Unexpected Deposition of Brown Fat in Mammary Gland During Postnatal Development

Abstract

Mammary fat tissue is crucial for mammary ductal morphogenesis in both fetal and adult mice. There are two kinds of adipocytes, the energy-storing white and the energy-dissipating brown adipocyte. The precise identity of the types of adipocyte in the mammary gland has never been investigated but was always assumed to be only white fat. In this study, we show that both white and brown adipocytes are present in the postnatal mammary gland. The amount of brown adipose tissue (BAT) examined by histology and electron microscopy correlates with the transcript levels of uncoupling protein 1, which is a mitochondrial carrier expressed exclusively in BAT. Uncoupling protein 1 mRNAs are the highest during prepuberty, decrease upon puberty, and are finally undetectable in the adult mammary gland. The analysis of a BAT-depleted mouse model showed that depletion of mammary BAT in early postnatal development induces epithelial differentiation. Alveolar structures were formed along all ducts and were functional since they produced β-casein. However, mammary transplantation experiments indicated that a systemic effect was responsible for epithelium differentiation. Our data suggest that BAT negatively regulates the differentiation of mammary epithelial cells in a systemic manner during prepubertal ductal outgrowth.

DEVELOPMENT OF THE mammary gland is initiated in the embryo, but the major part of its growth occurs postnatally. After a quiescent neonatal growth period, the rudimentary epithelial ductal tree commences extension at about 3 wk of age. Large club-shaped terminal end buds (TEBs) appear, grow out through the fat pad, and disappear when the fat pad is laced with a ductal tree at around 9 wk of age. During this last mature virgin period, the highly mitotic TEBs are replaced by mitotically quiescent terminal end duct (TED) and alveolar buds. During pregnancy, the alveolar buds develop into mature alveoli and become functionally active during lactation, secreting milk products including casein. The lobulo-alveolar units are maintained after parturition during lactation until weaning, when the gland rapidly regresses to the virgin-like state in a process known as involution (1–3).

Epithelial/mesenchymal cell interactions are necessary for the proper ductal morphogenesis throughout embryonic and postnatal mammary gland development (4–6). Adipocytes are the most abundant mesenchymal mammary cell type followed by fibroblasts, migrating macrophages and eosinophils, endothelial cells, and nerve cells (7, 8). Numerous studies have demonstrated the crucial role of mammary fat tissue in the morphogenesis of the mammary parenchyma in both fetal (9, 10) and adult (11, 10) mice. The requirement of the mammary fat pad for epithelial cell growth, morphogenesis, and the limitation of the size of the mammary tree was first demonstrated by DeOme et al. (12). Transplantation studies utilizing chimeras of mammary-specific mesenchymes with mammary epithelial cells have clearly demonstrated that fat pad precursors are essential for mammary-specific pattern formation (10). For example, chimeras of salivary mesenchyma and mammary epithelium grown either in vitro (13) or in vivo (9) have resulted in salivary-like epithelial morphogenesis. In vitro studies defined more precisely the capacity of adipocytes to induce epithelial cell growth and morphological differentiation by using either the nonmammary mouse preadipocyte cell line 3T3-L1 (14–18), medium conditioned from other adipocyte cell lines (19), excised mammary gland fat pad (20), or adipocytes isolated from the mammary gland (21).

There are two kinds of adipose tissue; the energy-storing white adipose tissue (WAT) and the energy-dissipating brown adipose tissue (BAT) (22). The thermogenic action of brown adipocytes is due to a unique mitochondrial protein, called uncoupling protein (UCP). UCP is a proton translocator in the inner mitochondrial membrane and functions as a facultative uncoupler of the mitochondrial respiratory chain (23). BAT, in contrast to WAT, is highly innervated, and BAT thermogenesis is under the control of the sympathetic nervous system. Thermogenesis is acutely activated by noradrenaline acting on β3-adrenergic receptors (24) that are specifically expressed in both BAT and WAT in rodents (25). The difference in function of these adipocytes is associated with a difference in structure and anatomic localization. BAT is mainly in the interscapular, subscapular, axillary, and suprasternal regions. Brown adipocytes can be distinguished from white adipocytes easily by their multilocular lipid inclusions and numerous well developed mitochondria, whereas white adipocytes are characterized by one large lipid inclusion and very few mitochondria (26). The precise identity of the type of adipocyte in the mammary gland has never been investigated but was always assumed to be WAT (5). In this study, we established that BAT is also a component of the mammary fatty stroma, which is temporally regulated in early postnatal development. Taking advantage of different strains of mice displaying variable amount of BAT, the chemical induction of BAT, and a transgenic mouse model depleted in BAT, we investigated the role of brown fat in mammary gland development.

RESULTS

The determination of the type of adipocytes in mammary glands has never been investigated, and it has always been assumed that the mammary gland fat was composed of WAT. The present study focused on the identification of the mammary brown fat and its implication during mammary gland development.

To identify BAT in mammary tissue, three separate techniques were used. First, hematoxylin/eosin staining of longitudinal mammary gland sections showed typical white adipocytes with their unilocular cytoplasmic fat droplet and their nucleus pushed away at the periphery, and also typical multilocular brown adipocytes with a central nucleus placed among the cytoplasmic fat droplets (Fig. 1, A and B). The characteristic highly developed web of vessels in BAT was also observed in the mammary BAT with numerous eosin-stained red blood cells lined up in the microvessels between adipocytes (Fig. 1B, arrowheads), which were very few in the mammary WAT (Fig. 1A). Interestingly, BAT, strongly colored by red carmine dye in the stained mammary gland whole mount, was consistently organized around the growing duct at prepuberty between the nipple and the lymph node (dotted lines), while the WAT was evenly spread in the whole fatty stroma (solid line) (Fig. 1C). The identity of brown adipocytes was confirmed by electron microscopy and immunohistochemistry using a brown adipocyte-restricted protein UCP-1 antibody. In the mammary gland, white adipocytes display a reduced cytoplasm due to the large unilocular fat droplet with few mitochondria (Fig. 1D), whereas brown adipocytes are characterized by a high density of mitochondria with numerous cristae due to their high respiratory function (Fig. 1E). By immunohistochemistry, UCP-1 protein was specifically expressed in the cytoplasm of mammary brown adipocytes, while it was absent in white adipocytes (Fig. 1, F and G).

Fig. 1.

White and Brown Adipocytes Are Components of the Mammary Fatty Stroma A and B, Hematoxylin/eosin staining of 5-μm sections of different areas of mammary glands from 3-wk-old 129SvEv mice. A, Unilocular white adipocytes in the upper area of the mammary gland; B, multilocular brown adipocytes localized between the nipple and the lymph node (LN). C, Mammary gland whole mount stained with red carmine from a 3-wk-old SvEv mouse shows the dense brown fat area in between the nipple and the lymph node, indicated with the dashed line as opposed to the white fat delimited with the solid line. D and E, Electron micrographs of white adipocytes (D) and brown adipocyte (E) in the mammary gland. Note the numerous mitochondria enriched in cristae in brown adipocyte cytoplasm. Three large fat droplets from three different unilocular white adipocytes are seen in panel D and indicated by asterisks, whereas the cytoplasm of only one multilocular brown adipocyte is shown in panel E. F and G, Immunohistochemistry for UCP-1 in the mammary gland showing white adipocytes (F) and brown adipocytes (G). Note the restriction of UCP-1 protein for the cytoplasm of brown adipocytes. Magnification: A and B, ×400; F and G, ×1000; bars, 2 mm in panel C and 2 μm in panels D and E.

To show that the presence of BAT is universal between different strains of mice, we analyzed the transcript level of UCP-1 in three different strains at various time points during their mammary gland development (Fig. 2). UCP-1 transcripts were consistently abundant in mammary glands in 2- and 3-wk-old mice, while at 5 wk of age they decreased considerably and were completely absent at 8 wk consistent with the histological observations. Transcripts were not detected again during pregnancy, lactation, or involution (data not shown). The amount of UCP-1 transcripts correlated perfectly with the density of BAT within mammary glands from different mouse backgrounds (not shown).

Fig. 2.

Mammary UCP-1 Transcript Levels Are Consistently Abundant During Early Postnatal Development Five to 10 mice from three different strains (C57Bl/6. C3H, C57Bl/6, 129/SvEv) were killed at 2, 3, 5, and 8 wk of age. The fourth abdominal mammary glands were processed for RNA isolation. Amounts of UCP-1 transcripts are shown by Northern blot for two mice per age. Total mammary RNA (10 μg) was analyzed, and the 28S ribosomal RNA was used for the RNA loading control of each lane (bottom panels).

Taken together, the data show that BAT is one of the components of the fatty stroma of the mammary gland and that this tissue is spatially and temporally regulated, since it is mainly present between the nipple area and the lymph node area during early postnatal development.

To further investigate the role of BAT in mammary gland development, we analyzed the mammary phenotype of 4 wk-old mice, in which the formation of BAT was induced chemically or reduced by transgenic means. Excessive amounts of BAT in the mammary glands was obtained by treating the mice daily with one sc injection of a specific β3 adrenergic agonist from 12 d of age to 4 wk of age. The mammary BAT induction was evaluated by analyzing the amount of UCP-1 transcripts in mammary glands of β3 adrenergic agonist-treated mice compared with the saline-treated mice. Indeed, it has been previously reported that levels of UCP-1 mRNA correlate perfectly with the number of brown adipocytes (27). Transcript levels were dramatically induced in agonist-treated mammary glands compared with control mammary glands (Fig. 3A). Staining of mammary gland sections with hematoxylin/eosin confirmed the abundance of BAT in treated mice (not shown). Whole mounts of inguinal mammary glands from both groups were stained to examine the numbers of branches and TEBs as well as the ductal length (Fig. 3, B and C). Even though the mammary BAT was dramatically induced in treated mice, the numbers of branches and TEBs and the ductal lengths were not affected (Fig. 3, D–F).

Fig. 3.

Overproduction of Mammary BAT Does Not Affect Ductal Morphogenesis Mice were treated daily with the β3-adrenergic agonist CL 316,243 at 1500 h at a dose of 1 mg/kg body weight from d 12 to 4 wk of age. Control mice (CTL) were treated the same way with saline solution. Mice (15–20 mice per group) were killed at 4 wk of age, and their mammary glands were processed as described in Fig. 2. A, Amounts of UCP-1 transcripts are shown by Northern blot for three mice per group. Total mammary RNA (10 μg) was analyzed. Note the dramatic induction of UCP-1 transcript levels in β3-adrenergic agonist-treated mice. B and C, Typical mammary gland whole mounts of 4-wk-old mice from saline (CTL, panel B)- and β3-adrenergic agonist (β3 agonist, panel C)-treated mice are represented. Pictures were taken at the same magnification. D–F, Graphs represent means ± sem of branching numbers (D), TEB numbers (E), and ductal lengths (F) of mammary gland whole mounts obtained in panels B and C from mice treated with β3-adrenergic agonist (β3) or saline (CTL). Statistical evaluations using sem were performed with a two-tailed t test. No statistical differences were found between groups.

The partial ablation of BAT in mammary gland tissue was assessed by overexpressing the toxin a-chain under the promoter of UCP-1 that specifically kills brown adipocytes where UCP-1 expression is restricted. In this mouse model (UCP-DTA mice), it has been previously shown that UCP-1 protein content in interscapular brown fat, the largest BAT in the mouse body, was reduced by 68% at 16 d of life and this low amount was maintained as the mice aged (28). The effect of the BAT depletion on mammary ductal morphogenesis was analyzed by comparing mammary gland whole mounts from the transgenic and wild-type mice at 5 wk of age. Interestingly, the mammary gland morphogenesis of the transgenic mice was consistently affected. In 5-wk-old transgenic mice, the degree of branching complexity was enhanced, assessed by an increased number of TEBs and branches, with an obvious epithelial differentiation into premature alveoli; whereas in wild-type mice, alveoli were completely absent as expected at this early age (Fig. 4, A and B). High magnifications of the differentiated epithelial structures in transgenic mice were reminiscent of the alveoli developed at about d 5 of pregnancy (Fig. 4, C and D), whereas the ducts in wild-type mice had a smooth appearance normally seen at 5 wk of age (Fig. 4E). The identification of alveoli in transgenic mice is illustrated by the presence of round epithelial structures along the ducts (Fig. 4F, arrowheads) as seen at the beginning of pregnancy (Fig. 4G, arrowheads), compared with the continuous ductal structures in wild-type mice (Fig. 4H). In addition to the epithelial differentiation, branching numbers and TEB numbers were significantly increased in transgenic mice, without affecting the ductal length (Fig. 4, I–K). To analyze the depletion of mammary BAT in transgenic mice, UCP-1 transcript level was examined and found to be variably affected from mouse to mouse. In wild-type mice, the amount of UCP-1 transcript levels were consistently present at 5 wk of age, whereas they decreased variably in age-matched transgenic mice, as shown in Fig. 5A, indicating a variable degree of ablation of brown fat in mammary gland. To study the early differentiation of alveoli in transgenic mice, the presence of β-casein transcripts was determined, since β-casein is one of the first milk proteins produced during pregnancy by alveoli. β-Casein transcripts were never seen in wild-type mice as early as 5 wk of age, but were detected in mammary glands of two of three transgenic mice (Fig. 5A).

Fig. 4.

Epithelial Differentiation in Transgenic Mice Overexpressing DTA under the UCP-1 Promoter A and B, Typical mammary gland whole mounts of 5-wk-old transgenic (Tg, panel A) or wild type (WT, panel B) mice. C–E, High magnification of ductal morphogenesis of mammary gland whole mounts from 5-wk-old Tg mouse (C), d 5 pregnant WT mouse (D), and 5-wk-old WT mouse (E). Panels F, G, and H are the hematoxylin/eosin staining of 5-μm sections from the corresponding whole mount from panels C, D, and E, respectively. The alveoli structures developed in Tg mice (F) and commonly seen in WT pregnant mice (G) are indicated by arrowheads. I–K, Graphs representing means ± sem of branching numbers (I), TEB numbers (J), and ductal lengths (K) of mammary gland whole mounts obtained in panels A and B from Tg and WT mice. (***, Significant differences between WT and Tg mice, P < 0.0001, two-tailed t test.)

Fig. 5.

Northern Blot Analysis of UCP-1 and β-Casein Transcript Levels in Mammary Glands from 5-wk-old Wild Type (WT) and UCP-DTA Transgenic Mice (Tg) Each lane represents one mouse. Total mammary RNA (10 μg) was analyzed, and the 28s rRNA levels were used for the RNA loading control of each lane. Interscapular brown fat (BF) was used as a positive control for UCP-1 mRNA levels.

The presence of β-casein at early puberty in transgenic mice confirmed the precociously abnormal differentiated phenotype of epithelium into alveoli. However, the depletion of UCP-1 transcripts in the mammary tissue as shown in Northern blots does not occur in every transgenic mouse (Fig. 5A), even though the differentiated alveoli were consistently seen in mammary glands of every transgenic mouse. This observation led us to investigate a potential systemic effect in transgenic mice that affects mammary epithelium differentiation. For this purpose, we designed two kinds of mammary transplantation; the first consisted of transplanting a mammary gland from a 4-wk-old wild-type mouse into a 4-wk-old transgenic mouse in between the body and the skin; the second reciprocal one consisted of transplanting a mammary gland from a 4-wk-old transgenic mouse into a 4-wk-old wild-type mouse. Mice were killed 4 wk after transplantation. In the first case, epithelium from wild-type mice developed alveoli (Fig. 6, A and B) similar to those seen in mammary glands from age-matched transgenic mice (Fig. 6, C and D). In the second case, epithelium from transgenic mice lost their alveoli structures (Fig. 6, E and F) to give rise to normal smooth ducts typical of the normal mammary gland of wild-type age-matched mice (Fig. 6, G and H). The compilation of both transplantation experiments is in the favor of a systemic effect in transgenic mice inducing the abnormal mammary differentiated phenotype. To investigate whether the systemic effect can be directly related to a lack of major deposits of BAT in the whole body of transgenic mice, we examined histologically the density of brown adipocytes in interscapular BAT (IBAT), the largest BAT deposit in mice. Interestingly, IBATs from every transgenic mouse were consistently formed of unusual brown adipocytes with enlarged cytoplasmic lipid droplets (Fig. 6I) compared with the numerous and small droplets seen in wild-type mice (Fig. 6J), suggesting that the tissue was hypoactive. These data indicated that the differentiated phenotype in transgenic mice is due to a systemic effect probably caused by atrophy of the general BAT from the whole body.

Fig. 6.

Systemic Effect On Mammary Ductal Morphogenesis in UCP-DTA Transgenic (Tg) Mice The abdominal mammary gland from a 4-wk-old wild-type (WT) mouse was transplanted in the abdomen between the body and the skin of a 4-wk-old Tg mouse (WT-Tg, panels A–D), and the inverse manipulation was also performed (Tg-WT, panels E–H). Four weeks after transplantation, mammary gland whole mounts were processed from the transplanted WT mammary gland (WT, panels A and B) or the control mammary gland of the transgenic mouse host (Tg CTL, panels C and D) in the case of the WT-Tg transplantation. Similarly, for the Tg-WT transplantation, whole mounts from the transplanted Tg mammary gland (Tg) or the control mammary gland of the WT mouse host (WT CTL) are represented (panels E–H). Note the differentiation into alveoli (open arrowheads) of the WT transplanted mammary gland (A and B) and the dedifferentiation of the Tg transplanted mammary gland (E and F), indicating a systemic effect on the mammary epithelial differentiation in Tg mice. Photographs were taken at the same magnification. I and J, Hematoxylin/eosin staining of 5-μm sections of IBAT from 5-wk-old Tg (I) and WT (J) mice. Note the enlarged lipid droplets in the cytoplasm of brown adipocytes from Tg mice indicated with solid arrowheads. Magnification: panels B, D, F, H, I, and J, ×400.

Differentiation of the mammary gland is regulated, at least in part, by prolactin (PRL) and GH (2). Consequently, in an attempt to examine the systemic defect related to the precocious differentiated mammary gland phenotype in transgenic mice, the serum levels of PRL and GH were kindly measured by Dr. A. F. Parlow (National Hormone and Peptide Program, Harbor-UCLA Medical Center). None of these hormone levels were significantly affected in transgenic mice vs. wild type. Moreover, the average weight of 5-wk-old transgenic mice was in the normal range compared with wild-type mice. This was consistent with normal GH levels in transgenic mice because a perturbation of GH levels would affect the weight of mice (see Table 1).

Table 1.

Hormonal and Weight characteristics of 4-wk-old Wild-Type (WT) Transgenic (Tg) Mice

	WT Mice	Tg Mice	Two-Tailed P Value
PRL	33.87 ± 4.58 (8)	3766 ± 8.95 (6)	9.9497a
GH	30.37 ± 14.14 (8)	17.66 ± 5.84 (6)	>0.9999a
Weight	17.45 ± 0.27 (6)	19.43 ± 0.72 (6)	0.0931b

Weight in grams, PRL and GH in nanograms per milliliter. Means are indicated ± se for a Mann-Whitney test. (n) indicates the number of mice. a, Not significant. b, Not quite significant.

DISCUSSION

Mutual and reciprocal epithelial-mesenchymal interactions are critical for fetal and postnatal mammary gland development (4–6). Adipocytes are the major part of the mammary mesenchyma. The presence of brown adipocytes in mammary gland was an unexpected finding, since it has always been assumed that mammary gland fat was only composed of white adipocytes (5). This observation was the result of the compilation of experiments performed in the mammary gland, which examined the microscopic and ultrastructural morphology of adipocytes, as well as the presence of UCP-1 protein and mRNA by immunohistochemistry and Northern blot, respectively. BAT was consistently detected in mammary glands of three different strains of mice in the early postnatal development, being most abundant at 2–3 wk of age, decreasing as mice aged, and becoming finally undetectable at 8 wk of age. Mammary BAT was restricted to the area between the nipple and the lymph node, where the rudimentary ductal tree starts to arborize at 3 wk of age after a quiescent postnatal period. Consequently, the presence of BAT in mammary gland coincides temporally and spatially with the postnatal burst of ductal outgrowth established before the first estrous cycles (at about 5–6 wk of age).

The origin of these transitory brown adipocytes within the mammary WAT is unknown, but they seem to fall into the description of the convertible BAT (29, 30). Indeed, thermogenic adipose tissue has two forms: BAT and convertible adipose tissue (CAT). Brown adipocytes have high mitochondria content and express UCP-1 throughout the entire life of small rodents, chiropterans, and insectivores. However, in other endotherms and in humans, CAT participates as thermogenic tissue only during the early postnatal period. Both BAT and CAT start to develop in utero, although in some animals (hamsters, marsupials) or in some particular areas (thoraco-periaortal and medio-perirenal areas in rats) development of thermogenic adipose tissue starts after birth, as seen in the mouse mammary gland in this study. In certain conditions, cold exposure or β3-adrenergic stimulation, the induction of BAT in rat is mainly due to the direct conversion of white adipocytes into multilocular brown fat cells (31). The specific spatial and temporal appearance of BAT in the mammary gland suggests a unique function in the postnatal mammary gland development.

The crucial role of mammary fat in ductal morphogenesis is well documented. In general, mammary adipocytes induce growth and differentiation of mammary epithelium. These experiments take into account the whole mammary fat including white and brown adipocytes in vivo (12, 10) and ex vivo (20), or specifically white adipocytes in cell culture (14–18). The acquisition of hormone sensitivity of the epithelium and hormone-dependent differentiation (secretion of casein) requires interaction between mammary epithelium and other specific nonepithelial cells (15). More precisely, in association with lactogenic hormones (insulin, hydrocortisone, and PRL), adipocyte-related factors induce secretory differentiation of the ducts, whereas ductal morphogenesis was lactogenic hormone independent (18).

Interestingly, before the demonstration of mammary BAT existence, IBAT has been used as a fat matrix in transplantation experiments to show that mammary epithelium was capable of growing in brown fat. Transplants of segments of both bud-free ducts and TEBs deriving from rat or mouse mammary glands can regenerate a branching system of ducts when grown in the interscapular fat pad of a syngenic host (32–34). We also performed transplantation of ductal segments into the dissected IBAT or a dissected cleared mammary fat pad of syngenic host replaced between the skin and the body of mice. As observed previously, the implanted epithelium was capable of organizing and growing as normal ducts with regular spacings and a competition for spaces among themselves in both IBAT and mammary cleared fat (data not shown). This is in line with the normal mammary gland development observed in mice treated with the adrenergic agonist CL316,243, where BAT was overproduced within the mammary WAT. The increased formation of BAT in mammary glands was expected since the emergence of brown adipocytes in traditional deposits of white fat after treatment with CL316,243 has been documented in mice and rats (27, 35–39). Altogether, these experiments indicate that in the presence of only, or mainly, BAT, epithelial outgrowth is similar to that found in the mammary gland fat pad including both white and brown adipocytes. Consequently, excess or the unique presence of BAT does not alter the mammary branching morphogenesis.

The next issue was to examine whether a depletion of mammary BAT affects ductal outgrowth. To address this question, we analyzed a transgenic mouse model expressing the toxin-a chain under the UCP-1 promoter (UCP-DTA mice), where the brown adipocytes are specifically killed. Depletion of BAT in UCP-DTA mice consistently induced a complex mammary ductal outgrowth, an epithelial differentiation into alveoli, and induction of the β-casein gene. However, transplantation experiments indicated a systemic effect in the transgenic mice affecting mammary ductal hypertrophy/differentiation. Our data suggest that BAT negatively regulates the differentiation of mammary epithelial cells in a systemic manner during prepubertal ductal outgrowth. The total absence of mammary UCP-1 transcripts in wild-type mice during pregnancy and lactation (data not shown), when epithelial differentiation obviously occurs, also supports this conclusion.

The systemic effect on mammary gland development in UCP-DTA mice remains to be understood. The mammary defects in these mice occurred early during postnatal development, before the abnormal hyperinsulinemia, hyperglucosemia, hypertrigliceridemia, and hypercholesterolemia observed at 22–26 wk of age which was due in part to a significant decrease of insulin receptors in fat and muscles (40, 41). An early modification reported in UCP-DTA mice is a 30% increase in total body lipid detected at 6 wk of age, inducing obesity as well as hyperphagia, which begins at 5–7 wk (28). Mammary gland development requires a well regulated hormonal system, which is also required for mouse fertility and growth (42). Nevertheless, these mice are fertile and able to lactate (Ref. 28 and personal communication by Jennifer Merriam from The Jackson Laboratory), which is not surprising since the mammary defect found is a premature differentiation of epithelium into functional alveoli producing β-casein. In addition, transgenic females display normal serum levels of GH and PRL; nevertheless, this does not rule out the contribution of other secondary endocrine effects in transgenic mice. These findings suggest that there may be a soluble factor secreted by the BAT of the whole body that may affect mammary epithelium differentiation. In transgenic mice, the depletion of IBAT, the largest part of BAT in the mouse, is severe since the UCP-1 protein content in IBAT is reduced by 68% at 16 d of life, which is maintained as the mice age (28). In line with this previous report, the transformation of the IBAT into a more WAT-like structure was consistently observed in every UCP-DTA mouse we studied, even though the ablation of BAT in mammary glands was variable. Considerable evidence links fat content with reproductive functions such that reproduction is coupled to nutritional status (Refs. 43–46 for reviews). Our data suggest that BAT and its unique function of thermogenesis could be a component of such a regulatory system. The specific location of BAT could also play a role in mammary gland development in mice exposed to the cold, because of the surface location of mammary glands. Such a mechanism might also improve reproductive functions.

In conclusion, we demonstrate that BAT is one of the components of the fatty stroma of the mammary gland and that this tissue is spatially and temporally regulated. Our data suggest that BAT negatively regulates the differentiation of mammary epithelial cells in a systemic manner during prepubertal ductal outgrowth, when the ductal outgrowth should be prevalent over ductal differentiation.

While this paper was under review, the presence of brown fat in the mammary gland was also reported by another group (47).

MATERIALS AND METHODS

Animals

The transgenic FVB/N-TgN (UCP-DTA) mice and their control FVB/N mice (28) and the standard strains C57Bl/6, C57Bl/6.C3H, and 129 SvEv were purchased from The Jackson Laboratory (Bar Harbor, ME). Mice were maintained under pathogen-free conditions in a barrier facility at the Albert Einstein College of Medicine (New York, NY). To induce BAT, C57Bl/6 mice were injected sc every day with the β3-adrenergic agonist, CL 316,243, at a dose of 1 mg/kg body weight at 1500 h (not exceeding 25 μg/d) (39). All studies were performed under NIH guidelines for the care and treatment of experimental laboratory rodents

RNA Isolation and Northern Blots

Total RNA from mammary glands was isolated by the method of Chomczynski and Sacchi (48). Ten micrograms of total RNA were separated by formaldehyde-agarose gel electrophoresis, transferred to nylon filters, and probed with a ³²P-dCTP-labeled cDNA probe for the UCP-1 (GenBank/U63419/MMU63419) or β-casein (49) using the methods described previously (7).

Histology/Immunohistochemistry

Whole-mount mammary glands were fixed with formalin overnight onto a glass slide and then embedded with paraffin wax. To identify white and brown mammary adipocytes, 5-μm sections were stained with hematoxylin/eosin. White adipocytes have one large cytoplasmic fat droplet (unilocular adipocyte), and brown adipocytes have numerous small cytoplasmic fat droplets (multilocular adipocytes). Sections (5-μm) were immunostained with the goat polyclonal IgG anti-mouse UCP-1 (Santa Cruz Biotechnology, Inc., Santa Cruz, CA), developed using a peroxidase detection kit (Vector Laboratories, Inc., Burlingame, CA), and counterstained with hematoxylin as described previously (7).

Electron Microscopy

Mammary glands from 3-wk-old C57Bl/6 females were dissected out, cut in small pieces, and fixed in fixative including 2.5% gluteraldehyde in 0.1 m Cacodylate buffer for 1 h at room temperature (RT). Mammary gland pieces were then stored in fresh fixative at 4 C overnight until processing by conventional methods.

Mammary Gland Whole-Mount Preparation

The fourth (abdominal) mammary glands were surgically removed, stretched onto a glass slide, and fixed in 75% ethanol/25% acetic acid overnight. They were then washed in 70% ethanol, washed 5 min in distilled water, and stained overnight in an alum carmine solution (Sigma, St. Louis, MO). The following day, mammary glands were dehydrated through a graded series of ethanol solutions, defatted in toluene (Sigma), and stored in methyl salicylate (Sigma). To determine ductal lengths, mammary gland whole-mount preparations were measured from the nipple area to the tip of the three longest ducts through the lymph node (1 U = 0.25 mm). The number of branches represents the mean of branching number along the three longest ducts from the nipple area to the migration front. TEBs were counted in the whole mammary gland (7). Statistical evaluations were performed with a two-tailed Student’s t test.

Mammary Gland Transplantation

The entire fourth (abdominal) mammary gland from UCP-DTA mice was removed at 4 wk of age and transplanted in between the body and the skin between the third and the fourth mammary gland of 4-wk-old wild-type mice. The opposite transplantation was also performed. Mice were killed 4 wk after transplantation, their transplanted mammary glands as well as their own fourth mammary glands as a control were removed, and the corresponding whole mounts were prepared and stained for analysis.

RIA

GH and PRL measurements were performed in the laboratory of Dr. A. F. Parlow. Statistical evaluations were performed with a two-tailed Mann-Whitney test.

Acknowledgments

We gratefully acknowledge Dr. A. F. Parlow (National Hormone & Peptide Program, Harbor-UCLA Medical Center, Torrance, CA) for performing GH and PRL measurements. We thank Dr. Philipp Scherer for helpful comments on the manuscript. We especially thank Dr. Kurt Steiner (Wyeth-Ayerst Laboratories, Inc., Philadelphia, PA) for critical readings of the manuscript and for providing CL 316,243. We are grateful to Mr. Jim Lee for his technical assistance in the mouse facility.

This work was supported by NIH Grant RO1HD-30280 (J.W.P), Department of Defense 17-97-1-7153 postdoctoral fellowship (V.G.-E.), and Einstein NCI Cancer Center Grant P30-13330. J.W.P. is the Sheldon and Betty E. Feinberg Senior Faculty Scholar in Cancer Research.

Abbreviations:

 
• BAT,

  Brown adipose tissue;

 
• CAT,

  convertible adipose tissue;

 
• IBAT,

  interscapular BAT;

 
• PRL,

  prolactin;

 
• TEB,

  terminal end bud;

 
• UCP,

  uncoupling protein;

 
• WAT,

  white adipose tissue.