inhibition of rat mammary tumorigenesis by human chorionic gonadotropin associated with increased...

BRIEF COMMUNICATION

Inhibition of Rat Mammary Tumorigenesis byHuman Chorionic Gonadotropin AssociatedWith Increased Expression of Inhibin

Pramod Srivastava, Jose Russo, and Irma H. Russo �

Breast Cancer Research Laboratory, Fox Chase Cancer Center, Philadelphia, Pennsylvania

This work was designed to test whether the suppression of 7,12-dimethylbenz[a]anthracene (DMBA)±induced ratmammary carcinomas by human chorionic gonadotropin (hCG) is associated with the synthesis of inhibin. For thispurpose, virgin rats received 8 mg DMBA/100 g body weight; 20 d later they were injected daily with 100 IU of hCG for40 d (DMBA�hCG group). Age-matched untreated (control), hCG-, and DMBA�saline-treated rats were used forcomparisons. Mammary tissues were collected for histopathological and mRNA analyses after 5, 10, 20, and 40 d ofhCG injection and 20 d after treatment. None of the animals in the control and hCG-treated groups developedmammary tumors. DMBA-treated rats developed a high incidence of both microscopic lesions, i.e., intraductalproliferations and ductal carcinomas in situ, and palpable tumors. In DMBA�hCG-treated rats, the incidence ofmicroscopic and palpable tumors was markedly reduced. In these animals, a- and b-inhibin immunoreactivity waselevated in the non-tumoral mammary glands in association with lobule formation and in the tumors. Inhibin A and BmRNAs were also elevated in the mammary tissue, and c-myc and c-jun were induced by the hormonal treatment.DMBA alone did not modify the expression of these genes. Our findings indicate that inhibin production and geneactivation are associated with both mammary gland differentiation and tumor regression. Mol. Carcinog. 26:10±19,1999. # 1999 Wiley-Liss, Inc.

Key words: mammary carcinogenesis; gonadotropins; tumor suppressor; tumor inhibition

INTRODUCTION

Studies of rodents in our laboratory have demon-strated that mammary cancer can be induced withthe chemical carcinogen 7,12-dimethylbenz[a]an-thracene (DMBA) only in young nulliparous females[1,2]. Completion of pregnancy before carcinogenexposure prevents carcinoma development [3±8].This preventive effect is in great part mediated bychorionic gonadotropin [3±6], a glycoprotein hor-mone secreted ®rst by the fertilized egg and later bythe placenta [9]. Administration of human chor-ionic gonadotropin (hCG) to virgin rats inhibitsboth the initiation and progression of DMBA-induced mammary carcinomas. We have demon-strated that the protective effect of both pregnancyand hCG is in great part mediated by the inductionof mammary gland differentiation [3±6]; this phe-nomenon results in inhibition of cell proliferation,increased DNA repair capabilities of the mammaryepithelium, and decreased binding of the carcino-gen to the DNA [1,2,10,11]. In addition, lobulardevelopment is associated with increased synthesisof inhibin, a gonadal protein with tumor suppressoractivity [12±16]. We previously reported [17] thatthe inhibition of mammary tumorigenesis by hCGtreatment is associated with the activation of genesthat are known to be responsible for programmedcell death and apoptosis, such as interleukin-1-b-

converting enzyme, testosterone-repressed prostatemessage 2, p53, c-myc, and bcl-XS [17].

Inhibins are heterodimeric glycoproteins belong-ing to the inhibin/activin family and are composedof a common 18-kDa a subunit and one of two 14-kDa b subunits (bA or bB), which form the hetero-dimers inhibin A (a-bA) and inhibin B (a-bB),respectively [18±21]. Inhibins were ®rst recognizedas gonadal hormones. The ovary is the major site ofinhibin synthesis in cycling females. Inhibin a andb cDNAs have been isolated from rat ovaries, andtheir expression has been characterized during therat estrous cycle [22]. During pregnancy, high levelsof circulating inhibin have been detected in rats andother primates [18±22]. Although inhibins areknown to selectively suppress the synthesis andrelease of the pituitary follicle-stimulating hormone[19], they have been recognized more recently to bepart of a large group of morphogenesis- and

MOLECULAR CARCINOGENESIS 26:10±19 (1999)

# 1999 WILEY-LISS, INC.

�Correspondence to: Breast Cancer Research Laboratory, FoxChase Cancer Center, 7701 Burholme Avenue, Philadelphia, PA19111.

Received 25 September 1998; Revised 12 April 1999; Accepted16 April 1999

Abbreviations: DMBA, 7,12-dimethylbenz[a]anthracene; hCG,human chorionic gonadotropin; TGF, transforming growth factor;i. p., Intraperitoneal; IDP, intraductal proliferations; DCIS, ductalCarcinoma in situ; DAB, 3,3 0-diaminobenzidine.

differentiation-related proteins the transforminggrowth factor (TGF)±b superfamily, which alsoincludes the b-homodimer activin [20,21]. Ourobservations that hCG stimulates the synthesis ofinhibin in the mammary epithelium (a phenom-enon related to lobular development [12,13,15] aswell as in human breast epithelial cells treated withhCG in vitro [14], probably by acting as a localmediator of hCG action either as an autocrine orparacrine growth-regulating factor, led us to testwhether inhibin was also involved in the regressionof DMBA-induced rat mammary carcinomas.

Our objective was to determine whether activa-tion of inhibin gene transcription is associated withthe induction of differentiation of the mammarygland and inhibition of mammary tumor progres-sion. Because inhibin synthesis could be mediatedby activation of speci®c transcription factors, wealso measured the levels of early-response genessuch as c-myc, c-fos, and c-jun. Both fos and jun aremembers of the genes known as cellular immediate-early genes and encode proteins that function intranscriptional regulation [23±25]. In the studydescribed here, we combined mRNA analysis andimmunohistochemical techniques to determinewhether lobular development, inhibin expression,activation of early-response genes, and the protec-tive effect of hCG against chemically inducedmammary carcinogenesis are related.

MMAATTEERRIIAALLSS AANNDD MMEETTHHOODDSS

EExxppeerriimmeennttaall PPrroottooccooll

Three hundred virgin Sprague-Dawley ratsobtained from Taconic Farm (New York, NY) werehoused four to a cage in environmentally controlledclean-air rooms with a 12-h light/12-h darknesscycle. They were fed pellets of Purina Certi®edRodent Chow 5002 and tap water ad libitum. The

animals were divided into the following groups:control, rats that received a daily intraperitoneal(i.p.) injection of saline for 40 d, starting at age 65 d;hCG, rats treated with a daily injection of 100 IU ofhCG (Sigma Chemical Company, St. Louis, MO) for40 d, starting at age 65 d; DMBA, rats inoculatedintragastrically with a single dose of 8 mg of DMBA(Eastman Kodak, Rochester, NY) per 100 g of bodyweight, dissolved in corn oil, when they were 45 dold, followed by a daily saline injection adminis-tered for 40 d from ages 65 to 105 d; andDMBA�hCG, rats that received the same dose ofDMBA when they were 45 d old, followed by a dailyinjection of 100 IU of hCG for 40 d, from ages 65 to105 d (Figure 1). Eleven animals each from thecontrol, DMBA, and hCG groups and 16 animalsfrom the DMBA�hCG group were killed when theywere 45 d or 65 d old (0 time control) and after 5, 10,20, and 40 d of hCG treatment and 20 d aftertreatment, when the animals were 65, 70, 75, 85,105, and 125 d old, respectively. Tumorigenicresponse was evaluated by biweekly palpation. Alltumors were sequentially numbered, their locationwas recorded, and the rate of tumor growth wasdetermined by biweekly measurement of tumor sizewith a Vernier caliper. The animals were killed at thetimes indicated above for collection of mammarytissues, tumors, and ovaries as described elsewhere[17]. All mammary glands and palpable tumors wereprocessed for histopathologic examination. Allmicroscopic lesions, such as intraductal prolifera-tions (IDP) and ductal carcinomas in situ (DCIS), aswell as palpable tumors, were classi®ed according topreviously established criteria [26].

PPrreeppaarraattiioonn ooff PPoollyy((AA))��,, PPrroobbeess,, aanndd NNoorrtthheerrnnBBlloott AAnnaallyyssiiss

Total RNA was isolated from frozen mammaryglands and tumors by the guanidinium isothiocya-

FFiigguurree 11.. EExxppeerriimmeennttaall pprroottooccooll ffoorr tteessttiinngg tthhee eeffffeecctt ooff hhCCGG oonnDDMMBBAA--iinndduucceedd mmaammmmaarryy ttuummoorr pprrooggrreessssiioonn iinn vviirrggiinn SSpprraagguuee--DDaawwlleeyy rraattss.. CCoonnttrrooll,, ddaaiillyy ii..pp.. iinnjjeeccttiioonn ooff ssaalliinnee;; hhCCGG,, ddaaiillyy iinnjjeeccttiioonnooff 110000 IIUU ooff hhCCGG ffoorr 4400 dd,, ssttaarrttiinngg wwhheenn tthhee aanniimmaallss rreeaacchheedd tthhee aaggeeooff 6655 dd ((hhCCGG ssttaarrtt,, aarrrrooww)) aanndd eennddiinngg wwhheenn tthheeyy wweerree 110055 dd oolldd((hhCCGG ssttoopp,, aarrrrooww));; DDMMBBAA aanndd DDMMBBAA��hhCCGG,, aaddmmiinniissttrraattiioonn ooff 88 mmgg

DDMMBBAA//110000 gg bbooddyy wweeiigghhtt wwhheenn tthhee rraattss wweerree 4455 dd oolldd ((DDMMBBAA,, lloonnggaarrrrooww)),, ffoolllloowweedd bbyy aa ddaaiillyy ii..pp.. ssaalliinnee ((DDMMBBAA ggrroouupp)) oorr hhCCGG iinnjjeeccttiioonn((DDMMBBAA��hhCCGG ggrroouupp)),, ssttaarrttiinngg wwhheenn tthhee aanniimmaallss wweerree 6655 dd oolldd..AAnniimmaallss wweerree kkiilllleedd aatt tthhee aaggeess iinn tthhee uuppppeerr lliinnee,, aass ddeessccrriibbeedd iinnMMaatteerriiaallss aanndd MMeetthhooddss.. ((RReepprriinntteedd ffrroomm SSrriivvaassttaavvaa eett aall.. [[1177]],, wwiitthhppeerrmmiissssiioonn ffrroomm OOxxffoorrdd UUnniivveerrssiittyy PPrreessss))..

HHCCGG,, MMAAMMMMAARRYY TTUUMMOORRIIGGEENNEESSIISS,, AANNDD IINNHHIIBBIINN EEXXPPRREESSSSIIOONN 1111

nate±phenol±chloroform procedure described byChomczynski and Sacchi [27] and by us [17]. TotalRNA (15±30mg) and polyadenylated RNA (3±5 mg/lane) were electrophoresed through formaldehydeagarose gels and transferred to nylon membranes(Stratagene, La Jolla, CA) followed by ultravioletcrosslinking. The rat inhibin A and inhibin Bplasmid DNA was kindly provided by Dr. Teresa K.Woodruff (Genentech, Inc., San Francisco, CA);c-fos, c-jun, c-myc, and b-actin cDNAs were pur-chased from the American Type Culture Collection(Manassas, VA). The probes were labeled with[a- 32P]dCTP with a random-primer labeling kit(GIBCO/BRL, Gaithersburg, MD). Riboprobes wereprepared for detection of inhibin A and B by in vitrotranscription of the linearized template DNA with[a- 32P]UTP and puri®ed with a Biospin column(BioRad, Richmond, CA). The nylon membrane blotwas prehybridized and hybridized as previouslydescribed [17]. Then, the membranes were exposedto a Kodak X-omat ®lm with an intensifying screenat ÿ70 �C. The human b-actin probe was used toverify load consistency, RNA integrity, and normal-ization of speci®c transcripts. The b-actin mRNA wasfound not to be affected by the hormonal andchemical carcinogen treatments used in this study(data not shown). Relative mRNA contents for eachexperimental group were determined by scanninglaser densitometry (Molecular Dynamics, Sunny-vale, CA) of autoradiographs and were expressed asan average of two separate runs.

IImmmmuunnoohhiissttoocchheemmiiccaall AAnnaallyyssiiss

Synthetic peptides were synthesized at the FoxChase Cancer Peptide Synthesis facility, and rabbitantisera directed against the inhibin a subunit(amino acids 1±16) and inhibin b subunit (aminoacids 93±104) were developed by Charles River East-Acres Biologicals (Southbridge, MA). Immunocyto-chemical procedures were performed by usingstandard methods [13]. Brie¯y, tissue sections weresequentially incubated in normal goat serum,inhibin primary antibodies, goat anti±rabbit bioti-nylated secondary antibody, and streptavidin-bio-tin-peroxidase (Vectastain ABC kit for rabbit; Vector

Laboratories, Burlingame, CA) using 3,3 0-diamino-benzidine (DAB) for color reaction. The sectionswere counterstained with hematoxylin. The nega-tive controls consisted of sections incubated withpre-immune serum or immune serum absorbed withthe corresponding inhibin peptide. Immunostain-ing was evaluated by examination of slides under anOlympus BHTU bright-®eld microscope and wasvisually graded according to the intensity of thecytoplasmic brown staining, as previously described[13].

RREESSUULLTTSS

EEffffeecctt ooff hhCCGG oonn TTuummoorr PPrrooggrreessssiioonn

None of the animals in the control or hCG groupsdeveloped mammary tumors. Microscopic lesions,namely IDP and DCIS as well as palpable tumors,were detected only in DMBA-treated animals(DMBA and DMBA�hCG groups) (Table 1). Animalsthat were injected with DMBA followed by saline

TTaabbllee 11.. EEffffeecctt ooff hhCCGG oonn TToottaall NNuummbbeerr ooff EEaarrllyy MMaammmmaarryy LLeessiioonnss aanndd IInnvvaassiivvee CCaarrcciinnoommaass IInndduucceedd bbyy DDMMBBAA

Mammary lesions �

IDP DCIS Invasive carcinomas y

Treatment Total no. No./animal Total no. No./animal Total no. No./animal

DMBA z 297 5.4 811 14.7 193 3.5DMBA�hCG§ 61 0.7 126 1.5 78 0.9

�The number of lesions found in the mammary glands of animals killed at 70, 75, 85, 105, and 125 d of age.yThe values are the number of invasive carcinomas found in tumors large enough to be detected by palpation.zThe number of lesions in a total of 55 animals at risk that were killed at ®ve different times.§The number of lesions in a total of 80 animals at risk that were killed at ®ve different times.

FFiigguurree 22.. HHiissttooggrraamm sshhoowwiinngg iinnhhiibbiittiioonn ooff aaddeennooccaarrcciinnoommaass bbyyhhCCGG ttrreeaattmmeenntt iinn rraattss.. TThhee nnuummbbeerr ooff ppaallppaabbllee iinnvvaassiivvee aaddeennooccaarrccii--nnoommaass iinn tthhee DDMMBBAA ggrroouupp rreeaacchheedd aa mmaaxxiimmuumm ooff 88..2277//aanniimmaall iinn tthhee112255--dd--oolldd aanniimmaallss,, wwhheerreeaass tthhee hhCCGG ttrreeaattmmeenntt rreedduucceedd tthheemm ttoo11..8811//aanniimmaall.. GGrroouuppss aanndd ttrreeaattmmeennttss aarree aass iinnddiiccaatteedd iinn FFiigguurree 11.. DD,,DDMMBBAA..

1122 SSRRIIVVAASSTTAAVVAA EETT AALL..

developed the highest number of microscopiclesions (Table 1). This group also developed thehighest incidence of palpable tumors, whichreached 100% in the 105-d-old animals andremained at the same percentage in the 125-d-oldanimals, although the number of tumors per animalincreased from 7.45 in the former to 8.54 in thelatter. The hormonal treatment signi®cantly re-duced the development of IDP and DCIS (Table 1and Figure 2). The incidence of palpable tumors wasalso reduced to 81.25% in both the 105- and 125-d-old animals. The number of tumors per animaldecreased to 1.93 and 1.81 tumors per animal,respectively. The percentage of adenocarcinomainhibition was highly signi®cant in those animalskilled at the ages of 105 and 125 d (Figure 2).

EEffffeecctt ooff hhCCGG TTrreeaattmmeenntt oonn MMaammmmaarryyGGllaanndd DDeevveellooppmmeenntt

The basic morphology and development of themammary gland under hCG treatment were eval-uated according to criteria previously established[1,2]. The administration of hCG alone or afterDMBA treatment increased the concentration oftype 2 and 3 lobules, while the number of the moreundifferentiated structures, such as terminal endbuds and type 1 lobules, decreased (Figure 3). Theconcentration of type 1 lobules decreased slightly inthe mammary glands of control animals as aconsequence of aging, but this value was notmodi®ed in the DMBA group (Figure 3). Type 2lobules were nonexistent in the 45-d-old animals;

FFiigguurree 33.. EEffffeecctt ooff hhCCGG oonn lloobbuullaarr ddeevveellooppmmeenntt ooff tthhee mmaammmmaarryyggllaanndd.. TTEEBB,, tteerrmmiinnaall eenndd bbuuddss;; LLoobb 11,, ttyyppee 11 lloobbuulleess;; LLoobb 22,, ttyyppee 22

lloobbuulleess;; LLoobb 33,, ttyyppee 33 lloobbuulleess..


FFiigguurree 44.. ((aa)) AAnnaallyyssiiss ooff iinnhhiibbiinn AA mmRRNNAA lleevveellss iinn tthhee mmaammmmaarryyggllaannddss ooff eeaacchh ooff tthhee ffoouurr ggrroouuppss ddeessccrriibbeedd iinn FFiigguurree 11:: ccoonnttrrooll ((cc)),,hhCCGG ((hh)),, DDMMBBAA ((DD)),, DDMMBBAA��hhCCGG ((DD��hh)).. RRaattss wweerree kkiilllleedd aatt 7700,, 7755,,8855,, 110055,, aanndd 112255 dd ooff aaggee.. bb--aaccttiinn wwaass uusseedd ttoo ddeetteecctt tthhee aammoouunntt ooffRRNNAA llooaaddeedd iinn eeaacchh llaannee.. ((bb)) CCoorrrreessppoonnddiinngg hhiissttooggrraamm ooff tthhee rreellaattiivveeiinnhhiibbiinn AA mmRRNNAA eexxpprreessssiioonn sshhoowwnn iinn ppaanneell aa.. TThhee rreellaattiivvee lleevveellss ooffttrraannssccrriippttss ffoorr eeaacchh eexxppeerriimmeennttaall ggrroouupp wweerree ddeetteerrmmiinneedd bbyy

ssccaannnniinngg llaasseerr ddeennssiittoommeettrryy ooff aauuttoorraaddiiooggrraapphhss aanndd eeqquuaalliizzeedd bbyyddeetteeccttiioonn ooff bb--aaccttiinn,, aanndd tthhee ssttaattiissttiiccaall ssiiggnnii®®ccaannccee ooff ddiiffffeerreenncceesswwaass ddeetteerrmmiinneedd bbyy SSttuuddeenntt''ss tt tteesstt.. TThhee ssiiggnnii®®ccaannttllyy ddiiffffeerreenntt ggrroouuppsswweerree ccoonnttrrooll aanndd hhCCGG--7700 dd vveerrssuuss hhCCGG--8855 dd ((PP << 00..0022)) aanndd 110055 dd((PP << 00..0033));; DDMMBBAA--7700 dd vveerrssuuss DDMMBBAA--110055 dd ((PP << 00..0022));; aanndd ccoonnttrroollaanndd DDMMBBAA��hhCCGG--7700 dd vveerrssuuss DDMMBBAA��hhCCGG--8855 dd ((PP << 00..0011)),, 110055 dd((PP << 00..000088)),, aanndd 112255 dd ((PP << 00..0011))..

FFiigguurree 55.. ((aa)) AAnnaallyyssiiss ooff iinnhhiibbiinn BB mmRRNNAA iinn tthhee mmaammmmaarryy ggllaannddssooff ccoonnttrrooll aanndd hhCCGG--,, DDMMBBAA--,, aanndd DDMMBBAA��hhCCGG--ttrreeaatteedd rraattss.. TThheeaabbbbrreevviiaattiioonnss aanndd aaggeess ooff aanniimmaallss kkiilllleedd aarree ddeessccrriibbeedd iinn FFiigguurree 44aa..((bb)) CCoorrrreessppoonnddiinngg hhiissttooggrraamm ooff tthhee rreellaattiivvee iinnhhiibbiinn BB mmRRNNAAeexxpprreessssiioonn sshhoowwnn iinn ppaanneell aa aanndd eeqquuaalliizzeedd bbyy ddeetteeccttiioonn ooff bb--aaccttiinn..

TThhee ssttaattiissttiiccaall ssiiggnnii®®ccaannccee ooff ddiiffffeerreenncceess wwaass ddeetteerrmmiinneedd bbyySSttuuddeenntt''ss tt tteesstt.. TThhee ssiiggnnii®®ccaannttllyy ddiiffffeerreenntt ggrroouuppss wweerree ccoonnttrrooll--7700 dd vveerrssuuss hhCCGG--8855 dd ((PP << 00..000000000011));; ccoonnttrrooll aanndd hhCCGG--110055 dd((PP << 00..0011));; ccoonnttrrooll aanndd DDMMBBAA��hhCCGG--7700 dd vveerrssuuss DDMMBBAA��hhCCGG--8855 dd((PP << 00..0044)),, 110055 dd ((PP << 00..0022)),, aanndd 112255 dd ((PP << 00..000000000011))..


these lobules increased progressively with agingin the control animals but remained low in theDMBA-treated group (Figure 3). In the hCG andDMBA�hCG groups, the type 2 lobules developedin a biphasic pattern, reaching a peak in the 85-d-old animals and progressively decreasing as theconcentration of type 3 lobules increased. In the125-d-old animals, the number of type 2 lobulesincreased again at the expense of the number of type3 lobules, which had regressed after hormonaltreatment was stopped (Figure 3).

IInnhhiibbiinn AA aanndd BB aanndd TTGGFF--bb mmRRNNAA EExxpprreessssiioonn

The expression of inhibin A and B was very low inthe mammary gland of control animals at all theages tested (Figures 4 and 5). Inhibin A and Bexpression in the mammary glands of hCG- andDMBA�hCG-treated animals was slightly elevatedat 5 d of treatment initiation (70-d-old group),reaching a maximum at 85 and 105 d of age; theirvalues remained elevated up to 20 d after treatment(125-d-old group) (Figures 4a and b and 5a and b).The expression of inhibin A transcript increasedfrom 1.5- to 3.5-fold (Figure 4a and b) and theexpression of inhibin B transcript from 1.5- to 3.0-fold (Figure 5a and b). DMBA treatment alone didnot modify the expression of either inhibin A orinhibin B transcripts, except for a slight increase at105 d of age (Figures 4 and 5). Both the TGF-a andTGF-b genes were abundantly expressed in themammary glands of control animals. Neither hCGnor DMBA treatment modi®ed these genes' expres-sion at any of the time points tested (data notshown).

EExxpprreessssiioonn ooff cc--mmyycc,, cc--ffooss,, aanndd cc--jjuunn

HCG treatment given alone or after DMBAinduced an increase in the expression of c-myc,which became evident as early as 5 d after initiationof treatment, with further increases by 10 and 20 dof treatment. Maximal expression was observed at20 d, when the animals were 85 d old (Figure 6b). Aslight reduction was seen by 40 d of treatment and20 d after treatment, although the values werealways signi®cantly higher than in control animals.The increase in c-myc expression was sevenfoldgreater in DMBA�hCG-treated animals than inanimals treated with hCG alone, whereas DMBAtreatment alone did not modify c-myc expression(Figure 6b).

Because maximal expression of c-myc wasobserved at 20 d of treatment, that time point wasused to assess the levels of c-fos and c-jun expression(Figure 6a and b). c-fos gene expression was notsigni®cantly affected by treatment with hCG,DMBA, or DMBA�hCG (Figure 6a and b). Theexpression of c-jun, however, was elevated in bothhCG-treated groups; the increases were twofold andthreefold in the hCG and DMBA�hCG groups,respectively (Figure 6a and b). Even though the rateof increase was higher in the group treated withDMBA�hCG than in the one treated with hCGalone, this effect could not be attributed to thecarcinogen, because DMBA treatment alone did notmodify c-jun expression (Figure 6a and b).

IImmmmuunnoohhiissttoocchheemmiiccaall SSttuuddiieess

None of the inhibin subunits tested wereexpressed in the young virgin rat mammary gland(Figure 7a), whereas hCG induced the expression ofboth inhibin a and b subunits after as few as 5 d oftreatment. The expression reached its maximalconcentration at 20 d of treatment. The immunor-eactivity was limited mainly to the cytoplasm of theepithelial cells lining the secretory alveoli; thehighest intensity of immunocytochemical reactivitywas observed around milk secretory products (Fig-ure 7b and c). DMBA treatment did not affect theproduction of inhibin a or b subunits (Figure 7d),whereas hCG given after DMBA treatment inducedthe same effect as did hCG alone (Figure 7e and f).Twenty days after hCG withdrawal, when theanimals were 125 d old, inhibin a and b subunitswere still expressed in the mammary glands oftreated animals. Immunoreactive inhibins wereconcentrated in the cytoplasm around the nuclei(Figure 7h), whereas the age-matched controlsremained negative (Figure 7g). In 125-d-old ratstreated with DMBA�hCG, the expression of inhi-bins in the mammary gland was signi®cant; therewas cytoplasmic accumulation with formation ofcoarse clumps (Figure 7j). In contrast, the reactivityobserved in the mammary glands of animals treated

FFiigguurree 66.. AAnnaallyyssiiss ooff cc--ffooss aanndd cc--jjuunn mmRRNNAA eexxpprreessssiioonn iinn tthheemmaammmmaarryy ggllaanndd.. ((aa)) CCoonnttrrooll ((llaannee 11)) aanndd hhCCGG-- ((llaannee 22)),, DDMMBBAA-- ((llaannee33)),, aanndd DDMMBBAA��hhCCGG--ttrreeaatteedd ((llaannee 44)) ggrroouuppss ooff rraattss kkiilllleedd oonn dd 2200 ooffhhCCGG ttrreeaattmmeenntt,, wwhheenn tthheeyy wweerree 8855 dd oolldd.. PPoollyy((AA))�� RRNNAA wwaass iissoollaatteeddaanndd hhyybbrriiddiizzeedd wwiitthh cc--ffooss,, cc--jjuunn,, aanndd bb--aaccttiinn.. ((bb)) HHiissttooggrraamm sshhoowwiinnggrreellaattiivvee cc--mmyycc,, cc--ffooss aanndd cc--jjuunn mmRRNNAA eexxpprreessssiioonn iinn tthhee mmaammmmaarryyggllaanndd ooff ccoonnttrrooll ((CCoonntt)) aanndd hhCCGG--,, DDMMBBAA--,, aanndd DDMMBBAA��hhCCGG((DD��hhCCGG))±±ttrreeaatteedd aanniimmaallss kkiilllleedd aatt tthhee ttiimmee iinnddiiccaatteedd iinn ppaanneell aa..RReellaattiivvee mmRRNNAA ccoonntteennttss ffoorr eeaacchh eexxppeerriimmeennttaall ggrroouupp wweerreeddeetteerrmmiinneedd iinn aauuttoorraaddiiooggrraapphhss bbyy ssccaannnniinngg llaasseerr ddeennssiittoommeettrryy aannddeeqquuaalliizzeedd bbyy ddeetteeccttiioonn ooff bb--aaccttiinn,, aanndd tthhee ssttaattiissttiiccaall ssiiggnnii®®ccaannccee ooffddiiffffeerreenncceess wwaass ddeetteerrmmiinneedd bbyy SSttuuddeenntt''ss tt tteesstt.. TThhee ggrroouuppssssiiggnnii®®ccaannttllyy ddiiffffeerreenntt ffoorr cc--jjuunn wweerree ccoonnttrrooll vveerrssuuss hhCCGG ((PP << 00..0044))aanndd ccoonnttrrooll vveerrssuuss DDMMBBAA��hhCCGG ((PP << 00..0011)).. FFoorr cc--mmyycc,, tthheeyy wweerreeccoonnttrrooll vveerrssuuss hhCCGG ((PP << 00..0044)) aanndd ccoonnttrrooll vveerrssuuss DDMMBBAA��hhCCGG((PP << 00..0022))..


with hCG alone was more evenly distributed in thecytoplasms of the epithelial cells (Figure 7h). These®ndings con®rmed at the protein level the increasesobserved in inhibin mRNAs in northern blots. Themammary gland of animals treated with DMBAalone did not express inhibins at any of the timepoints tested (Figure 7i).

The tumors in animals treated with DMBA alonedid not express inhibin immunoreactivity (Figure 8aand b). Invasive adenocarcinomas in DMBA�hCG-treated rats, on the other hand, presented a strongreaction against inhibin a and b subunits that wasexpressed in the epithelial component of the tumors(Figure 8c±f). The reaction was diffusely distributedin the cytoplasm of the cells, with a tendency toaccumulate in the apical portion of the cellssurrounding a lumen (Figure 8c±f).

DDIISSCCUUSSSSIIOONN

The results of this study revealed that hCGtreatment of rats previously inoculated with DMBAinhibited the development of early mammarylesions, i.e., IDP and DCIS, as well as invasivecarcinomas. DMBA�hCG-treated animals developedfewer and smaller tumors, which in addition showedan increased expression of inhibin. The tumorsdeveloped by animals treated with DMBA alone, onthe other hand, were larger and more numerous anddid not express this tumor-suppressor factor.

The administration of hCG alone or after DMBAalso induced lobular differentiation of the mam-mary gland that was accompanied by increasedexpression of inhibin mRNA and protein. Themammary gland of rats treated with DMBA alone,

FFiigguurree 77.. IInnhhiibbiinn iimmmmuunnoorreeaaccttiivviittyy iinn tthhee rraatt mmaammmmaarryy ggllaanndd..((aa)) 8855--dd--oolldd ccoonnttrrooll ((1166XX));; ((bb aanndd cc)) hhCCGG--ttrreeaatteedd 8855--dd--oolldd aanniimmaallss1166XX aanndd 4400XX,, rreessppeeccttiivveellyy;; ((dd)) DDMMBBAA--ttrreeaatteedd 8855--dd--oolldd aanniimmaall ((1166XX));;((ee aanndd ff)) DDMMBBAA ffoolllloowweedd bbyy hhCCGG ttrreeaattmmeenntt iinn 8855--dd--oolldd aanniimmaallss ((1166XXaanndd 4400XX,, rreessppeeccttiivveellyy));; ((gg)) 112255--dd--oolldd ccoonnttrrooll aanniimmaall ((1122XX));; ((hh)) hhCCGG--ttrreeaatteedd 112255--dd--oolldd aanniimmaall ((4400XX));; ((ii)) DDMMBBAA--ttrreeaatteedd 112255--dd--oolldd aanniimmaall

((1166XX));; ((jj)) DDMMBBAA ffoolllloowweedd bbyy hhCCGG ttrreeaattmmeenntt iinn 112255--dd--oolldd aanniimmaall((1166XX)).. PPoossiittiivvee iimmmmuunnoorreeaaccttiivviittyy wwaass ddeemmoonnssttrraatteedd bbyy tthhee yyeellllooww--bbrroowwnn ccyyttooppllaassmmiicc rreeaaccttiioonn wwiitthh DDAABB aafftteerr tthhee sseeccttiioonnss wweerreeiinnccuubbaatteedd wwiitthh rraabbbbiitt aannttii±±bb--iinnhhiibbiinn ssuubbuunniitt;; hheemmaattooxxyylliinn wwaass uusseeddaass aa ccoouunntteerrssttaaiinn..


in contrast, did not differ in their degree ofdevelopment from that of untreated controls. Theseobservations con®rm our previous results [13,17].The expression of inhibin A and B transcripts wassigni®cantly increased in the group of animals killedat the ages of 85 d (P < 0.0 and P < 0.000001,respectively), 105 d (P < 0.008 and P < 0.02, respec-tively), and 125 d (P < 0.01 and P < 0.000001, respec-tively). Interestingly, the administration of hCG toanimals previously treated with DMBA resulted in agreater increase in inhibin mRNA expression than inanimals treated with hCG alone, but the differenceswere statistically signi®cant only for the animalskilled at 85, 105, and 125 d of age. Although thereason for this phenomenon is not clear, it wasinterpreted as possibly associated with a greaterelevation in serum levels of estrogen and progester-one in the carcinogen-treated animals [28,29]. Theincreased expression of inhibin A and B mRNAs wasaccompanied by increased protein synthesis, whichwas detected by immunohistochemical procedures.The positive immunocytochemical reaction withinhibin antibodies in the mammary epithelium wasobserved not only during the administration of thehormone but also after cessation of treatment, eventhough the lobular structures showed regression at amorphological level. Because increased expressionof inhibin preceded lobular formation, the presence

of inhibin could be considered a good marker ofearly differentiation. It could also be an indicator ofprevious hormonal effects, because it was stillpresent in the mammary gland after the cessationof hormonal treatment, at times when the morpho-logical phenotype had totally disappeared. Inhibinswere not detected in the mammary glands ofcontrol and DMBA-treated animals or in the tumorsinduced by this carcinogen. In addition to reducingthe number of IDP, DCIS, and palpable invasivecarcinomas, treatment with hCG after DMBAresulted in the immunocytochemical expression ofinhibin in the tumors in the treated animals. Intumors in DMBA�hCG-treated animals, inhibinwas localized in the cytoplasm of the cells, in adistribution similar to that observed in normalmammary epithelium, although the reaction wasmore intense in the tumoral tissue. Its cytoplasmiclocalization indicated that inhibin is a secretoryproduct that is speci®cally induced by hCG. Theabsence of reactivity in tumors developed byanimals treated with DMBA alone indicated thatthe synthesis of inhibin by the neoplastic epithe-lium was speci®cally induced by this hormone. Theproduction of inhibin by the tumors could beconsidered, therefore, to be a marker of differentia-tion induced by hCG, a postulate supported by theoverall characteristics of the tumors, i.e., smaller

FFiigguurree 88.. ((aa aanndd bb)) PPaappiillllaarryy--ccrriibbrriiffoorrmm aaddeennooccaarrcciinnoommaa rreemmoovveeddffrroomm aa DDMMBBAA--ttrreeaatteedd 8855--dd--oolldd aanniimmaall aanndd iimmmmuunnoohhiissttoocchheemmiiccaallllyyrreeaacctteedd wwiitthh pprree--iimmmmuunnee rraabbbbiitt sseerruumm ((aa)) aanndd rraabbbbiitt aannttii±±aa--iinnhhiibbiinn((bb)) ((1100XX)).. ((cc)) WWeellll--ddiiffffeerreennttiiaatteedd aaddeennooccaarrcciinnoommaa ooff aa DDMMBBAA��hhCCGG--ttrreeaatteedd aanniimmaall eexxcciisseedd aatt dd 2200 ooff hhoorrmmoonnaall aaddmmiinniissttrraattiioonn ((8855--dd--oollddaanniimmaall)).. ((dd aanndd ee)) SSeeccttiioonnss ooff aaddeennooccaarrcciinnoommaa eexxcciisseedd 2200 dd aafftteerr

cceessssaattiioonn ooff hhCCGG ttrreeaattmmeenntt ((112255--dd--oolldd aanniimmaall)) aanndd iimmmmuunnoorreeaacctteeddwwiitthh rraabbbbiitt aannttii±±aa--iinnhhiibbiinn aannttiibbooddyy.. ((ff)) SSeeccttiioonnss iimmmmuunnoorreeaacctteedd wwiitthhrraabbbbiitt aannttii±±bb--iinnhhiibbiinn aannttiibbooddyy ((4400XX)).. IInn aallll sseeccttiioonnss,, iimmmmuunnoossttaaiinniinnggwwaass vviissuuaalliizzeedd wwiitthh DDAABB,, tthhee hheemmaattooxxyylliinn wwaass uusseedd aass aa ccoouunntteerr--ssttaaiinn..


size, more differentiated appearance, and activationof apoptotic genes [17]. However, the possibilitythat inhibin might exert an inhibitory effect on thegrowth of neighboring cells through a paracrinemechanism needs to be ruled out [12]. The synthesisof inhibin in tumors developed by DMBA�hCG-treated animals may also indicate that the tumor-suppressing properties described for this protein [16]may be linked to the protective effect of hCG inmammary carcinogenesis. However, the fact thatthe animals used in these studies were intact andthat their ovaries responded with an increasedsecretion of estrogen and progesterone does notallow us to rule out that these hormones might beinvolved in these processes, as other authors haveindicated [30]. Further studies are required to dissectmore clearly the ultimate mechanisms involvedin the inhibition of tumor development by hCGtreatment.

An increase in the levels of inhibin a and bsubunits has been also observed in the immortalizedhuman breast epithelial cell line MCF-10F and inthe breast cancer cell line MCF-7 after hCG treat-ment, accompanied by a dose-dependent inhibitionof cell growth [14]. Although the gonads areconsidered to be the primary source of circulatinginhibin, our ®ndings indicated that the mammaryepithelium is an additional extragonadal source ofinhibin synthesis [16,31±34], thus suggesting adiverse function of inhibins outside the reproduc-tive system.

Inhibins are structurally related to the TGF-bsuperfamily of proteins [35], and enhanced expres-sion of inhibins has been reported in regressinghuman breast cancer cells after hormone with-drawal [36]. Here we did not ®nd any change inthe expression of the TGF-a and TGF-b genes afterany period of treatment, which indicated that hCGspeci®cally stimulated inhibin gene expression. Theinhibition of mammary tumor development byhCG, which was accompanied by increased inhibinsynthesis, acquires greater signi®cance with the®nding that inhibin has been demonstrated to bea critical regulator of gonadal stromal cell prolifera-tion and the ®rst secreted protein identi®ed to havetumor suppressor activity [16]. Administration ofhCG either alone or after DMBA signi®cantlyincreased c-myc and c-jun mRNA expression,whereas c-fos expression was not affected by thehormonal treatment. These ®ndings are in agree-ment with the reported association between induc-tion of terminal differentiation and increased levelsof c-jun mRNA [37,38] and are further supported bythe absence of changes in gene expression in themammary glands and tumors [17] of animals treatedwith DMBA alone. Nevertheless, the expression of c-jun, as well as of other genes such as c-fos and c-myc,is also required for cells to transit G1 phase andinitiate DNA synthesis [38±40], indicating that

these genes function as components of a relativelycommon signal transduction cascade that operatesin several different pathways. Our observation thathCG treatment resulted in the induction of thesegenes and simultaneous induction of inhibin pro-duction led us to speculate that the hormonaltreatment causes the interaction of c-jun with thetranscription factor AP-1, which might ®nallyregulate the synthesis of inhibin. We are pursuingthese studies.

Chorionic gonadotropin has a direct effect on theovary, acting through a luteinizing hormone/hCGreceptor that is present in luteal, granulosa, theca,and stromal cells [41]. On binding to the luteinizinghormone/hCG receptor, hCG stimulates the pro-duction of estrogen and progesterone by the corpusluteum through a mechanism involving the pro-duction of cyclic AMP [41,42]. The possibility thatthe effects of hCG on the mammary gland arereceptor-mediated is supported by our observationof a direct effect of this hormone on human breastepithelial cells in vitro [43] and the demonstrationby others of the presence of a receptor in the humanbreast and in human breast carcinoma cells [44,45].Thus, the effects of hCG on the mammary gland aremultiple and involve different pathways. However,the ultimate mechanisms through which thishormonal treatment induces mammary gland dif-ferentiation, gene activation, and inhibition oftumor progression are just starting to be unraveled.Our study indicated that the production of inhibinis associated with the growth and differentiation ofthe mammary gland and with tumor regression.

AACCKKNNOOWWLLEEDDGGMMEENNTTSS

This work was supported by grant CA 64896 fromthe National Cancer Institute, National Institutes ofHealth.

RREEFFEERREENNCCEESS

1. Russo J, Russo IH. In¯uence of differentiation and cellkinetics on the susceptibility of the mammary gland tocarcinogenesis. Cancer Res 1980;40:2677±2687.

2. Russo J, Russo IH. Toward a physiological approach to breastcancer prevention. Cancer Epidemiol Biomarkers Prev1994;3:353±364.

3. Russo IH, Koszalka M, Gimotty PA, Russo J. Protective effectof chorionic gonadotropin on DMBA-induced mammarycarcinogenesis. Br J Cancer 1990;62:243±247.

4. Russo IH, Koszalka M, Russo J. Human chorionic gonado-tropin and rat mammary cancer prevention. J Natl CancerInst 1990;82:1286±1289.

5. Russo IH, Russo J. Chorionic gonadotropin: A tumoristaticand preventive agent in breast cancer. In: Teicher BA, editor.Drug resistance in oncology. New York: Marcel Dekker Inc;1992. p 537±560.

6. Russo IH, Koszalka M, Russo J. Effect of human chorionicgonadotropin on mammary gland differentiation andcarcinogenesis. Carcinogenesis 1990;11:1849±1855.

7. Grubbs CJ, Juliana MM, Hill DL. Suppression by pregnancyof chemically induced preneoplastic cells of rat mammarygland. Anticancer Res 1986;6:1395±1400.


8. Sinha DK, Pazik JE. Tumorigenesis of mammary glandby 7,12-dimethylbenz[a]anthracene during pregnancy:Relationship with DNA synthesis. Int J Cancer 1981;27:807±810.

9. Simpson ER, MacDonnald PC. Endocrine physiology of theplacenta. Annu Rev Physiol 1981;43:163±188.

10. Russo J, Tay LK, Russo IH. Differentiation of the mammarygland and susceptibility to carcinogenesis. Breast CancerRes Treat 1982;2:5±73.

11. Russo IH, Russo J. Mammary gland neoplasia in long termrodent studies. Environ Health Perspect 1996;104:938±967.

12. Russo IH, Russo J. Role of hCG and inhibin in breast cancer.Int J Oncol 1994;4:297±306.

13. Alvarado MV, Ho T-Y, Russo J, Russo IH. Human chorionicgonadotropin regulates the synthesis of inhibin in theovary and mammary gland of rat. Endocrine 1994;2:1107±1114.

14. Alvarado MV, Alvarado NE, Russo J, Russo IH. Humanchorionic gonadotropin inhibits proliferation and inducesexpression of inhibin in human breast epithelial cells. In Vitro1994;30A:4±8.

15. Alvarado MV, Russo J, Russo IH. Immunolocalization ofinhibin in the mammary gland of rats treated with humanchorionic gonadotropin (hCG). J Histochem Cytochem1993;41:29±34.

16. Matzuk MM, Finegold MJ, Su J-G, Hsueh AJW, Bradley E.Inhibin is a tumor-suppressor gene with gonadal speci®cityin mice. Nature 1992;360:313±319.

17. Srivastava P, Russo J, Russo IH. Chorionic gonadotropininhibits rat mammary carcinogenesis through activationof programmed cell death. Carcinogenesis 1997;18:1799±1808.

18. De Jong FH. Inhibin. Physiol Rev 1988;68:555±607.19. De Jong FH, Sharpe RM. Evidence for inhibin-like activity in

bovine follicular ¯uid. Nature 1976;263:71±72.20. Vale W, Hsueh A, Rivier C, Yu J. The inhibin/activin family of

hormones and growth factors. Handbook of ExperimentalPharmacology 1990;95:211±248.

21. Robertson DM. Transforming growth factor-b/inhibinfamily. Baillieres Clin Endocrinol Metab 1991;5:615±634.

22. Woodruff TK, Mayo K. Regulation of inhibin synthesis in therat ovary. Annu Rev Physiol 1990;52:807±821.

23. Ransone LJ, Verman IM. Nuclear proto-oncogenes fos andjun. Annual Review of Cell Biology 1990;6:539±557.

24. Curran T, Morgan JI. Memories of fos. BioEssays 1990;7:255±258.

25. Lau LF, Nathan D. Expression of a set of growth-relatedimmediate early genes in Balb/c 3T3 cells. Coordinateregulation with c-fos and c-myc. Proc Natl Acad Sci USA1987;84:1182±1186.

26. Russo J, Russo IH, van Zwieten MJ, Rogers AE, Gusterson B.Classi®cation of neoplastic and non-neoplastic lesions of therat mammary gland. In: Jones TC, Mohr U, Hunt RD, editors.Integument and mammary glands of laboratory animals.Berlin: Springer Verlag; 1989. p 275±304.

27. Chomczynski P, Sacchi N. Single step method of RNAisolation by acid guanidinium thiocyanate-phenol-chloro-form extraction. Anal Biochem 1987;162:156±159.

28. Tahin Q, Mgbonyebi OP, Russo J, Russo IH. In¯uence ofhormonal change induced by the placental hormonechorionic gonadotropin on the progression of mammarytumorigenesis Abstract. Proceedings of the AmericanAssociation for Cancer Research 1996;37:1622.

29. Russo J, Russo IH. Human chorionic gonadotropin in breastcancer prevention. In: Ethier S, editor. Endocrine oncology.Totowa, NJ: Humana Press Inc., in press.

30. Nandi S, Guzman RC, Yand J. Hormones and mammarycarcinogenesis in mice, rats, and humans: A unifyinghypothesis. Proc Natl Acad Sci USA 1995;92:3650±3657.

31. Petraglia F, Sawchenko P, Lim ATW, Rivier J, Vale W.Localization, secretion, and action of inhibin in humanplacenta. Science 1987;237:187±192.

32. Petraglia F, Garuti GC, Calza L, et al. Inhibin subunits inhuman placenta: Localization and messenger ribonucleicacid levels during pregnancy. Am J Obst Gynecol 1991;165:750±758.

33. Meunier H, Rivier C, Evans R, Vale W. Gonadal andextragonadal expression of inhibin A and B subunits invarious tissues predicts diverse functions. Proc Natl Acad SciUSA 1988;85:247±251.

34. Haji M, Nishi Y, Tanaka S, et al. Evidence for the secretion ofinhibin-like immunoreactivity from cultured human adrenalcells. J Endocrinol 1991;128:13±16.

35. Blauer M, Kohonen J, Leivonen I, Vilja P, Touhimaa P. Inhibinalpha and beta A subunit immunoreactivity in the chickenembryo during morphogenesis. Development 1992;115:1079±1086.

36. Kyprianou N, English HF, Davidson NE, Issacs JT. Pro-grammed cell death during regression of MCF7 humanbreast cancer following estrogen ablation. Cancer Res1991;51:162±166.

37. Miller Jr WH, Dmitrovsky E. Retinoic acid audits rearrangedreceptor in the treatment of acute promyelocytic leukemia.In: De Vita V, Hellman S, Rosenberg, SA, editors. Importantadvances in oncology. Philadelphia: J. B. Lippincott Co;1993. p 81±90.

38. Iwai SA, Kosaka M, Nishinu Y, Sumi T, Sakuda M, NishimuneY. Changes in Hox 1.6, c-jun and Oct-3 gene expressions areassociated with teratocarcinoma F9 cell differentiation inthree different ways of induction. Exp Cell Res 1993;205:39±43.

39. Draetta G. Cell cycle control in eukaryotes. Trends BiochemSci 1990;15:378±383.

40. Evan GI, Littlewood TD. The role of c-myc in cell growth.Current Opin Genet Dev 1993;3:44±49.

41. Hunzicker-Dunn M, Birnbauer L. The involvement ofadenylyl cyclase and cyclic AMP-dependent protein kinasesin luteinizing hormone action. In: Ascoli M, editor. Luteiniz-ing hormone action and receptors. Boca Raton, FL: CRCPress, Inc; 1985. p 57±134.

42. Wang Z, Liu X, Ascoli M. Phosphorylation of the lutropin/choriogonadotropin receptor facilitates uncoupling of thereceptor from adenylyl cyclase and endocytosis of thebound hormone. Mol Endocrinol 1997;11:183±192.

43. Srivastava P, Russo J, Mgbonyebi OP, Russo IH. Growthinhibition and activation of apoptotic gene expression byhuman chorionic gonadotropin in human breast epithelialcells. Anticancer Res 1998;18:4003±4010.

44. Meduri G, Charnaux N, Loosfelt H, Jolivet A, Spyratos F,Brailly S, Milgrom E. Luteinizing hormone/human chorionicgonadotropin receptors in breast cancer. Cancer Res 1997;57:857±864.

45. Lojun S, Bao S, Lei ZM, Rao CV. Presence of functionalluteinizing hormone/chorionic gonadotropin (hCG) recep-tors in human breast cell lines: Implications supporting thepremise that hCG protects women against breast cancer.Biol Reprod 1997;57:1202±1210.


inhibition of rat mammary tumorigenesis by human chorionic gonadotropin associated with increased...

Documents