Mammary gland development occurs over a long period of time, starting in fetal life and following successive bursts of development and involution, during either ovarian or pregnancy-lactation-involution cycles. At birth, the mammary gland consists of a small number of rudimentary ducts embedded in subcutaneous white adipose tissue. During puberty, the mammary gland develops to become a branched epithelial network of ducts that can support larger alveolar development during pregnancy and allow subsequent milk production during lactation (Howlin et al.,2006). Mammary epithelial growth and differentiation are tightly modulated by several hormonal and metabolic signals (Hennighausen and Robinson,2005). Thus, age and body weight may be important factors for mammary development, so that nutrition during the major developmental steps of this organ may be of critical importance (Sejrsen,1994).
In order to understand the impact of nutrition on mammary gland development and lactation, changes to the nutritional status of various animal models have been made in order to alter the metabolic environment. In cattle, an increased growth rate due to a high feeding level during puberty was able to reduce mammary epithelial cell proliferation in areas of active ductal expansion and thus limit mammary development and subsequent milk potential (Davis Rincker et al.,2008; Sejrsen et al.,2000). The effects of obesity on lactogenesis have also been demonstrated in the rat, where this pathology was shown to affect the chances of a successful outcome to pregnancy and lactation (Rolls et al.,1984). Similarly, in mice, diet-induced obesity resulted in lactation failure. More specifically, obese mice displayed marked abnormalities in alveolar development of the mammary gland during pregnancy, together with a marked decrease in major milk protein expression (Flint et al.,2005). Moreover, during earlier stages of development (puberty), obesity was shown to disrupt mammary ductal growth by reducing the branching frequency and width of ducts (Kamikawa et al.,2009). In humans, obesity is considered to be a major worldwide health issue and a predisposing risk for the morbidity of type 2 diabetes, hypertension, and cardiovascular diseases. Obesity has also been strongly correlated with an increased risk of mammary tumorigenesis (Stoll,2000). Furthermore, in the United States, Chapman and Pérez-Escamilla noted that women who were overweight or obese at the time of childbirth were at significant risk of failing to initiate successful lactation, or were no longer breastfeeding at two days postpartum (Chapman and Perez-Escamilla,1999). Similarly, a European study of obese women (BMI >26 kg/m2) found an association between obesity and an early cessation of breastfeeding (Riva et al.,1999) and a decrease in the normal prolactin response to suckling (Rasmussen and Kjolhede,2004).
All these studies described the effects of obesity at puberty or after puberty on mammary gland development and lactation. The present study was designed to analyze precisely the effects on mammary gland development at mid-pregnancy of a high-fat/high-sugar diet administered prior to puberty, and to develop a nutritional model that more closely resembled the human obesity condition during adolescence. Mammary gland development has been described extensively in the rabbit (Denamur,1963), together with secretion of the main milk proteins (Houdebine et al.,1985; Devinoy et al.,1988). Furthermore, fat development in this species is abdominal and interscapular and can thus be easily evaluated. The rabbit also appears to be a relevant model for human physiology since the rabbit-primate phylogenetic distance is the same as the rodent-primate distance (Graur et al.,1996). However, because rodent sequences have evolved more rapidly, rabbit gene sequences are more similar to human sequences than to those of rodents (Graur et al.,1996), which might be of considerable importance to further analyses of altered gene expression profiles in human pathologies. By subjecting female rabbits to a high-fat/high-sugar obesogenic diet from an early age (before puberty) and by evaluating their mammary development on Day 14 of pregnancy, we were able to show that an obesogenic diet induced morphological and functional changes in the mammary tissue at mid-pregnancy, which could perhaps lead to lactation defects.
Body Weight and Food Intake
Before the age of 19 weeks, the weight of the rabbits did not differ statistically between females receiving the control (C) or high-fat/high-sugar (OD) diets started at 8 weeks of age (Fig. 1). Thereafter, during the 4 weeks before mating, body weight became significantly higher in OD animals than in C animals. Before puberty, which occurs at between 9–13 weeks of age, and before mating (at between 19–23 weeks), the total energy intake was significantly increased in OD rabbits compared to C-fed animals. By contrast, OD rabbits displayed no difference in calorie intake between 14–18 weeks and during early to mid-pregnancy (24–25 weeks) (Table 1), because the food intake (g) decreased in this OD group.
Table 1. Energy Intake (total kcals per rabbit) per Each Five-week Period from the Age of Puberty to Mid-pregnancy in Control (n=5) and OD (n=4) Animalsa
Data are means ± SEM.
2734 ± 117
2800 ± 151
2244 ± 138
2270 ± 185
3146 ± 116
3098 ± 200
2800 ± 114
1816 ± 218
C versus OD
P < 0.002
P < 0.001
Number of Fetuses, Fetal and Placental Growth, and Adipose Tissue
Nine female rabbits were mated at 24 weeks of age (4 OD and 5 C animals) and were then sacrificed on Day 14 of pregnancy. The number of fetuses per doe did not differ significantly between the two groups (Fig. 2A), nor did the weight of placentas (Fig. 2C). However, the mean fetal weight was significantly lower in the OD group (0.245±0.012g in OD vs. 0.279± 0.009g in C rabbits, P < 0.03, Fig. 2B).
As shown above, body weight was significantly higher in OD animals as from 19 weeks of age. It could be inferred that the increase in body weight was the result of increased body fat mass, since the TOBEC measurement was greater in OD animals (7.5±0.3% fat in OD vs. 6.6±0.4 in C, P = 0.03). Differences in the weights of major organs (kidney, liver, and muscle) were not observed. In rabbits, adipose tissue is mainly perirenal and interscapular. After sacrifice, both types of adipose tissue were dissected and weighed. Perirenal adipose tissue was significantly heavier in the OD group (247.6±39.2g in OD vs. 141.6±9.9g in C rabbits, P < 0.05, Fig. 2D), whereas no significant difference in the weight of interscapular adipose tissue was observed.
Mammary Gland Development
Histological examinations of thin sections of mammary tissue from control and OD animals were performed on Day 14 of pregnancy (Fig. 3 and see Supp. Fig. S1, which is available online) and revealed striking differences between the two groups. Alveolar secretory structures were isolated in control animals, whereas they invaded the fat pad in OD rabbits, as observed in both global sections (Fig. 3Aa, Ab) and at higher magnification (Fig. 3Ba and Bb). Moreover, the ducts were empty in C rabbit mammary glands observed at both low and high magnification (Fig. 3Ac and Bc, respectively) whereas in the OD group they were characterized by abundant products within the lumen (Fig. 3Ad and Bd).
Morphological differences were evaluated using a quantitative analysis based on the ratio between alveolar areas per section and total section areas. This analysis revealed a significantly increased mammary epithelial content in OD rabbits (29.93±2.90% of epithelial tissue) compared to C animals (17.94±1.18%, P < 0.001). Mammary epithelial cell proliferation was quantified using Ki-67 staining. No significant difference was observed between C and OD animals (1.51±0.49% and 3.17±1.75% in C and OD rabbits, respectively).
Electron microscopy analysis also revealed morphological features that differed as a function of diet. In C females, the mammary epithelial cells (MEC) surrounding the lumen (L) contained large lipid droplets (ld) at the apex (Fig. 4A). The lumina contained a small quantity of granular, electron-dense material in which a few casein micelles and rare lipid droplets were embedded (Fig. 4A). By contrast, the lumina of OD mammary glands contained a huge amount of casein micelles as well as large lipid droplets buried in a fibrillar material (Fig. 4B).
These findings support the observation that a high-fat/high-sugar diet had markedly modified mammary gland development and induced an abnormal secretion of large lipid droplets and numerous micelles, which were detected in the alveolar and ductal lumina.
Expression and Localization of Milk Proteins and Lipids
In order to investigate the biochemical consequences of the abnormal mammary development caused by a high-fat/high-sugar diet, milk protein accumulation in mammary gland extracts was evaluated using a Western blot procedure. AlphaS1-casein (the major milk protein in rabbit milk), κ-casein (the protein that enables the formation of casein micelles), and Whey Acidic protein (WAP, the major protein of the whey fraction) were studied. As shown in Figure 5, Western blot analysis of three out of four protein extracts from OD mammary glands revealed clear bands. Some of them co-migrated with those observed in milk and corresponded to αS1- and κ-caseins and WAP. Such bands were barely observed in the mammary gland of control animals for equivalent amounts of protein loaded onto the gels (15 μg), as confirmed by BCA quantification and Coomassie blue staining (Supp. Fig. S2). For caseins, different band patterns were observed in milk and the OD group. They reflected either polymorphisms, largely described in this species (Chanat et al.,1999), or potential post-translational modifications as predicted from genomic sequences. These results are in agreement with previous studies showing that for the major caseins, several bands can be detected using immunoelectrophoresis in rabbit mammary extracts (Baranyi et al.,1995; Bösze et al.,2000) and are not observed in milk.
The localization of these proteins was examined by immunohistochemical analyses using the same αS1-, κ-casein, and WAP-specific antibodies as above. As expected at mid-pregnancy, positive staining for each milk protein was detected inside the alveolar lumena of both C and OD animals (Fig. 6). However, a strong positive signal was observed specifically in the lumena of ducts from OD rabbits (Fig. 6B, E, H), while the ducts of C animals displayed a reduced signal (Fig. 6A, D, G). These results suggest an accumulation of secreted milk proteins in the mammary glands of OD rabbits. Specific lipid staining (BodiPy) revealed numerous fat globules in the mammary ducts of OD animals but not in control rabbits (Fig. 7), confirming the presence of secreted lipids in the lumena of OD rabbits.
The present study reports the effects on mammary gland development in the rabbit of a high-fat/high-sugar diet started before puberty. This model was chosen because it could mimic human obesity during adolescence. Indeed, 24 weeks of age in rabbit development can be considered to be approximately 18 years of age in humans, and in our experiments the cafeteria diet was started before puberty and was withdrawn at a young adult age. Female rabbits fed this obesogenic diet grew initially at the same rate as the controls but they had become significantly heavier by 19 weeks of age. This increase appeared to be due entirely to increased amounts of body fat adipose tissue as compared to controls. The rabbits were mated at 24 weeks of age and sacrificed on Day 14 of pregnancy. Their fertility and prolificacy did not appear to be affected. Even if the target of our study was not to examine reproductive phenomena in detail, fetal intra-uterine growth retardation was observed, despite a similar placental weight, thus illustrating a possible decrease in placental efficiency in OD rabbits (Picone et al.,2011). These observations are in agreement with similar reports on different species in which obesity led to low birth weight (Rolls et al.,1984; Flint et al.,2005).
Histological and biochemical analyses revealed clearly that an obesogenic diet had strongly modified mammary development at mid-pregnancy. The mammary tissue of OD animals was characterized by ducts filled with abundant secretion products and developed alveoli invading the fat pad. Preliminary data (not shown) tended to indicate that this altered mammary phenotype was already observed at earlier stages of pregnancy and could be related to a modification of the hormonal status of the animals. Indeed, it is known that ovarian and pituitary hormones act according to a precise chronology to control mammary epithelial growth and differentiation. Any modification to this endocrine equilibrium may impact lobulo-alveolar formation during pregnancy and subsequent lactation (Sternlicht et al.,2006). A similar abnormal mammary phenotype had previously been reported in transgenic mice over-expressing a constitutively active prolactin (PRL) receptor (Gourdou et al.,2004), suggesting that the alterations observed in the mammary glands of OD rabbits may be related to a modification of the PRL pathway. However, no differences in the localization of the PRL receptor, or in the phosphorylation of Stat5, were observed between the two groups during preliminary immunolocalization and Western blot experiments (data not shown), but we cannot exclude an alteration to either the expression of the prolactin receptor or its signaling pathway in response to acute PRL stimulation in the OD group (Nevalainen et al.,2002). Furthermore, obesity has been shown to strongly modify the expression profile of numerous adipokines, which are integrated in a communication network between numerous organs, including the mammary gland (Ronti et al.,2006). In particular, leptin has been shown to act as an autocrine and paracrine factor to influence the development and differentiation of mammary gland (Hu et al.,2002; Silva et el.,2008). Considering that leptinemia is correlated to fat mass, high levels of leptin could be expected in OD rabbits. Such a hypothesis is supported by preliminary results obtained in mammary glands of rabbits during early pregnancy (data not shown). Moreover, previous studies showed that leptin administration during pregnancy and lactation was able to rescue lactation failure observed in leptin-deficient ob/ob mice (Chehab et al.,1996; Malik et al.,2001). These findings clearly demonstrate that leptin is required during pregnancy and after parturition to ensure development of the mammary gland and the establishment of lactation (Malik et al.,2001). The mechanism by which leptin regulates mammary growth, and in particular the existence of a direct and/or indirect effect on mammary epithelial cells, is still largely unknown even though many hypotheses have been formulated (Kamikawa et al.,2009; Thorn et al.,2006,2010).
At mid-pregnancy in the present case, the proportion of the mammary gland occupied by epithelial tissue was higher in OD rabbits, thus demonstrating enhanced mammary tissue development in these animals. However, at this stage, proliferation of the mammary epithelial cells has not been observed, as shown by Ki-67 immunostaining, which exhibited no difference between the two groups. Previous studies by our group had established that the higher levels of Ki-67 labeling in rabbit mammary gland were observed during the first week of pregnancy (data not shown). It is thus likely that increased epithelial development might be the result of an intense proliferation that had occurred earlier in pregnancy.
However, our data contrasted with some findings previously reported using different animal models, although the diets used were similar. In mouse models fed a high-fat/high-sugar cafeteria-type diet starting at or after puberty, impaired mammary development was observed before pregnancy (Kamikawa et al.,2009), at mid-pregnancy and during lactation, together with reduced milk protein expression leading to lactation failure (Flint et al.,2005). In our case, the diet was started before puberty, whereas in the rodent model the obesity-inducing diet was introduced at or after puberty. One possibility to explain this discrepancy between the findings may have been the age of the animals when the high-fat/high-sugar diet was introduced, which could be crucial. Moreover, possible differences among species can cause speculation, related to specific endocrine profiles and subsequent hormonal regulation and adipose tissue distribution. Indeed, the adipose tissue distribution we observed differed quite markedly from that observed in rodents fed the cafeteria-type diet. Indeed, in rodents fed a cafeteria-type diet (Flint et al.,2005), mammary adipose tissue was intra-epithelial whereas in the rabbit it remained peri-epithelial. The rabbit phenotype appears to be closer to the human situation where adipose deposits in specific regions of the body (including the mammary gland) can be observed, even in the absence of massive obesity.
In cattle, several studies have demonstrated the deleterious effects on mammary growth and milk composition of prepupertal feeding with a high-energy diet. In particular, elevated nutrient intake reduces prepubertal mammary development by impairing epithelial cell proliferation (Brown et al.,2005; Sejrsen and Purup,1997). However, when a high-energy diet was introduced during gestation, lactation efficiency increased (Park,2005), thus revealing the importance of the reproductive status of animals in the event of nutritional changes.
Although it is not possible to infer future milk production from the data presented here, we demonstrated important qualitative and quantitative variations in milk protein levels in high-fat/high-sugar-fed rabbits at mid-pregnancy, and in particular a precocious expression of αS1- and κ-caseins and WAP. This accelerated milk protein expression, and the increased mammary duct lipid content, could lead to subsequent lactation deficiencies. Indeed, lactation failure has been induced by over-expressing WAP in the pig and mouse at early stages of pregnancy (Shamay et al.,1992; Burdon et al.,1991,1999).
In conclusion, the data presented in this report strongly support a critical effect on mammary gland differentiation of high-fat/high-sugar nutrition starting before puberty and continuing throughout pregnancy. However, the differences that we observed at mid-pregnancy may have reflected the effects of diet on mammary development during puberty, or during pregnancy, or both. Thus, a better understanding of the importance of both these periods to mammary gland differentiation is clearly an important future goal. Moreover, it is tempting to speculate that the accelerated mammary gland development at mid-pregnancy that was demonstrated during this study may have important effects on milk quantity and/or composition, a factor that is currently under study.
Animals and Experimental Design
All animal studies were carried out in compliance with French regulations on animal experimentation and with the authorization of the French Ministry of Agriculture (Animal Health and Protection Directorate, accreditation number 78-119).
Ten female rabbits (New Zealand White, 1077-INRA) were housed individually in an indoor facility under controlled light conditions (8 hr light/16 hr darkness except from the week before mating when they were under 16 hr light/8 hr darkness) and temperature (18°C). Before puberty, at the age of 8 weeks, the rabbits were randomly divided into two groups: one group was fed ad libitum with a custom-made standard laboratory diet designed to satisfy their needs (Control Diet, C; De Blas and Mateos, 2010) whereas the other group was fed an obesogenic cafeteria-type diet (OD) based on the control diet and adapted from that described by Guo and Jen (1995) (Table 2). The obesogenic diet, increased in energy levels (+18% compared to the control diet), was supplemented with lard (60 g/1,000 g, +276% fat compared to the control diet) and beet sugar (100 g/1,000 g, +269% carbohydrates compared to the control diet). Food intake and body weight were monitored on a weekly basis. Adiposity was measured in vivo at 11, 14, and 19 weeks of age using total body electrical conductivity (TOBEC), as recently established in a rabbit model (Fortun-Lamothe et al.,2002). After 16 weeks of the diet (24 wk of age), nine females were mated (one female died, an event not linked to the diet) and on Day 14 of pregnancy they were sacrificed. Mammary tissue, placenta, fetuses, liver, kidneys, and adipose tissue (perirenal and interscapular) were collected and weighed. Mammary samples were subsequently stored, fixed, or snap-frozen until use.
Table 2. Composition and Energy Content of the Control (C) and Obesogenic Diets (OD)
Theoretical chemical composition
Digestible energy (Kcal/kg)
Rough cellulose (%)
Acid detergent fiber (%)
Neutral detergent fiber (%)
Crude protein (%)
Digestible protein (%)
Histological and Ultrastructural Analyses
Abdominal mammary glands were dissected and the adipose tissue and muscle were removed. In each animal, the left lower mammary gland was excised and divided into several samples. For histology and immunohistochemistry, mammary samples were fixed in 4% paraformaldehyde for 10 min at 4°C. The samples were then cryoprotected in 40% sucrose, embedded in TissuTek (Sakura, Torrance, CA), and kept in liquid nitrogen before storing at −80°C. Five- micrometer-sections, lying at least 100 μm apart, were mounted on slides. For histology, the slides were stained with Hematoxylin and Eosin (H&E, Sigma, Saint Quentin fallavier, France) and then either examined under bright-light microscopy or digitalized using a Hamamatsu nanozoomer (Tokyo, Japan). The latter technique enabled observation of the entire section. Five sections per rabbit were processed and the images were analyzed using both the Hamamatsu Nanozoomer Digital Pathology Virtual Slide Viewer and Image J software. Areas occupied by mammary epithelial tissue were measured and divided by the whole section area to generate the proportion of epithelial tissue.
For Transmission Electron microscopy experiments, the tissues were fixed with 2% glutaraldehyde in 0.1 M sodiumcacodylate buffer, pH 7.2, for 4 hr at room temperature, and then post-fixed with 1% osmium tetroxide containing 1.5% potassium cyanoferrate, gradually dehydrated in ethanol (30 to 100%), and embedded in Epon (Delta Microscopy, Ayguesvives, France). Thin sections (70 nm) were collected on 200-mesh copper grids, and counterstained with lead citrate before examination with a Zeiss EM902 electron microscope operated at 80 kV (Carl Zeiss, France). Images were acquired with a charge-coupled device camera (Megaview III) and analyzed with ITEM Software (Eloïse, France) at the MIMA2 facilities (UR1196 Génomique et Physiologie de la Lactation, INRA, Plateau de Microscopie Electronique, F78350 Jouy-en-Josas, France).
The main milk protein and prolactin receptors were localized by immunohistochemical analyses using the following dilutions of primary antibodies: guinea pig anti-rabbit-αS1-casein (1:500), guinea pig anti-rabbit WAP (1:500) antibodies (Grabowski et al.,1991), sheep antiserum rabbit κ-casein (1:400, Baranyi et al.,1995), and S46 serum developed against purified PRL-R (Waters et al.,1995). Briefly, 5-μm frozen sections were treated in 50 mM ammonium chloride for 30 min followed by permeabilization in 2% BSA, 0.05% saponin, and 0.05% sodium azide in PBS 1× for 1 hr. Primary antibodies were diluted in the same buffer and then added to the tissue sections for 1 hr at room temperature. Antibody binding was visualized with fluorescence-labeled secondary antibodies (anti-guinea pig FITC-conjugated, 1:300, and anti-sheep TRITC-conjugated, 1:300, Jackson Immunoresearch, West Grove, PA), applied to the sections in PBS 1× for 45 min. DAPI (4,6-diamidino-2-phenylindole) was diluted 1:500 in PBS 1× and applied for 3 min at room temperature.
To evaluate epithelial cell proliferation, mammary gland sections were stained with monoclonal mouse anti-rat Ki-67 (1:100, Dako, Trappes, France) and revealed with fluorescence-labeled secondary antibody (anti-mouse FITC-conjugated, 1:300, Jackson Immunoresearch).
For lipid staining, frozen sections were incubated for 20 min with 0.5 μ g/ml Bodipy 493/503 (Molecular Probes, Cergy Pontoise, France). For observations, the slides were mounted in Vectashield mounting medium (Vector Laboratories Inc., Burlingame, CA) and observed with a Leica DMRB microscope coupled with a DP50 Olympus Camera (Tokyo, Japan).
Western Blot Analysis
Protein extracts from each rabbit mammary gland were prepared in 50 mM Tris-HCl (pH 8), 137 mM NaCl, 2.7 mM KCl, 1% NP40, and 10% glycerol, containing protease and phosphatase inhibitors (Complete Mini and PhoSTOP, Roche, Meylan, France). Their protein concentration was determined using the BCA Protein Assay kit (Thermo Scientific, France). Proteins (15 μg) were separated on 10% SDS-polyacrylamide gel and transferred onto nitrocellulose filters (Protran, Schleicher & Schuell, France). The membranes were saturated for 1 hr in blocking solution (TBS-T:50 mM Tris pH7.4, 200 mM NaCl, 0.1% Tween 20, 5% defatted milk) and incubated overnight with the primary antibodies at 4°C. After incubation, the membranes were washed with TBS-T and incubated at room temperature with horseradish peroxidase-linked secondary antibody for 45 min. Immune complexes were detected using the ECL kit for autoradiography (GE Healthcare, France). For each experiment, equal protein loading between the different wells and the different gels was verified by Coomassie blue staining. The following specific antibodies were used: guinea pig anti-rabbit-αS1-casein (1:5,000), guinea pig anti-rabbit WAP (1:5,000) antibodies (Grabowski et al.,1991), and sheep anti-rabbit κ-casein (1:5,000) serum (Baranyi et al.,1995). The secondary antibodies used for immunoblotting were: anti-guinea pig (1:5,000) and anti-sheep horseradish peroxidase (HRP)-conjugated (1:10,000) (Sigma, France).
Experimental data are presented as means±SEM (standard error of the mean). Statistical analyses were performed to detect significant inter-group differences using either unpaired Student's t-test when the sample size was >30, or the Mann-Whitney U-test when the sample size was <30 (Georgin and Gouet, 2000). P ≤ 0.05 was considered to be significantly different.
We are grateful to Dr. Laurence Lamothe from INRA for her help in designing the obesogenic diet and for the staff of the Unité Commune d'Expérimentation Animale of INRA, and in particular, Michel Baratte and Gilbert Boyer, for animal care. We thank Dr. Jean Djiane for his precious advice. We are also grateful to Dr. Louis-Marie Houdebine for the gift of guinea pig anti-rabbit αS1 casein.