Integrin adhesion molecules in the human endometrium Correlation with the normal and abnormal menstrual cycle

Integrin adhesion molecules in the human

endometrium. Correlation with the normal and

abnormal menstrual cycle.

B A Lessey, … , S M Albelda, C A Buck

J Clin Invest. 1992;90(1):188-195. https://doi.org/10.1172/JCI115835.

Integrins are a class of cell adhesion molecules that participate in cell and cell-substratum interactions and are present on essentially all human cells. The distribution of nine different alpha and beta integrin subunits in human endometrial tissue at different stages of the menstrual cycle was determined using immunoperoxidase staining. Glandular epithelial cells expressed primarily alpha 2, alpha 3, and alpha 6 (collagen/laminin

receptors), while stromal cells expressed predominantly alpha 5 (fibronectin receptor). The presence of alpha 1 on glandular epithelial cells was cycle specific, found only during the secretory phase. Expression of both subunits of the vitronectin receptor, alpha v beta 3, also underwent cycle specific changes on endometrial epithelial cells. Immunostaining for alpha v increased throughout the menstrual cycle, while the beta 3 subunit appeared abruptly on cycle day 20 on luminal as well as glandular epithelial cells. Discordant luteal phase biopsies (greater than or equal to 3 d "out of phase") from infertility patients exhibited

delayed epithelial beta 3 immunostaining. These results demonstrate similarities, as well as specific differences, between endometrium and other epithelial tissues. Certain integrin moieties appear to be regulated within the cycling endometrium and disruption of integrin expression may be associated with decreased uterine receptivity and infertility.

Research Article

Find the latest version:

http://jci.me/115835/pdf

Integrin Adhesion Molecules in

the Human

Endometrium

Correlation with theNormal and Abnormal MenstrualCycle

Bruce A. Lessey,*Laszlo Damjanovich,' ChristosCoutifaris,*ArthurCastelbaum,*Steven M. Albelda,*andClaytonA.Buck$

*Department ofObstetrics and Gynecology, and DepartmentofMedicine, UniversityofPennsylvania,

and §Wistar Institute ofAnatomy and Biology, Philadelphia, Pennsylvania 19104

Abstract

Integrinsare aclassofcelladhesionmolecules that participate

incell-celland cell-substratum interactions andarepresenton

essentially allhumancells. Thedistribution of nine differenta

and,f integrin subunitsinhuman endometrial tissueatdifferent

stages of the menstrualcyclewas determined using immuno-peroxidase staining. Glandular epithelial cells expressed pri-marilya2,a3,anda6

(collagen/laminin

receptors), while

stro-mal cells expressed predominantly a5 (fibronectin receptor).

Thepresenceof_a,onglandularepithelial cellswascycle

spe-cific, found only during thesecretoryphase.Expression of both subunits of the vitronectinreceptor,

a,83,

also underwentcycle specific changesonendometrialepithelial cells. Immunostain-ing fora,increasedthroughout the menstrual cycle, while the

0

subunitappeared abruptlyoncycle day20onluminalaswell asglandular epithelial cells. Discordant luteal phase biopsies (23 d "outof phase") from infertility patientsexhibited

de-layedepithelial 3immunostaining.Theseresults demonstrate

similarities, aswell as specific differences, between endome-trium andotherepithelialtissues.Certainintegrinmoieties ap-pear to be regulated within the cycling endometrium and disruption of integrin expression maybe associated with

de-creased uterine receptivity and infertility. (J. Clin. Invest.

1992.90:188-195.) Keywords: collagen* laminin*fibronectin. vitronectin * infertility

Introduction

Over the pastdecade, investigatorshavecometorecognizethe

importanceofthe extracellularmatrix(ECM)1_{indirectingthe} growth, differentiation and function of the overlying epithe-lium (1). The interaction betweencell and substratum is

me-diatedbyseveralclasses of cell adhesionmolecules,oneofthe

mostimportant beingtheintegrins (2-5).Thisdiversefamily

ofglycoproteinreceptorsisexpressedonthecellmembraneas heterodimericaand i8subunitsinvolvedinboth cell-cell and

cell-substratum adhesion. Specific recognitionandbindingof

extracellular matrix (ECM) components such as fibronectin

(FN),laminin(LM),andcollagen (Col)transmitinformation tothecytoskeletalstructure(6),aninteraction that mayhave

Addressreprintrequest toDr. BruceA.Lessey, Universityof

Pennsyl-vania,422CurieBlvd., Philadelphia,PA 19104.

Receivedfor publication 29 August 1991 andin revisedform 17

January1992.

1.Abbreviationsused in thispaper:Col, collagen; ECM, extracellular

matrix; FN, fibronectin; LM, laminin; LPD,lutealphase dysfunction; OOP,outofphase.

major roles in promoting hormone responsiveness and geno-mic activation ( 1).

Although extensive information exists about specific inte-grin proteins (7), little is known concerning the distribution of these receptors in the female reproductive tract. Endometrium, composed of glandular epithelium and associated mesen-chyme (stroma), maintains complex temporal and spatial functions in response to the cyclic hormonal milieu. The phe-nomena of cyclic endometrial replenishment, implantation, gestation, and parturition may require specific cell-cell and cell-substratuminteractions.

As a first step in understanding the function of integrins in the human endometrium, we have examined their distribution throughout normal and abnormal menstrual cycles. Patients with inadequate corpus luteum function (luteal phase dysfunc-tion; LPD) have delayed endometrial maturation and an

asso-ciatedinfertility(8, 9). Since appropriately primed endome-triumis thought to be necessary for normal cycle fecundity and uterinereceptivity,weexamined integrindistribution in

endo-metrial biopsies from infertile women with significant endo-metrial dyssynchrony, and compared the results with the pat-tern observed in concordantsamples. Fromthese data it

ap-pears that some integrins normally undergo spatial and temporal changes in expression within the cycling

endome-trium,andthatadisruptionofthispattern may be associated withcertain types of infertility in women.

Methods

Humansamples. Endometriumwasobtainedfrom 35 reproductive

agewomen atthetime of hysterectomyattheHospitalofthe Univer-sity ofPennsylvania.Tissuewasobtained fromtheearlyproliferative

(day5) throughlatesecretoryphase (day28),andallhysterectomies

wereperformed for benign disease. Endometrial biopsies were

per-formedon womenaspartof their evaluationforinfertilityatthe Uni-versity of Pennsylvania.Useofendometrialtissuehad beenapproved bytheinstitutional reviewboard oftheUniversityofPennsylvania.All

patientswerecycling normallyandnonehadreceivedhormonesforat

least 3mo beforesurgery. Datingof the endometriumwasassessed

accordingtothecriteria ofNoyesetal.(10).Endometrialbiopsieswere

evaluated in thecontextoftimingofovulation and/ortheonsetofthe next menstrualperiod.Sampleswerejudgedas"outofphase"if histo-logicdatingwasdelayedby 3or moredaysrelativetothepredicted day

of the menstrual cycle. Proliferative endometrium was categorized

basedonhistologyandonlastmenstrualperiod. Sampleswere

trans-portedonicetothelaboratorywheretheyweresnap frozenondryice andstoredat-70'C.

Antibodies. Monoclonalantibodies (mAbs) PlH5, PlB5, PlD6

specifictoa2, a3, and a5 subunits were generously provided by Drs. Elizabeth Wayner(University of Minnesota, Minneapolis, MN) and William Carter (Fred Hutchinson Cancer Research Center, University ofWashington, Seatle, WA).Dr.Martin Hemler (Dana-Farber Cancer Institute, Boston, MA)kindly suppliedmAbsTS2/7 andB-5H10 di-rectedagainstthe_aland a4subunits, respectively. GoH3,aspecific

mAbdirectedagainsta6 wasdonatedby Dr. ArnoudSonnenberg

(Cen-tral Laboratory of The Netherlands Red CrossTransfusion Service,

J.Clin. Invest.

© TheAmerican SocietyforClinicalInvestigation,Inc.

0021-9738/92/07/0188/08 $2.00

Amsterdam).Drs. Joel Bennett and James Hoxie(Universityof Penn-sylvania) provided mAb SSA6 specificto the 33 subunit, and mAb LM142againsta,wasprovidedbyDr. David Cheresh(Scripps Clinic,

La Jolla,CA). The 4antibodywasgenerously supplied byDr. Steven Kennel (OakRidgeNationalLaboratory,OakRidge, TN).These anti-bodies havepreviouslybeen characterized(1 1-16).

Immunohistochemistry. Immunoperoxidase staining was

per-formedon cryostatsections ofendometrium throughout the menstrual cycle. Serial cryosections 4-8 ALthick were placed onto poly-L lysine

coatedslides,fixed in -20'C acetone for 10 min,and stained using Vectastain Elite® ABC kits (VectorLaboratories, Burlingame, CA).

Diaminobenzidine(DAB; Sigma Chem.Co., St. Louis, MO) was used

asthe chromagen.Primary antibody was placed on cryosections after

blocking with 1% BSA inPBS,andallowedtobind at room tempera-turefor 1 h. A PBS pH 7.2-7.4rinsewasfollowedby secondary anti-bodyconsisting of biotinylatedgoatanti-mouseantibody for30 min. After a PBS rinse, the endogenousperoxidaseswere quenchedwith a 30-minincubation with 0.3% H202in absoluteethanol, followed bya 30-min rehydration inPBS. Avidin:biotinylated horseradish

peroxi-dase macromolecularcomplex (ABC)wasthenincubatedonthe sec-tionsfor30 minbeforeadding diaminobenzadine for3 min to com-plete the reaction. Somesamples were treatedwith 1:200dilutionof fluorescein-labeled anti-mouse antibodies for 1 hfor

immunofluores-cent microscopy. Samples were subsequently washed in PBS and

mounted. Theresulting stainingwasevaluated on aNikonmicroscope

atlow (IOOX) andhigher(400X)magnificationwith orwithout

fluores-cence.Stainingwasjudgedasabsent(-),weak(±),moderate(+),or strong(++).Examples ofeach are presented inFig. 3(see Results).

Photomicrographswere madeusingKodak T-MAX100 ASAfilm.

Cell harvest andNP-40 extraction. Four samplesof endometrium

wereobtainedforfurtherevaluationofthe 33subunit inproliferative (n=2) and latesecretory(n=2)endometrial epithelium.Eachsample was placed in DMEM (Sigma ChemicalCo.), supplemented with 10% fetal bovine serum (Flow Laboratories, Inc., McLean, VA) glucose

(4,500mg/liter),Hepesbuffer (25 mM), L-glutamine (584 mg/liter),

andsodium bicarbonate (3.7g/liter). Using modifications ofthe pro-ceduresdescribed byGurpideandcolleagues ( 17), endometrium was minced in aplastic petri dish before incubation with 6 mg of collage-nase(type lA,550 U/mg; Sigma Chemical Co.) for 2 h at 37°C. The

resulting suspensionwassuccessivelypassed through a250-yrmand a

38-gmseive (Newark WireClothCo.,Newark, NJ). The coarsesieve

removedundigested material,while the secondretainedthe glandular elements andexcluded theindividualstromal and blood cells. After

thoroughrinsing,theglandularelements wereobtainedby

backwash-ing with10 to 20 mlofDMEM. Theisolatedglandular structures were thentransferredto a 1.5-mlmicrofugetubeandcentrifugedthreetimes

(82 g)for 2 minwithintermittent washes with PBS. Membrane

ex-tracts wereprepared byaddingsmall volumes( 100-200,ll)of10 mM Tris-acetate, pH 8.0, 0.5% NP-40, 0.5 mMCa2" (TNC)with 2 mM PMSF to thefinal pellet, pipetted and incubated on ice for 15 min. The lysate wascentrifuged for 5 min at 16,000 g in a microcentrifuge. The

resulting supernatant was called NP-40 extract and was frozen at

-70'Cuntil use. Aportionof the original, undigested tissue was cryo-sectionedfor immunohistochemical localizationof

_03.

TableI.DistributionofIntegrinsby Ligand Specificity

Ligand specificity

Collagen/laminin Fibronectin/vitronectin

a,/03 NA

Integrinsubunit _a2/01 aJ/,B

a3/01 a'101

CY6101 av/03

ad(4

Gelelectrophoresisand immunoblots. Theproteinconcentrationof

eachNP-40extractandanextractofplatelets(positive control)was

determinedusingthetechniquedescribedby Lowryetal.(18).

Sam-pleswithequalamountsofproteinwereaddedtoelectrophoresis

sam-ple buffer(62.5mMTris base,2%SDS, 10%glycerol,pH6.8).

Sam-pleswereanalyzed bySDS-PAGEusing6%polyacrylamidegels,using nonreducingconditions ( 19).Thegelwastransferredtonitrocellulose

usingatransferapparatus(Bio-radLaboratories, Richmond, CA)and

blockedwith 4%BSAin PBSwith0.2% Naazide for1 h. Afteraddition ofprimary antibody(SSA6 supernatant)for 2h,thegelswere

devel-opedusing an alkalinephosphatase-conjugated secondary antibody

(PromegaCorp., Madison, WI)accordingtomethodspreviously

de-scribed(20).

Results

Integrin distributioninnormalendometrium. The distribution

ofa2, a3, a6, and

₍₄

subunits of

_integrins

that

_recognize

pri-marily collagen (Col) and/orlaminin(LM;TableI)isshown

inFig. 1,A-D.These subunitswerepresentonglandular

epithe-lium(arrows) throughoutthe menstrualcycle.Their

distribu-tion within the endometriumwastypicalof thatseenfor most

epithelial tissues. The a2 and a3 subunits were distributed around the entire circumference of thecells,while thea6and

₍₄

subunits

appeared

to be localizedat the basolateral

surface,

adjacenttothe basementmembrane(BM)of the endometrial

glands. Theexpressionof thesesubunitsbythemesenchyme

(asterisks)

wasless

pronounced.

While moderate

staining

was

seenfora6onstromal cells

(Fig.

1 C),verylittle

staining

was

noted for

#4.

Theexpressionof thea4anda5subunits ofintegrinsknown tobind fibronectin(Table I)was

quite

restricted. Thea4 sub-unit was undetectable above

background

staining

(compare

Fig. 1 EwithFig. 2 C)ineither

epithelium

or

mesenchyme.

Thea5subunit

(Fig.

1F),

representative

ofthe classic

fibronec-tinreceptor,was not seen onthe

epithelial

components,but

was strongly expressed in the mesenchyme which is rich in fibronectin.

The

intensity

of

immunostaining

for three other subunits of

integrins

wasfoundtochange in acycle-dependentmanner.

Immunostaining

for

_aI

in the

proliferative

phase

(Fig.

2

A)

was

only

slightly

above

background

levels

(Fig.

2C).The

intensity

of

staining

increasedthroughoutthesecretoryphase

(Fig.

2

B).

This intensecircumferential

staining

wasfoundonglandular

andluminal

epithelium

onallsamplesfrom

day

15to28.

Like-wise, avwasweakly expressedonboth theepitheliumand

mes-enchyme

in the

proliferative

phase

(Fig.

3A)and

staining

in-creasedgradually

during

thesecretory

phase

tothe level noted in

Fig.

3 B.Duringthe

proliferative

phase 3

staining

wasonly

present on the mesenchymal cells

(Fig.

3 C). Increased

(3

staining

wasapparentontheendometrial

epithelium only

after

day 19of the menstrualcycle(Fig. 3D)ontheluminalaswell

asglandular epithelium,andwasalso present ina

pericellular

distribution. In contrast, thebasalislayerdidnot

_{significantly}

stain for either avor

(3.

This

changing

patternof

epithelial

av and

(3

throughout the cycle was studied in 35 endometrial

samplesandis

depicted graphically

in

Fig.

4,Aand B.

Integrins indiscordantendometrium. Thepresenceof the

epithelial

(3

subunit appeared to be a consistent internal

markeroflutealphasematuration,and the

timing

of 3

expres-sion correlatedwith the

putative

periimplantation

period.

To

Figure 1. The immunohistochemical profile of six integrin subunits in normal endometrium. Immunohistochemical staining of thecollagen/

laminin receptor subunits:a2 (A),a3(B),a6(C), andfh(D) showsprominentstaining ofepithelium (arrows) and microvessels (arrowheads)

without significant stromal staining (asterisks) fora2,a3,and 4.Note basolateral staining a6 and basal staining for . Staining for fibronectin

receptor subunits a4(E),a5 (F) show predominant mesenchyme staining (asterisks) with decreased epithelial staining (arrows). The immuno-reactions (areas ofdark staining) were developed by the avidin-biotin-peroxidase complex using diaminobenzidine as a chromagen. For greater sensitivity, no counterstain wasapplied. x 125.

clinical evaluation of endometrial biopsies, immunostaining

wasperformed on luteal phaseendometrial samples from

cy-cles that showed evidence of maturational delay. Endometrial biopsies from25 women who hadconcordance of menstrual

andhistologic dating ("normal" group)werecomparedto 12

biopsies that were identified as 23 d out ofphase (OOP),

based on either the time of ovulatory or the subsequent

menses. Samples wereimmunostainedfor

a,,

ac, and

_f

sub-units.Allbiopsieswereperformedondays 20-24ofthe

men-strual cycle. Inall instances, immunostaining forthese three

antigenswaspresent onendometrial epithelia fromthe normal

Figure2.Photomicrographsoftlheimmunohistochemicalstainingfor theintegrinsubunit_a,inproliferativevs.

_secretory

endometrium. Thestainingin theglandular epithelium(arrows)waslargelyabsent in theproliferative phase(A),butquitepronouncedin allsections after menstrualcycle day14(B;day20endometrium).The

mesen-chyme(asterisks)didnotstain for_a,.Themicrovasculature

(arrow-heads) stainingwasalsopronounced,butdidnotchange throughout

the menstrualcycle. Thestainingnoted in

secretory

endometrial

glandswassignificantly higherthan thatofbackground

(C).

x125.

staining

waspresent,but

epithelial

133

staining

wasabsent.The

comparisonOf

₁₃₃

staining

intensity

in thetwogroupsisshown

in

Fig.

5 A.

Accompanying

photomicrographs

Of

₀₃

immuno-staining

fromoutofphase biopsies

(QOP;

B)andnormal"in

phase"biopsies(C)areincluded, whichdemonstrates the

dis-crepancy seen in 13 staining.Insubsequent treatment cycles,

twoOOP patientsunderwent repeatbiopsy duringa

normal-ized cycleatwhich time immunostaining for epithelial 3 was

present. Thissuggeststhat the lackof

_#3

was not anintrinsic

defectin the OOPgroup.Rather, the discordant biopsies that lacked

#3

hadnot yetestablished the midluteal phenotype of normal day 20-24 endometrium.

To further demonstratethatimmunohistochemical stain-ing accurately reflected changes intheexpression of 13 subunit

on endometrial epithelium, immunoblots (Western blots)

wereperformedon samplesofenriched endometrial glandular

elements fromproliferativeand secretoryphase.Asshown in Fig. 6A,proliferative phase epithelialstructureshadlittleto no

immunostainingat95 K (lanes 2 and 3), compared with the

positivecontrol(plateletextract;laneI)or tosamples from the secretory phase (lanes 4 and 5) which showedstrongstaining

for13. The isolated endometrial glands appearedas tubular

structuresfree of surroundingstroma(Fig. 6 B).

Immunofluo-rescentstaining for13from samplescorrespondingtolanes 3

and 4 (midproliferative phase and day 23, respectively) are

shown inFig.6, C andD.Notethe absence of glandular stain-ing in the proliferative sample, while both glandular and lu-minal immunostaining is obvious from the secretory phase. These data confirm that the expressionof epithelial13in

hu-manendometrium isacycle-specific phenomenon.

Discussion

Theendometrium undergoeswell-delineatedchanges in its

his-tology throughoutthe menstrualcycle. We examined the

ex-pression of integrins withinthecontextofthesecyclicchanges andfoundan interestingpatternof distributionfor these cell adhesion molecules. A summaryof theexpression of integrin

subunits throughout the menstrualcycleis shown in Table II.

Collagen/lamininreceptors(TableI)characterized bya2, a3,

and a6 were uniformly expressed throughout the menstrual

cycle, witha distribution similartothat seen for mostother

epithelia(21-25). Thepericellular distribution ofa2 anda3

subunitswasdistinctly different fromthatofa6subunit.

Char-acteristic ofa LMreceptor, a6 wasconcentratedonthe basolat-eralsurface. The

₁₄

subunit that pairswith 6(16) was also founddistributedon the basolateral surface ofthe epithelial cells,andits distribution appearedeven morerestrictedtothe basal pole.Given thepresenceofstromala6stainingwithout

14,

it is likely that 6pairswith

1l3

in the mesenchyme. While theligandfor

a611

has been wellcharacterized asthe E8

frag-mentof laminin(26), thefunctioning of

a614

is lessdefined (26, 27). It appears likely, however, that

a614

also binds to

laminin (28). The significance of this segregation of these

laminin receptors between cell types is unknown. The

a5013

integrin,amajor fibronectin receptor, was alsouniformly

ex-pressed throughout the menstrual cycle. Unlike theCol and

LM receptors, thedistribution of

a51l

waslimitedtothe

mesen-chyme.

Forthe most part, thedistribution of these subunits

corre-spondstothatwhich wouldbepredictedfrom their functionas

both cell-cell and cell-matrix receptors(29), and from the

reported distribution of ECMcomponents. It has been postu-lated that thesedifferencesindistributiondenotedefinedroles

fortheseintegrins (30-32).Theintegrinscomposedofa2, a3,

'4v

Orr'.

Figure3. Immunostain-ing ofa, and 3in

prolif-erativephase vs.secretory

phaseendometrium. The

staining intensity of a, in theproliferativephase (A) wasjudged as"+" for the stromal cells

(as-terisks)and "±"for glan-dulara, (arrow).

Immu-nostainingfora,,in day

22endometrium(B)

demonstrates asignificant increasein glandular

staining (exampleof "++"staining intensity). Likewise,thestainingfor

/3wasabsent in

prolifera-tiveepithelium (C;

arrows) and wasnotably

increased inthisday 22

secretory endometrium

(D).

X125.

anda6 areknowntobindtoColand LM, componentsfound in

BM, and thus,theirexpressionlargely onendometrial epithe-lial cellsisnot unexpected. Twoofthesesubunits, a2 anda3,

are known topromote epithelial cell-cell adhesion (29, 30),

and this function is consistent with their pericellular rather

than basaldistribution. Theintense-staining fora5in the

mes-enchyme islikewise paralleledby arich distribution

offibronec-tin inthis compartmentof theendometrium(33).

Incontrast tothereceptors that wereuniformlyexpressed

atall stagesofendometrial development, the distribution of

several subunits forintegrins was found to change with the

advancement of the menstrual cycle. The

_a,

subunit, a

Col/

LMbindingreceptor, waspresent onsecretorybut not prolifer-ativeendometrialepithelium. This subunit has previously been documentedonsecretorystageendometrium (34), and its

ex-pression in human endometrial explantswasshowntobe

pro-gestin dependent (35). The function of

a,

inendometrial epi-thelium remains poorly understood, though the pericellular

distributionmayportendarole incell-celladhesion.

Thea, integrin subunit is known for its abilitytoassociate withanumber

of/f

subunits, forming integrinscapable ofbind-ing to fibronectin, fibrinogen, vitronectin, thrombospondin, andvonWillebrand's factor (36-38). The temporalpatternof distribution for this subunit in endometriumwasrather

inter-esting. Immunostainingwasfirst detected before thesecretory

phasewithanincrease inintensitythroughout the cycle.While it is not possible to tell which /3 _molecule the av subunit is

pairedwith usingimmunohistochemistry,itsabilitytobindto

/1, /3, /5,or/6(36-41)suggests that av, may havemultiple

functions in humanendometrium.

Onesubunit knowntopairwith a, is /3, whosepatternof immunostainingonendometrial epithelial cellswasquite

un-usual. Westernblotanalysisappears to confirm

immunohisto-chemicalstainingthatendometrial epithelialcells possess little

ifany/3until the midlutealphase. Epithelial

a,03

is remark-ablefor severalreasons.First, /3is notcharacteristically

pres-a

el

1.E -I -T ..l.-.

a1 - a

... .. ...

o 5 10 15 20 25

c

Al

I5

ic

DayofCycle

2.5 - - -

-1.5

1'_-. I-. -

--

I-0 5 10 15 20 25

Day ofCycle

Figure4.Relativeintensityofstaining for theepitheliala, and 3

subunits in 35 endometrialsamples throughout themenstrualcycle. Thepatternofexpression fora, is shownin A,showingagradual

increaseinstaining throughout the menstrual cycle.Contrast thisto

thepatternfor 3inB,showingamoreabrupt riseinthisintegrin

subunitaroundday20 of themenstrualcycle.As described in the Methodssection, sampleswerestaged accordingtothelast menstrual cycle,aswellasbythehistologiccriteriaofNoyesetal.( 10).Sections

wereassignedascoreof 0 (-; negative), I(±;weak), 2(+;

moder-ate),or3(++; strong), byablindedobserver, andconfirmedbya

secondobserver.

A

3

0

0

Co

0

c

C

c

c

as

2

1

0

OOP

Ni

B

c

o.

a

,1

Figure5. _{Staining intensity}ofepithelial 3in 12

infertility

patients

with_delayedendometrialmaturation. Endometriumwascollected

asdescribed in the Methods_section,fromwomenundergoing evalu-ation for

_infertility.

The_biopsieswereseparatedintotwogroups based

onthecorrelationbetween_histologiccriteria and the menstrualcycle

dating

basedonthetimeof ovulationand/orthesubsequent

men-strual period.Patients with endometrialbiopsies 3ormoredaysout

of_{phase (OOP group)}werecomparedwith 25 endometrialbiopsies

thatwerein_{phase(Normal)}andshown inA.Sectionswereassigned

a scoreofO_(-;negative), 1(±; weak),2(+;moderate),or3(++;

strong),basedontheintensityof_{epithelial #3 staining.} Examplesof

immunohistochemicalstainingofan outofphasebiopsy(B)anda

normal in_{phase sample (C)}is includedto contrasttheepithelial#3

staining

in each group. x400.

enton

_epithelial

cells.

Second,

the

_abrupt

appearanceof the

/3

subunit afterday 19suggeststhatexpressionof the vitronectin

receptoris

_regulated

in human endometrium.

Third,

the

in-creasedepithelial

a(/A3

staining in normal cycles correlatesto a

putative "implantation window" thoughttoexist within the secretory phase. In studies that examined the conceptofan

implantation window,Navot and co-workers(42) usingdonor

oocytesin prepared recipients reporteddata on success of in

vitrofertilization (IVF) cycles based on the time ofembryo

transfer. Though the numbersaresmall, when the data is

ex-pressedas pregnancy pertransfer, the highestsuccess ratewas

noted around day 19ofthe menstrualcycle.Itis important to

pointoutthat thephysiologicbasis for this window hasnotyet beenestablished. These datasuggest apossible role for integrins in the initial interaction between maternal and embryonic cells, andmayparticipate in the establishment ofan

endome-trialperiod of receptivity.

Available ligands present on the developing embryo are

consistent with this hypothesis. Vitronectin ispresent on fetal

membranetissues (43), and evidence that vitronectinand fi-bronectin are involved in trophoblast attachment and

out-growth has previouslybeen demonstratedinthe mouse blasto-cyst(44). Trophoblast attachment in vitrocanbe blockedwith

RGD-containing peptides, providing additional evidence for

integrin involvement in trophoblast adhesion (45). Recently, data have accumulatedsuggesting that trophoblastselaborate analternatively splicedformoffibronectin,oncofetal fibronec-tin (46).Feinbergetal. (47) havesuggested that this form of

FN is aspecific markerfor the extravillous trophoblast and

mayprovide thematrix for trophoblast-endometrial interac-tion. Thus,epitheliala1,/3 could playarole inearly fetal

attach-ment, whilemesenchymal

_a,(/3?)

participates in maintaininga

fetal/maternal interfaceintheongoing pregnancy.

Involvement of

_aCf3

inimplantation is consistent with the

observation that this integrin is lacking in midluteal biopsies

from women whoseendometrial progression is retarded.

Im-munostainingfor/3in normalendometrium providesauseful internal landmarkof luteal phase development, and its delayed

expressioninthesepatientsmayprovidea newdiagnostic

mo-dalityforthe evaluation ofLPD. Traditionally, endometrial

biopsiesareperformedon orafter postovulatory day 10 (day 25ofthecycle).Sincepatients withLPDwhoundergobiopsy laterthanday24 mayeventually "catch up" andexpressthis

antigen, the earlier timing of endometrial biopsies as

per-formed in this study will likelybe acritical consideration in

establishing the use of/3 testing for this disorder. Once the

specificcellulareventsrelatingtoembryo attachmentare

bet-terdefined,theuseof endometrialbiopsy duringtheputative

window of endometrial receptivity(day20-24)may be more

appropriate.

Theprocesswherebyintegrinsareregulated in the

endome-triumhas yettobeestablished. While thetimingofal

expres-sionappearstobeprogestin dependent,thesituationfora,,and

3 appearstobemorecomplex.Adecliningratio ofestradiolto

progesterone may correlate better than progesterone alone,

with the expression ofone orboth ofthesesubunits.Weand others have noteda decrease in epithelial (but not stromal)

progesterone receptorsthatis maximal by day 20 of the

men-strualcycle (48, 49),independentevidence thatmidluteal

epi-thelial cellsundergodynamic changesinsteroidresponsiveness

thatmay accountfor theonsetof

/3

expression.Basedon

re-centstudies (50, 51), it would not besurprisingtofind that growth factors also playarole in theexpression ofcertain endo-metrialintegrins.

's. IM lw

41V,.a ..;4Tt

.:. ..-j

i4..

.1

A

C

- 205

.

.

-92

1 2 3- 4 5 69

B

D

Figure 6.Immunoblotanalysisofproliferativeandsecretory endometrium,stainedforthe(3subunit.(A)Immunoblotofplateletextract(lane 1),compared with twosamples from the early andmid-proliferativephase(lanes 2,3)andfrom the lutealphase(lane4and5;days23 and 26,respectively) demonstratesabandat- 95,000 mol wt,correspondingto(33.Samplesof endometriumwerepartiallydigestedwithcollagenase

and theglandular elementsobtained (B),usingamodification of the methods ofSatyaswaroopetal.(17).Theglandsappearashollowstructures free fromsurroundingstroma.Immunofluorescenceofsamplesfrom lanes 3 and 4(Cand D,respectively)correspondstothe absenceor pres-enceofthe95-kD band inA.x400.

Insummary,integrinexpression by the endometrium isan

interesting and dynamicprocess. Theendometrial epithelium

expressesintegrinsubunitssuchas

#3

that have not been noted

for mostother epithelia. Integrin expressionappearstobe

pre-dictable,based on the time in the menstrualcycle, makingthe

use ofintegrins of diagnosticuse as stage-specific markers in

Table II. Distribution ofIntegrin SubunitsinNormal

Endometriumduringthe Menstrual Cycle

Col/LM FN/VN

al a2 a3 a6 134 a4 a5 av 13 Epithelial

Proliferative 0 0 0 ** 0* Q Q Q 0

Early

secretory 0 0 0 0 0 0 0 a 0

Late

secretory 0 0 0 0 0 0 0 0 0

Stromal

Proliferative 0 0 0 0 0 0 0 0 0

Early

secretory 0 0 0 a 0 0 0 0 0

Late

secretory 0 0 0 0 0 0 a 0 0

*Correspondsto+or++staining.

oCorrespondsto ±staining.

o Correspondsto-staining.

*_{Signifiesbasolateraldistribution ofstaining.}

thistissue. Finally, an understanding of the function of

inte-grinsthroughout the menstrual cycle,and a means toregulate

their expressionmay prove tobe of future therapeuticuse.

Acknowledgments

We would also like to express our appreciation to Drs. Elizabeth Wayner,William Carter, Steven Kennel, MartinHemler,Arnoud Son-nenberg, Joel Bennett, JamesHoxie,andDavid Cheresh for

theirgener-ousandunconditional contributions ofantibodies,and Aracelis Po-leski for hertechnical assistance.

This project was supported in part by grant IN-135-J from the American Cancer Society and B.R.S.G. S07-RR-05415-29 awarded by the Biomedical Research Support Grant Program, Division of Re-search Resources, National Institutes of Health (B. A. Lessey); grants HL-01587 (S. M. Albelda), and HL 39023 (C. A. Buck) from the National Heart, Lung and Blood Institute; grants CA-19144 (C. A. Buck) and CA-108 15 (C. A. Buck) from the National Cancer Institute; and grantsfrom the Berlex Foundation, (B. A. Lessey), W. W. Smith Charitable Trust, The Mellon Foundation and the American Cancer Society (S. M. Albelda).

References

1.Getzenberg,R.H., K. J.Pienta,and D. S.Coffey.1990. The tissue matrix: celldynamicsand hormoneaction. Endocr. Rev. 11:399-417.

2.Albelda,S. M., and C.A.Buck. 1990.Integrinsand other cell adhesion molecules.FASEB(Fed.Am.Soc.Exp.Biol.)J. 4:2868-2880.

3. Ruoslahti, E., andM. D.Pierschbacher. 1987. Newperspectivesin cell adhesion: RGD andintegrins.Science(Wash.DC).238:491-497.

4. Hynes, R. 0. 1987. Integrins: a familyof cell surface receptors. Cell. 48:549-554.

5.Buck, C. A., andA.F.Horowitz. 1987.Integrin,atransmembrane glyco-protein complexmediatingcell-substratumadhesion. J. CellSci.Suppi. 8:231

6. Burridge, K., F. Karl, T. Kelly, G. Nuckolls, and C. Turner. 1988. Focal adhesions: transmembrane junctions between the extracellular matrix and the

cytoskeleton.Annu. Rev. Cell Biol. 4:487-525.

7. Hemler, M. E. 1990. VLA proteins in the integrinfamily:structures, func-tions, andtheirrole onleukocytes. Annu. Rev.Immunol.8:365-400.

8. Soules, M. R., R. 1. McLachlan, M. Ek, K. D. Dahl, N. L. Cohen, and W. J. Bremner. 1989. Luteal phasedeficiency:characterizationof reproductive hor-mones over the menstrual cycle. J. Clin. Endocrinol.Metab.69:804-812.

9.McNeely, M. J., and M. R.Soules. 1988. The diagnosis of luteal phase

deficiency:acritical review.Fertil. Steril.50:1-15.

10. Noyes, R. W., A. T.Hertig,and J. Rock. 1950. Dating the endometrium.

Fertil. Steril. 1:3-8.

11.Albelda, S. M., M.Daise,E. M. Levine, and C. A. Buck. 1989.

Identifica-tion and characterizaIdentifica-tionof cell-substratumadhesion receptors on cultured hu-manadult large vessel endothelial cells. J. Clin. Invest. 83:1992-2002.

12.Carter,W.G.,E.A.Wayner, T.S. Bouchard,and P. Kaur.1990. The role

ofintegrinsa2,f8 anda38,1incell-cell and cell-substrateadhesionofhuman

epider-mal cells. J.Cell Biol.110:1387-1404.

13. Cheresh, D. A., and J. R. Harper.1987.Arg-Gly-Asprecognitionby acell

adhesionreceptorrequires its 130-kD a-subunit.J.Biol.Chem. 262:1434-1437. 14.Hemler, M. E., C. Huang, Y. Takada, L. Schwarz, J. L. Strominger, and M. L.Clabby. 1987. Characterization ofthe cell surface heterodimer VLA-4 and relatedpeptides.J.Biol.Chem.262:11478-1 1485.

15.Sonnenberg, A., H. Janssen, F. Hogervorst, J.Calafat,and J.Hilgers.

1987. A complexofplateletglycoproteinsIcandIlaidentifiedby a rat monoclo-nalantibody. J.Biol.Chem. 262:10376-10383.

16.Sonnenberg,A.,C.J.Linders,J. H. Daams, andS.J.Kennel. 1990. The

a6:f,(VLA-6) anda684protein complexes: tissue distributionandbiochemical properties.J.Cell Sci. 96:207-217.

17.Satyaswaroop,P.G.,R.S. Bressler,M. M.DeLaPena,and E.Gurpide. 1979. Isolation andcultureof human endometrium glands.J. Clin. Endocr. Metab. 48:639-641.

18. Lowry,0.H., N. J.Rosebrough,A.L.Farr,and R. J.Randall. 1951.

Proteinmeasurementwith Folinphenolreagent. J.Biol. Chem. 193:265-271.

19.Laemmli,U. K.1970.Cleavage ofstructuralproteins during assemblyof the headof bacteriophageT4.Nature(Lond.). 227:680-685.

20.Albelda, S.M., P. D.Oliver,L. H.Romer,andC. A. Buck.1990. Endo-CAM: a novelendothelialcell-celladhesionmolecule. J.Cell Biol. 110:1227-1237.

21.Peltonen, J., H.Larjava, S. Jaakkola,H.Gralnick,S. K.Akiyama,S. S. Yamada, K. M.Yamada,and J.Uitto. 1989. Localization ofintegrinreceptors

forfibronectin, collagen,andlaminin inhumanskin.Variableexpressionin basal and squamous cellcarcinomas.J.Clin.Invest.84:1916-1923.

22.Choy,M.-Y., P.I.Richman,M. A.Horton,andT. T.MacDonald. 1990.

Expression ofthe VLAfamily of integrinsin humanintestine.J.Pathol.

160:35-40.

23.Simon,E.E., and J. A. McDonald. 1990. Extracellularmatrixreceptors in thekidneycortex. Am.J.Physiol. 258:F783-F792.

24.Albelda,S. M. 1991.Endothelialandepithelialcelladhesion molecules.

Am.J.Respir.CellMol. Biol.4:195-203.

25.Hertle,M.D., J. C. Adams, and F. M. Watt. 1991.Integrinexpression during human epidermal development in vivo and in vitro. Development (Camb.).112:193-206.

26.Sonnenberg,A., P. W.Moddermann,and F.Hogervorst. 1988. Laminin

receptor onplateletsinthe VLA-6family.Nature(Lond.). 336:487-489.

27. Lotz, M. M., C. A.Korzelius,and A. M.Mercurio.1990. Human colon

carcinomacells usemultiplereceptors to adhere tolaminin:involvement ofa644 anda26,.Cell Regul. 1:249-257.

28. deLuca, M., R. N. Tamura, S.Kajiji,S. Bondanza, P. Rossino, R. Can-cedda,P.C.Marchisio,and V.Quaranta.1990.Polarizedintegrin mediates

hu-mankeratinocyteadhesion to basallaminin.Proc.Natl. Acad.Sci USA. 87:6888-6892.

29.Larjava, H.,J.Peltonen,S. K.Akiyama,S.S. Yamada, H. R.Gralnick,J.

Uitto,and K. M. Yamada. 1990. Novelfunctionforfl, integrins in keratinocyte cell-cellinteractions. J.Cell Biol. 110:803-815.

30.Carter,W.G.,P.Kaur, S. G.Gil,P. J.Gahr,and E. A. Wayner. 1990.

Distinctfunctions for integrinsa3ftIinfocaladhesions anda6fllI/bullous pem-phigoid antigen in a new stable anchoring contact (SAC) of keratinocytes. Rela-tiontohemidesmosomes.J.CellBiol. 111:3141-3154.

31.Marchisio,P.C., S. Bondanza, 0. Cremona, R. Cancedda, and M. De Luca. 1991. Polarized expression of integrinreceptors(a6#4,a2pi,afi,,and

a435)and their relationship with the cytoskeleton and basement membrane ma-trix in cultured human keratinocytes. J. Cell Biol. 112:761-773.

32.Lampugnani, M. G., M. Resnati, E. Dejana, and P. C. Marchisio. 1991. The roleof integrins in the maintenanceof endothelialmonolayer integrity. J. Cell Biol. 112:479-490.

33.Aplin, J. D., A. K. Charlton, and S. Ayad. 1988. An immunohistochemi-cal study of human endometrial extracellular matrix during the menstrual cycle andfirst timester ofpregnancy.CellTissue Res. 253:231-240.

34.Tabibzadeh, S.S. 1990. VLA-1 associated with endometrial biopsies.

Fer-til.Steril.54:624-631.

35.Tabibzadeh, S. S., and P. Satyaswaroop. 1990.Progestin-mediated

induc-tionofVLA-1 inglandular epitheliumof humanendometrium,invitro. The 72nd AnnualMeetingof theEndocrine Society. 199:700. (Abstr.)

36.Lawler, J., R.Weinstein, and R. 0.Hynes. 1988. Cell attachmentto

thrombospondin:theroleofArg-Gly-Asp,calcium,andintegrinreceptors. J.

Cell Biol. 262:17249-17298.

37.Cheresh,D. A., J. W.Smith,H. M.Cooper,and V.Quaranta. 1989. A novelvitronectin receptor receptorintegrin(avflx) is responsiblefor distinct

adhesivepropertiesof carcinoma cells. Cell. 57:59-69.

38. Bodary, S. C., and J. W. McLean. 1990. Theintegrin ,11subunit associates with the vitronectinreceptor av subunit toforma novelvitronectinreceptor in a humanembryonic kidney cell line. J.Biol. Chem.265:5938-5941.

39.Ramaswamy,H.,and M. E. Hemler. 1990.Cloning, primarystructure, andproperties ofanovel humanintegrinbetasubunit. EMBO(Eur.Mol.Biol. Organ.)J.9:1561-1568.

40.Sheppard,D., C. Rozzo, L.Starr,V.Quaranta,D. J.Erle,and R.Pytela. 1990. Completeaminoacidsequenceofanovel integrin subunit(#,6)identified

inepithelial cells using the polymerasechainreaction.J.Biol.Chem.

265:11502-11507.

41.Busk, M., R. Pytela, V. Quaranta, and D. Sheppard. 1991.The novel integrinav/,B6is afibronectin receptor onairwayepithelial cells. Am. Rev. Re-spir. Dis. A402.(Abstr.)

42.Navot, D., P. A. Bergh, M.Williams,F.J.Garrisi,I.Guzman,B.Sandler, J.Fox, P. Schreiner-Engel, G. E.Hofmann,and L.Grunfeld.1991. Aninsight

into earlyreproductiveprocesses through the in vivo modelofovumdonation. J.

Clin. Endocrinol. & Metab.72:408-414.

43. Haymen, E.G., M. D.Pierschbacher,Y.Ohgren,and E.Ruoslahti.1983. Serumspreading factor (vitronectin) ispresent at the cellsurfaceand intissues. J.

Cell Biol. 80:4003-4007.

44. Armant, D.R., H. A. Kaplan, H. Mover, and W. J. Lennarz. 1986.The

effect ofhexapeptidesonattachment and outgrowthofmouseblastocysts

cul-turedinvitro:evidence fortheinvolvement ofthe cellrecognition tripeptide

Arg-Gly-Asp. Proc.Natil. Acad.Sci. USA.83:6751-6755.

45.Kao,L.-C., S. Caltabiano, S.Wu, J. F.Strauss, III,and H. J.Kliman. 1988.

The humanvillous cytotrophoblast: interactions withextracellular matrix

pro-teins,endocrinefunction,andcytoplasmic differentiationin the absenceof

syn-cytium formation. Development (Camb.). 130:693-702.

46.Zhu, B.C.-R.,S. F.Fisher,H.Pande, J. Calaycay, J. E.Shively,and R.A.

Laine. 1984. Human placental(fetal) fibronectin: increased glycosylation and

higherprotease resistance than plasmfibronectin.J.Biol. Chem. 259:3962-3970. 47.Feinberg,R.F., H. J. Kliman,andC. J. Lockwood. 1991. Isoncofetal fibronectinatrophblast glueforhumanimplantation.Am.J.Pathol. 138:537-543.

48.Lessey,B. A., A. P.Killam,D. A.Metzger, A. F. Haney, G. L. Greene, and K.S.McCarty, Jr. 1988.Immunohistochemicalanalysisof human uterine estro-gen and progesterone receptors throughout the menstrual cycle. J.Clin. Endo-crinol. Metab.67:334-340.

49.Garcia, E.,P.Bouchard, J. de Brux, J. Berdah, R.Frydman,G.Schaison,

E.Milgrom, and M. Perrot-Applanat. 1988. Use of immunocytochemistry of progesterone andestrogen receptors for endometrial dating. J. Clin. Endocrinol.

Metab.67:80-87.

50. Ignotz, R. A., J. Heino, and J. Massague. 1989. Regulation ofcell adhesion receptors bytransforminggrowthfactor-,B. J. Biol. Chem. 264:389-392.

51.Heino,J., R. A. Ignotz, M. E.Hemler,C. Crouse, and J.Massague. 1989. Regulation of cell adhesion receptors by transforming growth factor-,B.J.Biol.