• No results found

Patterns of expression of vascular endothelial growth factor (VEGF) and VEGF receptors in mice suggest a role in hormonally regulated angiogenesis

N/A
N/A
Protected

Academic year: 2020

Share "Patterns of expression of vascular endothelial growth factor (VEGF) and VEGF receptors in mice suggest a role in hormonally regulated angiogenesis"

Copied!
10
0
0

Loading.... (view fulltext now)

Full text

(1)

Patterns of expression of vascular endothelial

growth factor (VEGF) and VEGF receptors in

mice suggest a role in hormonally regulated

angiogenesis.

D Shweiki, … , H Gitay-Goren, E Keshet

J Clin Invest. 1993;91(5):2235-2243. https://doi.org/10.1172/JCI116450.

Vascular endothelial growth factor (VEGF) is a secreted endothelial cell-specific mitogen. To evaluate whether VEGF may play a role in angiogenesis, we have determined the spatial and temporal patterns of expression of VEGF and VEGF receptors during natural angiogenic processes taking place within the female reproductive system. Four angiogenic processes were analyzed: neovascularization of ovarian follicles, neovascularization of the corpus luteum, repair of endometrial vessels, and angiogenesis in embryonic implantation sites. During all processes, VEGF mRNA was found to be expressed in cells surrounding the expanding vasculature. VEGF was predominantly produced in tissues that acquire new capillary networks (theca layers, lutein cells, endometrial stroma, and the maternal decidua, respectively). VEGF-binding activity, on the other hand, was found on endothelial cells of both quiescent and proliferating blood vessels. These findings are consistent with a role for VEGF in the targeting of angiogenic responses to specific areas. Using in situ hybridization, we show that VEGF is expressed in 10 different steroidogenic and/or steroid-responsive cell types (theca, cumulus, granulosa, lutein, oviductal epithelium, endometrial stroma, decidua, giant trophoblast cells, adrenal cortex, and Leydig cells). Furthermore, in some cells

upregulation of VEGF expression is concurrent with the acquisition of steroidogenic activity, and expression in other cell types is restricted to a particular stage of the ovarian cycle. These findings suggest that expression of VEGF is […]

Research Article

Find the latest version:

(2)

Patterns

of

Expression

of Vascular Endothelial Growth Factor

(VEGF)

and

VEGF

Receptors

in

Mice

Suggest

a

Role in

Hormonally Regulated Angiogenesis

Dorit Shweiki, Ahuva Itin, Gera Neufeld,* HelaGitay-Goren,* and EliKeshet

DepartmentofMolecularBiology, The HebrewUniversity-HadassahMedicalSchool,Jerusalem 91010; and*Department ofBiology, Technion, Haifa 32000,Israel

Abstract

Vascularendothelialgrowth factor(VEGF)isasecreted endo-thelialcell-specific mitogen.To evaluate whether VEGFmay play arole inangiogenesis,wehavedeterminedthespatialand temporalpatternsofexpressionof VEGF and VEGF receptors

duringnaturalangiogenicprocessestaking placewithin the fe-malereproductivesystem. Fourangiogenicprocesseswere ana-lyzed: neovascularization of ovarianfollicles, neovasculariza-tionof the corpus luteum, repairofendometrial vessels,and

angiogenesisin embryonic implantation sites.Duringall pro-cesses, VEGF mRNA was found to be expressed in cells surrounding theexpanding vasculature. VEGFwas

predomi-nantly produced in tissuesthatacquirenewcapillarynetworks (thecalayers, luteincells,endometrial stroma, and the mater-naldecidua,respectively).VEGF-bindingactivity,onthe other hand,wasfoundonendothelial cells of bothquiescentand pro-liferating blood vessels. Thesefindings areconsistent witha roleforVEGFin thetargetingofangiogenicresponsesto spe-cificareas.Usinginsituhybridization, weshow that VEGF is expressedin10differentsteroidogenicand/or

steroid-respon-sivecell types(theca, cumulus,granulosa,lutein,oviductal epi-thelium, endometrialstroma,decidua, giant trophoblast cells,

adrenalcortex,andLeydigcells). Furthermore,insomecells

upregulationofVEGFexpressionisconcurrentwith the

acqui-sition ofsteroidogenic activity, and expression in other cell types is restricted to a particular stage oftheovarian cycle.

Thesefindingssuggest thatexpressionof VEGF ishormonally

regulated.We propose thatexcessiveexpressionofVEGF dur-ing gonadotropin-induced ovulationmaycontributetothe devel-opmentof ovarian hyperstimulationsyndromesbyvirtue ofthe vascular permeabilization activityof thisfactor.(J. Clin. In-vest. 1993. 91:2235-2243.) Key words: neovascularization * ovary *steroidogenesis*vascularpermeability* insitu hybrid-ization

Introduction

Angiogenesis,theformation ofnewbloodvessels as extensions ofexistingvessels, isacomplexmultistepprocess. Angiogene-sis is thoughtto beinitiatedby localactivationof genes

encod-ingdiffusible angiogenic factors(or,alternatively,by releaseof

Address correspondence to Dr. Eli Keshet, Department of Molecular Biology, TheHebrewUniversity-HadassahMedical School, Jerusalem 91010, Israel.

Receivedforpublication 7August 1992 and in revised form 16 De-cember 1992.

preformedfactorsfromtheir stores). The secretedangiogenic factors, in turn, initiate a cascade of downstream responses, thatresult in the growthof new capillariestowardsthe sites emitting the angiogenic signal. Since the turnover rates of vas-cularendothelial cells are extremely low, angiogenesis is gener-ally a quiescent process in the healthy adult organism. A marked exception isthefemalereproductivesystem,where the need for additional vasculature is constantly imposed by the cyclic evolution of transient structures and by the cyclic repair ofdamaged tissues. Specifically, in every estrous cycle, the se-quential development of ovarian follicles and corpora lutea

(CL)' is accompanied by concomitant development of elabo-ratecapillarynetworks.Thelatter areformedinthe periphery ofeachgrowing follicle,or inthevicinity of luteincells, respec-tively (3). Similarly, in the uterus, blood vessels of the func-tional endometrium undergo cyclic extensions and repair, compensatingforvessels that have beendamagedorsloughtoff (4). Extensive neovascularization ofthe endometrium also takesplaceupondecidualtransformationin response to embry-onicimplantation.Theincreaseindecidual mass is accompa-nied byinduction of neovascularizationandformation ofan anastomosing network traversing the maternal decidua and converging on the embryo (5). This vascular network supplies the maternalcontribution of the future placenta. These four processes representindependent examples ofprogrammed self-limitingangiogenesis withinthefemale reproductivesystem.

Studies from manylaboratories, using naturallyavascular tissues as model systems, have shown that a relatively large number of factors are capable of eliciting an angiogenic re-sponse (forreviews, seereferences 6-8). Yet the identity of factor(s)thatinitiateneovascularizationundernatural settings isnotknown. Thereproductivesystemisanattractivesystem tocriticallyevaluate therelevance ofputative angiogenic fac-tors tonaturalangiogenesis. It has theadvantagesthat angio-genesiscanbereadilyinduced and synchronizedby gonadotro-pins, and that itoccursover arelativelyshortperiodof timein awell-characterized schedule. Furthermore, vascular extension follows known routeswithrespect tothe vesselsof originand thetargettissue (9). Thus, specimenscanberetrievedat succes-sivestagesoftheongoingprocess.Elucidating spatialand tem-poral patterns of expression for each candidate angiogenic growthfactor, in conjunctionwith the pattern of expression of itscognate receptors, defines the framework of its potential in vivoaction.

Vascular endothelial growth factor (VEGF), is a heparin-binding, secreteddimeric glycoprotein (for review, see refer-ence 10). VEGFdisplayanangiogenicactivity inthe cornea

1.Abbreviations usedinthispaper:CL,corporalutea;OHSS, ovarian

hyperstimulation syndrome; VEGF,vascular endothelialgrowth

fac-tor.

VascularEndothelial GrowthFactorandAngiogenesis 2235

J.Clin. Invest.

© The American Society for Clinical Investigation, Inc. 0021-9738/93/05/2235/09 $2.00

(3)

h

cII

.

4

fcl

.'.

B

4t

I'

C'

...::"1

C

t,Jvwv

.4tv

D

'

y

v

-I.,.

;-/

tit'tll

.i

(4)

andchorioallantoic assays( 1 1-13),induces themigration of monocytes in vitro (14), and is amitogen with atarget cell specificity that seems to be restricted to vascular endothelial cells ( 15, 16). Recent in situanalysis by Phillipsetal.(17) have shown thatVEGF-binding activityisexpressedonendothelial cells comprising the vascular network of the corpus luteum, while the ligand is produced by the lutein cells(10). These findings prompted us to further investigate whether VEGF may play a role in theangiogenic processesdiscussedabove. Here, we show that in thereproductive tract, VEGFis specifi-cally expressed in thevicinityofproliferatingendothelium,in a pattern thatmayprovidedirectionalitytotheextending vascu-lature. VEGF receptors,on the otherhand,areconstitutively expressed in the endothelium, regardless of its proliferative status.

Itismost likelythat thecyclicangiogenicwaves inthe fe-male reproductive system arecoordinated by hormones. It is assumed, therefore, that these processes are mediated by an angiogenic factorsthat are, eitherdirectly or

indirectly,

hor-moneinducible. To examine whetherexpression of VEGFis regulated by hormones underphysiological conditions,we ex-tended the in situ analysis to steroidogenic cells outside the femalereproductivesystem. Interestingly, all cell types shown here to expressVEGFaresteroidogenicand /or steroid-respon-sive cells.

Methods

Animals,organs,and tissues. Hyperstimulation of ovaries wasinduced in immature (25-d-old) female rats by injecting 25 IU of pregnant mare's serum gonadotropin (G-4877; Sigma Immunochemicals, St. Louis,MO). 5 IU of human chorionic gonadotropin (GG-2; Sigma Immunochemicals)wasinjected46hlater. Whole reproductive tracts (includingovaries, oviducts, anduteri) were withdrawn atthe indi-catedtimes, representingsuccessive stages in the ovarian cycle.

Foranalysis of embryonicimplantation sites, 8-w-oldC57BL/6 female micewerehyperstimulated and matedwith C57BL/6males. Embryos were stagedas0.5-d postcoitum atnoonof the day in which the vaginal plugwasobserved. Whole implantation sites(including maternal, extraembryonic, and embryonictissues)werecollected.

Pseudopregnancywasinducedin mice byinjectinggonadotropins, matingwithavasectomized male, andintrauterine injectionofoil at 3.5 dpostcoitum(18).

Tissues werefixedovernight in 4%paraformaldehydein PBS. Fixed specimenswerethenincubatedin 0.5Msucrose in PBS before embed-ding inTissue-Tek OCTembedding medium (Miles Inc., Kanakee, IL) and insituhybridization.

In situhybridization. Thisprocedurewasperformed essentiallyas

describedby Hoganetal.( 19).Inbrief, 10-mm thick frozensections

werecollectedonpoly(L-lysine)-coated glass slides, refixed,and dehy-drated ingradedethanol solutions. Beforehybridization,sectionswere

pretreated successivelywith 0.2MHCl,pronase(0.125 mg/ml),4%

paraformaldehyde,and aceticanhydridein triethanolamine buffer. Hy-bridizationwascarriedout at50°Covernightinasolutioncontaining 50%(vol/vol). formamide/0.3M NaCl and35S-labeledRNAprobe (2

X 108cpm/ ml). Washingwasperformedunderstringentconditions that includedanincubationat50°C for> 14 h in 50%formamide/0.3

MNaCl anda30-min incubationat37°C with RNaseA(20mg/ml). Autoradiography was performed usingKodak NTB-2 nuclear track emulsion with exposure for 5-9 d. In the initialexperiments,control

hybridizationswitha riboprobeinthe "sense" orientationwerealso carriedout.

Hybridizationprobes.ThefollowingDNAfragmentswerecloned

ontothepolylinkerofaPBSvector(Stratagene,LaJolla, CA)toserve astemplatesforsynthesisofspecificcRNAs: A 1.8-kb-longcDNA

frag-mentcontaining - 3'twothirds of thecoding region, aswellasthe entire3'-untranslatedregionofmouseVEGF;an8.8-kb-long

EcoRI-SalI fragmentderived from a human vonWillebrand factor cDNA clone(20).

Constructs in PBSvector werelinearizedby digestionwith the

ap-propriaterestriction endonucleasetoallowsynthesisofan35S-labeled complementaryRNA in either the antisenseorsenseorientation

(us-ingT3orT7 RNApolymerase).RNAprobeswerefragmented bya

mild alkalinetreatmentbeforeusefor in situhybridization.

Insitu VEGF-binding analysis.Recombinant human VEGF(165 amino-acidlong form)wasproducedinSF9 cells infected witha bacu-loviruscontainingacloned VEGF cDNA under control of the poly-hedrin promoter. Production and purification of the factor were

performedaspreviouslydescribed(21 ),with the addition ofa cation-exchange chromatographystep. VEGFwasiodinatedbythe chlora-mine-T methodasdescribedbyVaismannetal.(22).

Bindingof[1251JVEGFtofrozen sections was performed as de-scribedbyMarchettietal.(23),withsomemodifications.Briefly,

or-ganswereincubatedovernightin0.5 Msucrosein PBS before embed-dingin Tissue-Tek OCTembeddingmedium. 10-mm thick frozen

sec-tionswerecollectedongelatine-coated glassslides. Afterpreincubation (PBS,30min, roomtemperature),slides were incubated(1 h,room temperature)in PBS solutioncontaining5ng/mlof[1251]VEGF (spe-cificactivity: 150,000 cpm/ng), 1 mg/ml heparine,and0.2%gelatine. Slides were subsequently washed with ice-cold PBS and, briefly(I

min),withwater.Sectionswerethen fixed for 10 min in 2.5% glutaral-dehydesolution in 50 mMphosphatebuffer(pH 6.5),washed in phos-phate buffer,airdried,andautoradiographed usingKodak NTB-2

nu-clear track emulsion. Exposuretime was 1-2 d.

Immunocytochemicalstaining.Sectionswerestained with a poly-clonal rabbit anti-humanvonWillebrand factor( 1:200)(DakoCorp., Carpinteria, CA)for 90minatroomtemperature.Theprimary

anti-Figure 1. Expression of VEGF mRNA during development of ovarian follicles and corpus luteum. In situ hybridization with a VEGF-specific probe. Left and right figures show the same section photographed under bright field and dark field illumination, respectively. (A) Early stages in folliculardevelopment. Arrows point at theca layers of preantral follicles and follicles with a small antrum. (B) Developed follicles showing highlevel of expression in cumulus cells. (C) Immediate preovulatory follicle. (D) An early stage in CL development. (E) Fully developed CL. v, blood vessels; is, ovarian interstitial tissue; t, theca; o, oocyte; c, cumulus; g,granulosa;a,antrum;cl,corpus luteum;lu,lutein cells.

Figure2. [1251]VEGF binding sites within the ovary. Ovary sections, representing growing follicles, were assayed for[1251 ]VEGF binding as

de-scribed in Methods. (A)Asection showing multiple transverse sections through large medullar vessels (v), interstitial tissue (is), and a number of growing follicles. Bright field (top) and dark field (bottom) photographs. Labeling is mostly in large vessels and in the periphery of follicles. (B)Alarger magnification of the area boxed in A. Note that autoradiographic grains are detectable primarily on the endothelial lining of the blood vessel. (C and D) Colocalization of von Willebrand factor immunoreactivity and['251]-VEGFbinding sites. (C) Immunostaining with

the endothelial cell-specific von Willebrand factor (see Methods for details). Note staining (brown reaction products) on large vessels (V), as well as on capillaries embedded in the theca layers of the follicle (tv). (D) An adjacent serial section analyzed for [1251 ]VEGF binding (bright fieldand darkfield photographs). Note colocalization of binding sites with von Willebrand factor immunoreactivity.

(5)

body was detectedusingtheavidin-biotin peroxidase system (Vecta-stain ABC; Vector Labs, Burlingame,CA)withdiaminobenzidine as the chromogen.

Results

Insitu hybridization of sectioned tissuesenabled us to deter-mine the sites of VEGF productionunder naturalsettingsthat preserve thecorrect interactions between different cell types. Organswereretrievedattimes where neovascularizationof tis-sues is ongoing, immediately withdrawn into a fixative, and processedforinsituhybridization analysis using

VEGF-speci-fic,

"IS-labeled

antisense ribopobrobes. Mouse VEGFcDNA

was cloned in ourlaboratory from a mouse brain cDNA li-brary. The 1.8-kb-longcDNA fragment used throughoutthis

study

contained 3'twothirdsof the

coding region,

aswellas

theentire3'-untranslated region (the VEGF-specificprobe de-tected only the expected VEGF-containingbandsin prelimi-narygenomic DNAblotting experiments [data not shown]). Insitu analysis ofmRNA waspreferredoverin situ immuno-detection ofthe encoded protein because the localization ofthe mRNA unequivocally identifies the producer cells, whereas VEGFisknown to be secreted andmightalso be sequestered elsewhere in the tissue (24). To define potential targets for locally released VEGF,weidentifiedcellsexpressing functional VEGFreceptors byperforming insitu ligand binding analysis. Tothisend,

'25I-labeled

VEGF165(produced in abaculovirus vectorsystem)wasused.

Neovascularization ofovarianfollicles andcorpusluteum.

Tofollow VEGF expression during follicular development,we usedfollicle-stimulating hormone-primedfemalerats. Inthis

experimental system, a largecohort of folliclesisinducedto

develop in a roughly synchronous manner (25). Whereas smallpreantralfollicles havenospecialvascularsupplyof their own,concomitantwithfurther growth follicles acquire individ-ualvascular sheaths. Thefollicularnetworksoriginatesat ves-selsresidingin themedulla,extends towards eachgrowing folli-cle, andformstwoconcentriccapillarynetworks embedded in the thecaexterna and theca interna, respectively. The whole process is completed within 1 d from the onset of follicular growth. Focusing on preantral follicles and on follicles with small antrum (i.e., follicles undergoing neovascularization),

VEGFwasexpressed in the interstitialtissueof theovaryand in the theca layers (Fig. 1 A). The VEGF producedand se-creted bythesecells islikelyto act on

nearby

cells, provided

thatthetarget cellsexpressthe

appropriate

receptors. Recent work has shown that thevast majority ofVEGFbinding ap-pears tobeassociated with endothelial

cells,

ingeneral,andto correspondtothe patternofovarianvascularization,in

particu-lar(10, 26). Wehavedetermined the cellpopulationsthat bind VEGFat thetime of follicular neovascularization by in situ receptor-ligand binding analysis. In agreement with the pre-vious findings,VEGF receptors weredetectableonthe endothe-liumoflargemedullar vessels(from which the thecal networks originate),oncellsinterspersedin the stroma(presumably en-dothelial cells),and in capillaries arranged in theperipheral

thecalayers ofgrowingfollicles(Fig. 2).Thus, itappearsthat theprimarytargetoftheligand,which is produced alongthe path ofcapillaryextensiontowardgrowing follicles,isthe endo-thelium. Furthermore, expression ofVEGFinthe thecalayers

(but notin the inner granulosa) suggeststhat thelocally

se-creted protein maytarget capillary extension to theperiphery ofthe follicle.

Concomitantlywith furthergrowth and maturation of folli-cles, the site of VEGF expression shifts to additional ovarian compartments(Fig. 1, B-E). Thefirstcellsinner to thetheca toexpressVEGFare cumulus cells engulfing the oocyte (Fig. 1 B). In the granulosa compartment, high levels of VEGF mRNA are detectable only at the immediate preovulatory stage (Fig. 1 C). Shortly after ovulation, as granulosa cells transformtolutein cellsand developmentofthe corpus lutem ensues,thepredominantsite ofVEGFexpressionis luteincells

(Fig. 1, Dand E). Fig. 1 thus portrays a dynamic pattern of VEGF expression, changing in parallel to the gonadotropin

stimuli receivedby different steroidogenicovarian cells and in

paralleltochanges in their steroidoutputs.

Withrespect toangiogenesis,the pattern of VEGF

expres-sionshown inFig.1 suggests that VEGF may also play a role in the second angiogenicprocess takingplacewithin the ovary, namely, neovascularization ofthe corpus luteum. Before

ovu-lation,thethecal capillarynetwork is sufficient to provide the granulosa cells residingonlyfewcell layers away. Upon devel-opmentof thecorpusluteum, however,lutein cells proliferate and gradually fill theentirespace ofthe formerantrum. To

satisfythe addedperfusionrequirements,capillaries invade the corpusluteum and form an elaborate network whereby each lutein cellresides in close proximityto bloodcapillaries. A role for VEGF in the development ofthe CL vasculature is

sug-gested bytheabundantproductionofVEGF by all lutein cells

duringtheperiodwhereconcurrentproliferationof luteincells andendothelialcell takes place (Fig. 1 D).Consistentwith its roleas aparacrine endothelialcell mitogen, it has been shown that theCL vasculaturepossesses aVEGF-binding activity (ref-erence 10andourdatanotshown). Notably, luteincells con-tinuetoexpress VEGF in thefully developed,functional CL

(Fig. 1E). The latter observation confirmsthefindingsof

Phil-lipsetal. (17).

Angiogenesis in theendometrium. Throughout the repro-ductive lifespan,theendometrium undergoes cyclic changes in

its secretoryactivityandstructure, in correlation with the cy-clicgrowthandmaturation ofovarianfollicles,and under con-trol of ovarian hormones. Theendometrial uniquesystemof blood vessels alsoundergoescyclicdevelopment. Specifically, after thepartialdestruction of the endometriumatthe end of eachcycle, thedistal portion of the spiralarteries undergoes

regeneration. Lengtheningandcoilingof arteries is under the

influenceofestrogensandprogesterone,beginningatthe prolif-erativephaseof the endometrium andcontinuing duringthe secretoryphase.

Toevaluate whether VEGFmayplay a role in neovascular-ization oftheendometrium,weanalyzeditspatternof expres-sion in the reproductivetrack during different stages of the

cycle(Fig. 3).VEGFwasfoundtobeexpressed in the

(6)

I

p

'A

.

Figure3.Expressionof VEGF in the oviduct anduterus.InsituhybridizationwithaVEGF-specificprobe: (A)asectionthroughtheoviduct;(B and C)sections ofthe uterus,withdrawnatearlyproliferating phaseandsecretory phase, respectively.m, muscularlayer;mu, mucosallayerof theoviduct;e,epithelium;en,endometrium.

clearly showing a cycle-dependent pattern of expression of VEGFin theendometrium,suggestthatexpression of VEGF is hormonally regulated. VEGF produced in the endometrial stroma may serve as a sourceofangiogenic activity that sup-ports theextensionof stromal vessels.

Angiogenesis in the embryonicimplantation site. The

se-quential influence ofestrogens and progesterone onthe stro-malcells of theendometrium makes them capable of undergo-ing transformation into decidual cells. The stimulus for the transformation is theimplantationof the blastocyst. Develop-mentofthedecidua involves extensive proliferation of decid-ual cells.Accordingly,anelaborate networkofblood vessels is formed within the decidua.At7-9 dpostcoitum, angiogenesis

isongoing and is visibleasmultipleendothelial cell cords tra-versing thedecidua. The anastomosingnetwork formed con-verges on theembryo.Thisspatialarrangementof the vascula-turereflects the fact that, lateron,thedecidua basalis will de-velopinto thematernalcontribution ofthe placenta (seeFig.4 Cforaschematic illustration of theimplantationsite ofa9.5-d embryo). The uniquespatialarrangementof the decidual vas-cular network isreadily visualized through in situ

hybridiza-tion with theendothelial cell-specific probe von Willebrand factor cDNA(Fig. 4D).

TodeterminewhetherVEGFmayplay a roleinformation oftheunique vascular network of the decidua, we analyzed implantation sites ofmouseembryosat9.5 dpostcoitum with respect to VEGF and VEGF receptorexpression.Our findings suggest thefollowing scenario: decidual transformation leads to the upregulation ofVEGFexpression in alldecidual cells (Fig. 4 A). Note that the deepest layers oftheendometrium (the so-called stratum basale)remainsnonexpressing. Thisis consistent withthe factthatdecidual transformationincludes allbut the deepest layer of the endometrium. Moreover, we

show thatupregulationofVEGFexpressiontakesplace concur-rently with decidual transformation. To show this point, we made useofapseudopregnancymodelsysteminwhich decid-ual transformationwasfocallyinducedthroughapplicationof oil to the hormonally conditioned, mechanically stimulated uterus(see Methods fordetails).Asshown inFig.4B, VEGF mRNAis only detectable in cells that have undergone decidual

transformation,while the bulk of nontransformed uterine cells remained nonexpressing(decidualcellsaredistinguishableby their large size relative to nontransformed endometrial cells, andarearranged in clusters around the presumed locations of oildroplets).VEGFproduced bythedeciduamaypotentially

supporttheproliferation ofinterspersedendothelial cells, es-sential for theexpansion ofthe decidual vasculature. To ac-countfor thedirectionalityof the vascular network, however, one has toevokeanonhomogenous distribution of VEGF(or,

alternatively, the involvement of additional angiogenic fac-tors). Therefore,weexamined whether VEGF is producedin additionalcell typeswithin the implantation site. Indeed, we found thathigheststeadystatelevelsof VEGFmRNA are pro-duced intheextra-embryonic giant trophoblasts cells(Fig. 4 F). VEGF vastly produced, and presumably secreted by giant

trophoblastcellssurrounding the embryo, islikelyto form a gradient of angiogenic activity directingthegrowth and/or mi-gration of endothelial cells towards the embryo. Consistent with this assignment,we showthat cordsofendothelial cells targeted to this site display a potent VEGF-binding activity

(Fig. 4E). Noteworthy, thesyncytiotrophoblast isthesite of synthesis ofbothsteroidsandpeptidehormones.

ExpressionofVEGF inothersteroidogeniccells.Strikingly,

all steroidogenic cell types included in the in situ survey de-scribedabovewerefoundtoexpress VEGF mRNA.Therefore,

we wished to determine whether thepropertyofVEGF

(7)
(8)

sion is sharedby other steroidogenic celltypes. In additionto theovaryandplacenta,twomajor steroidogenicorgans arethe adrenalgland and thetestes.

Intheadrenalgland,steroidogenesis is confinedtothe cor-tex where, under the influence ofACTH, different cortical zonesspecializein thesynthesisofdifferent steroids.Asshown inFig. 5A,VEGFmRNAisabundantly expressed in all cells of the adrenalcortex.Theadrenalmedulla,whichproduces

non-steroidhormones, doesnotexpressVEGF.

In the testes, specialized interstitial cells termed Leydig

cells, elaborate the male steroid hormone testosterone. As shown inFig.5B,VEGFexpressionin thetestis is confinedto Leydig cells, seenscattered in theinttersticesamongthe semi-nefrous tubules. Forcomparison,noexpression wasdetected in the tubules.

Inconclusion, itappearsthat VEGF isefficiently expressed

in all thesteroidogeniccelltypesexamined.

Discussion

The search for angiogenic factors has yielded a considerable number of factors that elicitanangiogenicresponsein model systemsand/or factors that fulfill certain criteria ofangiogenic factors(e.g., endothelial cell mitogens orchemoattractants).

The apparent redundancy in putative angiogenic factors has been previously discussed, and the

possibilities

that redun-dancy reflectseither artifactual promiscuityof theassay sys-tems, orindirect effectsofsomeangiogenic

factors,

or a situa-tion where different tissueselaboratedifferent

angiogenic

fac-tors, hasbeenconsidered(e.g.,seereference27). Yet,theissue ofphysiologicalrelevance of each candidate factortoanygiven

naturalprocessremainsakeyunresolved issue inangiogenesis

research. Ideally, this issue should be addressedthrough

analy-sisof naturalangiogenesis.Bestsuitableareprogrammed pro-cesses, where specimens that represent all stages of

ongoing

angiogenesis, including initiating eventsand downstream re-sponses, canberetrieved for in situanalysis. Theobjective of

thepresent study was toanalyzein situ

angiogenic

processes

taking placein thereproductivesystem,with theaim of

identi-fyinganangiogenicfactor that isexpressedin close spatiotem-poralproximitytothe

forming

vasculature.

The female reproductive system undergoesa number of programmed neovascularization processes coupled with the

cyclicevolution and decline of ovarian and uterinestructures.

Cyclingin theovary,ingeneral, isinitiatedthroughthecyclic

secretion of

gonadotropins

from extraovarian tissues. Func-tionalcompartmentalizationwithinthe ovary,withrespectto

gonadotropin receptors and steroidogenic capabilities gener-ates,inturn, awavelikesteroidogenesiswithin theovary. Ovar-ian steroids also govern endometrial cycles. It ismostlikely,

therefore, that the angiogenic waveswithin the reproductive system are coordinated by gonadotropins and/orby locally producedsteroids. This requires that the expression ofthe ini-tiating angiogenic factor will be hormone responsive (although hormoneresponsivenessmaybeindirect).Wehaveexamined successivestagesinfour independentangiogenicprocesses tak-ing place in thereproductivesystem.Ourfindingscanbe sum-marized by generalizing that, in all four processes, VEGFis indeed expressed in spatial and temporal proximity to the forming vasculature. Specifically, during follicular neovascular-ization, VEGF was produced by interstitial and peripheral theca layers;duringformation of the CL vasculature, VEGF productionwaspredominantlydetectedinluteincells;during neovascularization oftheendometrium, VEGF was detected throughout the endometrial stroma; and during vasculariza-tionofthedecidua,VEGFwasexpressed in thedeciduabasalis. These sites of VEGF expression colocalize with the sites of residence ofthe new vasculature. VEGF receptors were pri-marily detectedonendothelial cells(e.g.,Figs.2 and 4 E). It should be noted, however, that thelimited sensitivity ofthein situreceptor-ligand bindingassay mayhave precluded detec-tion of VEGFreceptors onadditional celltypes(28). These resultssuggestthat,during theseprocesses,VEGFacts primar-ilyonnearby endothelialcells. SinceVEGF receptors are ex-pressed both in endothelial cells participating in neovasculariza-tion,aswellasinquiescent endothelium (reference26 andour

unpublisheddata), itappearsthat the locally releasedligand

determinesthe area of endothelial cell growth bycreating a microenvironment thatsupporttheproliferation ofnearby en-dothelial cells. Furthermore,unevendistributionof extracellu-lar VEGF in the neovascularized tissue may target capillary growth towards cellsproducing highestlevelsoftheangiogenic growth factor.Forexample,we suggestthat themassive produc-tion of VEGF by thesyncytiotrophoblast playsanimportant role intargetingthedecidual network towards theimplanted embryo(Fig.4). Althoughcircumstantialinnature, webelieve thatourfindingssupportthethesis that VEGF mediates hor-monally regulatedangiogenesis.

Itshould be noted that in thesystems analyzed, VEGFwas alsoexpressedattimes whereneovascularizationwas not

ongo-ing;(e.g., intheca andluteincellsafter completion of

angio-genesis). Thus, VEGFmayplayadditional roles in these sys-tems, possibly in the maintenance of the vasculature, or in

controllingitspermeability.Asimilarobservation, namely, ex-pression ofVEGFduringembryonic angiogenesis,but persis-tence ofVEGF expression in the fully developed organ was madebyBrieretal. (29) in thecaseof brainandkidney devel-opment.

Whatmechanism s) accountfor the activation of VEGF in eachofthedifferentcelltypes? And,morespecifically,canthe activation ofVEGF in the respective celltypes beaccounted

Figure 4.Expression of VEGF and VEGFreceptorsinthe decidua and embryonic implantation site. (A) Decidual tissue of a 9.5-dembryo. Hybridization withaVEGF-specificprobe. (B) Uterusofa pseudopregnant female mouse hybridized with a VEGF-specific probe. Note thatonly

afraction of endometrial cells have undergone decidual transformation.Hybridization with a VEGF-specific probe. (C) A schematic viewof

animplantationsiteat9.5-d postcoitum. (D) Theformingvascular network visualized through in situ hybridization with an endothelial cell-specificvonWillebrand factorcDNAprobe.Note theorientation ofendothelial cell cords traversing the decidua basalis and convergingon theembryo. (E) Specific binding of '25I-labeledVEGF toendothelialcell cords. The region of the implantation site shown is the same as inD. (F)Insituhybridization withaVEGF-specificprobe. The section shown includes the embryo, extraembryonic tissues, and a portion of the de-ciduaundergoing neovascularization.m,myometrium;sb,stratum basale;db,decidua basalis; en, endometrial cells; d, decidual cells; e, endo-thelial cells;gt,giant trohoblast cells;em,embryo;dc, deciduacapsularis; u, uterine lumen.

(9)

forby hormonalregulation?Together, this in situ study shows the in vivo expression of VEGF in 10 differentcell types: theca, cumulus, granulosa, lutein,oviductaland uterineepithelium,

endometrialstroma,decidua basalis, syncytiotrophoblast, adre-nalcortex,and Leydig cells.Noteworthy, incertain cell types (e.g.,granulosaandendometrialstroma)expressionofVEGF 02 is temporally restricted, detectable only at some stages of the

ovarianorendometrial cycle butnot atothers.Thesefindings suggestthat,atleast in thesecells,expression ofVEGF is hor-monallyregulated, ratherthan merelyreflectingcell type speci-< ficity.All 10expressing celltypes areeither steroid hormone-Q¢ responsive orsteroid hormone producers, or both. Excluding theestrogen-responsive oviductalandendometrial epithelium, allVEGF-expressing cells produce steroids inresponse to the respectivetrophic hormone. It is thus possible that VEGF is luteinizinghormone (LH)- or ACTH- inducible in a similar mannerthat LHinducesthetranscriptionof mRNAs encoding

steroidogenic enzymes. Consistent with this hypothesis, we

0

.0 find a temporal correlation and colocalization of VEGF

< mRNA and mRNA ofthe key steroidogenic enzyme cy-tochromeP450side-chain

cleavage

in the

theca,

cumulus and CLcompartmentsof theovary(datanotshown).Also, wefind that whenpartial decidualization is induced in the pseudopreg-nancy

model,

VEGFexpressionco-localize with cells express-ing P450 side-chain cleavagemRNA(Schiff, R.,Y. Arensburg, A.

Itin,

E.

Keshet,

and J.

Orly,

manuscript

in

preparation).

Alternatively, expression of VEGF might be steroid-inducible.

W, Inthiscase,expression in steroidogenic cells mightbeinduced

CZ by steroidsacting autocrinically. Obviously, different

mecha-.g

.~

nisms of activation might be operating in the different cell types shown hereto express VEGF. In any event, thestudy

*a: reported hereprovidesasuitable invivo framework for further

*l O invitrostudies addressingtheissueof hormonalregulationof =

.=

VEGFexpression.

Certain steroids areknown to possess

angiostatic

activity

.E;;g (30).Yet, the molecular basisforthe

antiangiogenic activity

of steroids is

poorly

understood. We

hope

that further

analysis

of

natural,self-limiting

angiogenic

processes, in whichboth the "on" and "off" switches are likely to be regulated by

hor-o .X° mones,will shed

light

onthis issue.

Finally,

VEGF is knownto possess apotent vascular

permeabilization activity (31, 32).

Thus,overstimulation ofVEGF expression

might

leadto ex-cessive vascular

permeability.

The ovarian

hyperstimulation

protocol used in these experiments, leadingto high levels of VEGF expression in a large number of follicles andcorpora lutea, issimilartotheoneused in ovulationinduction therapy,

where the recruitment andmaintenance ofalarge numberof

ovarian follicles is inducedbygonadotropins follicle-stimulat-ing hormone combined with an unphysiological dose of

hu->c

manchorionic

gonadotropin

(LH).

A main

complication

of

o

ovulation induction therapy is the ovarian hyperstimulation

°0 za syndrome (OHSS). Interestingly, the underlying problem in OHSS istheshift of

body

fluids fromthevascularbed intothe extravascularspace. Wespeculatethat excessive

production

of VEGFmightcause oratleastcontributetoOHSS. This pro-posal is currently underinvestigation.

Acknowledgments

This workwassupported bygrantsfrom the CharlesH. RevsonFund,

(10)

Ministry forHealth(toE.Keshet)andbyagrant fromtheUSA-Israel Binational Fund (toG.Neufeld).

References

1.Gospodarowicz, D., and K. K. Thakral. 1978. Production of corpus luteum

angiogenicfactorresponsibleforproliferationofcapillariesand

neovasculariza-tionofthe corpus luteum. Proc. Natl.Acad.Sci. USA.75:847-851.

2. Koos, R. D., and W. J. Le Maire. 1983. Factors that mayregulatethe growth andregression ofbloodvessels in the ovary. Semin. Reprod. Endocrino!.

1:295-307.

3. Findlay, J. K. 1986.Angiogenesisinreproductivetissues.J.Endocrinol. 111:357-366.

4.Christianes,G. C. M. L., J. J.Sixma,andA. A.Haspels. 1982.Hemostasis

inmenstrualendometrium:areview.Obstet.Gynecol. Survey. 37:281-303.

5.Motro,B.,A.Itin,L.Sachs, andE.Keshet. 1990. Patternofinterleukin 6

expressioninvivo suggestsarole inangiogenesis.Proc.Natl. Acad. Sci. USA.

87:3092-3096.

6.Folkman, J., and M.Klagsbrun. 1987.Angiogenicfactors. Science(Wash.

DC).235:442-447.

7. Risau, W. 1990.Angiogenicgrowth factors. Progress in growth factors research.Neuroendocrinol.Immunol. 2:71-79.

8. Klagsbrun, M., and P.A. D'Amore. 1991. Regulators ofangiogenesis.

Annu. Rev.Physio!.53:217-239.

9. Bassett, D. L., 1943. The changes in the vascular pattern ofthe ovary of the

albinoratduringthe esrouscycle.Am.J.Anat.73:251-291.

10. Ferrara, N., K. Houck,L.Jakeman, and D. W. Leung. 1992. Molecular andbiological propertiesof the vascular endothelial growth factor family of pro-teins.Endocr.Rev. 13:18-32.

11.Connoly, D. T., D. M. Heuvelman, R. Nelson, J. V. Olander, B. L. Eppley, J. J.Delfino, N. R. Siegel, R. M. Leimgruber, and J. Feder. 1989. Tumor vascular

permeabilityfactor stimulates endothelial cell growth and angiogenesis. J. Clin. Invest.84:1470-1478.

12. Leung, D. W.,G. Cachianes, W.-J. Kaung, D. V. Goeddel, and N. Ferrara. 1989. Vascular endothelial growth factoris a secreted angiogenic mitogen. Science(Wash.DC). 246:1306-1309.

13. Plouet, J., J.Schilling,andD.Gospodarowitz.1989.Isolation and

charac-terization ofanewlyidentifiedendothelialcellmitogenproduced by AtT-20 cells. EMBO(Eur. Mol.Bid.Organ.) J. 8:3801-3806.

14.Clauss, M., H. Gerlach, J. Brett, F. Wang, P. C. Familletti, Y.-C. E. Pan, J. V.Olander, D. T. Connolly, and D.Stern.1990.Vascular permeability factor:

Atumor-derivedpolypeptidethat inducesendothelial cell and monocyte procoag-ulantactivity,and promotes monocyte migration.J.Exp.Med. 172:1535-1545. 15.Gospodarowitz, D., J. A. Abraham, and J. Schilling. 1989. Isolation and

characterization ofvascularendothelial mitogen produced by pituitary-derived

folliculostellate cells. Proc.Natl. Acad.Sci. USA. 86:7311-7315.

16. Conn, G., D. D. Soderman, M.-T. Scaeffer, M. Wile, V. Hatcher, and

K.A.Thomas. 1990.Purification ofaglycoproteinvascularendothelialcell mi-togenfroma ratglioma-derivedcell line. Proc. Nat!. Acad. Sci. USA.

87:1323-1327.

17.Phillips, H. S., J. Hains, D. W. Leung, andN.Ferrara. 1990. Vascular endothelial growth factor is expressed in rat corpus luteum. Endocrinology

127:965-967.

18.Finn, C. A. 1965. Oestrogen and the decidual cell reactionofimplantation

inmice. J.Endocrinol.32:223-229.

19. Hogan, B., F. Constantini, and E. Lacy. 1986.Manipulatingthe Mouse Embryo: A Laboratory Manual. Cold SpringHarbor Laboratory Press, Cold

SpringHarbor, NY. 219-244.

20. Bonthron, D., E. C.Orr,L.M.Mitsock, D. Ginsburg, R. I.Handin,and S.H.Orkin.1986.Nucleotide sequenceofpre-pro-von Willebrand factor cDNA. NucleicAcidsRes. 14:7125-7127.

21. Cohen, Z., H. Gitay-Goren, G.Neufeld,and B.-Z. Levi. 1992.Highlevels

ofbiologically activevascularendothelialgrowth factorareproduced by the

Baculovirusexpression system.GrowthFactors. 7:131-138.

22.Vaismann, N., D. Gospodarowicz, and G.Neufeld.1990.

Characteriza-tion of the receptor for vascular endothelial growth factor. Biol. Chem.

265: 1946 1-19466.

23. Marchetti, B., M. C. Morale, and G. Pelletier. 1990.Sympatheticnervous system controlofratthymus gland maturation: autoradiographic localization of

theb2-adrenergicreceptor in thethymusand presenceofsexualdimorphism duringontogeny. Prog. inNeuroendocrinol.Immunol. 3:103-115.

24.Dvorak,H.F., T. M.Sioussat,L.F.Brown, B. Berse, J.A.Nagy,A.Sortel,

J.Manseau, L. Van De Water, and D. R. Senger. 1991.Distribution ofvascular

permeability factor(vascularendothelialgrowthfactor)intumors:concentration

in tumor blood vessels. J. Exp.Med. 174:1275-1278.

25.Gates, A. H. 1971. Maximizing yield and developmentaluniformityof eggs. In Methods in MammalianEmbryology.J.C. Daniel and W. H.Freeman, editors.SanFrancisco,CA. 46-76.

26. Jakeman, L. B., J. Winer, G. L. Bennett, C. A. Altar, and N. Ferrara. 1992.

Binding sites forvascularendothelialgrowthfactorarelocalizedonendothelial

cells in adult rattissues.J.Clin.Invest.89:244-253.

27.Zetter, B.R. 1988.Angiogenesis, stateofthe art.Chest. 93:159S-166S.

28. Gitay-Goren, H., S. Sofer, 1. Vlodavsky, and G. Neufeld. 1992. The

bind-ing ofvascularendothelial growth factor to its receptor is dependent on

cell-sur-face associatedheparin-likemolecules.J.Biol.Chem.267:6093-6098. 29.Brier, G., U.Albrecht, S. Sterrer,and W. Risau. 1992.Expression of vascular endothelial growth factor during embryonic angiogenesis and endothe-lial celldifferentiation.Development. 114:521-532.

30.Folkman, J., and D. E. Ingber. 1987. Angiostatic steroids:methodof discoveryandmechanismofaction.Ann.Surg.206:374-384.

31. Senger, D. R., S. J. Galli, A. M. Dvorak, C. A. Perruzzi, V. S. Harvey, and

H. F.Dvorak. 1983. Tumor cells secrete a vascular permeability factor that pro-motesaccumulationof ascitesfluid.Science(Wash. DC). 219:983-985.

32.Keck, P. J., S. D. Hauser, G. Krivi, K. Sanzo, T. Warren, J. Feder, and

D.T.Connolly. 1989. Vascular permeability factor, an endothelial cell mitogen related to PDGF. Science (Wash. DC). 246:1309-1312.

References

Related documents

Clarke &amp; Wright (1964) generalized the truck dispatching problem to a linear optimization problem which is commonly used in the field of logistics and transport: i.e., how

subtilis spores by fusing laccase to the spore coat proteins CotC and CotG using two linker peptides; a short, flexible peptide linker with the amino acid sequence GGGGS and a

Preven- tive measures to protect young players from direct ball contact include the use of batting helmets and face protectors while at bat and on base, the use of special equipment

Here, we present data demonstrating that similar to αβ T-cells, age-associated differences in γδ T-cell frequency and differentiation phenotype are predominantly present in

 If we increase the node density then data dropped due to buffer overflow is low in TORA and at node density 30, OLSR has lower data dropped in Normal flow as well as jelly

(2006) Existence, Uniqueness of Weak Solutions and Global Attractors for a Class of Nonlinear 2D Kirchhoff-Boussinesq Models.. (2009) Global Attractor for a Class of

In situations where the cost of successful election campaigns has risen, where funding (and notably state funding) is inadequate, and where party finance laws are

associated with increased risk of lower extremity neuropathy in the lithotomy position, presumably due to less organic padding to protect nerves, as well as brachial plexus injuries