Mechanisms of Hormonal Prevention of Breast Cancer

13
23 Mechanisms of Hormonal Prevention of Breast Cancer DANIEL MEDINA, a LAKSHMI SIVARAMAN, a SUSAN G. HILSENBECK, a,b ORLA CONNEELY, a MELANIE GINGER, a JEFFREY ROSEN, a AND BERT W. O’MALLEY a a Department of Molecular and Cellular Biology, b Department of Medicine-Breast Center, Baylor College of Medicine, Houston, Texas 77030, USA ABSTRACT: Reproductive history is a consistent risk factor for human breast cancer. Epidemiological studies have repeatedly demonstrated that early age of first pregnancy is a strong protective factor against breast cancer and pro- vides a physiologically operative model to achieve a practical mode of preven- tion. In rodents, the effects of full-term pregnancy can be mimicked by a three- week exposure to low doses of estrogen and progesterone. Neither hormone alone is sufficient to induce protection. The cellular and molecular mecha- nisms that underlie hormone-induced refractoriness are largely unresolved. Our recent studies have demonstrated that an early cellular response that is altered in hormone-treated mammary cells is the initial proliferative burst induced by the chemical carcinogen methylnitrosourea. The decrease in prolif- eration is also accompanied by a decrease in the ability of estrogen receptor- positive cells to proliferate. RNA expression of several mammary cell-cycle– related genes is not altered in hormone-treated mice; however, immunohis- tochemical assays demonstrate that the protein level and nuclear compartmen- talization of the p53 tumor suppressor gene are markedly upregulated as a consequence of hormone treatment. These results support the hypothesis that hormone stimulation, at a critical period in mammary development, results in cells with persistent changes in the intracellular regulatory loops governing proliferation and response to DNA damage. A corollary to this hypothesis is that the genes affected by estrogen and progesterone are independent of alve- olar differentiation-specific genes. Suppressive subtractive hybridization-PCR methods have identified several genes that are differentially expressed as a consequence of prior estrogen and progesterone treatment. Future experi- ments are aimed at determining the mechanisms of hormone-induced upregu- lation of p53 protein expression as part of the overall goal of identifying and functionally characterizing the genes responsible for the refractory phenotype. KEYWORDS: rat; mammary gland; estrogen; progesterone; estrogen receptor; cell proliferation; p53; gene expression; subtractive hybridization Address for correspondence: Dr. Daniel Medina, Department of Molecular and Cellular Biology, Baylor College of Medicine, Houston, TX 77030. Voice: 713-798-4483; fax: 713- 790-0545. [email protected]

Transcript of Mechanisms of Hormonal Prevention of Breast Cancer

Page 1: Mechanisms of Hormonal Prevention of Breast Cancer

23

Mechanisms of Hormonal Prevention of Breast Cancer

DANIEL MEDINA,

a

LAKSHMI SIVARAMAN,

a

SUSAN G. HILSENBECK,

a,b

ORLA CONNEELY,

a

MELANIE GINGER,

a

JEFFREY ROSEN,

a

AND BERT W. O’MALLEY

a

a

Department of Molecular and Cellular Biology,

b

Department of Medicine-Breast Center, Baylor College of Medicine, Houston, Texas 77030, USA

A

BSTRACT

: Reproductive history is a consistent risk factor for human breastcancer. Epidemiological studies have repeatedly demonstrated that early ageof first pregnancy is a strong protective factor against breast cancer and pro-vides a physiologically operative model to achieve a practical mode of preven-tion. In rodents, the effects of full-term pregnancy can be mimicked by a three-week exposure to low doses of estrogen and progesterone. Neither hormonealone is sufficient to induce protection. The cellular and molecular mecha-nisms that underlie hormone-induced refractoriness are largely unresolved.Our recent studies have demonstrated that an early cellular response that isaltered in hormone-treated mammary cells is the initial proliferative burstinduced by the chemical carcinogen methylnitrosourea. The decrease in prolif-eration is also accompanied by a decrease in the ability of estrogen receptor-positive cells to proliferate. RNA expression of several mammary cell-cycle–related genes is not altered in hormone-treated mice; however, immunohis-tochemical assays demonstrate that the protein level and nuclear compartmen-talization of the p53 tumor suppressor gene are markedly upregulated as aconsequence of hormone treatment. These results support the hypothesis thathormone stimulation, at a critical period in mammary development, results incells with persistent changes in the intracellular regulatory loops governingproliferation and response to DNA damage. A corollary to this hypothesis isthat the genes affected by estrogen and progesterone are independent of alve-olar differentiation-specific genes. Suppressive subtractive hybridization-PCRmethods have identified several genes that are differentially expressed as aconsequence of prior estrogen and progesterone treatment. Future experi-ments are aimed at determining the mechanisms of hormone-induced upregu-lation of p53 protein expression as part of the overall goal of identifying andfunctionally characterizing the genes responsible for the refractory phenotype.

K

EYWORDS

: rat; mammary gland; estrogen; progesterone; estrogen receptor;cell proliferation; p53; gene expression; subtractive hybridization

Address for correspondence: Dr. Daniel Medina, Department of Molecular and CellularBiology, Baylor College of Medicine, Houston, TX 77030. Voice: 713-798-4483; fax: 713-790-0545.

[email protected]

Page 2: Mechanisms of Hormonal Prevention of Breast Cancer

24 ANNALS NEW YORK ACADEMY OF SCIENCES

INTRODUCTION

Reproductive history is a consistent risk factor for human breast cancer.

1,2

Whereas many aspects of reproductive history confer increased risk, early age offirst pregnancy is a strong protective factor. The protective effect of early first preg-nancy has been repeatedly demonstrated in numerous epidemiological studies andprovides a physiologically rational model to achieve practical prevention of breastcancer in humans.

3–5

In rodents, as in humans, a full-term pregnancy markedly inhibits subsequentchemical carcinogen-induced mammary tumorigenesis. This result has been demon-strated by numerous investigators over the span of the last 30 years in both rats

6–15

and mice

16

(T

ABLE

1). The protective effect of pregnancy was observed even whenthe carcinogen was administered 100–130 days after the first parturition, indicatinga long-lasting alteration in the state of sensitivity of the mammary gland.

6

The pro-tective effects of pregnancy can be mimicked by administration of estrogen andprogesterone.

14,15

The experiments have been performed in outbred Sprague-Dawley rats, inbred Lewis and Wistar-Furth rats and inbred mice; these have exten-sively validated the rodent model for the human paradigm.

T

ABLE

1. Summary of pertinent experiments that demonstrate the protective effect ofpregnancy hormones on mammary tumorigenesis

Tumorigenesis (Control/Treated)

Hormone Treatment Carcinogen (days)

a

a

Interval in days between parturition (or weaning) and treatment with chemical carcinogen.

b

Rats were 180–190 days when administered carcinogen.

c

MC, mammary carcinoma; MC incidence in control rats/MC incidence in treated rats.

d

All experiments used rats except this experiment, which used two different mouse strains

MC Incidence

c

% Inhibition Ref.

Pregnancy DMBA (21 d) 71/0 100 6

Pregnancy/Lact. DMBA (50 d) 71/11 84 6

Pregnancy DMBA (21 d) 48/6 88 7

Pregnancy DMBA (15 d) 70/14 80 8

E+P (20

µ

g/4 mg) MNU (21 d) 61/13 79 9

E+P (5

µ

g/2 mg) MNU (21 d) 61/19 69 14

hCG (10 Iu) DMBA (21 d) 44/18 52 10

Pregnancy/Lact. DMBA (100 d)

b

38/16 58 11

Pregnancy DMBA (130 d)

b

38/18 53 11

Pregnancy MNU 80/17 79 12

E+P (30 mg/30 mg) MNU 55/0 100 13

E+P (20

µ

g/20 mg) MNU (28 d) 56/10 82 15

Pregnancy

d

DMBA (15 d) 70/25 64 16

Page 3: Mechanisms of Hormonal Prevention of Breast Cancer

25MEDINA

et al.

: HORMONAL PREVENTION OF CANCER

Despite the extensive data on the tumor biology of hormone-induced refractori-ness to breast cancer, the cellular and molecular mechanisms that underlie hormone-induced refractoriness are unknown. The results reported herein address this issue.

ANIMAL MODEL

All experiments were performed using the same basic protocol (see F

IGURE

1).Female Wistar-Furth (WF) rats, 35 days old, were purchased from Harlan Sprague-Dawley, Inc. (Chicago, IL). The WF female rat is an inbred strain of rats sensitive toboth MNU- and DMBA-induced mammary carcinogenesis.

15,17

Forty-five-day-oldfemale rats which were cycling regularly were administered a primary dose of 0.1ml solution of 2.5

µ

g estradiol benzoate (EB) subcutaneously. The animals weredivided into two groups; one group was maintained as age-matched virgins and thesecond group was hormone-stimulated, either via a normal full-term pregnancy orby an estrogen-progesterone pellet (20

µ

g E/20mg P) for 21 days. The rats experi-enced a full 28-day regression before being treated with one or two doses of MNU(50mg/kg body weight, i.p.).

TUMORIGENESIS

Hormonal stimulation of Wistar-Furth female rats results in a resistant state tosubsequent MNU-induced mammary tumorigenesis (see F

IGURE

2). After pregnancyand mammary gland involution, 119- to 122-day-old parous rats and age-matchedvirgins received two doses of MNU (50mg/kg body weight) one week apart. At the

FIGURE 1. Experimental paradigm for studying the protective effect of estrogen andprogesterone treatment. Forty-five-day-old rats were administered estradiol benzoate(E2B) by s.c. injection; and then a subcutaneous pellet of estrogen and progesterone (E/P)3 days later. The pellets were replaced after 10 days to provide hormone treatment for a totalof 21 days. After 21 days the pellet was removed and the mammary gland allowed to returnto a resting state approximating a 28-day cycle of involution. At completion of 28 days of“involution” mammary glands are removed for investigation.

Page 4: Mechanisms of Hormonal Prevention of Breast Cancer

26 ANNALS NEW YORK ACADEMY OF SCIENCES

end of 200 days post-carcinogen treatment, 9/20 (45%) of the virgin rats developedadenocarcinomas whereas 0/20 (0%) of the parous rats had adenocarcinomas. Thus,parity conferred resistance to chemical carcinogen–induced mammary tumors in WFrats analogous to that previously observed in SD and Fischer 344 rats. In the virginrats age-matched to the hormone pellet-containing rats, mammary tumor incidencewas 56% (26/46) with a mean tumor latent period of 118 days. Rats which receivedthe high dose E/P pellets exhibited a tumor incidence of only 10% (2/20) with amean latent period of 164 days (

p

<

0.05). Rats which received the medium- andlow-dose E/P pellets exhibited a significant reduction in tumor incidence of only30% (6/20, 6/20) with mean tumor latent periods of 159 and 169 days, respectively(

p

<

0.05) (F

IG

. 2). The apparent growth rate of tumors was significantly slower inthe E/P treated rats compared with the age-matched virgins. The tumors were mea-sured by vernier caliper for a period of up to six weeks. At the third measurement,the mean diameter was 16 mm for 23 tumors arising in age-matched control rats ver-sus only 10 mm for nine tumors arising in the rats which had received the medium-and low-dose hormone pellets. A positive control of virgin WF rats administeredMNU at 55 days was also included. These animals responded to MNU with a 95%(21/22) mammary cancer incidence with a mean tumor latency of 118 days. Addi-tionally, multiple mammary tumors were frequently observed in these rats (35 pal-pable tumors in 21 rats) in contrast with the mature age-matched virgins where only27 tumors in 23 tumor-bearing rats were noted. Administration of either 20

µ

gE or20 mgP alone for 21 days followed by MNU treatment did not induce a protectiveeffect.

FIGURE 2. The percent mammary tumors are plotted for different treatment groups.AMV, age-matched virgin (to parous rats); AMV2, age-matched virgin to estrogen/proges-terone-treated groups; HH, high hormone concentration (20 µgE/20 mgP); MH, mediumhormone concentration (10 µgE/10 mgP); LH, low hormone concentration (5 µgE/5 mgP);YV, young (50-day-old) virgin rats.

Page 5: Mechanisms of Hormonal Prevention of Breast Cancer

27MEDINA

et al.

: HORMONAL PREVENTION OF CANCER

CELLULAR MECHANISMS

Both cell proliferation and cell death are coordinately regulated during normalmammary gland morphogenesis and involution. The high levels of proliferation thatoccur in the large number of terminal end buds (TEBs) in the virgin gland are corre-lated with increased sensitivity to carcinogenesis. A differential regulation of theseprocesses in the parous/hormone-treated gland and age-matched virgin glands couldexplain the resistance of the gland to carcinogenesis. Apoptosis and cell proliferationlevels were examined in the parous-involuted gland, the E/P-treated involuted glandand the AMV gland at the time of carcinogen treatment and at intervals up to 40 daysafter carcinogen treatment.

In these experiments we examined the distal and proximal portions of the glandseparately because of the regional differences in morphology. F

IGURE

3 shows onlythe data for the distal portions of the mammary gland since marked differences inproliferation were not observed in the proximal portions of the gland. The BrdUlabeling index (LI) at 97 days (day 0 in F

IGURE

3) in the distal portions of the glandwas 0.8% in the parous-involuted gland and 1.8% in the AMV (

p

<

0.04). Therefore,the proliferation at 97 days was low in both groups of animals. However, at day 8post-MNU treatment, the LI was still low in the parous-involuted gland (1.3%) andthe E/P-treated involuted gland (1.5%), but significantly higher in the AMV (5.7%,

p

<

0.01). The BrdU-LI decreased to baseline levels (1–2%) in all groups at 20 daysafter carcinogen treatment.

Among all groups (distal versus proximal, hormone-treated versus AMV, days 0,3, 8) there were no differences in apoptotic indices with respect to time after MNU

FIGURE 3. Cell proliferation in virgin and hormone-treated glands before and afterMNU. AMVD, age-matched virgin-distal; PARD, parous-distal; EPD, estrogen/progesterone-treated-distal.

Page 6: Mechanisms of Hormonal Prevention of Breast Cancer

28 ANNALS NEW YORK ACADEMY OF SCIENCES

treatment, either in distal or proximal portions of the gland or in age-matched virginversus hormone-treated glands. The amount of apoptosis was uniformly low witha mode of 0.5%. These data suggest that hormones induce a block to MNU-inducedproliferation and perhaps, a reduction in steady-state proliferation of the post-involuted untreated gland.

As proliferation is regulated by the levels of estrogen and progesterone, we exam-ined the status of estrogen and progesterone receptors as well as BrdU-labeled cellsin the rat mammary gland using a dual-labeling method with an epifluorescencemicroscope. The analysis was performed on paraffin-embedded tissue using anantigen-retrieval protocol in glands derived from 45-day virgin, 96-day AMV, 96-day E/P-treated, 3-day and 7-day carcinogen-treated AMV and E/P rats. We askedtwo questions: (1) Is the distribution and percentage of ER/PR positive cells differentin AMV versus involuted gland before and after carcinogen-treatment? (2) Does thefrequency of ER/BrdU dual-labeled cells change in the two groups before and aftercarcinogen-treatment? The results are illustrated in T

ABLE

2 and F

IGURE

4.The percentage of ER

+

cells was similar between any two matched groups(T

ABLE

2) and did not change significantly from day 96 to 7 days post-carcinogentreatment. These data and those of Yang

et al

.

18

provide little support for the concept

FIGURE 4. Graphical representation of ER, BrdU labeling. The large boxes representthe total epithelial cells counted. The hatched area represents BrdU-negative, the solid arearepresents BrdU-positive. The respective area (hatched or solid) is directly proportional tothe percent of cells for that character. Thus, the solid area in the ER+ cells in the 45-day ratrepresents 1.4% of all cells counted in that group. The rat mammary gland has four popu-lations of epithelial cells, namely, ER+ proliferating and non-proliferating populations andER− proliferating and non-proliferating populations. A focus on the solid boxes (BrdU+)shows the relative changes in ER+ proliferating cells with gland development and with car-cinogen treatment.

Page 7: Mechanisms of Hormonal Prevention of Breast Cancer

29MEDINA

et al.

: HORMONAL PREVENTION OF CANCER

T

AB

LE

2.

Coe

xpre

ssio

n of

ER

and

pro

life

rati

on in

the

nor

mal

vir

gin

rat

mam

mar

y gl

and

and

afte

r ca

rcin

ogen

cha

llen

ge

a

At

leas

t fi

ve f

ield

s w

ere

capt

ured

per

tis

sue

sect

ion,

at

leas

t on

e pr

olif

erat

ing

cell

was

inc

lude

d in

eac

h fi

eld

of c

aptu

re,

and

at l

east

fiv

e an

imal

s w

ere

incl

uded

in

each

gro

up.

b

Exp

ecte

d is

the

per

cent

age

of d

oubl

e-la

bele

d ce

lls

that

wou

ld b

e ex

pect

ed (

E)

if t

he t

wo

vari

able

s w

ere

inde

pend

ent

and

was

cal

cula

ted

by m

ulti

plyi

ngth

e pe

rcen

tage

of

ER

+

and

Brd

U

+

cel

ls a

nd t

hen

divi

ding

by

100

for

each

sam

ple.

c

Obs

erve

d/ex

pect

ed (

O/E

) is

the

actu

al n

umbe

r of

dou

ble-

labe

led

cell

s co

unte

d/ex

pect

ed, a

s ca

lcul

ated

abo

ve.

d

O/E

giv

es a

rat

io i

ndic

atio

n of

whe

ther

rec

epto

r ex

pres

sion

and

pro

life

rati

on a

re n

egat

ive

or p

osit

ivel

y as

soci

ated

wit

h ea

ch o

ther

and

the

str

engt

h of

the

asso

ciat

ion.

In

the

form

er c

ase,

val

ues

less

tha

n 1

are

expe

cted

, and

in

the

latt

er c

ase,

val

ues

grea

ter

than

1 a

re e

xpec

ted.

e

p

val

ues

for

χ

2

tes

t fo

r in

depe

nden

ce.

Gro

up

a

# B

rdU

/%#

ER

/%#

doub

le/%

# D

AP

1E

xpec

ted

b

O/E

c

Ass

ocia

tion

d

p

val

ue

e

45d

AM

V (

enti

re)

176

(12.

1)46

3 (3

1.2)

20 (

1.4)

1452

3.8

0.36

neg

0.00

1

45d

(AM

V)

(TE

B)

136

(18.

2)32

3 (4

3.2)

42 (

5.6)

747

7.86

0.71

neg

0.00

1

96d

AM

V59

(4.

69)

425

(33.

8)31

(2.

47)

1256

1.58

1.56

pos

0.00

1

96d

E/P

15 (

0.98

)50

3 (3

2.9)

4 (0

.26)

1529

0.32

0.81

neg

0.61

96d

AM

V (

3d

MN

U)

77 (

2.5)

647

(21.

3)32

(1.

05)

3036

0.53

1.98

pos

0.00

01

96d

E/P

(3

d M

NU

)21

(1.

3)39

7 (2

4.9)

1 (0

.06)

1594

0.32

0.19

neg

0.03

96d

AM

V (

7d

MN

U)

174

(4.3

4)11

32 (

28.2

)60

(1.

5)40

131.

221.

23po

s0.

06

96d

E/P

(7

d M

NU

)39

(1.

66)

685

(29.

1)8

(0.3

4)23

540.

480.

71ne

g0.

23

Page 8: Mechanisms of Hormonal Prevention of Breast Cancer

30 ANNALS NEW YORK ACADEMY OF SCIENCES

that there is a significant difference in ER

+

cells in the parous-involuted gland com-pared to the AMV gland. Of more interest is the result that demonstrates the prolif-erating cells in the mature gland are ER

+

, a situation diametrically opposite fromthat occurring in the immature gland (compare 96 AMV with 45 day AMV). Thisresult indicates that ER

+

proliferating cells are already present in the adult virginmammary gland and form the majority of the proliferating cells. Furthermore, itwould appear that hormone stimulation prior to carcinogen exposure abolishes theco-localization of ER and proliferation. Finally, the results demonstrate that in theMNU-treated AMV gland, ER-positivity and BrdU proliferation are positively asso-ciated (i.e., the ER

+

cell is proliferating in proportion, sometimes greater, to theirpopulation frequency). In contrast, in the MNU-treated E/P gland, the ER

+

cell isgrowth arrested. Thus, these results extend our earlier studies and suggests that it isthe ER

+

cell that is growth arrested as a consequence of E/P treatment and strength-ens our view that some of the genes specifically affected by hormone treatmentinvolve growth-regulatory genes, (e.g., p53, Rb-related).

The studies on cell proliferation and hormone-receptor positivity have helped togenerate the following hypothesis. In the mammary glands of the early post-pubescent female, hormones induce a molecular switch in stem cells that results incells with persistent changes in the intracellular regulatory loops governing prolifer-ation and response to DNA damage. These molecular changes are manifested in atleast two cellular responses that distinguish between the two cell states: first, a blockin proliferation, particularly carcinogen-induced proliferation and second, a block inproliferation of estrogen-receptor positive cells.

MOLECULAR ALTERATIONS

We have used two approaches to identify some of the molecular alterations under-lying the refractory state. In one set of experiments, we have pursued the observationof Jerry

et al

.

19

that reproductive hormones upregulate expression of the tumorsuppressor gene, p53. In a second set of experiments, we have used suppression sub-tractive hybridization (SSH) methodology to identify new genes that are differential-ly regulated by E and P.

To determine whether the block in proliferation observed early after MNU treat-ment was associated with an E/P-induced change in the expression of p53, we com-pared the levels of p53 protein in AMV and E/P treated glands from 0 to 8 days afterMNU treatment by immunohistochemistry. The results of this analysis demonstrateda striking increase in both the level and nuclear localization of p53 in E/P treatedmammary glands compared to the AMV. The increase occurred as early as 3 daysafter E/P treatment and was sustained through the 21 days of stimulation and 28 daysof involution, namely, at the time of carcinogen treatment. The expression persistedthrough 3 days after MNU treatment, but decreased by day 8 post MNU treatment.Thus, the changes in p53 protein levels and localization preceded the observedincrease in proliferation that are induced by MNU and that are blocked by E/P. Theobserved increase in p53 protein at the time of carcinogen treatment was surprisinggiven our earlier demonstration that E/P treatment did not appear to alter the levelsof p53 mRNA expression. These data indicate that the effects of E/P on p53 are likely

Page 9: Mechanisms of Hormonal Prevention of Breast Cancer

31MEDINA

et al.

: HORMONAL PREVENTION OF CANCER

to occur at a posttranscriptional level and are twofold: involving (1) a change in therate of p53 protein synthesis and/or degradation and (2) changes in nuclear localiza-tion and hence activation of p53.

To address the functional activity of the stabilized p53 protein in the E/P-treatedmammary glands we examined the expression of a well-characterized downstreamtarget of p53, namely, p21 (CIP1/WAF1). p21 was induced in the E/P-treated mam-mary glands upon carcinogen challenge, but was absent in the AMV upon carcino-gen exposure. These results suggest that hormones stabilize p53, which provides acheckpoint against propagation of carcinogen-induced DNA damage.

FIGURE 5. Examples of three differentially expressed clones. (A) agarose gel analysisof poly A RNA from E +P-treated (E +P), and control (vir) glands. (B–D) Northern analysisof the gel shown in A. (B) Clone D.E10 (EST); (C) Clone G.B7 (unknown function); (D)Clone B.F2 (α-casein).

T

ABLE

3. Summary of sequencing result

Identity Ratio of clones sequenced

Known genes 65

%

Transferrin 16%

α-Casein 5%

Stearyl CoA desaturase 3%

κ-Casein 2%

ESTs 27%

Novel genes/unknown function 8%

Page 10: Mechanisms of Hormonal Prevention of Breast Cancer

32 ANNALS NEW YORK ACADEMY OF SCIENCES

In approach two, SSH-PCR identified genes that were differentially induced byE/P treatment. cDNA from AMV gland was used as a driver to subtract moleculescommon to populations of RNA from the E+P tester cDNA. The amplified and clonessequences that were differentially expressed were validated by Northern blot analy-sis. FIGURE 5 illustrates examples of three genes, two of which are unknown, D.E10and G.B7 and one of which is predicted, α-casein. A summary of the sequencing

TABLE 4. Summary of the types of genes isolated by SSH-PCR analysis of the E +Ptreated polyA RNA

Function Gene

Transcription factors Zinc finger protein

Ring zinc finger protein

Sp1 family of transcription factors

Sp3

SHARP-2

Cell cycle control/signal transduction Retinoblastoma associated protein

Retinoblastoma

CDC42

Nip-3 (E1B/Bcl-2 binding protein)

Homologue of T16/S1-5

Homologue of MSS4

Interferon gamma receptor

Nap-1

Helicase Helix-destabilizing protein

Translational repressor NAT-1

Kinases Jak-1

Protein phosphatase-1

Tyrosine phosphatase Prl-1

Extracellular matrix component Decorin

ATP-dependent Metalloprotease

Heparan sulfate proteoglycan core protein

Epithelial glycoprotein

Intercellular signaling Vascular cell adhesion molecule

Homologue of MAGUK

Cell adhesion molecule

Growth factor Mast cell growth factor

Follistatin-related protein

Page 11: Mechanisms of Hormonal Prevention of Breast Cancer

33MEDINA et al.: HORMONAL PREVENTION OF CANCER

results are shown in TABLE 3. Interestingly, many of the differentially expressedgenes are genes involved in alveologenesis and lactogenesis (e.g., caseins, transfer-rin), but 27% of the events were ESTs and 8% novel genes. Sustained expression ofthe genes involved in mammary differentiation have also been reported from studiesin the parous mouse.20 TABLE 4 shows a summary of the types of genes isolated bySSH-PCR. The genes fall into several categories. One gene of interest is theretinoblastoma-binding protein (RbAp46), a molecule implicated in the regulation ofcellular proliferation and differentiation.21,22 Overexpression of this gene suppressesthe growth of cells in culture.22 This protein interacts with breast cancer tumor sup-pressor genes23 and is a component of chromatin-associated protein complexes.24

Thus, RbAp46 is a good candidate for a gene that would lead to persistent changesin gene expression. In situ hybridization demonstrated that the increase in geneexpression was localized to the mammary parenchyma (see FIGURE 6). The SSH-PCR approach promises to be a valuable tool for identifying genes that are specifi-cally upregulated by estrogen and progesterone and that may be casually related tothe resistant state.

SUMMARY AND CONCLUSIONS

The molecular basis for the hormone-mediated prevention of carcinogen-inducedbreast cancer still defies explanation. We have approached this issue using a well-defined model system where a short-term exposure to estrogen and progesterone mim-ics the protective effect of pregnancy. The focus of our research efforts has been onthe time period immediately preceding and following carcinogen-induced initiation

FIGURE 6. Localization of RbAp46 by in situ hybridization. (A) 12-day pregnantgland-antisense (magnification: 4 × ); (B) 12-day pregnant gland—sense control; (C) 12-day pregnant gland—antisense (magnification: 4× ); (D) non-parous virgin gland—anti-sense (magnification: 4 × ).

Page 12: Mechanisms of Hormonal Prevention of Breast Cancer

34 ANNALS NEW YORK ACADEMY OF SCIENCES

of tumorigenesis. We have defined a set of cellular and molecular alterations that occurin the hormone-pretreated gland and which we think are causally related tothe inhibition of tumorigenesis. At the cellular level, these alterations include a blockin carcinogen-induced proliferation, specifically, a block in proliferation of ER-positive mammary epithelial cells. At the molecular level, these alterations include anupregulation of the p53 tumor suppressor gene, a known G1 checkpoint gene, andupregulation of a limited set of known and unknown genes identified by SSH-PCR.Our aim is to identify and functionally characterize these genes with the long-termgoal to test the hypothesis that estrogen and progesterone treatment results in the per-sistent differences in a subset of regulatory molecules between the involuted and theage-matched virgin gland. These differences in gene expression are causally related tothe protective effect of hormone treatment.

REFERENCES

1. HARRIS, J.R., M.E. LIPPMAN, U. VERONESI, et al. 1992. Breast Cancer Pt. 1. N. Engl. J.Med. 327: 319–328.

2. KELSY, J.L. & M.D. GAMMON. 1991. The epidemiology of breast cancer. CA Cancer J.Clin. 41: 146–165.

3. HENDERSON, B.E., R.K. ROSS & M.C. PIKE. 1991. Toward the primary prevention ofcancer. Science 254: 1113–1138.

4. MCMAHON, B., P. COLE, T.M. LIN, et al. 1970. Age at first birth and breast cancer risk.Bull. World Health Org. 43: 209–221.

5. ROSNER, B., G.A. COLDITZ & W.C. WILLETT. 1994. Reproductive risk factors in a pro-spective study of breast cancer: the nurses’ health study. Am. J. Epidemiol. 139:819–835.

6. MOON, R.C. 1969. Relationship between previous reproductive history and chemi-cally-induced mammary cancer in rats. Int. J. Cancer 4: 312–317.

7. RUSSO, J. & I.H. RUSSO. 1980. Susceptibility of the mammary gland to carcinogenesis,Am. J. Pathol. 100: 497–512.

8. GRUBBS, C.J., D.R. FARNELL, D.L. HILL, et al. 1985. Chemoprevention of N-nitroso-n-methylurea induced mammary cancers by pretreatment with 17β-estradiol andprogesterone. J. Natl. Acad. Sci. 74: 927–931.

9. SINHA, D.K., J.E. PAZIK & T.L. DAO. 1988. Prevention of mammary carcinogenesis inrats by pregnancy: effect of full-term and interrupted pregnancy. Br. J. Cancer 57:390–394.

10. RUSSO, I.H., M. KOSZALKA & J. RUSSO. 1990. Human chorionic gonadotropin and ratmammary cancer prevention. J. Natl. Cancer Inst. 82: 1286–1289.

11. RUSSO, I.H., M. KOSZALKA & J. RUSSO. 1991. Comparative study of the influence ofpregnancy and hormonal treatment on mammary carcinogenesis. Br. J. Cancer 64:481–484.

12. THORDARSON, G., E. JIN, R.C. GUZMAN, et al. 1995. Refractoriness to mammary tum-origenesis in parous rats: is it caused by persistent changes in the hormonal environ-ment of permanent biochemical alterations in the mammary epithelia?Carcinogenesis 16: 2847–2853.

13. GUZMAN, R.C., J. YANG, L. RAJKUMAR, et al. 1999. Hormonal prevention of breastcancer: mimicking the protective effect of pregnancy. Proc. Natl. Acad. Sci. USA96: 2520–2525.

14. GRUBBS, C.J., M. JULIANA & L.M. WHITAKER. 1988. Short-term hormone treatment asa chemopreventive method against mammary cancer initiation in rats. AnticancerRes. 8: 113–118.

15. SIVARAMAN, L., L.C. STEPHENS, B.M. MARKAVERICH, et al. 1998. Hormone-inducedrefractoriness to mammary carcinogenesis in Wistar-Furth rats. Carcinogenesis 19:1573–1581.

Page 13: Mechanisms of Hormonal Prevention of Breast Cancer

35MEDINA et al.: HORMONAL PREVENTION OF CANCER

16. MEDINA, D. & G.H. SMITH. 1999. Chemical carcinogen-induced tumorigenesis inparous, involuted mouse mammary glands. J. Natl. Cancer Inst. 91: 967–969.

17. ISSACS, J.T. 1986. Genetic control of resistance to chemically-induced mammary ade-nocarcinogenesis in the rat. Cancer Res. 46: 3958–3963.

18. YANG, J., K. YOSHIZAWA, S. NANDI, et al. 1999. Protective effects of pregnancy andlactation against N-methyl-N-nitrosourea-induced mammary carcinomas in femaleLewis rats. Carcinogenesis 20: 623–628.

19. KUPERWASSER, C., J. PINKAS, G.D. HURLBUT, et al. 2000. Cytoplasmic sequestrationand functional repression of p53 in the mammary epithelium is reversed by hor-monal treatment. Cancer Res. 60: 2723–2729.

20. CHODOSH, L.A., C.M. D’CRUZ, H.P. GARDNER, et al. 1999. Mammary gland develop-ment, reproductive history, and breast cancer risk. Cancer Res. 59: 1765s–1771s;discussion 1771s–1772s.

21. QIAN, Y.W. & E.Y. LEE. 1995. Dual retinoblastoma-binding proteins with propertiesrelated to a negative regulator of ras in yeast. J. Biol. Chem. 270: 25507–25513.

22. GUAN, L.S., M. RAUCHMAN & Z.Y. WANG. 1998. Induction of Rb-associated protein(RbAp46) by Wilms’ tumor suppressor WT1 mediates growth inhibition. J. Biol.Chem. 273: 27047–27050.

23. YARDEN, R.I. & L.C. BRODY. 1999. BRCA1 interacts with components of the histonedeacetylase complex. Proc. Natl. Acad. Sci. USA 96: 4983–4988.

24. ZHANG, Y., H.H. NG, H. ERDJUMENT-BROMAGE, et al. 1999. Analysis of the NuRD sub-unit reveals a histone deacetylase core complex and a connection with DNA methy-lation. Genes Dev. 13: 1924–1935.