A maternal high-fat diet during pregnancy in rats results in a greater risk of carcinogen-induced mammary tumors in the female offspring than exposure to a high-fat diet in postnatal life

Abstract

The association between a high-fat diet and breast cancer risk is controversial. We hypothesized that the exposure of rats to a high-fat diet in uterovia the maternal diet would result in a greater risk of carcinogen-induced mammary tumors than high-fat diet exposure in postnatal life. Rats were exposed to a high-fat diet with 40% of the energy source as safflower oil in utero (In utero group), at postnatal days 30–50 (Puberty group), postnatal days 150–170 (Adult group), postnatal days 1–230 (Postnatal group) or for their whole life from in utero (Whole group). Chow diet-fed rats were used as the Nonexposure group. Mammary tumor incidence was significantly higher in the In utero (60%), Postnatal (61%) and Whole (91%) groups than in the Nonexposure group (32%), but there was no significant difference between the Puberty (44%), Adult (44%) and Nonexposure groups. Arachidonic acid levels were 10 times higher in mammary tumor tissue than in the normal mammary gland across all groups and were positively correlated with tumor weight. We conclude that the timing, but not the duration, of high-fat diet exposure makes rats more susceptible to carcinogen-induced mammary tumors and that exposure in utero to a maternal high-fat diet during pregnancy is more important in increasing the risk of mammary tumors in the female offspring than exposure of the offspring to the same high-fat diet later in life. © 2009 UICC

Breast cancer is the leading cancer in women worldwide1 and about two-thirds of cases have hormone-dependent cancers.2–4 It has been suggested that hormones, diet and genetics may be associated with breast cancer risk.5 However, genetic studies have shown that no more than 10% of breast cancers result from defects in breast cancer gene 1 or 2.6–8 Although Caucasian women in North America and Western Europe have a higher incidence of breast cancer than Chinese and Japanese women, Asian-American women whose grandparents migrated to the United States have the same incidence of breast cancer as other American women.9, 10 A number of studies have suggested that lifestyle, especially diet, plays a more important role than genetic differences in causing breast cancer.11–14

There is substantial evidence from human and rodent studies for a positive correlation between high-fat intake and breast cancer risk,15–19 but the association has been challenged.20, 21 It has been proposed that exposure to nutrients in early life is associated with the risk of chronic diseases, such as coronary heart disease, stroke, hypertension, diabetes and cancer,22, 23 that breast cancer may have a fetal origin24–26 and that the timing of dietary fat intake may modify breast cancer risk.27 In this study, we hypothesized that the timing, but not the duration, of high-fat diet exposure was critical for breast cancer risk and that rats exposed to a high-fat diet in uterovia the maternal diet would have a greater risk of breast cancer than those exposed in postnatal life.

The study was designed to evaluate at which stage of life a high-fat diet exposure resulted in the greatest risk of breast cancer. We hypothesized that a maternal high-fat diet during pregnancy would result in a greater incidence of carcinogen-induced hormone-dependent mammary tumors in the female offspring than exposure to the same high-fat diet during postnatal life. A high-fat diet containing 40% of the energy source as safflower oil was fed to pregnant dams during gestation or to female rats during puberty, as an adult, throughout postnatal life, or from in utero throughout their whole life. We examined mammary tumorigenesis and the fatty acid composition of the mammary tumor and the normal mammary gland in both tumor-bearing and tumor-free rats. We also examined the effect of the high-fat diet on serum estradiol levels in the pregnant dams.

Material and methods

Animals and study design

Sprague-Dawley rats (7-weeks-old) were obtained from the National Laboratory Animal Center (Taipei, Taiwan) and housed in a humidity controlled room at 24°C ± 1°C on a 12 hr light–dark cycle with free access to tap water and diets. The protocols and animal treatments used in this study were approved by the Animal Care and Use Committee of the National Taiwan University College of Medicine. The study design is shown in Figure 1. Chow diet-fed 10-week-old female rats were mated and conception was confirmed by the presence of vaginal plugs. They were then divided into 6 groups and fed different diet regimens.

Figure 1. Study design for female rats fed a high-fat diet with 40% of the energy source as safflower oil at different life stages to evaluate whether the timing or the duration of high-fat diet exposure makes the offspring more susceptible to DMBA-induced mammary tumors. Nonexposure group, chow diet for both dams and pups. In utero group, the dams were fed the high-fat diet during pregnancy, then were changed to chow diet during lactation and the pups were maintained on the chow diet until sacrifice. Puberty group, chow diet for both the dams and pups, except the female pups were fed the high-fat diet from 30 to 50 days old. Adult group, same as the Puberty group, but fed the high-fat diet from 150 to 170 days old. Postnatal group, the dams were fed the chow diet during pregnancy and were switched to the high-fat diet during lactation and the pups were fed the high-fat diet. Whole group, high-fat diet for the dams and pups. All female pups (at 55 days old) were given 10 mg of DMBA by oral gavage to induce hormone-dependent mammary tumors. The presence of mammary tumors was checked by palpation once a week until sacrifice at 230 days old, 25 weeks after DMBA administration.

The dams in the Nonexposure (n = 5), Puberty (n = 6) and Adult (n = 7) groups were fed normal chow diet with 13.5% of the energy source as fat with linoleic acid (18:2n − 6) and n − 3 fatty acids accounting for, respectively, 30 and 3.4% of the total fatty acids (5001, LabDiet, Table I) during pregnancy and lactation. After weaning at postnatal day 21, the pups were maintained on the chow diet and, at 28-days-old, the female offspring were assigned to the Nonexposure (n = 31), Puberty (n = 34) or Adult (n = 32) groups and fed chow diet throughout life, except that the Puberty and Adult groups were fed a high-fat diet (Table I) with 40% of the energy source as safflower with 18:2n − 6 and n − 3 fatty acids accounting, respectively, for 24 and 0.4% of the total fatty acids, prepared in our laboratory, for days 30 to 50 or days 150–170, respectively.

Table I. Ingredients and Fatty Acid Composition of the High-Fat and Chow Diets

Ingredient (g/kg diet)	High-fat diet1	Chow diet (LabDiet 5001)
1 The composition of the high-fat diet was modified from that of the AIN 76 purified diet. The amounts of casein, methionine, fiber, vitamin mixture, mineral mixture and choline in the high-fat diet were adjusted to maintain the same nutrient/energy ratio as the AIN-76 purified diet. 2 LabDiet 5001 vitamin mix and mineral mix.28 3 AIN 76 vitamin mix, AIN 76 mineral mix.29 4 The data are presented as the mean ± SEM (n = 5). Fatty acids that account for less than 0.5% of the total fatty acids are not shown.
Fat		50 (ether extract)
Safflower oil	200	57 (acid hydrolysis)
Protein		239
Casein	238
DL-Methionine	3.5
Carbohydrate		487
Corn starch	150
Sucrose	294.3
Fiber		51
Alphacel	58.8
Vitamin mix		2.52
AIN 76 vitamin mix3	11.8
Mineral mix		702
AIN 76 mineral mix3	41.2
Choline chloride	2.4
Fatty acid composition (%)4
Myristic acid, 14:0	0.2 ± 0.1	2.4 ± 0.4
Palmitic acid, 16:0	8.1 ± 0.4	24.1 ± 1.1
Stearic acid, 18:0	3.0 ± 0.1	7.5 ± 0.5
Palmitoleic acid, 16:1n − 7	0.0 ± 0.0	3.0 ± 0.1
Oleic acid, 18:1n − 9	63.1 ± 1.2	28.1 ± 1.1
Vaccenic acid, 18:1n − 7	0.6 ± 0.1	1.9 ± 0.3
Linoleic acid, 18:2n − 6	24.0 ± 1.0	29.5 ± 0.9
α-Linolenic acid, 18:3n − 3	0.4 ± 0.0	1.6 ± 0.2
Eicosapentaenoic acid, 20:5n − 3	0.0 ± 0.0	1.1 ± 0.1
Docosahexaenoic acid, 22:6n − 3	0.0 ± 0.0	0.7 ± 0.1
Fat, % of energy	40	13.5 (ether extract)
Protein, % of energy	21	28.5
Carbohydrate, % of energy	39	58

In the Postnatal group, the dams (n = 6) were fed chow diet throughout pregnancy, then were switched to the high-fat diet during lactation and their offspring (n = 33) were maintained on this diet till sacrificed.

In the In utero group, the dams (n = 7) were fed the high-fat diet from day 1 of gestation throughout pregnancy, then were changed to the chow diet during lactation and the offspring (n = 37) maintained on this diet until sacrificed.

In the Whole group, the dams (n = 5) were fed the high-fat diet from day 1 of gestation throughout pregnancy and lactation and the offspring (n = 22) were maintained on this high-fat diet through their life.

Age of onset of puberty was measured by examining vaginal opening daily, starting on postnatal day 28.

Diet composition

The high-fat diet (Table I) containing 20% of oil by weight was modified from the AIN76 purified diet to maintain the same nutrition density.29 Instead of the corn oil used in most previous studies, which contains high levels of the essential fatty acid 18:2n − 6, but none of the other essential fatty acid, α-linolenic acid (18:3n − 3), the oil used was safflower oil, which contains both of these and is also oleic acid (18:1n − 9)-rich, in contrast to corn oil. All diet ingredients were obtained from MP Biomedicals (Ohio, USA), except the methionine and choline, which were from Sigma-Aldrich Chemical (St. Louis, MO), and the safflower oil, corn starch and sucrose, which were purchased from a local supermarket.

Tumor induction and mammary tumorigenesis

Mammary tumors were induced by single intragastric administration of 10 mg of 7,12-dimethylbenzanthracene (DMBA) (Sigma) (10 mg/ml in peanut oil) to the 55-days-old female offspring. The DMBA-induced mammary tumor is an adenocarcinoma resembling human hormone-dependent breast cancer.30–32 The rats were checked once a week for mammary tumors by palpation, and the appearance of palpable mammary tumors and the number of tumors were recorded. The latency was calculated as the average week of appearance of the first tumor. The rats were killed at 230 days old, i.e., 25 weeks after DMBA administration.

Serum estradiol levels in the pregnant dams

The effect of the high-fat diet on serum estradiol levels in the pregnant dams was examined using a separate group of Sprague-Dawley rats. Chow diet-fed 10-week-old female rats were mated and conception confirmed by the presence of vaginal plugs. The dams were then fed either the high-fat diet or the chow diet and blood collected by cardiac puncture at gestation day 19 and serum immediately prepared by centrifugation and stored at −80°C until analysis. Serum estradiol levels were measured using an EIA kit (Cayman Chemical, MI) according to the manufacturer's instruction.

Lipid analysis

The rats were anesthetized with CO₂ and decapitated. The mammary tumor and mammary glands were dissected, weighed, frozen in liquid nitrogen and stored at −80°C until analysis. Total lipids were extracted from aliquots of tissue homogenate, converted to their methyl esters and analyzed by gas chromatography.28 The fatty acid composition of the total lipids was expressed as the weight % of total fatty acids.

Statistical analysis

All statistical analyses were performed using the SAS program (version 9.1.3, SAS Institute, Cary, NC). A 2-sided p ≤ 0.05 was considered statistically significant. Continuous data are presented as the mean ± SEM, whereas proportions were computed for categorical data. The mean differences in the characteristics of the mammary tumor with continuous measurements (tumor weight, tumor number, etc.) between the 6 groups were analyzed by one-way ANOVA, followed by Tukey's test for multiple comparisons. The time to the occurrence of mammary tumor was analyzed using the log rank test and further analyzed using Cox's proportional hazards model. Basic model-fitting techniques for variable selection, assessment of goodness-of-fit and regression diagnostics were used in regression analysis to assure the quality of analysis results.

Results

Serum estradiol levels in the pregnant dams

Serum estradiol levels at gestation day 19 were significantly higher in rats fed the chow diet before and during pregnancy than in those fed the chow diet before pregnancy and the high-fat diet during pregnancy (Fig. 2).

Figure 2. Serum estradiol levels at gestation day 19 in pregnant rats fed the high-fat or chow diets. The data are presented as the mean ± SEM (n = 6 rats/group). * indicates a significant difference compared with the high-fat diet exposure group (p = 0.0003).

Puberty onset and body weight

Puberty onset, determined by examining vaginal opening, occurred significantly earlier in the In utero, Puberty, Postnatal and Whole groups than in the Nonexposure and Adult groups (Table II). There was no significant difference in time of puberty onset between the Nonexposure and Adult groups. The mean age of puberty onset ranged from postnatal day 37.0 to 39.4 in the In utero, Puberty, Postnatal and Whole groups and from postnatal day 42.7 to 43.5 in the Nonexposure and Adult groups, indicating that female rats exposed to the high-fat diet before puberty showed earlier vaginal opening than the chow diet ones.

Table II. Effect of the High-Fat Diet on Puberty Onset and DMBA-Induced Mammary Tumorigenesis

	Nonexposure	In utero	Puberty	Adult	Postnatal	Whole
The data are presented as the mean ± SEM. The different letters show significant differences between groups using Tukey's test.
Age at puberty (day)	42.7 ± 0.6^a	37.2 ± 0.7^b	37.0 ± 0.4^b	43.5 ± 0.5^a	39.4 ± 0.2^b	39.2 ± 0.8^b
Mean latency to the first palpable tumor (week after DMBA administration)	15.5 ± 1.2	16.4 ± 1.0	15.6 ± 1.8	16.6 ± 1.3	15.4 ± 1.3	14.1 ± 1.2
Tumor multiplicity (number of tumors/rat)	0.58^c (18/31)	1.05^bc (39/37)	0.97^bc (33/34)	0.75^c (24/32)	1.45^b (48/33)	2.32^a (51/22)
Number of tumors per tumor-bearing rat	1.8 ± 0.4	1.7 ± 0.2	2.2 ± 0.4	1.8 ± 0.3	2.4 ± 0.4	2.6 ± 0.3
Tumor weight (g) per tumor-bearing rat	5.6 ± 2.4^b	6.5 ± 1.9^b	17.9 ± 5.7^ab	11.7 ± 4.7^b	13.5 ± 4.0^b	30.8 ± 8.4^a
Tumor weight/body weight (%)	1.7 ± 0.7^b	2.1 ± 0.6^b	5.0 ± 1.5^ab	3.5 ± 1.3^b	3.7 ± 1.3^b	9.2 ± 2.3^a
Body weight (g)	290 ± 8^bc	306 ± 7^b	338 ± 6^a	353 ± 11^a	276 ± 5^c	351 ± 18^a

Mean body weight at sacrifice was significantly higher in the Whole, Puberty and Adult groups than in the Nonexposure, In utero and Postnatal groups, with no significant difference between the Nonexposure and In utero groups.

Mammary tumorigenesis

Tumor latency

The appearance of mammary tumors was examined by palpation once a week after DMBA administration. The first tumor appeared on week 3 in the Puberty group, week 7 in the Adult, Postnatal and Whole groups, week 9 in the Nonexposure group and week 10 in the In utero group (Fig. 3). There was no statistical difference in the mean latency to first tumor appearance (14.1–16.6 weeks after DMBA administration) between the 6 groups.

Figure 3. Percentage cumulative incidence of mammary tumors in rats exposed to the high-fat diet at the different life stages of In utero (n = 37), Puberty (n = 34), and Adult (n = 32) and for the different periods of Postnatal (n = 33) and Whole life (n = 22). Chow diet-fed rats were used as the Nonexposure group (n = 31). Appearance of mammary tumors was examined by palpation once a week after DMBA administration. * indicates a significant difference compared with the Nonexposure group.

Tumor incidence

The percentage of rats with mammary tumors at 25 weeks after DMBA administration was 90.9% (20/22) in the Whole group, 60.6% (20/33) in the Postnatal group, 59.5% (22/37) in the In utero group, 44.1% (15/34) in the Puberty group, 43.8% (14/32) in the Adult group and 32.3% (10/31) in the Nonexposure group (Fig. 3). The log rank test revealed that mammary tumor incidence over the whole 25 week period was significantly higher in the In utero (p = 0.0338), Postnatal (p = 0.0233) and Whole (p < 0.0001) groups than in the Nonexposure group. Tumor incidence tended to be higher, but not significantly different, in the Puberty and Adult groups than in the Nonexposure group. Tumor incidence was not significantly different between the In utero and Postnatal groups, but was significant higher in the Whole group than in the other 5 groups (p < 0.01).

The time to occurrence of mammary tumor was further analyzed using Cox's proportional hazards model for modeling the hazard ratio of mammary tumors for the different groups. Because the hazard ratios between the Puberty and Adult groups and the Nonexposure group were not significantly different from 1, the Puberty and Adult groups were not included in the subsequent regression analysis. Taking the Nonexposure group as the reference group in the fitted Cox's proportional hazards model, we found that the estimated hazard ratio was 2.2 in the In utero group (95% CI 1.0–4.6; p = 0.0435), 2.3 in the Postnatal group (95% CI 1.1–5.0; p = 0.0294) and 5.4 in the Whole group (95% CI 2.5–11.7; p < 0.0001). When the Nonexposure, Puberty and Adult groups were pooled together as the reference group, the results of the fitted Cox's proportion hazards model were similar.

Tumor multiplicity and tumor weight

Mammary tumor multiplicity (number of tumors per rat) at 25 week was significantly higher in the Whole group than in the other 5 groups, but there was no significant difference between the In utero and Postnatal groups or between the Nonexposure, Puberty and Adult groups (Table II). There was no significant difference in the average number of tumors per tumor-bearing rat at 25 week between the 6 groups, the values ranging from 1.7 to 2.6. The mean tumor weight per tumor-bearing rat and the tumor weight as a percentage of the body weight were significantly higher in the Whole group than in the other groups, except the Puberty group.

Fatty acid levels in the mammary tumor and mammary gland

There was no difference between the 6 groups in terms of the fatty acid composition of the mammary tumor or the normal mammary gland in tumor-bearing rats or tumor-free rats, so the data for the 6 groups were pooled. The fatty acid profiles of the mammary gland in tumor-bearing rats and in tumor-free rats were similar (Fig. 4). The mammary tumors showed significantly higher levels of arachidonic acid (20:4n − 6) (range 2.0 to 22.6% of total fatty acids, with a mean of 10.3%) than the mammary gland in both tumor-bearing and tumor-free rats (<1.0%, with a mean of 0.7%). The increase in 20:4n − 6 levels was accompanied by a significant decrease in 18:2n − 6 levels and in the sum of the monounsaturated fatty acids in the mammary tumor compared to the mammary gland in both tumor-bearing and tumor-free rats. Levels of docosahexaenoic acid (22:6n − 3), the sum of the n − 6 fatty acids and the sum of the saturated fatty acids were significantly increased in the mammary tumor compared to the normal mammary gland.

Figure 4. Fatty acid composition of the mammary tumor, mammary gland in tumor-bearing rats and mammary gland in tumor-free rats. The levels of 20:4n − 6 (a), 18:2n − 6 (b), the sum of the n − 6 fatty acids (c), 22:6n − 3 levels (d), the sum of the monounsaturated fatty acids (e) and the sum of the saturated fatty acids (f) are shown as the percentage of total fatty acids. The mammary tumors (n = 42), mammary glands of tumor-bearing rats (n = 29) and mammary glands of tumor-free rats (n = 21) from the 6 groups of rats (n = 2–10 per Group) were pooled. The different letters shows significant differences between tissues.

When we evaluated the association between 20:4n − 6 levels and tumor weight, surprisingly, a significant positive correlation (p < 0.0001) was found between mammary tumor 20:4n − 6 levels and tumor weight, with a Pearson correlation coefficient of 0.7939, especially at 20:4n − 6 levels greater than 10% of total fatty acids (Fig. 5).

Figure 5. Correlation of mammary tumor 20:4n − 6 levels and tumor weight. The values for the mammary tumor weight and 20:4n − 6 levels are the pooled values for all 6 groups (n = 42). The 20:4n − 6 levels are shown as the % of total fatty acids in the Nonexposure (n = 6), In utero (○, n = 8), Puberty (▾, n = 5), Adult (▵, n = 7), Postnatal (▪, n = 6) and Whole (□, n = 10) groups. 20:4n − 6 levels in the mammary tumor were positively correlated with tumor weight (r = 0.7939, p < 0.0001).

Discussion

To our knowledge, this is the first study demonstrating that the timing, but not the duration, of high-fat diet exposure makes rats more susceptible to DMBA-induced mammary tumors. We found that exposure in utero to a maternal high-fat diet during pregnancy was more important in increasing the risk of mammary tumors in the female offspring than exposure of the offspring to the same high-fat diet later in life. Moreover, mammary tumor 20:4n − 6 levels were positively correlated with tumor weight.

Although the length of exposure to the high-fat diet was the same (21 days) in the In utero, Puberty and Adult groups, the incidence of DMBA-induced mammary tumors in the In utero group was higher than that in the Puberty and Adult groups (60% vs. 44% vs. 44%, respectively), showing that the timing of high-fat diet exposure plays a critical role in determining the risk of development of DMBA-induced mammary tumors. Exposure only in utero to a maternal high-fat diet during pregnancy (In utero group) resulted in a significant increase in DMBA-induced mammary tumor incidence compared to similar exposure to the chow diet (Nonexposure group) (60% vs. 32%). Moreover, the incidence of mammary tumors was the same in the In utero group exposed to the high-fat diet for only 21 days and in the Postnatal group exposed to the high-fat diet for 230 days. In addition, mammary tumor incidence was much higher in the Whole group than in the Postnatal group (90% vs. 60%), in which the only difference in high-fat diet exposure was at the in utero stage. These results show that the timing, but not the duration, of high-fat diet exposure is important for DMBA-induced mammary tumor risk and that high-fat diet exposure in uterovia maternal intake has a greater effect on the incidence of DMBA-induced mammary tumors than high-fat diet exposure in later life. Thus, the in utero stage is a critical period for the effect of a high-fat diet on later breast cancer risk, and a maternal high-fat diet during pregnancy is more important for breast cancer risk in the female offspring than the same high-fat diet exposure later in life.

It has been proposed that breast cancer may have a fetal origin24–26 and that the timing of dietary fat intake may modify breast cancer risk.27 Exposure of the rat in utero to a maternal high-fat diet with 46% of the energy source as corn oil increases the incidence of DMBA-induced mammary tumors in the female offspring to 60% compared to the 30% seen using a low-fat diet with 12% of the energy source as corn oil.33 In addition, exposure of mice in utero to a maternal high-fat diet with 37–49% of the energy source as corn oil significantly increases the incidence of spontaneous mammary tumors in the female offspring compared to a maternal low-fat diet with 6–20% of the energy source as corn oil.34 These studies provide support for a maternal high-fat diet being an important risk factor for mammary tumorigenesis in the female offspring.

To focus on whether the timing or duration of exposure to a high-fat diet was more important, we examined the risk of mammary tumorigenesis in rats exposed to a high-fat diet in utero, during puberty, as an adult, throughout postnatal life or from in utero for the whole life and that in rats which had not been exposed to a high-fat diet. As the chow diet is the most widely used rodent diet, rats were fed this diet in the nonexposure group or during the nonexposure period. Despite the different ingredients and fatty acid composition of the low-fat diet used by Hilakivi-Clarke et al.33 and the chow diet used in this study, the tumor incidence in the chow diet-fed nonexposure group in our present study (32%) was similar to that in the low-fat diet group (30%) in the study of Hilakivi-Clarke et al., and the tumor incidence in the high-fat diet groups (exposed in utero) were also similar (both 60%) in these 2 studies, suggesting the fat content of the diet is the important factor.

It has been suggested that elevated maternal serum estradiol levels are a key factor in increasing susceptibility to later breast cancer development in the female offspring.24, 35 Dizygotic twins, who are exposed to a higher estrogenic environment than single fetuses in utero, show an increased breast cancer risk36–39 and pregnant women who take the synthetic estrogen diethylstilbestrol to prevent miscarriage have a higher breast cancer risk in the offspring.40 However, this association is challenged by the observation that, although maternal plasma estradiol levels are higher in pregnant Asian women than in pregnant American women,41 breast cancer risk in North American countries is higher than that in East Asian countries.17, 42–44 In animal studies, a high-fat corn oil diet resulted in increased plasma estradiol levels (22 vs. 16 pg/ml) in pregnant dams at gestational day 14 compared to a low-fat corn oil diet,33 and serum estradiol levels were not changed at gestation day 12 (∼30 pg/ml), but were significantly higher (64 vs. 31 pg/ml) at gestation day 19 in pregnant dams fed a high-fat corn oil diet compared to a low fat corn oil diet.45 However, serum estradiol levels at gestation day 18 were the same (125 pg/ml) in pregnant dams fed a high-fat or low-fat (n − 3) fatty acid-enriched menhaden oil diet, and were significantly higher in pregnant dams fed a high-fat (n − 3) fatty acid-enriched menhaden oil diet compared to a high-fat (n − 6) fatty acid-enriched corn oil diet (125 vs. 87 pg/ml).46 We found that serum estradiol levels at gestation day 19 were significantly higher in dams fed the chow diet containing 0.46% of the energy source as (n − 3) fatty acids than in those fed the high-fat safflower oil diet containing 0.16% of the energy source as (n − 3) fatty acids (38 vs. 18 pg/ml). It would be interesting to know whether a maternal (n − 3) fatty acid-enriched diet induces higher serum estradiol levels in pregnant rats. Whether a maternal high-fat diet has an estrogenic effect or causes epigenetic changes resulting in breast cancer susceptibility in the female offspring requires further study.

In breast cancers, 20:4n − 6 levels are significantly increased to 5.9% of total fatty acids in the phospholipids compared to 1.3% in the normal mammary gland in the same human47 or from 1.9 to 11.3% in the same rat.48 In the present study, mammary tumor 20:4n − 6 levels were increased, with a mean of 10.3% of total fatty acids compared to 0.7% in the normal mammary gland in tumor-bearing or tumor-free rats, suggesting that 20:4n − 6 levels play a role in tumor growth. Moreover, we found a positive correlation between 20:4n − 6 levels and tumor weight. These results are supported by the finding that 20:4n − 6 enhances growth.49, 50

Our study provides evidence that (i) the timing, but not the duration, of high-fat diet exposure is important in the development of mammary tumors, (ii) the in utero stage is a critical period for the promoting effect of high-fat diet exposure on the development of DMBA-induced mammary tumors, (iii) there is a positive association between mammary tumor 20:4n − 6 levels and tumor weight and (iv) rats exposed to a high-fat diet before puberty show earlier vaginal opening than chow diet-fed rats.

Acknowledgements

The authors would like to thank Dr. Fu-Chang Hu and Ms. Chia-Chi Cheng from the National Center of Excellence for General Clinical Trials and Research, National Taiwan University Hospital, and Dr. Chi-Rong Li from the Consulting Center for Statistics and Bioinformatics, College of Bioresources and Agriculture, National Taiwan University for guidance and assistance in the statistical analysis.