Regulation of mammary cell proliferation is complex and unique. Understanding the regulation of mammary gland development and the hormonal contribution to epithelial cell proliferation are important in assessing breast cancer etiology and election of treatment strategy.

2 .1 Control of Mammary gland development by steroid hormones

The contribution of ovarian steroids and pituitary peptides to mammary gland development and lactation has been well documented (Halban, 1900). Effects of estrogens are mediated by two distinct estrogen receptors, ERa and ERb, and the effects of progesterone are mediated by two isoforms of the progesterone receptor, PR-A and PR-B. The biological activities of 17b-estradiol (E2) through its receptors are broad but tissue-specific (reviewed in (Dickson and Stancel, 2000)). In mice, ERa is the main estrogen receptor in the mammary gland, as the level of ERb mRNA is very low (Bocchinfuso et al., 1999). Definite roles of each ER receptor in mediating E2 activities have been demonstrated by studies in ERa and ERb knockout mice, respectively (reviewed in Couse and Korach, 1999; Dickson and Stancel, 2000). ERa is required to mediate the activity of E2 in ductal outgrowth (Korach et al., 1996), whereas mammary glands of prepubertal ERb knock out mice are morphological normal. However, the organization and adhesion of epithelial cells in the lactating gland are altered (Forster et al., 2002).

PR-A and PR-B arise from the same gene. Selective ablation of PR-A and PR-B proteins in knockout mice has allowed the analysis of their function, specifically. The responses of the mammary gland and thymus to progesterone in mice with PR-A ablation are normal, but ovarian and uterine are dysfunctional leading to female infertility. On the other hand, PR-B ablation does not affect ovarian, uterine, or thymic responses to progesterone, but results in reduced mammary ductal morphogenesis (reviewed in (Conneely et al., 2002)). The function of PR is mainly in the epithelium -- absence of PR in donor epithelium, but not in the stroma, prevented lobulo-alveolar formation in response to E2 and progesterone in mammary gland transplants (Humphreys et al., 1997). Overexpression of the Wnt-4 gene, which encodes a secreted glycoprotein, overcomes the branching defects, indicating that the Wnt protein acts downstream of PR to mediate the process of side-branching (Brisken et al., 2000). These elegant studies provide significant insights into genetic regulation and signal pathways in mammary gland development. Details of steroid and peptide hormonal control of mammary gland development are described below.

2.2. Mammary gland development

Gertraud Robinson
Laboratory of Genetics and Physiology, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, MD

Development of the mammary gland can be divided into distinct stages, embryonic and prepuberty, puberty, pregnancy, lactation, and involution. While growth of the gland until puberty advances synchronously with the growth of the animal it is greatly increased during puberty. Transient alveolar development and functional differentiation occurs during estrous cycling but lactogenesis and milk secretion are found only in late pregnant and lactating animals. These processes are regulated by steroid and peptide hormones (reviewed in Hennighausen and Robinson, 1998 and Hennighausen and Robinson, 2001). Mouse strains deficient in specific genes have provided novel insights into molecular mechanisms of mammary gland development. These include growth factors, hormones, and their receptors as well as proteins that have not previously been associated with mammary gland development such as transcription factors, cell cycle regulators and cell surface proteins.

Embryonic and prepubertal phase. Mammary gland development begins in the mouse embryo late on day 10 of gestation. Cells in the ectoderm form small placodes along two lines that run on either side just ventral of the limbs. The placodes first protrude from the surface but soon sink into the dermis where they increase in size and form bulb shaped buds. The increase in size is only partially achieved by cell proliferation and the majority of cells are recruited from the surrounding epidermis. Development of the anlage is arrested in mice that are deficient in FGF10 signaling (Mailleux et al., 2002). A similar arrest of development also occurs in mice that lack the transcription factor LEF1 (van Genderen et al., 1994) suggesting that wnt signals are required for further epithelial development. Concomitant with formation of the epidermal component of the gland, changes also occur in the mesenchyme. A halo of concentrically oriented mesenchymal cells, which are more densely packed than the dermal mesenchyme surrounds the bud. On day 15 the mammary anlagen in male embryos undergo apoptosis. The connection of the epithelial stalk to the epidermis is severed and nipple formation does not occur. This process depends on PTHrP signaling which induces the synthesis of androgen receptors and a number of other molecules in the mammary mesenchyme (Dunbar et al. 1999, Foley et al., 2001). Around embryonic day 15 cell proliferation resumes at the tip of the bud, leading to the formation of a primary sprout, which starts to bifurcate and grow into a small system of ducts at the time of birth. Until the onset of puberty these ducts continue to grow at a similar speed as the newborn mouse. The primary sprout penetrates and grows toward deeper layers of the mesenchyme, the future fat pad that constitutes the stroma of the adult gland. Results from tissue recombination and transplantation studies have shown that EFGR and ERa in the stroma are required for growth of the epithelium during this period (Wiesen et al. 1999, Bocchinfuso et al., 2000).

Ductal development during puberty. The onset of secretion of the ovarian hormones estrogen and progesterone at puberty stimulates ductal growth. Large club shaped structures, the terminal end buds develop at the tips of the growing ducts. There are two distinct cell types: body cells, which give rise to mammary epithelial cells, and cap cells, which are precursors of myoepithelial cells (Humphreys et al., 1996) in the TEB (teminal end buds). These terminal end buds are areas of very active cell proliferation and cell death, the processes that lead to elongation and lumen formation. Secondary and tertiary ducts develop from the primary ducts and form a ductal system that penetrates the entire fat pad in the mature animal around 10 weeks of age. A single layer of luminal epithelial cells lines the mammary gland ducts and the myoepithelial cells form a collar around the primary ducts but are discontinuous around secondary and tertiary ducts and TEBs. The cyclical changes of estrogen and progesterone levels during each estrous cycle induce a peak in cell proliferation and formation of small side branches in the late proestrous and estrous phase and their regression by an increase in apoptosis in diestrus. These changes are accompanied by activation of Stat5a and transient expression of milk proteins in each cycle (Andres and Strange, 1999). The majority of stromal cells are adipocytes and fibroblasts. Rather limited information is available about the vascular system of the fat pad.

The role of estrogen and progesterone in ductal outgrowth during puberty was established by hormone ablation and reconstitution experiments (reviewed in (Imagawa et al., 2002)). Experiments with animals deficient in these receptors have confirmed these findings. In many cases the mutations affect viability and fertility and transplantation and recombination of wild type and mutant tissues are required to dissociate systemic from tissue intrinsic effects.

Transplantation of estrogen receptor ERa-deficient epithelium into wild type stroma and vice versa demonstrated a requirement for ERa in both compartments. Treatment with high doses of estradiol and progesterone induces normal ductal development of ERa-null epithelium but ERa-null stroma supports only rudimentary ductal development of wild type epithelium. Lack of progesterone receptor (PR) in these transplants suggests that epithelial ERa is required for induction of epithelial PR in the absence of stromal ERa (Mueller et al., 2002). Lack of both forms of the PR in epithelial cells inhibits ductal side branching while disruption of the A form of the PR (which contains a longer N-terminus than PR-B) does not affect epithelial development suggesting that PR-B mediates the progesterone signal in the mammary gland (Mulac-Jericevic et al., 2000). Normal ductal development also requires epithelial glucocorticoid receptor but the function of the glucocorticoid receptor in alveolar development can be compensated by an upregulation of the mineralocorticoid receptor during pregnancy (Kingsley-Kallesen et al., 2002). There is strong experimental evidence that the action of estrogen is mediated through IGF1 (Kleinberg et al., 2000). Direct involvement of this signaling pathway was obtained through transplantation of IGF-R1 deficient epithelium which displayed reduced cell proliferation in terminal end buds in virgin animals (Bonnette and Hadsell, 2001). Mammary epithelial cells also require the sodium-potassium cotransporter NKCC1 for proper ductal growth and morphogenesis indicating that dysregulation of ion transport affects epithelial cell proliferation (Shillingford et al., 2002a). Epithelial proliferation is also regulated by tissue interaction and depends on stromal signals. Absence of activin B, a member of the transforming growth factor (TGF)b family signaling molecules, in the stroma causes attenuated ductal and alveolar development (Robinson and Hennighausen, 1997). Lack of gelsolin, an actin binding protein that is involved in regulation of the actin cytoskeleton, in the stroma leads to a delay of terminal end bud development, slow ductal elongation and delayed lobuloalveolar development (Crowley et al., 2000). Furthermore, cells of the hematopoietic system play a role in mammary gland development. An inactivating mutation of colony stimulation factor (CSF)-1, a major growth factor for macrophages and eosinophils, reveals a role for these cells in ductal development. These cells are recruited around terminal end buds and are thought to be involved in tissue remodeling during ductal elongation (Gouon-Evans et al., 2000).

Alveolar development during pregnancy. During pregnancy the lobular compartment of the gland expands. Cell proliferation occurs in ducts and alveoli throughout pregnancy and continues into early lactation. At term the mammary fat pad is filled with a dense epithelial compartment consisting of alveolar lobules that are lined with secretory epithelial cells actively secreting milk proteins and lipid droplets into the luminal space. Synthesis of milk proteins starts at mid gestation with WDNM1 being one of the first to be expressed, followed by b-casein. Whey acidic protein and a-lactalbumin appear only shortly before birth (Robinson et al., 1995).

Prolactin and placental lactogens are the key hormones regulating alveolar development and functional differentiation of secretory cells during pregnancy. It is therefore no surprise that deletion of different components of the prolactin receptor (PrlR signaling pathway inhibits alveolar development. Epithelial transplants from mice deficient in the PrlR, Jak2 (the kinase associated with the receptor and responsible for signal transduction to Stat5) and both isoforms of Stat5 (Stat5a and Stat5b) develop normal ducts but fail to proliferate in response to stimulation by estrogen and progesterone. They do not form secretory alveoli and do not synthesize milk proteins (Brisken et al., 1999, Shillingford et al. 2002b, Miyoshi et al., 2002). Interestingly, there seems to be a dose dependence in this signaling pathway and even mice heterozygous for PrlR and Stat5 fail to nurse their first litters (Ormandy et al., 1997). A similar phenotype is also seen in epithelia lacking the transcription factor C/EBPb (Robinson et al. 1998, Seagroves et al., 1998). These cells appear to be unable to respond to hormonal stimuli and have more PR positive cells that are uniformly distributed along the ducts while these receptors are found in patches in wild type glands. This indicates a defect in cell-to-cell communication and paracrine signaling (Seagroves et al., 2000). Reduced alveolar development and lack of proliferation upon hormone stimulation is also found in cyclinD1 deficient epithelium (Fantl et al. 1995, Sicinski et al., 1995). Inactivation of the helix-loop-helix protein Id2 inhibits cell proliferation and disrupts epithelial cell differentiation at a stage equivalent to day 12 of pregnancy (Mori et al. 2000, Miyoshi et al., 2002). RANKL (osteoprotegerin-ligand, OPGL, Tnfsf 11, osteoclast differentiation factor, TRANCE) originally identified as a factor produced by osteoclasts is also produced by mammary epithelial cells and is required for cell proliferation and full functional differentiation. Inactivation of the ligand RANKL as well as the receptor RANK leads to lactation failure as a result of decreased cell proliferation and increased cell death (Fata et al., 2000). As shown by transplantation the RANKL signal is required in epithelial cells and involves activation of PKB/Akt. The importance of NF-kB dependent signals in alveolar development has been identified through an engineered inactivatable mutation of the IkB kinase which activates NF-kB (Cao et al., 2001). This mutation results in downregulation of cyclin D1 expression, reduced cell proliferation and incomplete differentiation of epithelial cells. Transgenic expression of cyclin D1 can restore the defect suggesting that cyclin D1 plays a dual role as proliferation and differentiation factor in the mammary gland. Alveolar development is compromised in plasminogen-deficient animals indicating that proteolytic tissue remodeling in required in this process (Lund et al., 2000). In several cases the inactivation of genes results in precocious development of alveoli indicating that proper development is determined by a delicate balance of stimulating and inhibiting signals. SOCS1 is a negative regulator of PrlR signaling. SOCS1-deficient mice exhibit accelerated lobuloalveolar development and haploinsufficiency for SOCS1 is able to rescue PrlR haploinsufficiency demonstrating again the dosage sensitivity of this signal (Lindeman et al., 2001). Caveolin-1has been identified as another modulator of prolactin signaling that interacts with the receptor in a manner similar to SOCS1. Stat5a activation, lobuloalveolar growth and milk protein synthesis are accelerated in caveolin-deficient mice (Park et al., 2002). The adhesion factor P-cadherin is expressed in myoepithelial cells. Its absence leads to formation of alveolar buds, which express casein in virgin mice followed by hyperplasia and dysplasia in older mice (Radice et al., 1997).

Cell death and tissue remodeling during involution. Lactation is terminated when a lack of suckling causes milk stasis and a reduction of prolactin release from the pituitary. Involution occurs in two phases. The first phase is reversible and is characterized by induction of proapoptotic genes and alveolar cell death. During the second phase the alveoli collapse, tissue protease activity is induced and massive remodeling of the alveolar compartment leads to establishment of a ductal system with small side branches only. Early events are a decrease of Stat5a phosphorylation and phosphorylation of Stat3. That the rapid phosphorylation of Stat3 at the onset of involution may trigger apoptosis is supported by the finding that deletion of Stat3 in alveolar epithelial cells delays involution (Champan et al. 1999, Humphreys et al., 2002). Several factors that regulate apoptosis in the immune system are also involved in mammary gland involution. Absence of the transcription factor IRF-1 (interferon regulatory factor) accelerates apoptosis indicating that it is a survival factor (Chapman et al., 2000). Inactivating mutations of Fas, a member of the TNF receptor family that is involved in mediating apoptotic signals, and its ligand FasL cause a decrease in cell death (Song et al., 2000). TGFb3 is upregulated within a day of involution and appears to induce apoptosis as suggested by a delay in involution in TGFb3-deficient tissue (Nguyen and Pollard, 2000). Similarly, cell death is reduced in tissue deficient in Smad3, one of the mediators of TGFb signaling (Yang et al., 2002). This further supports the role of TGFb signals in the early phase of involution. Proteases that degrade the extracellular matrix are active during the later stages of involution. Lack of plasminogen is associated with reduced apoptosis and remodeling of the glands (Lund et al., 2000) while the lack of TIMP-3, an inhibitor of metalloproteinases accelerates involution (Fata et al., 2001).

The creation of mice with deletions of specific genes has allowed us to identify signaling pathways and molecular mechanisms participating in cell determination, cell proliferation, survival and death and cell differentiation in the mammary gland. These results provide a framework for the development of better animal models and more direct therapeutic interventions.