Growth control and differentiation in mammary epithelial cells.

Growth and differentiation of the mammary gland are controlled by various hormones and other environmental factors. The role of hormones and growth factors in mammary development is discussed with regard to animal species, physiological stages, and the various experimental systems in vitro. In the female embryo, mammary morphogenesis is induced by the mesenchyme and is hormone independent, whereas androgens cause the partial necrosis of mammary epithelium in the male. Ductal growth during adolescence requires estrogen and prolactin or growth hormone. During pregnancy, progesterone participates in the development of the lobuloalveolar structure of the gland. After parturition, changes in the hormonal environment lead to production and secretion of milk. Proliferation and differentiation of mammary epithelium can be induced in culture systems. Insulin and epidermal growth factor (EGF) stimulate mammary cell proliferation in vitro. EGF is required for the optimal growth of the mammary gland during pregnancy. EGF also appears to play an important role in mammary tumorigenesis in certain mouse strains. Production of milk proteins can be induced in vitro by the synergistic interactions of prolactin, insulin, and glucocorticoids and is inhibited by EGF and progesterone. Complete or partial sequencing of several milk protein genes and comparative analysis have led to identification of a sequence of high homology and conservation in the 5' flanking region that is likely to be involved in the regulation of milk protein gene expression.


Introduction
In recent years, biomedical studies have focused on understanding the mechanisms regulating cell growth and differentiation. For this purpose, the mammary gland serves as a very suitable experimental system because of its characteristic pattern of morphological and functional development.
Most tissues and organs undergo massive growth in the early stages of development, i.e., during embryogenesis and in the early period of postnatal life. By contrast, mammary tissue expresses its maximum growth potential after the animal has reached maturity, that is, following the onset of pregnancy and during lactation. Moreover, this growth potential is maintained throughout the reproductive life of the animal, and, to some extent, even thereafter. The cycle of proliferation-differentiation-regression is repeated at each gestation, and can be reproduced in culture systems in vitro. The characteristic physiological pattern of mammary gland development is regulated by several factors; endocrine control is the most obvious and perhaps the best understood, but several studies have provided evidence that paracrine, autocrine, and intracellular factors, as well as the extracellular matrix, can also affect the growth and differentiation processes.
The availability of molecular biological techniques has given investigators a new possible approach to the understanding of the complex mechanisms of regulation of mammary gland function. Studies on the structure and expression of milk protein genes have provided valuable information to the overall picture. A deeper understanding of how growth and differentiation of the mammary tissue are regulated could complement the knowledge of developmental processes and also provide invaluable information for clinical treatment and prevention of mammary cancer.
Several excellent reviews (1-7') giving an extensive overview of the growth and development of the mammary gland have been published. In this review, we have focused on the recent progress made in this field, with special attention to the interactions between the different regulatory factors and their mechanisms of action.

Outline of Mammary Gland Development
The mammary tissue from different species has a very similar structural organization, although hormonal re-quirements for the growth and expression of the specific function of the gland vary among various strains or species. The following is a brief account of developmental changes of the mammary gland.

Embryogenesis
The embryonal development of mammary tissue appears to be comparable in all species (8,9). In the mouse, the development in the male and in the female follows the same pattern up to about the day 13 of embryonal life; at day 11, an area of raised ectoderm is formed on both sides of the trunk, and the neighbouring ectodermal cells begin to congregate around this area to form the mammary band (Fig. 1A). The migration of the epidermal cells to preferential centers of congregation gives rise to the individual buds (8,9) (Fig. 1B).
The critical difference between male and female embryos in the development of the mammary tissue occurs between days 13 and 15. In the female, at day 15 of embryonic life, the mammary bud cells undergo a phase of rapid proliferation, giving rise to the mammary cord, a  12-13; (C) mammary cord, day 15. band of epithelial cells elongated into the mesenchyme but still connected to the epidermis by a collar of ectodermal cells (Fig. 1C). At the end of gestation, the mammary epithelium in the female embryo consists of a cord of cells, embedded in the mesenchyme and partially canalized. The cord opens at the apex, where the future nipple will be formed, and begins to show some branching at the distal end.
In the male embryo, a series of events takes place between days 13 and 15 as a result of the androgen production by the fetal testes (10). The mesenchyme condenses all around the mammary bud, which becomes isolated into the subepidermal tissue. The bud eventually becomes a rudimentary epithelial cord, but no further development occurs. In the mouse, the mesenchymal condensation leads eventually to the partial or total destruction of the mammary bud; in rat and rabbit embryos the necrosis of the mammary rudiment does not occur.

From Birth to Maturity
The prepubertal development of the mammary gland occurs in two distinct phases. Up to day 22 to 23 of life, the development of the gland is mainly due to an increase in connective tissue and deposition of fat. The increase in size parallels the increase in body weight, and is therefore referred to as isometric growth; it does not seem to depend on hormones. During the 2 weeks preceding puberty, the mammary gland grows at a rate about three times greater than that of the body weight. This phase is referred to as allometric growth, and can be prevented by ovariectomy (2).
Sexual maturation in the mouse occurs during week 4 to 6 of postnatal life. During this phase the mammary epithelium in the female grows extensively and extends throughout the fat pad. Very few alveoli are formed at this time; the growth pattern is three-dimensional and can be simulated in cultures of mammary epithelium into a collagen matrix (11). Elongation of the ducts occurs as a result of rapid growth in end buds, which, by turning and branching, give rise to the characteristic tree-shaped pattern of the mammary ductal system. The end bud tip is covered with a monolayer of epithelium, the cap cells, which are characterized by a relative lack of specialized features. These cells have been defined as a stem cell population; as the duct elongates, the cap cells gradually differentiate to provide new myoepithelial cells for ductal morphogenesis and elongation (12). In human breast tissue, myoepithelial cells also originate from a precursor population within the epithelium after a number of cell divisions (13). During the different stages of mouse mammary gland development, myoepithelial cells undergo several changes in size and shape (14). The stroma surrounding the elongating end bud shows a significant increase in DNA synthesis (15).
As the mouse reaches sexual maturity at 6 to 8 weeks of age, the development of the ductal structure stops, in spite of the presence of the major hormonal stimuli (i.e., estrogen produced by the ovaries), for ductal growth. Some residual mitotic activity is observed at the end buds in the mature mouse gland during estrus (16). At this stage, the mammary gland shows ductal ramifications throughout the mammary fat pad; however, a clear zone of fat pad remains between the ducts. This interductal space will eventually be occupied by lobules during pregnancy.

Pregnancy
Following the onset of pregnancy, cell replication resumes. During the first period, ductal growth occurs extensively, and the ducts undergo further branching. The interstitial adipose tissue disappears progressively, and proliferating epithelial cells fill the interductal spaces. During the middle and late period of pregnancy, epithelial cells continue to multiply to fonn the lobuloalveolar structure. As pregnancy approaches term, the alveoli start to show secretory activity, and the epithelial cells become distended, with a granular cytoplasm (17,18) (Fig. 2). Increased vascularization of the gland and loss of fat in fat pad cells seem to be induced by the proliferating epithelial cells. In rat mammary gland, the proliferative activity is also higher during estrus and pregnancy than during nonestrus and lactation (19).

Lactation
Functional differentiation of mammary epithelial cells culminates in lactogenesis. Lactating tissue shows the typical features of a secretory epithelium. Cells are highly polarized, both functionally and morphologically, are connected by tight junctions, and the surfaces that face the cavity of the alveoli are covered with microvilli. A large part of the cytoplasm is occupied by rough endoplasmic reticulum; the well-developed Golgi apparatus is located in the apical region. Small droplets of fat are present all over the cytoplasm (Fig. 2). The epithelium and the surrounding myoepithelial cells are enclosed by the basal lamina, which provides a boundary between the epithelium and the stromal tissue.
In the mouse, milk yield is maximum at day 15 of lactation. After day 15 a sharp drop occurs, followed by a decrease in the synthetic activity of the gland and in cell population. When lactation is extended by substituting younger pups to the original litter at day 15, cell number and differentiation are maintained even though at a lower degree (20).

Involution
Natural weaning is usually a gradual process, and the progressive decrease in the suckling stimulation results in involution of the gland. The major changes of involution take place in the epithelial cells, which begin to show large vacuoles containing casein micelles and large fat droplets. As weaning progresses, more vacuoles appear, enclosing a variety of subcellular organelles such as mitochondria and fragments of endoplasmic reticulum. Hydrolytic enzymes are released by the fusion of lysosomes with the vacuoles. This autophagic process leads to cell depletion and lysis. The cell residues are ingested by macrophages present in a quite large number in the tissue at this stage. However, not all of secretory epithelial cells are destroyed; some of them undergo a partial autophagocytosis but survive. The myoepithelial cells surrounding the alveoli are usually preserved, although their shape changes from starlike to a more compact form.

Determinants of Mammary Gland Development In Vivo
The regulation of mammary gland morphogenesis in vivo is a very complex phenomenon that involves both hormonal stimuli and environmental factors. Tb simplify the exposition, we have grouped under the section "Hormonal Determinants" all the factors acting in an endocrine, autocrine, or paracrine fashion through a specific receptor. Several factors of mesenchymal origin seem to affect epithelial growth in a paracrine fashion. We will limit our discussion of this subject to generic information; for an extensive treatment, we refer the reader to the specific papers in this journal.

Hormonal Determinants
In the female, mammary embryogenesis is induced by the underlying mesenchyme and is hormone independent. Destruction of the ovaries with X-ray radiation in the 13-day embryo does not affect mammary development (21). In the male, the mesenchyme-directed development is interrupted at day 13, when mammary rudiments acquire sensitivity to androgens; after day 15 they lose hormone responsiveness. Destruction of the fetal testes with X-rays on day 13 leads to mammary development similar to that in the female (21). Cell replication during adolescence is hormone dependent (22,23). Ductal growth requires estrogen and either growth hormone or prolactin; however, neither type of hormone is effective alone. Extensive growth can occur in the absence of glucocorticoid, but maximal growth requires the steroid. Ductal growth appears to be independent of progesterone, thyroid hormones, and the thymus (7,24). Insulin is not involved in ductal growth in vivo, as the cells in virgin mice are not responsive to the hormone (25). In some species, as in the rat, the requirements for ductal growth vary among the different strains (2,26,27).
17-p-Estradiol is an absolute requirement for ductal growth, and when administered to a mouse it can enhance DNA synthesis in the mammary epithelium. It has been reported that the effect of estrogen on DNA synthesis begins approximately 12 hr after treatment (28), and maximal stimulation is reached at 48 hr. Autoradiographic studies have shown that the increase in thymidine incorporation is mostly restricted to the adipose and connective tissues during the first 24 hr; at 48 and 72 hr the epithelium undergoes extensive replication. Thus estrogen exerts its effect initially on stromal or fat pad cells (28). It would be interesting to study the relationship of this initial response to the subsequent proliferative response of the epithelium.
There seems to be a synergism between estradiol and cAMP in stimulating the growth of mammary epithelium (29); however, it is not clear whether the presence of ovarian hormones is necessary for cAMP to affect growth and morphogenesis of the gland because of the conflicting reports (30,31).
When the animal reaches sexual maturity, the mammary tissue enters a phase of dormancy during which only a modest replicative activity is observed at the end buds during estrus (16). However, estriol and progesterone administered daily to mature mice stimulate DNA synthesis and mitosis in resting mammary epithelium (32).
The maintenance of duct cells has different hormonal requirements from ductal growth; apparently pituitary hormones are not necessary when either adrenal or ovarian activity is present (7). Lobuloalveolar growth, which occurs mainly during pregnancy, requires estrogen, either growth hormone or prolactin, and progesterone (1,33,34). Adrenal steroids and thyroxine seem to enhance the formation of lobules (7,35). Even though mouse mammary adipose and connective tissues have high affinity binding sites for estrogens and progesterone, the largest amount of progesterone receptors is present in the epithelial component (36). Moreover, regulation of progesterone receptor concentration by estrogen and its changes during the different developmental phases appears to be restricted to the mammary epithelium (36). Ovarian steroids are also required for maintenance of alveoli.
Onset of lactation coincides temporally with a decrease in progesterone level and elevation of prolactin and glucocorticoids. Whether the trigger for the onset of milk production is due to the decrease in circulating progesterone or to the increase in a positive stimulus (prolactin and glucocorticoid) is still an open question. However, lactating mammary tissue has no detectable progesterone receptors (37,38), and, in fact, progesterone is unable to inhibit established lactation. On the other hand, administration of estrogen and progesterone to a lactating rat stimulates the rate of mammary cell replication that is similar to that observed during pregnancy (19).
Several lines of evidence support a role of epidermal growth factor (EGF) in mammary epithelial proliferation during pregnancy. The level of EGF receptors varies during the physiological stages of the gland: it is relatively low in the virgin and lactating glands, but increases during pregnancy and reaches a peak on day 10 of gestation (39).
The levels of EGF in the submandibular gland and plasma increase substantially during pregnancy (40). As sialoadenectomy results in marked decrease in the plasma EGF, it appears that the submandibular gland serves as a major source of circulating EGF. Earlier studies have indicated that the amount of EGF in the submandibular gland is increased by androgens (41) and progestins (42). As the level of these steroid hormones rises during pregnancy (43), it is possible that the increase in submandibular EGF is elicited by the action of androgens and/or progestin. In addition, as will be described later, physiological concentrations of EGF stimulate proliferation of mammary epithelial cells in culture.
Further evidence for the involvement of EGF in mammary gland development during pregnancy is provided by the work of Okamoto and Oka (44); the mammary glands of lactating mice that have been sialoadenectomized pregestationally are smaller and produce a smaller amount of milk compared to that of sham-operated controls. As a consequence, sialoadenectomized mothers cannot properly support all the litter, and thus a substantial number of pups die within the first week after birth. EGF-replacement therapy given to the sialoadenectomized mice during pregnancy restores the growth of the gland, milk production, and pup survival rate (44,45).
Thyroid hormones seem to have a role in the maintenance of alveoli and ductal branching in the involuting gland; the mammary gland of C3H mice treated with 2-thiouracil to repress thyroid function during involution shows less ductal branching and no alveoli (35).
Reinitiation of growth in senescent mammary epithelium has been obtained by application of implants that release cholera toxin (a cAMP-active agent) in the gland. This implies that the growth potential is maintained in the epithelium throughout the life span and that the change in the hormonal milieu is responsible for the alterations in growth patterns during the different physiological stages (46).

The Role of Stromal Components
Two different kinds of mesenchymal cells are involved in the morphogenesis of mammary gland epithelium of the mouse: fat pad precursor cells and fibroblasts. The specific interaction between fat pad precursor cells and epithelial cells seems to be important for determining the shape of the ductal branching structure in the female embryo.
The mesenchyme also plays a key role in the interruption of development that occurs in the male embryo: it has been identified as the specific target tissue for testosterone action, although its response to the hormone requires the presence of the mammary epithelium (47). More recent studies suggested that the mammary epithelium could induce the formation of androgen receptors in the surrounding mesenchyme, thus controlling the development of androgen respohsiveness in the tissue (48,49).
During the massive development that occurs at sexual maturation, epithelial-adipocyte interactions appear to be required for the formation of end buds and the subsequent morphogenesis of fully structured mammary ducts White fat tissue from a different origin other than mammary fat can support ductal growth (50,51).
For the epithelium to become responsive to lactogenic hormones, certain environment-associated events must occur. In co-culture systems of mouse mammary epithelium and adipocytes or other fibroblasts, only 3T3L1 and 3T3C2 cells could support casein synthesis in the presence of the lactogenic stimuli, insulin-cortisolprolactin (52). There was evidence of specificity of the epithelium-stroma interaction during this stage because Swiss 3T3 cells, newborn foreskin fibroblasts, substrateattached materials, and tissue culture plastic did not support casein synthesis (52). These results indicate that the acquisition of hormone sensitivity and hormonedependent differentiation of mammary epithelium are influenced by the stromal milieu. In cultured mouse mammary explants, stroma cells appear to support the epithelial differentiation partly by providing type I collagen, an essential component of extracellular matrix (53).

Biochemical Markers of Differentiation In Vitro
Differentiation of mammary epithelium is usually evaluated by measuring the capacity of the mammary tissue to produce the major milk proteins, such as the casein family and a-lactalbumin.
Before the development of molecular biological techniques, casein production was evaluated by immunoprecipitation with a polyclonal antiserum (54,55,56), and a-lactalbumin was measured by enzymatic assays (57,58) or immunoprecipitation with the antibody (59,60). The more recent approach is focused on studying transcription and stability and accumulation of specific messenger RNAs for milk proteins; cDNAs are available for some of the known caseins, a-lactalbumin, and whey acidic protein (61)(62)(63)(64)(65).
The a-lactalbumin genes from man and rat have been completely sequenced (66); partial sequences are available for several casein genes (67). Very recently, Hall and coworkers have compared the promoter regions of two alactalbumin and five casein genes and have found a region of high conservation and homology spanning from 140 to 110 in all the sequences examined (66). The 5' flanking re-gion, which includes the consensus sequence, is likely to be a locus for hormonal regulation of milk protein genes expression (66). The exact meaning of this conserved sequence will be elucidated by studying the expression of milk protein genes in transfected cell lines and transgenic animals

Growth and Differentiation In Vitro
It has been well established that the growth and differentiation of mammary cells can be induced in vitro. The in vitro systems have provided invaluable information about the morphogenesis, growth, and differentiation of the mammary gland. The first model system for the study of mammary growth and differentiation in vitro was reported in 1957 by Elias (68); this mammary explant culture system has been widely used over the years. A whole mammary gland culture system has been established by Ichinose and Nandi (69); more recently, several methods for primary culture of mammary epithelial cells have been developed. Several nonplastic substrates, such as collagen gels (70)(71)(72)(73)(74)(75), extracellular matrix (76,77), and L3TL1 adipocytes (52,78) have been used for their capacity to support and promote mammary epithelial growth and differentiation. Cell culture offers the opportunity to distinguish the activity of the different cell populations in the mammary gland; whole gland cultures are particularly suitable in studying the structural changes occurring during the various phases of development.
Most of the studies in culture have been performed using mouse mammary tissue from animals in different physiological stages; data obtained from in vitro experiments involving mammary glands from several other species such as rat, rabbit, cow, and goat have revealed some differences in the requirements for morphological and functional development (7).

Explant Culture
Explant culture, the first system developed, has provided infornation about the embryogenesis of the mammary gland (5-7). Body wall fragments from the mammary region of mouse embryos at 9 to 10 days of gestation have been shown to develop mammary buds when cultured on a mixture of cock blood plasma and chick embryo extract (79). Under similar culture conditions, explants with preformed buds developed branched mammary ducts In organ culture of mammary tissue from mouse and rat, insulin extends cell viability of explants in serum-free medium and stimulates cell replication (25). Addition of glucocorticoid and insulin can induce the development of rough endoplasmic reticulum (80)(81)(82). The expression of the differentiated phenotype requires the addition of prolactin (5,7,(80)(81)(82)(83). On the other hand, explants from rabbits can synthesize casein in the presence of prolactin alone (84).

Whole Gland Culture
Whole gland cultures of immature mouse mammary gland revealed differential hormonal needs of cell subpopulations for maintenance and development. Primary duct cells can be maintained and proliferate without any added hormone, whereas secondary and tertiary ducts require insulin for maintenance and show an enchanced proliferation in the presence of mineralcorticoid and prolactin. Maintenance of end bud cells requires a complex combination of hormones, and proliferative activity of these cells in vitro is modest (85).
In mouse whole gland cultures, lobuloalveolar growth can be induced by insulin, prolactin, and glucocorticoid or aldosterone, but the animal has to be previously primed with estrogen and progesterone (69,86,87); residual amounts of steroids carried over into the culture could account for the independence of lobuloalveolar formation on the added ovarian hormones in vitro. On the other hand, rat mammary gland does not require the priming with ovarian steroids (88).
Following hormonally induced lobuloalveolar growth, the whole gland in culture responds to the lactogenic hormones by synthesizing casein. After the first round of development, differentiation, and regression in the whole gland in culture, Ibnelli and Sorof were able to induce a second complete cycle by supplementing the serum-free medium with EGF (89).

Cell Culture
As discussed earlier, several substrates have been tested for their ability to support cell maintenance and proliferation. One of the problems that investigators have encountered in establishing an efficient cell culture system is reproducing the complex stromal-epithelial interactions that are important for mammary growth in vivo. In explants and whole gland cultures these interactions are retained, at least to some extent. Differentiation of mammary epithelial cells in culture requires proper substrate such as collagen and extracellular matrix, as well as lactogenic hormones such as insulin, glucocorticoid, and prolactin (72)(73)(74)76).
The epithelial cells of the rat embryo mammary bud at day 16 are apparently committed as mammary cells, and they respond to the three hormone combination by synthesizing casein (90). However, it is still not known when they first acquire this potential. No data are available about mouse mammary cells in embryo, but explant cultures of immature mouse mammary gland can synthesize caseins (91) and accumulate lactose synthetase activity in the presence of appropriate hormonal stimuli (92). Rat end bud cells grown on collagen matrix with serum and cholera toxin require insulin, glucocorticoid, estrogen, and prolactin to express functional and ultrastructural differentiation (93).
In epithelial cell cultures from mature mouse mammary gland, ultrastructural development and differentiation occur in the same hormonal conditions as in the explant cultures. Growth hormone and placental lactogen can replace prolactin in this system (74,94).
Epidermal growth factor and glucocorticoids can affect the growth efficiency of rat mammary epithelial cell cultures on different substrates, apparently modifying the rates of type IV collagen synthesis and degradation (95). According to McGrath et al., EGF appears to influence the cell type and the colony shape in serum-free culture of normal rat epithelial cells (96).
Maintenance of normal human breast tissue in organ culture requires the presence of insulin, cortisol, and prolactin or serum. Addition of progesterone in combination with insulin, cortisol, and prolactin to a serumsupplemented medium provides a better condition for epithelial growth (97). According to Hillman and co-workers, serum-supplemented medium enriched with insulin and cortisol can support explants from normal mammary tissue up to 6 months (98).
More recently, a serum-free medium containing insulin, cortisol, EGF, and bovine pituitary extract has been found to be capable of supporting the growth of normal human mammary epithelium up to 10 to 20 passages. Substitution of the bovine pituitary extract with prolactin or prostaglandin E 1 allows cell replication for 3 to 4 passages (99). Stampfer et al. reported an optimal condition for mammary epithelial cell growth in the presence of insulin, cortisol, EGF, and steroids that allowed the culture to be maintained up to 3 months (100).
Hormonal requirement for in vitro maintenance varies depending on the substratum. Yang et al. report that epithelial cells from human mammary fibroadenomas require EGF and cortisol when cultured in three dimensions into collagen gel matrix, whereas cortisol alone is sufficient for two-dimensional growth on collagen-coated dishes (101). In short-term monolayer culture of normal human mammary epithelium, cortisol and insulin stimulate growth; however, the degree of stimulation depends on the culture substratum. Cells organized into clusters undergo either inhibition of growth or terminal differentiation after a few passages; all isolated cells differentiate. The inhibited cells can resume growth when growth units are disrupted. This situation could resemble the one in mammary buds: The inhibited cells might provide pools for subsequent multiple cycles of differentiation (102).

Factors Involved in Regulation of Growth and Differentiation In Vitro
A very large amount of experimental data are available today; however, the differences between the experimental systems used and the complexity of the regulatory network operating during mammary gland growth and differentiation makes a clear organization of the subject extremely difficult. In this section we try to summarize the recent data concerning some of the effects of the principal agents known to affect mammary development and function in vitro.

Insulin
In serum-free culture, insulin stimulates cell replication but does not seem to be necessary for ductal or alveolar growth (103)(104)(105). Addition of an adrenal steroid, usually cortisol, to the culture medium induces the epithelial cells to develop the subeellular organelles, such as the rough endoplasmic reticulum, necessary for the synthesis and secretion of milk components (80)(81)(82). However, it is only in the presence of prolactin that the tissue achieves complete and functional differentiation (5,7,(80)(81)(82)(83).
Lithium ion (58,106), as well as sodium orthovanadate (107), can mimic the stimulatory effects of insulin on DNA synthesis in explant culture. On the other hand, EGF (108) and serum (109) appear to be able to replace insulin in the induction of both DNA synthesis and RER formation.
By measuring parameters like glucose-6-phosphate dehydrogenase activity and DNA synthesis, in vitro experiments suggested that mouse mammary epithelial cells in vivo undergo alternate phases of insulin resistance and sensitivity (25). Cells from immature or mature virgin mice are insulin insensitive but become responsive after 1 day in culture, regardless the presence of exogenous insulin. Mammary epithelial cells of pregnant or lactating mice are insulin sensitive, but loose their sensitivity during postlactational involution (25). The modulation of insulin sensitivity in mammary tissue could therefore be a key issue in the regulation of mammary development.
Inagaki and Kohmoto have shown that mouse mammary cells possess insulin receptors with a high (Kd = 1 nM) and a low (Kd = 20 nM) affinity, and the amount of the high affinity receptor is higher during the first half of pregnancy (110). Direct evidence supports the hypothesis that the tyrosine kinase activity of the insulin receptor is essential for insulin action (111).
In culture with cortisol and prolactin, insulin can induce terminal differentiation in mouse mammary explants by enhancing transcription of casein genes (112). Neither serum (109) nor EGF (108) can substitute for insulin in the synthesis of milk proteins in vitro. Lithium also cannot substitute for insulin in the induction of the terminal mammary differentiation, although the sensitivity of the mammary cells to lithium ions is greater in midpregnant than in virgin or lactating animals (58,106).
A further proof of the involvement of insulin in terminal differentiation is the fact that tissue from virgin mice can undergo terminal differentiation in vitro only after it has acquired sensitivity to insulin (25,103).

Cortisol
The concentration of cortisol required for maximal expression of casein and a-lactalbumin is remarkably different. Optimal casein production occurs in the presence of 3 MM cortisol, whereas 30 nM is sufficient for alactalbumin. Moreover, cortisol concentrations ranging from 300 nM to 3 ,M cause progressive inhibition of alactalbumin accumulation (55,113). This inhibitory effect of cortisol is reversed by prostaglandins (114). Cortisol might have a role in regulating the transcription of some whey protein genes; sequences bearing some resemblance to glucocorticoid receptor binding sites have been found in the 5' flanking region of human a-lactalbumin gene (66), and potential binding sites for glucocorticoid receptor were also found in the 5' flanking region of the whey acidic protein genes of rat and mouse (115).
Cortisol can remarkably extend the half-life of casein messenger RNA (116). In explants from rat mammary gland cultured in the presence of insulin, cortisol, and prolactin, glucocorticoid withdrawal reduces the half-life of casein mRNA to 1 hr and also results in the decrease of transcription of the casein gene (117). The effect of cortisol on casein mRNA is specific, since actin mRNA transcription occurs normally with or without the glucocorticoid (117). Cortisol also affects the lactogenic response of mammary tissue by regulating prolactin binding to the epithelial cells (118). These findings support the hypothesis of a regulatory role for glucocorticoids on the differentiation of mammary epithelium (5,7).

Prolactin
Prolactin is the main determinant of functional differentiation of mammary epithelium (5,7,83). Added to a mouse mammary gland culture in combination with insulin and glucocorticoid, prolactin induces milk protein gene expression in mammary epithelial cells. Prolactin alone is sufficient to trigger the synthesis of casein in explant culture from pregnant goats, but the efficient maintenance of the cultured tissue requires insulin and cortisol (119).
The responsiveness of mammary cells to prolactin can be influenced by several factors. For example, cells from midpregnant mice cultured on floating collagen gels show high responsiveness to insulin, cortisol, and prolactin and synthesize large amounts of casein for prolonged periods. Mammary epithelium from virgin mice is less sensitive to the lactogenic stimulus in culture; however, pretreatment of the virgin mouse with progesterone for 2 weeks produces a transient increase in responsiveness. A similar effect is produced by applying a pituitary allograft to increase prolactin levels (120). These data support the concept that the mammary cells' sensitivity to prolactin during pregnancy is regulated by progesterone and prolactin levels. Mammary fat pad cells appear not to be involved in the process of functional differentiation of the epithelium (120).
The influence of pregnancy, and lactation on the level of prolactin receptors in the mammary gland has been investigated in several animal species. In general, receptor levels are high in the mammary gland from virgin animal, decrease during pregnancy, and increase again after or near delivery (121)(122)(123). Prolactin upregulates the number of its receptors, whereas progesterone antagonizes this effect of prolactin (124,125). The number of prolactin receptors on mammary cells from rabbit (122) and mouse (121) in different developmental stages varies in inverse relationship to progesterone levels in serum. Glucocorticoid increases prolactin receptors in mammary cells in culture (118,126).
After binding of prolactin to the receptor, the complex hormone-receptor is internalized; as in the case of EGF and its receptor, internalization results in a downregula-tion of the receptor number (122,127,128). It is not clear whether internalization is required for prolactin action. It appears thus that prolactin regulates its receptor level both in a positive (129) and in a negative (127) fashion.
Prolactin is not necessary for the morphological development of rough endoplasmic reticulum in mouse mammary gland, but it increases the RNA content in the membranes However, prolactin has been shown to be necessary in vivo for the complete structural differentiation of mammary epithelium in cows (130). Prolactin depletion causes only partial development of rough endoplasmic reticulum and the cellular area occupied by Golgi apparatus decreases by 11% compared to normal (130). Like insulin, prolactin also appears to stimulate both ribosomal and transfer RNA accumulation; glucocorticoid prolongs this effect.
Growth hormone and placental lactogen can substitute for prolactin in epithelial cell culture of mouse mammary gland. Only at concentrations as high as 50 ,g/mL can bovine growth hormone stimulate casein synthesis in goat mammary tissue in culture (131). Bovine growth hormone cannot mimic the effect of prolactin on DNA synthesis in goat mammary explants (131).
Explants from pregnant rats can synthesize a-lactalbumin in the absence of prolactin, but those from virgin animals cannot unless the rat is pretreated with progesterone or estrogen (132). The production of a-lactalbumin in response to prolactin has been also studied in human mammary gland culture. Apparently, a-lactalbumin synthesis responds to prolactin stimulation in normal breast but is independent (when it occurs) of prolactin in malignant tissue. These findings suggest that prolactin receptors could be somehow defective in malignant breast cells (133). In mouse mammary epithelial cells cultured on floating collagen gels, prolactin increases ion transport across the epithelial layer, and affects passive permeability (134,135).
Using several agents affecting calcium transport and distribution, Bolander showed the involvement of the calcium-calmodulin system in prolactin-induced differentiation. However, this system appears not to serve as a sole mediator of prolaction action, since the calcium ionophore A23187, which reproduces the effect of prolactin on calcium accumulation, fails to induce differentiation in terms of casein synthesis (136).
Very recent work with tubulin-binding drugs provided the evidence that the action of prolactin to induce casein gene expression needs integrity of the microtubules in the mammary cell (137). Colchicine desensitizes the mammary epithelium to prolactin action (138). Estrogens 17-p-Estradiol (E2) is required in vivo for both ductal growth during sexual maturation and lobuloalveolar development during pregnancy (7). In human breast cell primary culture, E2 and progesterone act as antagonists in regulating cell multiplication (139): Estrogen acts as an inducer of cell proliferation; progesterone shifts the growth pattern to differentiation. In normal human breast epithelium, E2 stimulates the growth of ductal system and progesterone affects the development of acini. Mammary cell cultures from lactating rats proliferate actively in response to estradiol and progesterone at a rate similar to that of mammary epithelium during pregnancy (19). Calaf et al. reported that insulin, cortisol, and 17-pestradiol shorten the length of the cell cycle in cultured mammary cells from human normal breast and suggested that the cells could be hormonally induced to reenter the cell cycle from Go. They also reported that estrogen can modify the length of the S phase and proposed that this would account for the increase in the rate of DNA synthesis caused by the ovarian hormone (140).
In explants from human mammary gland at resting stage, estriol and estradiol caused a qualitatively different profile of growth stimulation (141). The third naturally occurring estrogen, estrone, does not show any stimulatory activity on thymidine uptake in cell culture (142). Proliferating response of normal mammary epithelium to estradiol is potentiated by the presence of stromal cells (143).
Estradiol induces modifications in the plasma membrane of epithelial cells; for example, the number and length of microvilli are increased. This appears to be a specific effect of estradiol, as progesterone, dexamethasone, and dihydrotestosterone are ineffective (144).
Casein synthesis and lactose synthetase activity are stimulated by 17-p-estradiol in explant cultures of mammary gland from pregnant mice. Mammary tissue from rats that have been ovariectomized and adrenalectomized loses the capacity to synthesize casein and to accumulate the casein mRNA although it retains the ability to respond to individual hornones (145). These data provide the evidence for the role of estrogen in mediating the tissue sensitivity to the lactogenic stimuli. Recently, Sheffield and co-workers found that the mammary gland growth induced by ovarian steroids is accompanied by an increase in cAMP-dependent protein kinase activity and tyrosine-kinase activity (146). By contrast, cGMP-dependent protein phosphorylation is decreased by progesterone but not by estrogen (146). Protein kinase C activity shows development-related regulation; a decrease is observed during pregnancy and throughout lactation. In addition, inhibitors of protein kinase C enhance a-lactalbumin production in explant culture with insulin, cortisol, and prolactin (147). In vivo, variations in protein kinase C activity appear to be related to the sex steroids rather than to the peptide hormones (148). These findings indicate that the pattern of protein phosphorylation in mammary epithelial cells undergoes substantial changes during mammary growth and differentiation, suggesting that it may represent an important regulatory step.

Epidermal Growth Factor
As described earlier, the evidence for the involvement of EGF in the development of the mammary gland has been provided by several studies. EGF stimulates proliferation of the epithelium in explants, whole gland, and cell cultures (70,89,149,150). Thymidine incorporation is stimulated about 5-fold, and the number of epithelial cells increases by 30 to 40%. Like insulin, EGF action is mediated by a receptor that has tyrosine kinase activity. Normal mammary cells have specific EGF receptors with a high and a low affinity (Kd = 0.1 nM; 3.6 nM) (150). The occupancy of EGF receptor for a half-maximal stimulation of DNA synthesis was about 10% of total receptors. Specific EGF binding to mammary gland membranes decreases at the beginning of gestation, then rises around day 5 to reach a maximum at day 10. During the second half of pregnancy and throughout lactation, there is a constant decrease in specific binding of EGF to the membranes. Thus EGF receptor levels are high during the proliferative phases of mammary gland development and decrease when the gland reaches functional differentiation (39). Whether the mitogenic activity of EGF requires the activation of tyrosine kinase is not clear.
Mouse mammary epithelial cells in culture show a spontaneous hyperpolarizing response when cultured in the presence of EGF and produce a depolarizing response when incubated with insulin (151). The hyperpolarizing response is mediated by activity of a Ca2 + -dependent K + channel, whereas the ionic species involved in the depolarization response have not been identified (152). The hyperpolarization induced by EGF occurs prior to cell proliferation and could be involved in the mitogenic action of EGF.
12-O-tetradecanoylphorbol-13-acetate (TPA), a potent tumor promoter (153), can mimic the effect of EGF on mammary cell proliferation in vitro (154). The phorbol ester can increase the specific binding of EGF to its receptors on the mammary cell (39,150), but it is not clear whether the stimulation occurs through an increase in the number of receptors, or rather a higher affinity of the binding. It has been proposed that TPA could inhibit the downregulation of the EGF receptors (155). This effect could be mediated by the TPA-induced kinase C activation. TPA can inhibit the production of milk proteins as well. TPA and EGF decrease prolactin binding to the epithelial cells by 50%.
When EGF is added to a culture with insulin, cortisol, and prolactin, it inhibits the induction of milk proteins by 50%. This effect and the stimulation of DNA synthesis are both obtained by using a physiological concentration of the growth factor (70,150). In addition, these effects are specific, as fibroblast growth factor, multiplication stimulating activity, nerve growth factor, and plateletderived growth factor were ineffective. EGF has been also shown to inhibit the synthesis of k casein and its mRNA accumulation in the presence of insulin, aldosterone, cortisol, and prolactin (156). However, EGF can stimulate the accumulation of a-casein mRNA when prolactin is omitted in mouse mammary explants (156). EGF appears to be essential for the formation of lobuloalveolar structure and for the expression of differentiative potential of the mouse mammary gland. In ovariectomized and sialoadenectomized mice, replacement therapy with ovarian steroids is not sufficient to restore the responsiveness of the tissue to the lactogenic stimuli. Moreover, the stimulatory effect of EGF is abolished by the absence of either 17-p-estradiol or progesterone, indicating a synergistic action of EGF and ovarian steroids in inducing lobuloalveolar growth and in enhancing the tissue ability to synthesize casein (45). Mammary explants from sialoadenectomized mice synthesize less casein in response to the lactogenic stimuli (44).
EGF is very likely to play an important role in the appearance of spontaneous mammary tumors in certain mouse strains. EGF concentration in the submandibular gland starts raising after 30 weeks of age, and shortly thereafter the mammary tumor incidence increases to reach a plateau at week 52. Moreover, the tumor incidence in mice that have been previously sialoadenectomized drops from 62.5 to 13% at 52 weeks of age, and the latency in the appearance of the neoplasm is shifted as much as 14 weeks. Even when sialoadenectomy was performed on tumor-bearing mice, it resulted in a rapid inhibition of tumor growth, whereas the tumor resumed its growth upon administration of EGF (157).
These results indicate that EGF stimulates proliferation of both normal and neoplastic growth of mammary epithelial cells. It has also been reported that EGF-like substances are produced by breast cancer cells (158,159). In HBL100 cells, a human mammary epithelial cell line that binds both EGF and glucocorticoids, EGF appears to enhance tyrosine phosphorylation of the glucocorticoid receptor, while dexamethasone prevents EGF effect (160).
Transforming Growth Factor P TGF-P is a 25,000 molecular weight polypeptide found in cultured normal cells from connective tissues and various epithelia. TGJF-P can inhibit normal and cancerous growth, but little is known about its physiological role.
When it was implanted in slow-release plastic pellets into the developing mouse mammary gland, TGF-P inhibited markedly mammary growth and morphogenesis, and the inhibitory effect was fully reversible (161). These results suggest that TIF-p mimics a natural negative growth regulator or is itself such a regulatory agent, which antagonizes the mitogenic effect of growth factors such as EGF.

Thyroid Hormones
Thyroid hormones are not necessary for ductal growth but seem to stimulate lobular development (7). They are also important for maintenance of alveoli during regression of the gland (35).
In terms of differentiation, thyroid hormones seem to enhance the tissue responsiveness to prolactin in mouse mammary gland by activating the prolactin receptors both in vivo and in explant culture (162,163). In rabbit mammary tissue, thyroid hormone enhances prolactininduced casein synthesis, acting probably at a posttranscriptional level (164). In explants from pregnant goats, neither L-T3 nor progesterone appears to influence the synthesis of casein and total proteins stimulated by prolactin (165). Thyroid hormones also regulate the level of EGF receptors in vivo at several stages of development. Lack of thyroid activity exerts a negative effect on EGF binding to mammary epithelial membranes by decreasing the number of binding sites (166).
The activity of thyroid hormone-binding inhibitor (THBI), which decreases the binding of T4 to T4-binding globulin in serum, is reported to be higher in lactating rat mammary gland than in pregnant or virgin tissue (167). However, it remains unclear whether the higher THBI activity and the consequent increase of free T4 available to the lactating epithelium is necessary in the regulation of lactation or whether it is just a consequence of the full differentiation.

Retinoids
It has been reported that retinoids inhibit differentiation and proliferation in the mammary gland of experimental animal (168,169). Retinoic acid, but not retinyl acetate, can reverse the growth stimulation induced by ovarian hormones in lactating rat mammary gland in organ culture (19). By contrast, in mouse mammary explants, retinoic acid potentiates the stimulatory effect of epidermal growth factor on DNA synthesis and enhances the specific binding of EGF to its receptor (170).
Vitamin A is known for its anticarcinogenic activity in the mammary epithelium (171,172), but its involvement in the functional development and maintenance of the mammary gland needs to be clarified. Mammary explants from vitamin-A deficient rats show no abnormality in producing casein and a-lactalbumin under appropriate hormonal stimulation (173). In mammary explants from mice, retinoic acid has no effect on casein synthesis, but significantly decreases a-lactalbumin production in a dosedependent fashion (170).
Progesterone alone or in combination with estrogen induces a remarkable increase in cellular retinoic acid binding protein (CRABP) in vitro (174). The actions of retinoids in vivo may be dependent on the available amount of CRABP in the tissue, which varies as the function of the developmental stage and physiological condition of the animal.

Vitamin D3
The addition of vitamin D3 significantly increases casein production in mouse mammary explants cultured in the presence of insulin, cortisol, and prolactin (175). Moreover, Bhattacharjee et al. found that mammary gland explants from rachitic rats and mice show a decreased production of milk proteins when cultured in the presence of insulin, prolactin, and glucocorticoid (176). Addition of 1,25-dihydroxycholecalciferol to the culture does not reverse the reduction, whereas pretreatment of the animal with the vitamin for 10 days can correct the defect. DNA synthesis is not affected by vitamin D deficiency (176). The action of vitamin D3 is specifically directed to the epithelial cells, that have been shown to concentrate the hormone in their nuclei (177). In addition, mouse mammary epithelial cells have been shown to possess specific vitamin D receptors, the level of which is enhanced by insulin, prolactin, and glucocorticoid (178). These data strongly suggest that vitamin D3 at physiological concentrations may be involved in the control of lactogenesis.

Some Intracellular Regulatory Agents
Calcium. Another factor affecting mammary growth in short-term cultures is Ca2 + concentration. Reduction of Ca2 + to levels below 0.08 mM leads to resumption of cell growth, but under these conditions differentiation does not occur. The effect of Ca2 + is specific and reversible (179). These findings suggest a possible modulation of growth-differentiation pathway by alterations in calcium level by such agents as phosphatidyl inositol, vitamin D3, or prolactin which affect calcium metabolism (180).
Polyamines. Spermidine is essential for milk protein synthesis in vitro (181); it can substitute for glucocorticoid in the induction of a-lactalbumin and, partially, casein synthesis in cultured mouse mammary tissue. Addition of methylglyoxalbis(guanylhydrazone) (MGBG), an inhibitor of polyamine biosynthesis, to the culture medium, causes the inhibition of prolactin-stimulated RNA, casein, and lipid synthesis. The inhibition is reversed by addition of exogenous spermidine (182).
In mammary explant cultures pretreated with insulin and cortisol, spermidine in combination with prostaglandin can mimic the action of prolactin on casein synthesis (183). Spermidine levels, as well as ODC activity, are elevated in response to prolactin. The requirement for spermidine for milk protein synthesis appears to be different among various species, but it is generally accepted that spermidine plays an important role in lactogenesis.
Poly(ADP-ribose). A decreased amount of poly(ADPribose) in the mammary gland tissue has been observed in vivo during pregnancy and lactation, suggesting an inverse relationship with the differentiation process (184). Data obtained from in vitro experiments offer a possible explanation for this decrease: in explant culture insulin stimulates the activity of poly(ADP-ribosyl)glycohydrolase, an enzyme responsible for poly(ADP-ribose) degradation, while prolactin appears to inhibit the specific synthetase activity. However, the role of poly(ADP-ribose) in mammary differentiation is unclear, as pharmacological inhibition of poly(ADP-ribosyl) synthetase results in an increase in a-lactalbumin accumulation (184,185).

Concluding Remarks
In this paper we have reviewed the recent literature related to the control of growth and differentiation of mammary gland. Experimental systems employed to assess the contribution of hormones and growth factors have been quite diverse, involving both in vivo and in vitro systems, animals from different species and strains and, more recently, several cell lines derived from normal or neoplastic tissue.
The most widely used approach in studies in vivo is to induce a specific hormonal deficiency and then to observe the ability of the mammary tissue to undergo the physiological changes related to pregnancy and lactation. The deficient condition may be obtained by different methods, such as surgical removal of the main hormonal source or antibodies against the hormone or its receptor. Replacement therapy can provide further information about the function of the specific factor studied. These approaches have been used in the recent study of the role of EGF in mammary gland during pregnancy.
In vitro systems have been successfully used to study the direct involvement of various hormones. It is possible to maintain mammary cells in the various culture systems using chemically defined medium and to monitor the degree of differentiation of the mammary tissue by measuring the production of milk proteins and their respective mRNA. These simplified systems have helped, at least to some extent, to clarify the contribution of the individual regulatory factors.
The mode of interactions among hormones and growth factors is different and varies with the physiological condition of the gland. Estrogens may be considered as the main effectors of mammary epithelial growth; however, the cell sensitivity to estrogen is controlled by progesterone or vice versa. In addition, recent studies have revealed the importance of EGF as mammary cell growth factor during pregnancy. It is also noteworthy that its production is regulated by ovarian steroids.
On the other hand, prolactin plays the major role in the induction of differentiation. Its receptor is regulated by some other hormones and growth factors such as insulin, glucocorticoid, and EGF. From these studies it is becoming more evident that hormone and growth factor receptors are not just passive carriers of a message but can actually modulate the signal by participating in complex regulatory circuits involving intracellular transmitters. Receptors are perhaps the main communication system through which mammary cells can adapt their behavior to the surrounding environment, and defective receptors are likely to be responsible for some kinds of mammary hyperplasia.
In the past, knowledge of the specific roles of different regulators of mammary function was mainly limited to the description of their effects on cell phenotype and cell physiology. The more powerful techniques now available allow us to focus on understanding how intracellular mediators are related to exogenous stimuli; on the other hand several current studies are trying to elucidate when and how determinants of mammary cell growth and differentiation are interacting with the processes of DNA replication, transcription and message translation.