Cow's milk contains high levels of estrogens, progesterone and insulin-like growth factor 1 (IGF-1), all of which are associated with breast cancer. We investigated whether prepubertal milk exposure affects mammary gland development and carcinogenesis in rats. Sprague–Dawley rats were given either whole milk or tap water to drink from postnatal day (PND) 14 to PND 35, and thereafter normal tap water. Mammary tumorigenesis was induced by administering 7,12-dimethylbenz[a]anthracene on PND 50. Milk exposure increased circulating E2 levels on PND 25 by 10-fold (p < 0.001) and accelerated vaginal opening, which marks puberty onset, by 2.5 days (p < 0.001). However, rats exposed to milk before puberty exhibited reduced carcinogen-induced mammary carcinogenesis; that is, their tumor latency was longer (p < 0.03) and incidence was lower (p < 0.05) than in the controls. On PND 25 and 50, mammary glands of the milk-exposed rats had significantly less terminal end buds (TEBs) than the tap water-exposed controls (p < 0.019). ER-α protein levels were elevated in the TEBs and lobules of milk rats, compared to rats given tap water (p < 0.019), but no changes in cyclin D1 expression, cell proliferation or apoptosis were seen. IGF-1 mRNA levels were reduced on PND 50 in the mammary glands of rats exposed to milk at puberty. Our results suggest that drinking milk before puberty reduces later risk of developing mammary cancer in rats. This might be mediated by a reduction in the number of TEBs and lower expression of IGF-1 mRNA in the mammary glands of milk-exposed animals.
Bovine milk and dairy products are part of a daily diet for many people. However, milk contains detectable to high levels of various hormones and growth factors that have been proposed to be associated with increased breast cancer risk, including estrogens, progesterone, leptin and insulin-like growth factors (IGFs).1–4 Animal5–8 and some human studies9–12 indicate that milk increases breast cancer risk; however, most studies have reported no change in risk being associated with milk intake13–15 and some have reported a protective effect.16–18 The protective effect was seen mainly against premenopausal breast cancer.16, 17 Childhood milk or dairy consumption in humans has been associated with a reduced breast cancer risk17, 19–21; this intake may confer protection against both premenopausal and postmenopausal breast cancer.20, 21
A contributing factor to the conflicting results regarding milk consumption and breast cancer risk could potentially be a difference in response to hormones present in milk, depending on the developmental stage of the breast at the time of exposure. It has been suggested that timing of exposures to estrogenic compounds, including endogenous hormones and those originating from the diet, determines whether they increase, reduce or have no effect on breast cancer risk.22 In animal models, for example, in utero estrogenic exposures increase breast cancer risk, and also increase the number of terminal end buds (TEBs) that are the targets of malignant transformation in the rodent mammary gland, and delay their differentiation.23 Prepubertal estrogenic exposures, in contrast, reduce mammary cancer risk and are associated with a reduction in the number of TEBs, a reduction in cell proliferation and increased apoptosis within the TEBs.24, 25
Animal studies investigating the role of whole, low-fat or nonfat milk on mammary tumorigenesis have all focused on the effect of milk exposure after treatment with the mammary carcinogen 7,12-dimethylbenz(a)anthracene (DMBA).5–8 In these studies, milk feeding was initiated 24 hr to 1 week after DMBA administration and continued for 20 weeks. Milk was found to increase mammary tumor incidence, number of tumors and tumor volume regardless of the milk fat content, when compared to rats that were given tap water or liquid that was energy and nutrient balanced with milk.
In the present study, we asked whether prepubertal cow's milk consumption affected carcinogen-induced mammary tumorigenesis and whether it was associated with changes in biomarkers previously linked to altered breast cancer risk; that is, mammary gland morphology, cell proliferation and apoptosis, and expression of estrogen receptor (ER) α and insulin-like growth factor (IGF)-1 in the mammary gland.
DMBA: 7,12-dimethylbenz[a]anthracene; E2: 17β-estradiol; ER: estrogen receptor; IGF-1: insulin like growth factor; MAPK: mitogen activated protein kinase; PND: postnatal day; TEB: terminal end bud; VO: vaginal opening
Material and Methods
Animals and treatments
Pregnant Sprague–Dawley dams were obtained from Charles River Laboratories (Wilmington, MA) and housed individually in standard rat plexiglass cages, at a constant temperature and humidity, under a 12-hr light–dark cycle. All animals were fed the AIN93G (American Institute of Nutrition) semipurified diet ad lib throughout the experiment. The day after the pups were born, male pups were removed and females were cross-fostered to avoid any litter effect. Each nursing dam had a total of 10 female pups. The study was performed in accordance with the appropriate institutional and federal regulations.
When pups were 14 days of age, dams were divided to 2 groups (6 dams per group, n = 60 female pups per group) that were given either tap water (control rats) or commercial whole milk containing 4% fat, purchased from the local supermarket (milk-exposed rats). Fresh milk was provided for the milk group every day. Since a previous study found no differences in body weight or mammary tumorigenesis between rats given tap water or a liquid that had a nutrient composition similar to that of milk,7 we chose to give tap water for the control group.
After weaning at postnatal day (PND) 22, the pups were housed 3 per cage and continued on the same liquid they received before weaning until PND 35; all rats received tap water from that age onward.
Since pups nurse until they are weaned, although they start consuming food pellets at about PND 16 and may also occasionally drink directly from the drinking bottle, rat pups in the present study were exposed to cow's milk during the first exposure week mainly through their dam. However, after weaning at PND 21 they consumed the cow's milk or water directly from the drinking bottle, for a total of 2 more weeks.
Serum estradiol level
At PND 25, 5 control and 8 milk-exposed rat pups from each group were sacrificed and blood was collected by cardiac puncture. Serum was separated and kept at −80°C until use. The level of 17β-estradiol (E2) was determined using a EIA kit from Alpco Diagnostics (Windham, NH) according to the manufacturer's instructions.
Puberty onset, vaginal opening
From PND 25 to 42, rats were examined daily to evaluate vaginal opening (VO). VO typically occurs in the Sprague–Dawley rat around PND 32–34 and represents the initial stage of attaining sexual maturity. Rats were recorded positive for VO when the vagina exhibited complete canalization and patency.
Mammary gland morphology
Mammary gland morphology is indicative of the level of susceptibility to develop mammary cancer. In particular, we and others have found that an increase in the number of TEBs proceeds an increase in the risk of developing mammary tumors (reviewed in Ref. 23). Changes in mammary gland morphology between milk and tap water-exposed rats were assessed on PND 25 and 50 in the 4th mammary glands from 6 to 8 rats per group.
Analysis of mammary epithelial structures in whole mounts was based on visual evaluation and computer-assisted image analysis. We have developed a visual scale to assess growth patterns of mammary epithelial cells.26 In this study, the following characteristics of the coded mammary glands were evaluated double-blindly using a 5-point scale for a density of (0: no structures detected, 5: numerous structures) (i) alveolar structures between the lymph node and periphery of the epithelial tree and (ii) lobular structures between nipple and lymph node. In addition, the number of TEBs at the distal periphery of the epithelial tree (defined as zone C by Russo and Russo27) was counted.
We determined cell proliferation and apoptosis, and the expression of ER-α and cyclin D1 in the mammary glands of 50-day-old rats. Six rats per group were used and their 3rd left mammary gland were fixed in 10% buffered formalin overnight at 4°C, dehydrated with graded ethanol and embedded in paraffin. The embedded tissue was sectioned (5 μm) and mounted on silane-coated glass slides. Embedding and mounting were performed at the histopathology laboratory, Lombardi Comprehensive Cancer Center (http://lombardi.georgetown.edu/research/resources/histopathology.htm).
Cell proliferation—PCNA assay
Using proliferating cell nuclear antigen (PCNA) immunohistochemistry, cell proliferation was determined. Unless otherwise noted, all materials for the PCNA assay were provided in the Vectastain Elite ABC Kit (Vector Laboratories, Burlingame, CA). Sections were deparaffinized in xylene, hydrated through graded alcohols and heated in the microwave for antigen retrieval in Antigen Retrieval Solution for 20 min. Sections were then incubated in 3% H2O2 for 15 min to block endogenous peroxides. Sections were washed in phosphate-buffered saline containing 0.1% Triton X-100 for 20 min to reduce nonspecific binding, and blocked with Vectastain Blocking Serum for 20 min. Tissue sections were incubated overnight at 4°C with the primary antibody against PCNA (Santa Cruz Biotechnology, Santa Cruz, CA) at a dilution of 1:700. After several washes, sections were treated with secondary antibody (biotinylated, anti-rabbit IgG) for 1 hr at room temperature, followed by treatment with avidin and biotinylated horseradish peroxidase complex for 30 min at room temperature. Sections were washed, and antigen-antibody complex was visualized by incubation with the chromogen, 3,3′-diaminobenzidine for 1 min, then washed and counterstained for 45 sec with Vector's Hematoxylin QS Nuclear Counterstain. Proliferation index was determined by calculating the percentage of cells that had positive PCNA staining (only darkly stained cells were counted) separately in TEBs, lobule-alveolar structures (LAs) and ducts among 1,000 cells per structure and per gland. Slides were blindly evaluated with help of the Image Tool software.
Apoptosis—TUNEL (terminal deoxynuclotidyl transferase dUTP nick end labeling) assay
The nuclei containing degraded DNA in mammary gland sections were stained by using the TUNEL assay, an in situ apoptosis ApopTag Peroxidase detection kit (Millipore, Billerica, MA, S7101), as recommended by the manufacturer and as described previously.28 The proportion of cells undergoing apoptosis was determined by calculating the percentage of apoptotic cells through both positive staining and histological evaluation of at least 1,000 cells per structure (TEBs, LAs or ducts).
ER-α and cyclin D1 protein expression
For determination of ER-α and cyclin D1 protein expression, mammary tissue sections were initially handled as described in the PCNA assay. These sections were then incubated overnight at 4°C with primary antibodies against ER-α (MC-20, rabbit polyclonal IgG) at the ratio of 1:100, or cyclin D1 (DCS-6, mouse monoclonal IgG) (Santa Cruz Biotechnology) at the ratio of 1:700. After several washes, sections were treated with secondary antibody (biotinylated, anti-goat IgG and anti-mouse for ER-α and cyclin D1), and the following procedure was identical to the previously described procedure for PCNA staining.
IGF-1 mRNA expression
The 3rd right mammary glands from six 50-day-old rats per group were obtained for real time PCR analysis of insulin-like growth factor 1 (IGF-1) mRNA expression. Mammary tissue was collected and immediately placed on dry-ice. The RNA was purified using the RNeasy Lipid Tissue Mini Kit (Quiagen, Valencia, CA) according to the manufacturer's instructions. cDNA was reverse transcribed from 50 μg/ml of total input RNA using Taqman Reverse Transcription Kit as described by the manufacturer (Applied Biosystems, Foster City, CA). Real time PCR was performed with ABI Prism 7900 Sequence Detection System with PCR master Mix, primers and probes for IGF-1 (Rn_00710306_m1) (Applied Biosystems). The probe was conjugated to 6-carboxy-flourescein phosphoradidite (FAM dye) at the 5′ labeled end with a nonfluorescent quencher located at the 3′ labeled end. The 18S RNA from Applied Biosystems was used as an endogenous control. All assays were run on 384-well plates in triplicate for the target gene and the endogenous control. Results were assessed by relative quantification of gene expression using the ΔΔCT method.
At PND 50, 49 rats in the control group and 24 in the milk group were given 10 mg of 7,12-dimethylbenz[a]anthracene (DMBA) (Sigma Chemical Co., St. Louis, MO) by oral gavage. In our prior studies, 10 mg DMBA was shown to induce tumors in approximately two-thirds of the control group and thus enables assessment of both reduction and increase in tumorigenicity.24 Starting at 6 weeks after DMBA administration, animals were examined for tumors by palpation once a week. Tumor growth was measured using a caliper and the length, width and height of each tumor was recorded. The endpoints for data analysis were (i) latency to tumor appearance, (ii) the number of animals with tumors (tumor incidence) and (iii) the number of tumors per animal (tumor multiplicity). Animals were sacrificed when the tumor burden was ∼10% of the total body weight. All remaining animals, including those that did not develop tumors were sacrificed 18 weeks after DMBA administration.
The results of serum estradiol levels, IGF-1 mRNA, and some mammary tumor endpoints (latency and multiplicity) were analyzed using the t-test. The number of proliferating or apoptotic cells, and protein levels of ERα and cyclin D1 were determined using 2-way ANOVA, with milk exposure and mammary epithelial structures (lobules, TEBs and ducts) as independent variables. Mammary gland morphology at the ages of 25 and 50 days (density of alveolar buds and lobules, and number of TEBs) was also analyzed using 2-way ANOVA. Kaplan–Meier curves were used to compare differences in VO and tumor incidence, followed by the log-rank test. All tests were performed using the SPSS SigmaStat software, and differences were considered significant if the p-value was less than 0.05.
Serum estradiol levels and body weight
Serum E2 levels, measured on PND 25, were higher in the rats exposed to milk from PND 14 onward than in the control rats consuming tap water throughout life (t = 4.49, df = 11, p < 0.001) (Fig. 1a).
Body weights were determined on PND 50 and 75. No differences between the control and milk-exposed rats were seen (Fig. 1b).
Prepubertally milk-exposed rats exhibited earlier puberty onset, determined by assessing the age of VO, compared to the control group (Log Rank = 23.755, p < 0.001) (Fig. 2). The age at which 50% (30 of 60 rats) of the rats showed VO was 35.5 and 33.0 for the control and milk groups, respectively.
Tumor latency (the time between DMBA administration and appearance of the first detectable tumor per animal) was longer in the milk-exposed rats than in the controls rats (t = 2.19, df = 71, p < 0.03) (Table 1). The proportion of rats per group that developed tumors (tumor incidence) was lower in the milk-exposed animals than in the control group (Log Rank = 3.84, p < 0.05) (Fig. 3). The average number of tumors per animal (multiplicity) was lower in the rats that had milk during prepubertal life than in the controls, but the difference failed to reach statistical significance (t = 1.62, df = 71, p < 0.10) (Table 1).
Table 1. Effects of prepubertal milk exposure on DMBA-induced mammary tumorigenesis
Mammary gland morphology
Total number of TEBs were counted in the mammary glands of 25- and 50-day-old rats. The data indicated that mammary glands of rats exposed to milk before puberty contained significantly less TEBs than the glands of tap water controls (F(1,18) = 6.67, p < 0.019) (Fig. 4a). In accordance with previous reports,27 mammary glands of 25-day-old rats contained significantly more TEBs than the glands of the 50-day-old rats (F(1,18) = 39.72, p < 0.001).
The densities of alveolar buds and lobules in the mammary epithelial tree were determined blindly using a visual scale. No differences in alveolar bud density were noted between the rats exposed to tap water or those consuming milk before puberty (Fig. 4b). Density of lobules, however, was affected by age, with mammary glands obtained on PND 50 containing significantly more lobules than the glands obtained on PND 25 (F(1,18) = 37.81, p < 0.001) (Fig. 4c). In addition, the milk-exposed rats tended to have lower lobular density than the control rats (F(1,18) = 4.02, p < 0.060).
Cell proliferation and apoptosis
Prepubertal milk exposure did not affect mammary cell proliferation, assessed by PCNA staining, or the number of apoptotic cells, assessed using TUNEL, in lobulo-alveolar structures, TEBs or ducts, when compared to the control rats and determined on PND 50 (Fig. 5). Different epithelial structures, however, exhibited significantly different levels of PCNA staining (F(2,30) = 5.26, p < 0.011) or apoptotic cells (F(2,27) = 7.68, p < 0.002). TEBs contained more proliferating cells and cells that underwent apoptosis than ducts did (p < 0.008 and p < 0.002, respectively).
ER-α and cyclin D1 expression
ER α protein levels were determined by immunohistochemistry and quantitated using a visual scale that took into account the percentile of cells stained positive (score 0–5) and staining intensity (score 0–3). When these scores were combined, the mammary glands of milk-exposed rats on PND 50 contained more ER-positive cells than the glands of the control rats (F(1,27) = 33.83, p < 0.001) (Fig. 6a). This difference was seen in the lobules and TEBs, but not in the ducts (F for interaction F(2,27) = 3.70, p < 0.038). In the control rats, the first 2 structures contained fewer ER-α positive cells than the ducts (F(2,27) = 5.50, p < 0.01), whilst no differences across different structures were seen in the milk-exposed group.
Cyclin D1 protein levels were not affected by prepubertal milk exposure (Fig. 6b), and neither was the expression different in the different epithelial structures.
IGF-1 mRNA expression
Milk intake during prepuberty was associated with a significantly reduced IGF-1 mRNA expression in the mammary glands on PND 50 determined using real time PCR (t = 2.58, df = 9, p < 0.03) (Fig. 7).
The only source of nutrition for newborn mammals is species-specific milk, obtained via nursing. After weaning, milk is not consumed, except by humans who continue to consume milk and dairy products throughout their life. In addition, the constituents of cow, human and rat milk are all different.29, 30 We found that cow's milk intake before puberty onset reduced the risk of developing mammary tumors in rats. This result is in line with human studies suggesting that childhood consumption of cow's milk is associated with a reduced breast cancer risk.17, 19–21 Previous animal studies have shown that milk exposure in adulthood promotes the growth of carcinogen-induced mammary tumors.5–8 Findings from human studies are inconsistent.9–11, 13–18 These conflicting results in rats and women may reflect the fact that the effects of milk are dependent on the age when cow's milk is consumed.
It is becoming increasingly clear that the age when an individual is exposed to various dietary compounds or hormones determines how they affect breast cancer risk.22 For example, an exposure to estrogens before24 and shortly after puberty onset,31 particularly at the levels that mimic high pregnancy estrogenic environment,32 provides a life-long protection against mammary tumorigenesis in animal models. In contrast, estrogenic exposures after mammary tumor initiation promote the malignant growth.33, 34 Milk contains high levels of estrogens and other hormones and growth factors.35 Milk obtained from nonpregnant cows and cows during the first pregnancy trimester contains 0–60 ng/l estrogens (free and conjugated), and during the 2nd and 3rd trimester the levels increase to above 1,600 ng/l.1 Calculations for a typical Western dairy herd, done by using the dynamic SimHerd model,36 show that 42% of commercial milk is produced by pregnant cows, of which half is from cows in their 2nd and 3rd trimester. However, because of the high levels of estrogens in late pregnancy, estrogen levels in commercial milk are high.35 Estrogen levels are not dependent on milk's fat content due to conjugated estrogens, which compromise most of these hormones, being stored in the aqueous fraction of milk.4, 35
We measured circulating E2 levels on PND 25 and found that the levels were 10-fold higher in the milk-group than in the tap-water drinking controls. However, since VO indicating puberty onset occurred on average 2.5 days earlier in the milk-exposed pups than in the controls (PND 33 vs. 35.5), and we did not measure the content of estrogens in the milk, it is possible that the increase reflected at least partly the approaching onset of ovarian estrogen production. Nevertheless, our finding is in line with a previous study indicating that adult rats consuming milk had higher circulating levels of estradiol and estrone than rats kept on tap water.6 Further, consumption of milk in humans has been reported to lead to an increase in E2 levels.37 It is therefore possible that reduced susceptibility to mammary tumorigenesis in rats, and perhaps in humans, who consumed milk before puberty, is caused by an increase in prepubertal estrogenic activity.
In addition to reducing later mammary cancer risk, prepubertal estrogenic exposures, including an exposure to genistein that is a phytochemical in soy with estrogenic properties, alter mammary gland morphology and expression of ER-α.24, 38 Changes in the mammary gland include a reduction in the number of TEBs.23 We found that prepubertal milk exposure also reduced TEBs. Since TEBs are the sites of malignant transformation in a rodent mammary gland,27 and over 90% of human breast cancers originate from a similar structure in the human breast, called terminal ductal lobular units,39, 40 prepubertal milk exposure may reduce later mammary cancer risk by eliminating structures that give rise to cancer.
The role of ER-α in affecting breast cancer risk remains to be determined. Although binding of E2 to the ER-α and the subsequent activation of this receptor induces estrogen-mediated increase in breast cancer cell proliferation,33, 34 and some studies suggest that high levels of ER-α expression in the normal mammary tissue are predictive of high breast cancer risk,41 some studies link high mammary ER-α expression to low breast cancer risk.42 These contrasting findings may reflect multiple roles of this receptor in the mammary gland. On one hand, estrogens promote growth via ER-α, but on the other hand, ER-α is expressed in the differentiated luminal mammary cells43 and not in mammary stem cells that are proposed to be the cell of origin of breast cancer.44 We found that ER-α protein levels were significantly higher on PND 50 in the mammary glands of rats exposed to milk before puberty onset. This increase was not associated with any significant changes in mammary cell proliferation, suggesting that downstream targets of ER-α activation by milk do not include genes that are linked to increased cell proliferation. Consistent with this conclusion, no changes in mammary cyclin D1 expression was seen between the milk and control groups. Thus, the increase in ER-α expression by milk may reflect increased population of differentiated luminal cells.
Epidemiological studies indicate that women who entered puberty early exhibit increased breast cancer risk. It therefore seems contradictory that prepubertal estrogenic exposures that reduce breast cancer risk, accelerate puberty onset,24 also found here in rats drinking milk and exhibiting elevated E2 levels. However, since puberty onset is determined by multiple factors, some of which may be present in the milk and/or altered in an individual consuming milk, the origins of earlier VO in the present study is not known. Further, we have previously proposed that the link between early puberty onset and increased breast cancer risk reflect in utero hormonal environment22 that can both accelerate puberty onset and increase later mammary tumorigenesis, and not pubertal exposures.
Milk contains insulin-like growth factors, including both IGF-1 and IGF-2.45 The levels are further increased in cows given recombinant bovine growth hormone to improve milk yield.46 Children who consume milk have higher circulating levels of IGF-1 than those who do not.47–49 When IGF-1 levels have been measured in adult individuals who consumed milk during childhood but not regularly thereafter, it has been found that their levels are reduced, when compared to nonmilk drinkers.50, 51 We did not measure circulating IGF-1 levels in the current study, but if they are reduced in adulthood, this could explain a reduction in mammary tumorigenesis in the rats drinking milk before puberty. IGF-1 might play a role in the etiology of premenopausal breast cancer.52 Further, transgenic mice over-expressing IGF-1 in the mammary gland show increased susceptibility to carcinogen-induced mammary tumorigenesis.53 IGF-1 is also a potent mitogen in ER-positive breast cancer cell lines.54 IGF-1 enhances cancer progression through ERα activation via the mitogen-activated protein kinase (MAPK) pathway,55 the MAPK pathway being a key player in inducing cell proliferation closely linked to breast cancer.56
IGF-1 is expressed both in the stroma and epithelium, where it plays a role in mediating the proliferation of epithelial cells and in inducing normal ductal branching, respectively.57 IGF-1 is also important for TEB formation.58 We determined the expression of IGF-1 mRNA in the mammary glands of milk-exposed and control rats, and found that the expression was significantly reduced on PND 50. This finding is in agreement with lower number of TEBs in the milk-exposed rats, suggesting that the observed reduction in mammary tumorigenesis in the prepubertally milk-exposed rats may be related to the downregulation of IGF-1 mRNA in the mammary gland.
In summary, we found that prepubertal intake of cow's milk reduces later susceptibility to develop mammary tumors. The protective effect may have been caused by an increase in prepubertal estrogenic environment that is known to reduce later mammary cancer risk in rats.24 The protective effect might also be related to a long-lasting reduction in mammary IGF-1 expression and the number of TEBs.
The authors thank Dr. Walter C. Willett at Harvard School of Public Health for providing the idea for this study and for his comments concerning the manuscript.