inducesaligand-mediatedtransformedphenotype. Mol. Endorinol.5:452-459.

6.Papa, V.,V.Pezzino,A.Costantino,A.Belfiore,D.Giuffrida,L.Frittitta, G.B.Vannelli,R.Brand,I. D.Goldfine,and R.Vigneri. 1990. Elevated insulin receptorcontentin human breastcancer.J.Clin. Invest. 85:1503-15 10.

7.Lippman,M.E.,R. B.Dickson,E. P.Gelmann,N.Rosen,C.Knabbe,S.

Bates,D.Bronzert,K.Huff,and A. Kasid. 1987. Growthregulationof human breastcarcinomasoccursthrough regulated growthfactor secretion. J. Cell.

Bio-chem.35:1-16.

8.Osborne,C.K.,J.Bolan,M. E.Monaco,and M. E.Lippman.1976.

Hor-moneresponsivehuman breastcancerinlong-termtissue culture: effect of insu-lin. Proc.Natl. Acad. Sci.USA.73:4536-4540.

9.Osborne,C.K,M. E.Monaco,M. E.Lippman,and C. R. Kahn. 1978.

Correlationamong insulinbinding, degradation,andbiological activityin

hu-manbreastcancercellsinlongtermtissue culture. Cancer Res. 38:94-101. 10.Cullen,K.J.,D.Yee,W.S.Sey,J.Perdue,B.Hampton,M. E.Lippman,

and N. Rosen. 1990.Insulin-likegrowthfactorreceptorexpressionand function in human breastcancer.Cancer Res. 50:48-53.

11.Kull,F.C.,S.Jacobs,Y. F.Su,M. E.Svoboda,J. J. VanWyk,and P.

Cuatrecasas.1983. Monoclonal antibodiestoreceptorsfor insulin and somato-medin-C.J.Biol. Chem. 258:6561-6566.

12.Furlanetto,R.W.,and J.N.DiCarlo.1984. Somatomedin-Creceptors

andgrowth effectsin human breastcancercells maintained inlong-termtissue culture.CancerRes.44:2122-2128.

13.Stewart,A.J.,M. D.Johnson,F. E. B.May,andB. R.Westley.1990. Role

of insulin-likegrowthfactorsand type I insulin-likegrowthfactorreceptorin

estrogen-stimulatedproliferationofhuman breastcancercells.J. Biol. Chem.

265:21172-21178.

14.Duronio,V.,andS. Jacobs. 1988.ComparisonofinsulinandIGF-I

recep-tors.InReceptorBiochemistryandMethodology.C.J. Venter and L. C.

Harri-son,editors.Alan R.Liss, Inc.,NewYork. 12B:3-18.

15.Jonas,A.H.1988.Heterogeneityofreceptorsfor insulin andinsulin-like

growth factorI:evidence forreceptorsubtype.InReceptorBiochemistryand

Methodology.C.J.Venter and L.C.Harrison,editors.Alan R.Liss,Inc.,New York.12B:19-36.

16.Soos,M.A.,and K.Siddle.1989.Immunologicalrelationshipsbetween receptorsfor insulinandinsulin-likegrowthfactorI.Biochem.J.263:553-563.

17.Soos,M.A.,J.Whittaker,R.Lammers,A.Ullrich,and K.Siddle.1990.

Receptors forinsulin andinsulin-likegrowthfactor-Icanformhybriddimers. Biochem.J.270:383-390.

18. Jacobs,S. 1990.InsulinlikegrowthfactorI receptor. InHandbook of Experimental Pharmacology.P. Cuatrecasas and S.Jacobs,editors. Springer-Ver-lag, Inc.,NewYork.92:267-286.

19.Yip,C. C.,C. W. T.Yeung, andM. L. Moule. 1980. Photoaffinity labelingof insulinreceptorproteinsof liverplasmamembranepreparations. Biochemistry. 19:70-76.

20.Stampfer,M.R.,and J. C.Bartley.1985. Induction oftransformationand

continuouscelllines fromnormal human mammaryepithelialcellsafter

expo-sure tobenzo[a]pyrene.Proc.Natl. Acad.Sci. USA. 82:2394-2398.

21. Forsayeth,J.R.,A.Montemurro,B. A.Maddux,R.DePirro,and I. D.

Goldfine.1987.Effect of monoclonal antibodiesonhumaninsulinreceptor

auto-phosphorylation, negativecooperativity,anddown-regulation.J.Biol. Chem. 262:4134-4140.

22.Yip,C. C.,C.Grunfeld,and I. D.Goldfine. 1991. Identificationand

characterization of theligand-bindingdomain of insulinreceptorbyuseofan

antipeptideantiserumagainstamino acidsequence241-251 ofthea-subunit. Biochemistry.30:695-701.

23.Hartmann,K. K.P.,V.Papa,E.J.Brown,U.Doerries,S.M.Rosenthal,

and I. D.Goldfine. 1990. A rapid, simpleone-step method for isolation of poly(A)+RNAfromcells inmonolayer.Endocrinology. 127:2038-2040.

24.Rosenthal,S.M.,A.Brunetti,E.J.Brown,P. W.Mamula,and I. D.

Goldfine. 1991.Regulationof insulin-like growth factorI receptorexpression duringmuscle celldifferentiation.J.Clin.Invest.87:1212-1219.

25.Sambrook, J.,E. F.Fritsch,and T.Maniatis.1989. MolecularCloning:A

LaboratoryManual.ColdSpringHarborLaboratory, Cold Spring Harbor,NY.

26.Roth,R.A.,D.J.Cassel,B. A.Maddux,and I. D.Goldfine.1983. Regula-tion ofinsulinreceptorkinaseactivitybyinsulinmimickersandaninsulin antago-nist.Biochem.Biophys.Res.Commun. 115:245-252.

27.Tollefsen,S.E.,K.Thompson,and D.J. Petersen. 1987.Separationofthe

highaffinityinsulin-likegrowthfactorI receptor from lowaffinity bindingsites

byaffinity chromatography.J. Biol.Chem. 262:16461-16469.

28.Forsayeth, J.,B.A.Maddux,andI.D.Goldfine. 1986.Biosynthesisand

processingof the human insulin receptor.Diabetes.35:837-846.

29.Lowry, 0. H.,N.J.Rosebrough,A.L.Farr,andR.J.Randall. 1951.

Proteinmeasurementwith the Folinphenolreagent.J.Biol.Chem. 260:4128-4135.

30. Pezzino, V., V. Papa,V.Trischitta, A. Brunetti, P. A. Goodman, M. K.

Humaninsulin receptor radioimmunoassay: applicability to insulin-resistant

states.Am.J.Physiol.257:E451-457.

31. Froesch, E. R.,C. Schmid,J. Schwander, and J. Zapf. 1985. Actions of

insulinlike growth factors.Annu.Rev. Physiol. 47:443-467.

32.Yip,C. C., H. Hsu, R. G. Patel, D. M. Hawley, B. A. Maddux, and I. D.

Goldfine. 1988. Localization ofthe insulin-bindingsite to thecysteine-richregion

oftheinsulin receptor a-subunit. Biochem. Biophys. Res. Commun.

157:321-329.

33. Hawley,D.M., B. A. Maddux, R. G. Patel, K. Y. Wong, P. W. Mamula, G. L.Firestone,A.Brunetti, E.Verspohl, andI. D. Goldfine. 1989. Insulin receptormonoclonal antibodiesthat mimic insulin action without activating

tyrosine kinase. J. Biol. Chem. 264:2438-2444.

34.McFarland, K. C., J. Ramachandran, and C. Grunfeld. 1989.

Peptide-basedradioimmunoassay for insulinreceptor detectionof insulin-stimulated downregulation in IM-9 lymphocytes. Diabetes. 38:978-990.

35. Kasuga,M., E. E. VanObberghen-Schilling,S. P.Nissley,and M. M. Rechler. 1988.Demonstration oftwo subtypesof insulin-likegrowthfactor re-ceptorsby affinity cross-linking. J. Biol. Chem. 256:5305-5308.

36. Kasuga, M.,F. A.Karlsson,andC.R.Kahn.1982.Insulin stimulatesthe

phosphorylation ofthe95,000-daltonsubunitofitsown receptor. Science (Wash.

DC).215:185-186.

37.Jacobs,S.,F.C.Kull,Jr., H.S.Earp, M. E.Svoboda,J. J.VanWhy, and P.

Cuatrecasas. 1983. Somatomedin-C stimulatesthephosphorylation ofthe (i-sub-unitof itsownreceptor.J.Biol.Chem. 258:9581-9584.

38.Rechler,M.M.,J.Zapf, S.P.Nissley,E. R.Froesch,A.C. Moses, J. M.

Podskalny,E. E.Schilling, andR. E.Humbel.1980. Interactions of insulin-like

growthfactorIandIIandmultiplication-stimulating activity withreceptors and serumcarrierproteins. Endocrinology. 107:1451-1459.

39. Kasuga, M., N.Sasaki,C.R. Kahn,S.P.Nissley,and M. M. Rechler.

1983. Antireceptor antibodiesasprobes of insulin-like growth factorstructure. J.

Clin.Invest.72:1459-1469.

40.Ebina,Y., L.Ellis,K.Jarnagin,M.Edery,L.Graf,E.Clauser,J.Ou,F.

Marsiaz,Y. W.Kan,I.D.Goldfine,et al. 1985.Thehumaninsulinreceptor cDNA: the structural basis forhormone-activated transmembrane signaling. Cell. 40:747-758.

41.Ullrich,A., A. Gray, A. W. Tam, T. Yang-Feng, M. Tsubokawa, C.

Col-lins,W.Henzel,T.LeBon, S.Kathuria,E.Chen,et al.1986. Insulin-like growth factorI receptorprimarystructure: comparison with insulinreceptor suggest

structural determinants that define functionalspecificity.EMBO(Eur. Mol.

Bio-chem.Organ)J.5:2503-2512.

42. Rechler, M. M., andS.P.Nissley. 1985.The natureandregulation ofthe

receptors for insulin-like growth factors. Annu. Rev. Physiol. 47:425-442.

43.Pezzino, V., G. Milazzo,L.Frittitta,R.Vigneri, 0. Ezaki,M.Kasahara,

R. LeBon,I.D.Goldfine,and Y. F.Yamaguchi. 1991. Radioimmunoassay for

humaninsulin-like growth factor-I receptor applicabilitytobreastcarcinoma specimensandcell lines. Metab. Clin. Exp. 40:861-865.

44.Treadway, J. L., B. D.Morrison, I.D.Goldfine,and J. E.Pessin.1989.

Assembly ofinsulin/insulin-likegrowth factorIhybridreceptorsin vitro.J.Biol. Chem. 264:21450-21453.

45.Moxham, C. P., V.Duronio, and S.Jacobs. 1989. Insulin-likegrowth factorI receptor,B-subunit heterogeneity.J.Biol. Chem. 264:13238-13244.

46.Alexandrides, T. K., andR.J.Smith. 1989.Anovelfetal insulin-like growth factor (IGF)I receptor. J.Biol.Chem. 264:12922-12930.