Sensitivity to DNA Damage Is a Common Componentof Hormone-Based Strategies for Protectionof the Mammary Gland

Yifan Tu,1,3 D. Joseph Jerry,1,2,3 Brooke Pazik,3 and Sallie Smith Schneider1,2,3

1Molecular and Cellular Biology Program and 2Department of Veterinary and Animal Sciences, University ofMassachusetts, Amherst, Massachusetts, and 3Pioneer Valley Life Sciences Institute Baystate MedicalCenter/UMass Biomedical Research Institute, Springfield, Massachusetts

AbstractAn early full-term pregnancy significantly reduces the

risk of getting breast cancer in women. In animals, this

protection can be mimicked by a short-term exposure to

physiologic doses of estrogen plus progesterone.

Sensitization of p53 and up-regulation of transforming

growth factor B are believed to be important aspects of

the mechanism by which protection is imparted. Little is

known, however, about the use of this pathway in

response to other chemopreventive agents. In this

article, we investigated the ability of retinoids, such

as 9-cis retinoic acid, all-trans retinoic acid, and

N-4-hydroxyphenylretinamide (4-HPR), to sensitize the

ductal epithelial cells of virgin mammary glands to DNA

damage responses. Using a whole-organ culture system,

we observed enhanced cell death in response to

;-irradiation in the virgin tissues treated with retinoids

for 72 hours. These retinoids were partially dependent

on p53 and transforming growth factor B to exert their

radiosensitizing effects. However, 4-HPR seemed to

sensitize other cells or activate these pathways in a

different manner as costimulation with ovarian

hormones and 4-HPR was additive, whereas coculture of

ovarian hormones and the natural retinoids did not

increase amount of death. Taken together, these data

suggest that sensitization of the mammary epithelium to

p53-dependent apoptosis is a common pathway, which

is engaged by retinoids as well as ovarian hormones.

(Mol Cancer Res 2005;3(8):435–42)

IntroductionBreast cancer is the most common cancer among women in

the Western countries. Risk factors include hormonal alterations

such as early menarche, late menopause, late first pregnancy, or

prolonged use of hormone replacement therapy (1). However,

the risk of developing breast cancer can be reduced. A full-term

pregnancy before 20 years of age is the most effective natural

protection against breast cancer in women. This protection

occurs universally among women with different ethnic back-

grounds. This indicates that the protective effect does not result

from environmental or socioeconomic factors, but from a

permanent biological change within the breast caused by

pregnancy (2-4). During pregnancy, the mammary gland is

exposed to the highest physiologic concentrations of ovarian

steroids (estradiol and progesterone), pituitary hormone (pro-

lactin and growth hormone), and placental (hCG as well as

placental lactogens) hormone, which cause proliferation and

differentiation of the mammary gland (5-7). In rats and mice,

the protection of pregnancy can be mimicked by short-term

exposure to physiologic doses of estrogen plus progesterone

(8). It has been hypothesized that hormone stimulation at a

critical stage of mammary development causes cells to change

both in the regulatory loops governing proliferation and in the

response to DNA damage (9).

Persistent molecular changes induced by parity or hormones

in mammary gland have been noted in several laboratories. With

the use of DNA microarrays, D’Cruz et al. (2) found that a full-

term pregnancy results in the persistent up-regulation of

transforming growth factor (TGF) h3, a growth-inhibitory

molecule. Ginger et al. (10) noted changes in the chromatin

remodeling gene RbAp46 in response to estrogen and

progesterone. Sivaraman and coworkers found that in Wistar-

Furth rats as well as in BALB/c mice, a 21-day estrogen +

progesterone treatment induced the up-regulation and nuclear

accumulation of p53 protein. The elevation of p53 was persistent

28 days after hormone withdrawal. Pregnancy induced a similar

pattern of expression (11). Minter and colleagues strengthened

these findings. They showed that p53-mediated apoptosis in

response to g-radiation was greater in proliferating or estrogen +

progesterone–treated mammary glands than in glands of virgin

mice (12). Based on these findings, a model has been proposed

that p53 activity is impaired during specific periods of mammary

gland development, causing the mammary epithelium to be

susceptible to cancer. However, prolonged ovarian hormone

stimulation, such as during pregnancy, serves to sensitize p53 to

the effects of DNA damage. The responsiveness of the p53

pathway to the changes of the endocrine environment associated

with pregnancy renders the mammary epithelium refractory to

tumors (13).

The main function of activated p53 is to arrest cell cycle and

induce apoptosis in response to DNA damage. Medina et al.

showed that the deletion of the tumor suppressor gene p53

Received 3/31/05; revised 7/19/05; accepted 7/21/05.Grant support: Rays of Hope at the Comprehensive Breast Center in theBaystate Health System.The costs of publication of this article were defrayed in part by the payment ofpage charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.Requests for reprints: Yifan Tu, University of Massachussetts at Amherst,Amherst, MA 01107. Phone: 413-695-1679. E-mail: yifantu@yahoo.comCopyright D 2005 American Association for Cancer Research.doi:10.1158/1541-7786.MCR-05-0038

Mol Cancer Res 2005;3(8). August 2005 435

results in enhanced tumorigenic risk. They found that the

mammary gland from the p53-wild-type and p53-null mice

exhibited the same dependence on ovarian hormones for

growth. However, estrogen and progesterone strongly enhanced

carcinogenesis in the p53-null mammary epithelium (14). This

suggests that p53 may be a critical factor for protection of the

mammary gland from further accumulation of mutations.

Furthermore, recent studies have indicated that pregnancy does

not impart protection in p53-null mice, suggesting that p53 is

required, in part, for the protective response imparted by

hormones (15).

Although an early full-term pregnancy can reduce the risk of

women developing breast cancer, it is not a practical approach

to breast cancer protection. Chemoprevention is defined as the

use of specific chemical substances (natural or synthetic) or

their mixtures to suppress, retard, or reverse the process of

carcinogenesis. Among these chemicals, the retinoids (deriva-

tives of vitamin A) have shown great promise in the area of

cancer chemotherapy and chemoprevention (16). The most

extensively studied native retinoids in cancer medicine are all-

trans retinoic acid (RA) and 9-cis RA. In humans, all-trans RA

has been used to treat acute promyelocytic leukemia with great

success. All-trans retinoids, alone or in combination with other

drugs, also showed activity in juvenile chronic myeloid

leukemia (17). Retinoids inhibit breast cancer growth and

augment the action of other breast cancer cell growth inhibitors

both in vivo and in vitro (18). However, retinoids cause serious

side effects at high doses (19), and thus more selective retinoids

have been designed (20). Among these is N-4-hydroxyphenyl-

retinamide (4-HPR), a synthetic retinoid that has been

examined as a cancer chemopreventive agent (21). A phase

III clinical breast cancer trial found a beneficial trend with 4-

HPR in the protection of premenopausal women from

contralateral and ipsilateral breast cancer (22).

FIGURE 1. The effect of retinoids on sensitivity to g-irradiation – induced death. A. Mammary glands from mice were treated in whole-organ culture(WOC) with different hormones in the media. After 72 hours of incubation in 37jC incubator, experimental whole-organ cultures (black columns ) were treatedwith 5 Gy g-irradiation, whereas control whole-organ cultures (white columns ) were not. Six hours after the g-irradiation, all tissues were processed forimmunohistochemical staining. TUNEL assays were used to identify the apoptotic cells. Compared with unirradiated treatments, ethanol cannot sensitizecells to death in response to irradiation (P > 0.05), whereas estrogen + progesterone or RAs can sensitize significantly more cells to apoptosis in response tog-irradiation (*, P < 0.05). However, there is no significant difference among the effects of drug treatments. Among the irradiated treatments (black columns ),estrogen + progesterone or RAs can sensitize more cells to death than ethanol. Without irradiation (white columns ), 9-cis RA can cause more baseline deaththan others. Exposure time: 72 hours. Concentrations: 9-cis RA 1 Amol/L, all-trans RA 10 Amol/L, 4-HPR 20 Amol/L, estrogen (E) 3.7 � 10�3 Amol/L,progesterone (P) 3.2 Amol/L. B. p53 wild-type mammary glands were treated with ovarian hormones or RAs for 72 hours. Glands of experimental groupswere g-irradiated (black columns ), whereas control group glands were not (white columns ). Six hours after irradiation, all tissues were fixed and processedand paraffin embedded. p53 staining was done and p53-positive cells were counted. Estrogen + progesterone or RA can induce more p53 nuclearsequestration in response to irradiation (*, P < 0.05). Exposure time: 72 hours. Concentrations: 9-cis RA 1 Amol/L, all-trans RA 10 Amol/L, 4-HPR 20 Amol/L,estrogen 3.7 � 10�3 Amol/L, progesterone 3.2 Amol/L. C. Tissues were treated as described in A except that the concentrations of all-trans RA and 4-HPRwere lowered to 1 Amol/L. TUNEL assays were used to identify the apoptotic cells. The results were very similar with those observed in A (*, P < 0.05).D. The same was done as in B except that the concentrations of all-trans RA and 4-HPR were lowered to 1 Amol/L. p53 staining was done and p53-positivecells were counted. Estrogen + progesterone or 1 Amol/L RA also can induce more p53 nuclear sequestration in response to irradiation (*, P < 0.05).

Tu et al.

Mol Cancer Res 2005;3(8). August 2005

436

Retinoids signal through multiple pathways and seem to exert

protection at numerous levels. Pathways shared by the natural

and synthetic retinoid family seem to include an increased

expression and activation of TGF-h family members (23, 24).

Activation of p53 is another pathway by which retinoids may

exert their effects. Numerous examples have been cited which

indicate that retinoid exposure leads to the induction of p53

stability and transcriptional activities (25-28).

In this article, we addressed the question of whether natural

or synthetic retinoids can sensitize the ductal epithelial cells to

p53-dependent death induced by irradiation. We showed that

DNA damage–induced death by these retinoids is partially

dependent on p53 and TGF-h.

ResultsRetinoids Sensitize Ductal Epithelial Cells to Apoptosis inResponse to Radiation

Ionizing radiation is a complete mutagen and is a known

carcinogen of human breast and rodent mammary glands (29).

Here, we treated mouse mammary glands with g-radiation to

damage DNA and impart a response.

In Fig. 1A, terminal deoxyribonucleotidyl transferase–

mediated dUTP nick-end labeling (TUNEL) assays were used

to detect apoptotic cells. In response to DNA damage, virgin

BALB/c-p53+/+ mammary glands treated with ethanol failed to

induce a significant apoptotic response following irradiation

compared with unirradiated tissues (P > 0.05). However, when

the glands were treated with estrogen + progesterone or with all-

trans (10 Amol/L), 9-cis RA (1 Amol/L), or 4-HPR (20 Amol/L)

for 72 hours, the percentage of apoptotic cells in the irradiated

glands was increased significantly compared with the irradiated

ethanol-treated glands. There was, however, no difference

among effects of drug treatments (P > 0.05). Interestingly, in

the absence of irradiation, 9-cis RA caused more baseline death

than others, suggesting that it can cause DNA damage or stress

response which can mediate death. However, irradiation still

significantly further increased death. All the treatments caused a

sensitization to radiation-induced death that was statistically

significant (P < 0.05). This suggests that all three RAs can

sensitize ductal epithelial cells to respond to DNA damage.

To determine if p53 can respond to radiation in virgin tissue

exposed to retinoids, we examined the nuclear accumulation of

FIGURE 1 continued. E.Photos were taken of theglands with the treatmentsdescribed in A and B. Thedark brown stained cells wereTUNEL- or p53-positive cells;the green cells were negativecells counterstained withmethyl green.

Sensitivity to DNA Damage and Hormone-Based Strategies

Mol Cancer Res 2005;3(8). August 2005

437

p53 protein in response to radiation through immunohistochemistry.

In the unirradiated tissues, nuclear accumulation of p53 was

not observed. However, in the 9-cis RA–, all-trans RA–, and

4-HPR–treated groups, subsequent radiation caused significant

nuclear accumulation of the p53 protein (P < 0.05) compared

with the ethanol-treated group. No significant difference was

found when comparing the efficiencies of all-trans RA and

estrogen + progesterone or 9-cis RA and estrogen +

progesterone, but 4-HPR was found to be more effective than

estrogen + progesterone (P < 0.05). Because the pharmacologic

significances of high doses of retinoids (10 Amol/L all-trans

RA and 20 Amol/L 4-HPR) were not clear, we investigated the

effects of retinoids at lower doses (1 Amol/L all-trans RA and

4-HPR). In Fig. 1C and D, we can see the same effect of

retinoids at lower doses.

4-HPR and Estrogen + Progesterone Work AdditivelyBecause all three retinoids, as well as the ovarian hormones,

sensitize glands to cell death in response to DNA damage, we

were interested to determine whether overlapping or distinct

pathways were used to sensitize p53 and the cells to DNA

damage–induced cell death. If the pathways were distinct, then

we would expect to see an additive response. However, if they

were similar, we would expect to see no augmentation of the

response when both stimuli were presented together. We treated

virgin mammary glands with retinoids and ovarian hormones

for 72 hours, and then irradiated the tissue. Figure 2 shows that,

in response to radiation, estrogen plus progesterone sensitized

5.7% cells to apoptosis in ductal epithelial cells. All-trans RA

or 9-cis RA alone sensitized 7.1% and 7.6% cells to death,

respectively; however, the levels of apoptosis did not increase

when glands were treated with estrogen plus progesterone along

with all-trans and 9-cis RAs. In contrast, the response to

4-HPR with estrogen + progesterone seemed to be additive.

When we treated glands with estrogen + progesterone and

4-HPR, 9.4% cells were apoptotic, whereas only 6.9% cells

(statistically significant, P < 0.05) were apoptotic when glands

were treated with 4-HPR alone and only 5.7% cells (statistically

significant, P < 0.05) were apoptotic when glands were treated

with estrogen + progesterone. This suggests that 4-HPR

sensitizes cells to death through a distinct pathway from the

one used by the natural retinoids and ovarian hormones.

Retinoids Are Partially Dependent on p53 to Induce CellDeath in Response to Irradiation

Previous studies have suggested that the hormonal preven-

tion of breast cancer is mediated by active p53. The protein

level and nuclear compartmentalization of the p53 gene are

largely up-regulated as a result of estrogen and progesterone

stimulation followed by g-radiation (9, 12, 30). The require-

ment for p53 activity in the sensitizing effects of the retinoids to

DNA damage–induced death was tested using p53-null mice.

In Fig. 3, we can see that when 9-cis RA–, all-trans RA–, and

4-HPR–treated glands were irradiated, the apoptotic cell

numbers of p53-null glands were approximately half of those

of p53-wild-type glands exposed to the same retinoid treat-

ments. In the ethanol group, only 0.27% p53-null cells commit

apoptosis in response to irradiation, whereas retinoids can still

sensitize p53-null cells to commit apoptosis (around 3%

apoptotic cells), suggesting that 9-cis RA, all-trans RA, and

4-HPR may be only partially dependent on p53 to sensitize

cells to death, indicating other pathways to death are involved.

In the case of 4-HPR, we see higher apoptotic cells in p53-null

glands with no irradiation treatment than in p53-wild-type

glands without irradiation treatment, suggesting that p53 may

inhibit the pathway induced by 4-HPR alone (P < 0.05).

FIGURE 2. HPR and ovarian hormones work additively, whereas 9-cisand all-trans RAs do not. Mammary glands from mice were treated withovarian hormones or RAs or both. Seventy-two hours later, glands ofexperimental groups were g-irradiated (black columns, single treatmentlike estrogen + progesterone or RA followed by irradiation; stripedcolumns, interaction treatments RA + estrogen + progesterone followedby irradiation); control group glands (white columns ) were not irradiated.Six hours after irradiation, all tissues were fixed and processed andparaffin embedded. TUNEL assays were done and positive cells werecounted. 4-HPR + estrogen + progesterone can sensitize more cells todeath than 4-HPR alone in response to irradiation (*, P < 0.05). Exposuretime: 72 hours. Concentrations: 9-cis RA 1 Amol/L, all-trans RA 10 Amol/L,4-HPR 20 Amol/L, estrogen 3.7 � 10�3 Amol/L, progesterone 3.2 Amol/L.

FIGURE 3. Retinoids are partially dependent on p53 to sensitize cellsto apoptosis in response to g-irradiation. Mammary glands from p53 wild-type (+) and p53-null (�) mice were cultured in whole-organ culture andwere treated with three kinds of RAs for 72 hours. Glands of experimentalgroups were g-irradiated (black columns ); control group glands were not(white columns ). Six hours after irradiation, all tissues were fixed andprocessed and paraffin embedded. TUNEL assays were done andpositive cells were counted. Without p53, RAs induced significantlyless cell death in response to irradiation (*, P < 0.05). Exposure time:72 hours. Concentrations: 9-cis RA 1 Amol/L, all-trans RA 10 Amol/L,4-HPR 20 Amol/L, estrogen 3.7 � 10�3 Amol/L, progesterone 3.2 Amol/L.

Tu et al.

Mol Cancer Res 2005;3(8). August 2005

438

TGF-b Is Involved in 9-cis RA– and all-trans RA–Mediated Sensitivity to c-Irradiation

g-Radiation has been shown to activate TGF-h, and the

signaling through this pathway has in turn been shown to

activate p53 in mammary epithelium (31). TGF-h has been

shown to require estrogen + progesterone induction for p53

sensitization to irradiation (32). Retinoids have been shown to

induce the secretion of several isotypes of TGF-h from some

cell lines (33, 24). Furthermore, all-trans RA and its analogue

BMS453 were shown to increase active TGF-h in human

mammary epithelial cells. Growth arrest by BMS453 is

believed to occur predominantly through TGF-h–inducedmechanism (34). To investigate if TGF-h is involved in the

retinoid-mediated sensitivity to g-irradiation, we added a

neutralizing pan-TGF-h antibody to the retinoid-treated media

to block the function of TGF-h1, 2, and 3. In Fig. 4A, the 9-cis

RA–induced death in the absence of radiation is strikingly

reduced, suggesting the addition of this retinoid induces death

on its own through a TGF-h–mediated pathway. TUNEL

assays also indicate that in the presence of anti-TGF-hantibodies, the radiation-induced death of ductal epithelial cells

in mammary glands treated with 9-cis RA, all-trans RA, and

4-HPR was partially reduced (for 9-cis RA–treated group,

apoptotic cells dropped from 7.6% to 2%; for all-trans RA–

treated group, apoptotic cells dropped from f7.1% to 3.6%;

for 4-HPR–treated group, apoptotic cells dropped from 6.9% to

3.8%). In Fig. 4B, we observed that the neutralizing antibody to

TGF-h also reduced the number of p53-positive staining cells

in the retinoid-treated tissues, demonstrating that TGF-h is

required for the accumulation of p53 under these circumstances.

DiscussionHormone stimulation at a critical stage of mammary

development is an effective protection against breast cancer

(9). The mechanism of the protection seems to be related, in

part, to the sensitization of p53 in response to DNA damage.

Retinoid derivatives have been shown to be effective in

preventing contralateral and ipsilateral breast cancer in

premenopausal women. The mechanism differs among the

retinoid derivatives. Natural retinoids have been shown to

activate retinoid receptors and strongly induce growth arrest.

The growth arrest seems to be due to a rapid decline of cyclin

D1 protein through a proteolytic mechanism (35, 36). Some

studies show that retinoid treatment induces a senescence-like

phenotype in tumor cells (37). 4-HPR, however, inhibits cell

growth through the induction of apoptosis rather than

differentiation, likely resulting from an increase of reactive

oxygen species and apoptosis related proteins.

In the present study, we showed that both natural (9-cis and

all-trans RAs) and synthetic RAs (4-HPR) can sensitize ductal

epithelial to death in response to DNA damage caused by

g-radiation (Fig. 1A). More apoptotic cells were detected in

mammary glands treated with estrogen plus progesterone or

retinoids in response to radiation. Previous studies had

suggested that estrogen and progesterone treatment caused

accumulation and nuclear localization of p53 after irradiation

(9, 12). Our study extended these findings and showed

enhanced DNA damage–induced p53 staining in the nuclei

of the cells treated with retinoids as well as estrogen plus

progesterone (Fig. 1B).

Interestingly, 4-HPR was found to work additively with

ovarian hormones whereas 9-cis and all-trans RAs did not

show the additive effect (Fig. 2). This suggests that 4-HPR may

use a pathway that differs from that used by hormones to

sensitize cells to death in response to DNA damage, but 9-cis

and all-trans RAs may use overlapping pathways. Retinoids

have previously been shown to cause the accumulation of p53.

In a normal human mammary epithelial cell line, p53 protein

expression was enhanced after 72 hours of all-trans RA

treatment (38). The human carcinoma cell line NT2/D1 also

responds with an up-regulation of p53 that is retinoid receptor

dependent (25). Correlative data suggest that retinoids may

induce death through the activation of p53. 9-cis and all-trans

RAs were found to induce apoptosis in human p53-wild-type

soft tissue sarcoma cells lines HTB-82 and HTB-93. However,

9-cis and all-trans RAs were not able to sensitize cells to death

in the other soft tissue sarcoma cell lines with mutated p53 (39).

FIGURE 4. TGF-h is involved in 9-cis RA – and all-trans RA –mediated sensitivity to g-irradiation. A. p53 wild-type mice mammaryglands were treated with all-trans RA, 9-cis RA, or 4-HPR with or withoutneutralizing TGF-h antibody (anti –TGF-b ; + or �), which can block theeffect of TGF-h1, 2, and 3. Seventy-two hours later, glands ofexperimental groups were g-irradiated (black columns ); control groupglands were not (white columns ). Six hours after irradiation, all tissueswere fixed and processed and paraffin embedded. TUNEL assay wasdone and apoptotic cells were counted. In the absence of TGF-h, all-transRA, 9-cis RA, and 4-HPR sensitized fewer cells to death in response toirradiation (*, P < 0.05). B. The same treatments were done as in A. p53staining was done and p53-positive cells were counted. In the absence ofTGF-h, RAs induced less p53 nuclear accumulation in response toirradiation (*, P < 0.05). Exposure time: 72 hours. Concentrations: 9-cis1 Amol/L, all-trans RA 10 Amol/L, 4-HPR 20 Amol/L, anti –TGF-h 0.1 Ag/mL.

Sensitivity to DNA Damage and Hormone-Based Strategies

Mol Cancer Res 2005;3(8). August 2005

439

A study done with the subclones of the retinoid-sensitive

acute myeloblastic leukemia (OCI/AML-2) cell lines showed

9-cis RA had the most prominent effect of inducing apoptosis

and inhibiting cell growth compared with all-trans and 13-cis

RAs and that 9-cis RA was most effective at causing the

translocation of p53 from cytosol to nucleus (27). Because the

p53 pathway is involved in the protective response imparted

by hormones (14), and retinoids in cell lines have been shown

to induce p53 stability and transcriptional activity (38), we

further investigated the relationship between p53 and retinoids

in a whole-organ culture system. Mammary glands from wild-

type and p53 knockout mice were pretreated with retinoids

and then irradiated. The numbers of apoptotic cells were

compared with those of the p53-wild-type mammary glands

given the same treatments. From Fig. 3, we can see that

without p53, apoptosis was greatly reduced in response to

g-radiation. However, all three retinoids could still induce

death in f3% of ductal epithelial cells. This suggests that

9-cis RA, all-trans RA, and 4-HPR are able to sensitize cells

to irradiation-induced death through p53-dependent and

-independent pathways.

TGF-h inhibits cell proliferation and has been shown to be

associated with death in many kinds of epithelial cells. TGF-h1has been found to be rapidly activated in response to ionizing

radiation. Barcellos-Hoff et al. found that radiation-induced

apoptosis was a TGF-h1–dependent process in the mammary

epithelium. In addition, they found that glands from TGF-h1knockout animals or addition of TGF-h1 neutralizing anti-

bodies resulted in less p53 activation in response to g-radiation

(31). Previous studies in cell lines have indicated that retinoids

can also induce the activity of TGF-h (34, 40). All-trans RA

and its analogue BM453 were found to inhibit proliferation and

induce more total and active TGF-h in normal human

mammary epithelial cells (34). To investigate the involvement

of TGF-h in the cell death induced by RAs, we used

neutralizing TGF-h antibody to block the TGF-h functioning

to see if there were less apoptotic cells and p53-positive cells. In

Fig. 4A, we observed that with neutralizing TGF-h antibody,

the background death observed in response to 9-cis RA was

reduced, suggesting that 9-cis RA causes activation of TGF-hin the mammary gland and that this is responsible for the death

observed in the absence of radiation. Reduced cell death was

observed in response to radiation with all the retinoid

treatments. p53 staining on the same samples indicated that

the neutralizing TGF-h antibody decreased the p53-positive cell

numbers of 9-cis RA–, all-trans RA–, and 4-HPR–treated

tissues followed by irradiation. This suggests that all these

retinoids can stabilize or up-regulate p53 through TGF-h in the

gland and that this pathway is partially responsible for the

increased amount of death. Examples of other pathways which

might be involved in the induction of death include the up-

regulation of reactive oxygen species, retinoic acid receptor

h, and other apoptosis-related genes such as c-myc and c-jun

(36, 41).

In summary, our study showed that sensitization of the

mammary gland to radiation-induced responses may be a very

important pathway shared by numerous agents that act in

protecting the mammary gland. Retinoids, as well as ovarian

hormones, can sensitize p53 such that radiation will activate

TGF-h and result in the accumulation and activation of the p53

protein and death. Unlike ovarian hormones, retinoids may also

contain the ability to sensitize cells via other pathways,

particularly in the case of 4-HPR. The ability of 4-HPR to

enhance the death response in combination with other agents

suggests that it may use a different pathway or can affect cells

which are not responsive to the other agents. This is an

important observation because it indicates that 4-HPR may be

used in combination with other chemopreventive agents to

enhance protection of the mammary gland. This requirement for

both p53 and TGF-h in the death induced by 4-HPR suggests

that the activation of p53 by TGF-h is likely to be downstream

of the convergence of the retinoid pathway on the sensitization

of p53. Future studies will determine if ovarian hormones use

the retinoid pathway for the sensitization of p53 or if the

ovarian hormones and retinoids are acting on similar effectors.

In conclusion, we are able to show that multiple chemo-

preventive agents are able to sensitize p53 in the mammary

gland to death induced by radiation. This suggests that the

responsiveness of p53 is a critical protein target for chemo-

preventive strategies.

Materials and MethodsAnimals, Treatments, and Tissue Procedures

All procedures were conducted in accordance with the NIH

Guide for the Care and Use of Laboratory Animals. Virgin

BALB/c female mice, Trp53+/+ or Trp53�/�, 8 weeks of age,

were housed in plastic cages under a 12-hour dark/light cycle,

and were permitted free access to food and water. Fourth

inguinal mammary glands were removed from these mice and

were maintained in whole-organ culture for 72 hours. Then, the

glands were treated with 5 Gy of g-radiation using a 137Cs

irradiator. Six hours after g-radiation, tissues were processed

with gradient alcohol and paraffin embedded.

Whole-Organ CultureAge-matched 8-week-old virgin BALB/c-p53+/+ and BALB/

c-p53�/� mice were sacrificed with CO2 and fourth inguinal

mammary glands were isolated aseptically, cut in half

longitudinally, and then floated on siliconized lens paper in

60 mm tissue culture dishes containing 2 mL of medium (12).

Basic medium was made of serum-free DMEM/F12 buffer with

HEPES and NaHCO3, plus insulin (5 Ag/mL; Sigma, St. Louis,

MO), penicillin (100 units/mL), streptomycin (100 Ag/mL), and

gentamycin (10 Ag/mL; Life Technologies, Grand Island, NY).

Hormone-supplemented media contained 1 ng/mL (3.7 � 10�3

Amol/L) estradiol (Sigma) and 1 Ag/mL (3.2 Amol/L)

progesterone. RA-supplemented medium was prepared by

adding 1 Amol/L 9-cis RA, 20 Amol/L 4-HPR (Sigma), or

10 Amol/L all-trans RA (Sigma) to the basic medium,

respectively. The retinoid plus neutralizing TGF-h antibody–

supplemented medium was prepared by adding 0.1 Ag/mL

pan-specific TGF-h antibody (R&D systems, Inc., Minneapolis,

MN) to the RA-supplemented medium at respective concen-

trations. Negative control medium was prepared by adding a

similar percentage of ethanol to the basic medium. Whole-

organ cultures were maintained in an incubator supplied with

5% CO2 in air. Seventy-two hours later, whole-organ cultures

Tu et al.

Mol Cancer Res 2005;3(8). August 2005

440

were subjected to 5 Gy of g-radiation using a 137Cs irradiator,

whereas control whole-organ cultures were not irradiated. Six

hours after irradiation, tissues were fixed in 10% neutral-

buffered formalin and then embedded in paraffin.

ImmunohistochemistryTo detect p53, 4-Am-thick slides were baked at 37jC

overnight and were deparaffinized in xylene, rehydrated in

graded ethanols, rinsed in TBS, then subjected to microwave

antigen retrieval in 0.1 mol/L citrate buffer for 5 minutes with

two sets of 1-minute intervals. Slides were stained by a Dako

autostainer with Dako Envision Anti-rabbit kit (Dako Cytoma-

tion, Carpinteria, CA). The primary antibody used was rabbit

polyclonal CM5 anti-p53 antibody (1:200, Novacastra, New-

castle upon Tyne, United Kingdom). Slides were counterstained

with methyl green, dehydrated through 100% ethanol and

xylene, and coverslipped. Internal negative controls were

treated identically, omitting primary antibody labeling on the

tissues from age- and treatment-matched p53-wild-type mice. A

minimum of 3,600 mammary gland epithelial cells were

counted per treatment (1,200 cells per culture/triplicate cultures

per treatment). Differences in the percentage of cells with p53

staining with or without g-radiation were determined using a

test of two portions with a 95% confidence interval. Differences

of the effects among drug treatments were tested by ANOVA

method and Tukey’s multiple comparison procedure at 0.05

a level.

TUNEL AssayFour-micrometer-thick slides were baked at 37jC overnight

and were deparaffinized in xylene, rehydrated in graded

ethanol, rinsed in TBS, and incubated in 3% hydrogen peroxide

in methanol. Terminal deoxynucleotidyl transferase dUTP was

used to label the nick ends of fragmented DNA (FragEL DNA

Fragmentation Kit, Oncogene Research Products, Cambridge,

MA). Sections were stained with 3,3V-diaminobenzidine and

counterstained with methyl green. A minimum of 3,600

mammary gland epithelial cells was counted per treatment

(1,200 cells per culture/triplicate cultures per treatment).

Differences in the percentage of TUNEL-positive cells with

or without g-radiation were determined using a test of two

portions with a 95% confidence interval. Differences of the

effects among drug treatments were tested by ANOVA method

and Tukey’s multiple comparison procedure at 0.05 a level.

Acknowledgments

We thank Dr. Lou Roberts and Dr. Jenny Zhang for critical reading of themanuscript.

References1. Ellisen LW, Haber DA. Recent progress in the understanding and treatment ofbreast cancer. 2001, published online.

2. D’Cruz CM, Moody SE, Master SR, et al. Persistent parity-induced changes ingrowth factors, TGF-h3, and differentiation in the rodent mammary gland. MolEndocrinol 2002;16:2034 –51.

3. Russo IH, Russo J. Developmental stage of the rat mammary gland asdeterminant of its susceptibility to 7,12-dimethylben(a)anthracene. J Natl CancerInst 1978;61:1439–49.

4. Thordarson G, Jin E, Guzman RC, Swanson SM, Nandi S, Talamantes F.Refractoriness to mammary tumorigenesis in parous rats: is it caused by persistent

changes in the hormonal environment or permanent biochemical alterations in themammary epithelia? Carcinogenesis 1995;16:2847–53.

5. Guzman RC, Yang J, Rajkumar L, et al. Hormonal prevention of breast cancer:Mimicking the protective effect of pregnancy, Proc Natl Acad Sci U S A 1999;96:2520– 5.

6. Nandi S, Guzman R, Yang J. Hormones and mammary carcinogenesis inmice, rats, and humans: a unifying hypothesis. Proc Natl Acad Sci U S A 1995;92:3650–7.

7. Russo IH, Russo J. Mammary gland neoplasia in long-term rodent studies.Environ Health Perspect 1996;104:938–67.

8. Sivaraman L, Stephens LC, Markaverich BM, et al. Hormone inducedrefractoriness to mammary carcinogenesis in Wistar-Furth rats. Carcinogenesis1998;19:1573–81.

9. Medina D, Sivaraman L, Hilsenbeck SG, et al. Mechanisms of hormonalprevention of breast cancer. Ann N Y Acad Sci 2001;952:23 – 35.

10. Ginger MR, Gonzalez-Rimbau MF, Gay JP, et al. Persistent changes in geneexpression induced by estrogen and progesterone in the rat mammary gland. MolEndocrinol 2001;15:1993–2009.

11. Sivaraman L, Conneely OM, Medina D, O’Malley BW. p53 is a potentialmediator of pregnancy and hormone-induced resistance to mammary carcino-genesis. Proc Natl Acad Sci U S A 2001;98:12379–84.

12. Minter LM, Dickinson ES, Naber SP, et al. Epithelial cells cycling predictsp53 responsiveness to g-irradiation during post-natal mammary gland develop-ment. Development 2002;129:2997–3008.

13. Jerry J, Minter LM, Klaus AB, et al. Hormonal control of p53 andchemoprevention. Breast Cancer Res 2002;4:91 –4.

14. Medina D, Kittrell F, Shepard A, Contreras Alejandro, Rossen JM, Lydon J.Hormone dependence in premalignant mammary progression. Cancer Res 2003;63:1067–72.

15. Medina D, Kittrell FS. P53 function is required for hormone-mediatedprotection of mouse mammary tumorigenesis. Cancer Res 2003;63:6140–3.

16. Dragnev KH, Rigas JR, Dmitrovsky E. The retinoids and cancer preventionmechanisms. Oncologist 2000;5:361 –8.

17. Sacchi S, Russo D, Avvisati G, et al. All-trans retinoic acid in hematologicalmalignancies. Haematologica 1997;82:106–21.

18. Yang Q, Sakurai T, Kakudo K. Reviewed in: Retinoid, retinoic acid receptorh and breast cancer. Breast Cancer Res Treat 2002;76:167–73.

19. Verma AK. Retinoids in chemoprevention of cancer. J Biol Regul HomeostAgents 2003;17:92 –7.

20. Camacho LH. Clinical applications of retinoids in cancer medicine. J BiolRegul Homeost Agent 2003;17:98 –114.

21. Malone W, Perloff M, Crowell J, Sigman C, Higley H. Fenretinide: aprototype cancer prevention drug. Expert Opin Investig Drugs 2003;12:1829 –42.

22. Veronesi U, De Palo G, Marubini E, et al. Randomized trial of fenretinide toprevent second breast malignancy in women with early breast cancer. J NatlCancer Inst 1999 Nov 3;91:1847– 56.

23. Bonewald LF, Oreffo R, Lee C, Park-Snyder S, Twardzik D, Mundy G.Effects of retinol on activation of latent transforming growth factor-h by isolatedosteoclasts. Endocrinology 1997;138:657 –66.

24. Tokuyama H, Tokuyama Y. Retinoids enhance IgA production bylipopolysaccharide-stimulated murine spleen cells. Cell Immunol 1993 Sep;150:353 –63.

25. Curtin JC, Dragnev KH, Sekula D, Christie AJ, Dmitrovsky E,Spinella MJ. Retinoic acid activates p53 in human embryonal carcinomathrough retinoid receptor-dependent stimulation of p53 transactivation function.Oncogene 2001 May 3;20:2559 –69.

26. Tokumoto YM, Tang DG, Raff MC. Two molecularly distinct intracellularpathways to oligodendrocyte differentiation: role of a p53 family protein. EMBOJ 2001 Sep 17;20:5261 –8.

27. Koistinen P, Zheng A, Saily M, Siitonen T, Mantymaa P, Savolainen ER.Superior effect of 9-cis retinoic acid (RA) compared with all-trans RA and 13-cisRA on the inhibition of clonogenic cell growth and the induction of apoptosis inOCI/AML-2 subclones: is the p53 pathway involved? Br J Haematol 2002 Aug;118:401 –10.

28. Sun SY, Yue P, Wu GS, et al. Mechanisms of apoptosis induced by thesynthetic retinoid CD437 in human non-small cell lung carcinoma cells.Oncogene 1999 Apr 8;18:2357–65.

29. Barcellos-Hoff MH, Ravani SA. Irradiation mammary gland stromapromotes the expression of tumorigenic potential by unirradaited epithelial cells.Cancer Res 2000;60:1254– 60.

30. Kuperwasser C, Pinkas J, Hurlbut G. Cytoplasmic sequestration and

Sensitivity to DNA Damage and Hormone-Based Strategies

Mol Cancer Res 2005;3(8). August 2005

441

functional repression of p53 in the mammary epithelium is reversed by hormonaltreatment. Cancer Res 2000;60:2723 –9.

31. Ewan KB, Henshall-Powell RL, Ravani SA, et al. Transforming growthfactor-h1 mediates cellular response to DNA damage in situ . Cancer Res 2002;62:5627 –31.

32. Becker KA, Lu S, Dickinson ES, et al. Estrogen and progesterone regulateradiation-induced p53 activity in mammary epithelium through TGF-h-dependentpathways. Oncogene. Epub 2005 Jun 6.

33. Cosgaya JM, Perona R, Aranda A. Retinoic acid induces secretionof transforming growth factors by PC12 pheochromocytoma cells. Oncogene1997 Feb 6;14:579 –87.

34. Yang L, Jacek O, Reczek P, Brown P. The retinoic acid receptor antagonist,BM453, inhibits normal breast cell growth by inducing active TGFh and causingcell cycle arrest. Oncogene 2001;20:8025 –35.

35. Sherr CJ. G1 phase progression: cycling on cue. Cell 1994;79:551–5.

36. Nurse P. Ordering S phase andMphase in the cell cycle. Cell 1994;79:547–50.

37. Chang BD, Broude EV, Dokmanovic M, et al. A senescence-like phenotypedistinguishes tumor cells that undergo terminal proliferation arrest after exposureto anticancer agents. Cancer Res 1999;59:3761 –7.

38. Seewaldt VL, Dietze EC, Johnson BS, Steven JC, Parker MB. Retinoic acid-mediated G1-S-phase arrest of normal human mammary epithelial cells isindependent of the level of p53 protein expression. Cell Growth Differ 1999;10:49–59.

39. Brodowicz T, Wiltschke C, Grunt TW, et al. Inhibition of proliferation andinduction of apoptosis in soft tissue sarcoma cells by interferon-a and retinoids.J Cancer 1999;80:1350– 8.

40. Herbert BS, Sanders BG, Kline K. N -(4-hydroxyphenyl) retinamideactivation of transforming growth factor-h and induction of apoptosis in humanbreast cancer cells. Nutr Cancer 1999;34:121– 32.

41. Sun SY, Yue P, Lotan R. Induction of apoptosis by N -(4-Hydroxyphenol)retinamide and its association with reactive oxygen species, nuclear retinoic acidreceptors, and apoptosis-related genes in human prostate carcinoma cells. MolPharmacol 1999;55:403–10.

Tu et al.

Mol Cancer Res 2005;3(8). August 2005

442