Sialomucin Complex (Rat Muc4) Is Regulated by Transforming Growth Factor β in Mammary Gland by a Novel Post-translational Mechanism*

Sialomucin complex (SMC, rat Muc4) is a heterodimeric glycoprotein complex consisting of a mucin subunit ASGP-1 (for ascitessialoglycoprotein-1) and a transmembrane subunit ASGP-2, produced from a single gene and precursor. SMC expression is tightly regulated in mammary gland; the level in lactating mammary gland is about 100-fold that in virgin gland. In rat mammary epithelial cells, SMC is post-transcriptionally regulated by Matrigel by inhibition of SMC precursor synthesis. SMC is also post-transcriptionally regulated by transforming growth factor-β (TGFβ). The repression of SMC expression by TGFβ is rapid, is independent of TGFβ-induced cell cycle arrest, and does not require new protein synthesis. Unlike Matrigel, TGFβ does not reduce SMC protein synthesis, as SMC precursor accumulation is equivalent in TGFβ-treated and untreated cells. Instead, SMC precursor in TGFβ-treated cells is more persistent and does not become processed as rapidly into mature ASGP-1 and ASGP-2, indicating that TGFβ disrupts processing of SMC precursor. These results indicate that SMC, a product of normal mammary gland and milk, is regulated by TGFβ by a novel post-translational mechanism. Thus, SMC is regulated by multiple post-transcriptional mechanisms, which serve to repress potential deleterious effects of overexpression.

TGF␤ 1 is a member of a family of growth factors that have been shown to have extensive effects on the maturation and function of normal mammary gland. For example, TGF␤ implants introduced into the mammary glands of subadult virgin mice can inhibit ductal development (1). In addition, overexpression of TGF␤1 in the mammary glands of transgenic mice inhibited lobuloalveolar development and milk protein production (2). TGF␤ can induce expression of extracellular matrix proteins by human mammary epithelial cells in culture (3). Further, TGF␤ can inhibit ␤-casein production by a post-transcriptional mechanism in mammary tissue explants from midpregnant mice (4,5), although the molecular aspects of this mechanism are not presently known. Thus, in addition to its effects on mammary gland patterning, TGF␤ appears to play a role in regulating accumulation of milk proteins during pregnancy.
TGF␤ also regulates expression of another milk protein, SMC (6), which was originally discovered as a highly overexpressed glycoprotein complex on the surface of rat ascites 13762 mammary adenocarcinoma cells (7,8). SMC consists of a peripheral O-glycosylated mucin subunit ASGP-1 (7)(8)(9)(10) and an N-glycosylated integral membrane glycoprotein ASGP-2 (8,11). The complex is transcribed from a single gene as a 9-kilobase pair transcript (12,13) and translated into a single large polypeptide, which is proteolytically cleaved early in its biosynthesis. The subunits remain stably associated during transit to the cell surface (14). Recent studies have demonstrated that SMC is the rat homolog of human MUC4 (15). Cloning and sequencing of full-length human MUC4 showed 60 -70% amino acid identities between human MUC4 and rat SMC in nonmucin regions of both the ASGP-1 and ASGP-2 (16,17). MUC4 and SMC differ in their repeat domains in that the sequence of SMC does not contain the 16-amino acid repeat cloned and sequenced in the original description of MUC4 (17). The high degree of similarity between MUC4␤, the human MUC4 analog of ASGP-2, and rat ASGP-2 provides strong evidence that they are homologous proteins. Several studies suggest that the twosubunit SMC is a multi-functional glycoprotein complex. Through its highly O-glycosylated tandem repeat domain, ASGP-1 can provide anti-recognition and anti-adhesive properties to tumor cells (9,10,18). Furthermore, SMC expression in tumor cells reduces their killing by natural killer cells (19). This anti-recognition property may be important to the high metastatic capacity of the 13762 ascites cells (7,9,20). ASGP-2 has two epidermal growth factor-like domains, which have all of the consensus residues present in active members of the epidermal growth factor family (12). Moreover, SMC has been shown to bind to and modulate phosphorylation of the receptor ErbB2 (21). Supporting the conclusion that ASGP-2 is a ligand is the observation that ErbB2 is constitutively phosphorylated in the 13762 ascites cells and associated with a multimeric complex of signaling components, including Src (22) and all of the components of the Ras to MAP kinase mitogenic pathway (23). Thus, the transmembrane subunit ASGP-2 is proposed to modulate signaling through the epidermal growth factor family of receptors via its interaction with ErbB2 (21,24), the critical receptor for formation of active heterodimeric class I receptor tyrosine kinases (25). This interaction may play a role in the constitutive phosphorylation of ErbB2 in the 13762 ascites cells (22) and the rapid growth of these cells in vivo.
Sialomucin complex expression has been described in a number of normal secretory epithelial tissues in the adult rat (26,27) and appears to have multiple and complex regulatory mech-anisms. SMC protein is abundant in lactating mammary gland, but its level is very low in the virgin gland. However, the transcript for SMC is present at high levels in the virgin gland and does not change during pregnancy (6), suggesting that SMC expression is post-transcriptionally regulated in normal rat mammary gland. SMC synthesis is induced rapidly in cultured primary mammary epithelial cells from either normal pregnant or virgin rats. When mammary cells are cultured in Matrigel, a reconstituted basement membrane that stimulates casein expression, SMC protein, but not transcript levels, are significantly reduced. This post-transcriptional regulation is achieved by a ϳ10-fold reduction in SMC precursor biosynthesis when the cells are cultured in Matrigel. Interestingly, Matrigel has no effect on either the level of SMC or its transcript in cultured 13762 mammary tumor cells. TGF␤1 can also regulate SMC levels in normal cultured mammary epithelial cells, but not the ascites tumors, by a post-transcriptional mechanism (6).
In the present study, we have characterized the mechanism of post-transcriptional regulation of SMC by TGF␤ in cultured primary mammary epithelial cells. TGF␤ inhibits induction of SMC expression when the cells are put into culture; the repression of SMC expression is rapid and is independent of TGF␤induced cell cycle arrest. The presence of TGF␤ does not affect the ratio of membrane-bound to soluble form of SMC produced, nor does it affect the rate of SMC turnover in these cells. Unlike Matrigel, which inhibits SMC precursor synthesis, TGF␤ has little effect on SMC precursor synthesis. Instead, TGF␤ alters the processing of SMC precursor into mature SMC (ASGP-1/ ASGP-2), a novel TGF␤ action, which appears not to be a consequence of the effects of TGF␤ on transcription.

EXPERIMENTAL PROCEDURES
Materials-The MAT-B1 ascites subline of the 13762 rat mammary adenocarcinoma was maintained by weekly passage (28). Anti-ASGP-2 polyclonal antiserum was prepared against purified ASGP-2 (14) and has been used extensively for immunoprecipitations in previous studies (6,14,21,26,27). The mouse monoclonal antibody 4F12 was elicited using purified SMC, recognizes an epitope in the N-terminal 53 amino acids of ASGP-2 and has been used extensively for immunoblots (6,21,26,27). Anti-C-Pep polyclonal antiserum used for immunoprecipitations was prepared against the C-terminal peptide of rat ASGP-2, N-SMNKFSYPDSEL-C (26). Anti-cyclin A polyclonal antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Antismooth muscle actin mouse monoclonal antibody was purchased from Sigma. TGF␤ was purchased from R&D Systems, Inc. (Minneapolis MN). Cycloheximide was purchased from Calbiochem (La Jolla, CA). Puromycin and tunicamycin were purchased from Sigma. Cell culture materials were obtained from Life Technologies, Inc., unless otherwise noted.
Cell Culture and Analysis-Primary mammary epithelial cell cultures were prepared from virgin Fischer 344 female rats by collagenase digestion of dissected mammary tissue. Briefly, mammary glands excised from virgin or pregnant female Fischer 344 rats were minced, resuspended in digestion medium comprising 1 mg/ml collagenase type II (Worthington Biochemical Corp., Freehold, NJ) and 100 units/ml penicillin, 100 g/ml streptomycin in Ham's F-12 medium (Life Technologies, Inc.) and incubated at 37°C with shaking for 45 min. Fully and partially digested epithelial cell clusters were pelleted and incubated a second time in digestion buffer at 37°C with shaking for 45 min. Digested epithelial cell clusters were pelleted, resuspended in PBS, and passed through a 520-m cell sieve to remove undigested material. Mammary epithelial cell clusters in the resulting filtrate were captured on a 70-m nylon membrane. Cell clusters were collected by rinsing the membrane with PBS and were subsequently washed three times in PBS prior to plating. Mammary epithelial cells were maintained in Ham's F-12 medium containing 10% FCS and 100 units/ml penicillin, 100 g/ml streptomycin. TGF␤ was added at a final concentration of 200 pM either at the time of plating or after 24 h of culture. Cells were cultured at 37°C in 5% CO 2 for 48 h prior to harvest. Cells were collected from culture on plastic dishes by scraping cells off the dish. Except where indicated, harvested cells were pelleted, washed with PBS, and lysed in 100 l of 1% SDS in water. Protein concentration of the cell lysates was determined by Lowry assay, and 5 g of total protein was loaded for immunoblot analysis.
Western Blotting-For Western blots, SDS-PAGE was performed under reducing conditions using 6% polyacrylamide gels and the mini-Protean II system (Bio-Rad). Resolved proteins were transferred to nitrocellulose membranes which were subsequently blocked with 5% nonfat dry milk in Tris-buffered saline with 0.5% Tween 20. After a 1-h incubation in primary antibody diluted in 1% bovine serum albumin/ Tris-buffered saline with 0.5% Tween 20, the membranes were incubated with horseradish peroxidase-conjugated goat anti-mouse IgG (Fcspecific; Pierce) diluted 1:20,000 in 1% bovine serum albumin/Trisbuffered saline with 0.5% Tween 20. Signals were detected with the Renaissance TM enhanced chemiluminescence kit (NEN Life Science Products).
Labeling of Mammary Epithelial Cells-Mammary epithelial cells were isolated from virgin rats and cultured on plastic in Ham's F-12 medium supplemented with 10% FCS. After 24 h TGF␤ was added to half the samples at a final concentration of 200 pM. After an additional 24 h cells were washed twice with PBS, starved for 30 min in Cys/Metfree Dulbecco's minimal essential medium supplemented with 100 units/ml penicillin, 100 g/ml streptomycin, 2 mM glutamine, and 10 mM Hepes, and incubated in 1 ml of labeling medium (starvation medium ϩ 550 Ci/ml

SMC (ASGP-2) Expression in Cultured MEC in the Presence
or Absence of TGF␤-We have shown previously that SMC/ Muc4 protein is induced rapidly when isolated mammary epithelial cells are cultured as a monolayer on plastic tissue culture dishes. Further, we demonstrated that TGF␤ posttranscriptionally regulates SMC in these cells. The aim of the current studies is to define the mechanism for post-transcriptional regulation of SMC by TGF␤. In all tissues studied to date, including mammary gland (8,17), ASGP-1 and ASGP-2 are present as a complex, allowing us to use immunoblotting of ASGP-2 for the analysis of SMC. Moreover, our monoclonal antibody 4F12, which recognizes an epitope in the N-terminal 53 amino acids of ASGP-2 is more sensitive and more specific than those for ASGP-1. This antibody recognizes both membrane-bound and soluble SMC (ASGP-2) and has been used extensively to study the expression of SMC (ASGP-2) in multiple tissues (26,27).
A time course was performed to characterize the expression pattern of SMC (ASGP-2) in the presence or absence of TGF␤ in cultured MEC. Isolated MEC from virgin rats were cultured on plastic in Ham's F-12 medium supplemented with 10% fetal calf serum with or without 200 pM TGF␤. Cells were harvested at times ranging from 0 to 24 h after plating and lysed, and total protein was quantified. SMC (ASGP-2) content was analyzed by immunoblotting with mAb 4F12, and actin was meas-ured as a loading control. In the absence of TGF␤, SMC (ASGP-2) appears at about 4 h after plating and reaches maximal levels only after 24 h (Fig. 1A). In the presence of TGF␤, SMC (ASGP-2) also appears at about 4 h after plating but levels off by about 12 h. The maximal level of SMC (ASGP-2) in MEC cultured in the presence of TGF␤ is about 50% of that in cells cultured without TGF␤ (Fig. 1B).
The specificity of the TGF␤ effect was studied by the addition of a neutralizing antibody to TGF␤. MEC were cultured on plastic in Ham's F-12 medium supplemented with 10% fetal calf serum in the presence or absence of 200 pM TGF␤ or a neutralizing antibody to TGF␤. 30 l of anti-TGF␤ antibody was incubated with the TGF␤ for 30 min at 4°C prior to addition to the culture. After 24 h the cells were analyzed for SMC (ASGP-2) by immunoblotting with mAb 4F12, and actin blotting was used as a loading control. In the presence of TGF␤, SMC (ASGP-2) levels were inhibited by approximately 50% (Fig. 2, A and B) as seen in Fig. 1A. However, in the presence of the neutralizing antibody, SMC (ASGP-2) levels were substantially less inhibited by TGF␤, indicating that the inhibition of SMC (ASGP-2) expression by TGF␤ is specific. TGF␤ is known to induce cell cycle arrest in epithelial cells, and the inhibition of SMC (ASGP-2) expression by TGF␤ may be one of the outcomes of cell cycle arrest. To determine if inhibition of SMC (ASGP-2) expression is a result of reduced cell number by TGF␤ treatment, MEC were cultured on plastic dishes in Ham's F-12 supplemented with 10% fetal calf serum in the presence or absence of 200 pM TGF␤. After 24 h, cells were harvested using an enzyme-free cell dissociation buffer and counted. Cells were lysed, and equal numbers of cells or equal amounts of total protein were analyzed by immunoblot with mAb 4F12. The inhibition of SMC (ASGP-2) expression is apparent when equivalent cell numbers (Fig. 2C) or equivalent total protein is analyzed (Fig. 1A). These results indicate that reduction of SMC (ASGP-2) levels by TGF␤ is not a result of reduction of cell number (or cell death).
To further investigate the relationship between SMC (ASGP-2) repression by TGF␤ and the cell cycle, the timing of SMC (ASGP-2) repression by TGF␤ was compared with that of TGF␤-induced cell cycle arrest. MEC from virgin rats were cultured on plastic dishes in Ham's F-12 medium supplemented with 10% fetal calf serum in the presence or absence of 200 pM TGF␤. Cells were harvested after 24 or 48 h of culture for immunoblot analyses with mAb 4F12, anti-cyclin A, and anti-actin antibodies. During the first 24 h of culture, very little cyclin A is produced by the MEC, a marker for progression through the cell cycle (29), suggesting that the cells are not cycling (dividing) in the presence or absence of TGF␤ (Fig. 3). However, during this time period, SMC (ASGP-2) levels are reduced in the TGF␤-treated cultures. During the second 24 h, cells cultured without TGF␤ produce cyclin A, indicating that they are cycling. Those cells cultured with TGF␤ produce less cyclin A, indicating that TGF␤ is causing cell cycle arrest. However, the reduction in SMC (ASGP-2) levels in the TGF␤treated cells are similar at the 24-and 48-h time periods. Thus, since TGF␤ reduces SMC (ASGP-2) levels when MEC are not cycling, the reduction of SMC (ASGP-2) levels by TGF␤ is independent of TGF␤-induced cell cycle arrest. Moreover, these data suggest that reduction of SMC (ASGP-2) by TGF␤ occurs by a different mechanism than TGF␤-induced cell cycle arrest.
TGF␤ can reduce SMC (ASGP-2) levels in cultured MEC in less than 24 h, suggesting that this is a rapid response. To determine more accurately how fast TGF␤ can reduce SMC (ASGP-2) levels, MEC were cultured for 24 h to induce high levels of SMC (ASGP-2). TGF␤ was then added to a final concentration of 200 pM to half of the cells, and samples were harvested 6 and 24 h later for immunoblot analyses. SMC (ASGP-2) expression was inhibited by TGF␤ within 6 h of its addition (Fig. 4A); the inhibition was more pronounced 24 h after addition of TGF␤. The relatively rapid effects suggest that new transcription and protein synthesis may not be necessary for TGF␤-mediated repression of SMC (ASGP-2) levels. To test this idea, MEC from virgin rats were cultured for 24 h, then 200 pM TGF␤ and/or 10 g/ml cycloheximide were added to the media. Cells were harvested after 6 h for immunoblot analyses with mAb 4F12. As demonstrated previously, SMC (ASGP-2) levels were reduced by TGF␤ within 6 h of its addition. The presence of cycloheximide, which inhibits new protein synthesis, did not reverse reduction of SMC (ASGP-2) levels by TGF␤ (Fig. 4B), indicating that no new protein synthesis is required for TGF␤ to reduce SMC (ASGP-2) levels.
Effect of TGF␤ on the Production of Soluble SMC (ASGP-2)-Normal mammary tissue produces both soluble and membrane forms of SMC (ASGP-2) in a ratio of ϳ60% membrane:40% soluble form (26). One possible effect of TGF␤ is alteration of the ratio of membrane-bound to soluble form of SMC (ASGP-2) by stimulating conversion of the membrane precursor to soluble form. Thus, in the presence of TGF␤, the detectable SMC (ASGP-2) in the cell would be reduced because it would be secreted from the cell. To test this possibility, MEC were cultured in the presence or absence of 200 pM TGF␤ for 48 h and lysed in radioimmune precipitation buffer (150 mM NaCl, 1% Nonidet P-40, 0.5% deoxycholic acid, 0.1% SDS, 50 mM Tris base, pH 8.0). The lysates were sequentially immunoprecipitated twice with anti-C-Pep, a polyclonal antibody that recognizes an epitope in the C-terminal (cytoplasmic) domain of SMC (ASGP-2), and once with polyclonal anti-ASGP-2. Two rounds of immunoprecipitation with anti-C-Pep will clear the cell lysate of membrane-bound form of SMC (ASGP-2) (26), while the polyclonal anti-ASGP-2 recognizes the remaining SMC (ASGP-2), the soluble form. This technique (with these antibodies) has been used to study the ratio of membranebound to soluble of SMC (ASGP-2) in multiple tissues (26,27). Immunoprecipitates were analyzed by immunoblotting with mAb 4F12, which recognizes both membrane and soluble SMC (ASGP-2) (26). The presence of TGF␤ does not affect the ratio of membrane to soluble form (Fig. 5A). Both treated and untreated cells produce ϳ55% membrane-bound and ϳ45% soluble form (Fig. 5B), and SMC (ASGP-2) soluble form was detected in the conditioned media from both treatment groups. The only difference was that the overall level of SMC (ASGP-2) produced in the TGF␤-treated cells was lower than that pro-duced in untreated cells. These data rule out the possibility that the apparent decrease in SMC (ASGP-2) levels in the cultured MEC is due to a shift of membrane SMC (ASGP-2) to the soluble form.
Effect of TGF␤ on Turnover of SMC (ASGP-2)-Since the effect of TGF␤ on SMC (ASGP-2) expression is rapid, another potential mechanism for its repression is the acceleration of SMC (ASGP-2) turnover. To investigate this possibility, virgin MEC cultured in the presence or absence of 200 pM TGF␤ were treated with 5 g/ml (final concentration) of cycloheximide or puromycin to inhibit new protein synthesis. Alternatively, MEC cultured in the presence or absence of TGF␤ were treated with 5 g/ml tunicamycin, a drug that inhbits N-glycosylation (30). We have found that treatment of MEC with tunicamycin inhibits new synthesis of SMC (ASGP-2), 2 and as a result, this drug can be used as an alternative (potentially less toxic) method for inhibiting SMC (ASGP-2) synthesis. Cells were harvested at times ranging from 0 to 24 h after addition of inhibitors. Protein concentrations were quantified by Lowry assay, and 5 g of total protein were subjected to immunoblot analysis with mAb 4F12. The stained bands were quantified by densitometry and the half-life of SMC (ASGP-2) in treated and untreated cells was estimated. Table I summarizes the estimated half-life of SMC (ASGP-2) in TGF␤-treated and untreated MEC for each inhibitor used. Thus, these data suggest that TGF␤ does not significantly change the turnover of SMC (ASGP-2) in normal cultured MEC. (ASGP-2) precursor and mature ASGP-2. Samples were subjected to SDS-PAGE and fluorography. Total protein synthesis was similar in both treated and untreated samples, indicating that TGF␤ did not inhibit total protein synthesis. The amount of accumulating SMC (ASGP-2) precursor detected in both treated and untreated samples was similar for all time points (Fig. 6A). Precursor accumulation was quantified by densitometry (Fig. 6B). These data demonstrate that similar levels of precursor were synthesized in the presence or absence of TGF␤, indicating that TGF␤ does not affect the rate of SMC (ASGP-2) precursor biosynthesis (message translation). Thus, the reduction of SMC (ASGP-2) levels by TGF␤ involves a different mechanism from that for Matrigel, which inhibits SMC (ASGP-2) precursor biosynthesis (6).

Biosynthesis of SMC (ASGP-2) in the Presence or Absence of TGF␤-To
Effect of TGF␤ on Processing of SMC (ASGP-2) Precursor-Since TGF␤ does not affect SMC translation or the turnover of the mature protein, another possibility is that TGF␤ could affect the processing of the SMC precursor into mature ASGP-1/ASGP-2. In order to test this possibility, a pulse-chase experiment was performed. MEC were cultured 24 h, and TGF␤ was added to half the cells to a final concentration of 200 pM. After an additional 24 h, the cells were pulse-labeled for 30 min with [ 35 S]Cys ϩ[ 35 S]Met. Following the pulse, the cells were washed in prelabeling medium twice and incubated in chase medium for times ranging from 1 to 8 h. TGF␤ was present in half the samples at a concentration of 200 pM throughout the labeling procedure. After the chase, cell lysates were immunoprecipitated with anti-ASGP-2 antibodies. Immunoprecipitates as well as an aliquot of non-immunoprecipitated cell lysate were subjected to SDS-PAGE and fluorography. Total labeled protein was similar for both samples with and without TGF␤, suggesting that protein synthesis is not inhibited by TGF␤ in these cells (Fig. 7A). The level of SMC (ASGP-2) precursor is similar for treated and untreated samples, again suggesting that TGF␤ does not inhibit the translation of SMC (ASGP-2) (Fig. 7A). To determine whether TGF␤ affects processing of SMC precursor into mature SMC (ASGP-2), the bands for SMC precursor and mature ASGP-2 were quantified by densitometry and the results were plotted (Fig. 7, B and C). In the absence of TGF␤, SMC precursor is processed rapidly into mature ASGP-2, such that Ͼ50% of the precursor is processed into mature ASGP-2 in 1 h (Fig. 7B). In the presence of TGF␤, the SMC precursor is processed more slowly; after 4 h, only about 50% of SMC precursor had disappeared. In addition, much less mature ASGP-2 accumulated in the TGF␤-treated samples (Fig. 7C). The fact that ASGP-2 appears more slowly than precursor disappears suggests that unprocessed precursor is being degraded. These results indicate that TGF␤ affects the processing of the SMC precursor into mature SMC (ASGP-2), causing the apparent reduction in SMC (ASGP-2) levels when cells are cultured in the presence of TGF␤. Once again, these   data point to a different mechanism of post-transcriptional regulation of SMC from that with Matrigel, which occurs by a reduction in SMC precursor synthesis. DISCUSSION SMC is expressed in a number of normal rat tissues and is developmentally regulated in normal rat mammary gland. Without tight regulation, overexpression of this protein could have profound deleterious effects on the mammary epithelium, including disruption of cell-cell and cell-matrix interactions. To achieve this precise regulation, a complex series of regulatory mechanisms has evolved, involving responses at several levels. Indeed, we are just beginning to elucidate factors and mechanisms involved in regulation of this protein in mammary epithelial cells. Here, we demonstrate the mechanism by which SMC (ASGP-2) is regulated in mammary epithelia by TGF␤ and provide evidence that this mechanism is different from that reported for regulation of SMC (ASGP-2) by Matrigel, a reconstituted basement membrane mimicking one type of extracellular matrix effect on the epithelium.
TGF␤ has numerous effects on the normal developing mammary gland. It inhibits the growth of primary mammary epithelial cells as well as that of several transformed mammary epithelial cell lines (31)(32)(33). TGF␤ can inhibit ductal growth in the virgin mouse mammary gland but does not influence alveolar morphognensis or DNA synthesis in the alveolar cells of pregnant mice. These data suggest that TGF␤ plays an important role in normal mammary gland patterning by controlling spacing of ducts to allow room for alveolar development during pregnancy, but does not affect alveolar development directly. In addition, TGF␤ can inhibit casein and SMC (ASGP-2) synthesis in pregnant mouse mammary organ explant cultures (4) and isolated virgin or mid-pregnant (data not shown) MEC (6), respectively. On the other hand, Sudlow and others (5) report that TGF␤ does not inhibit casein synthesis from lactating organ explant cultures or MEC from lactating mice. Taken together, these data suggest that TGF␤ controls synthesis and accumulation of milk proteins during pregnancy in addition to its role in development.
We had shown previously that SMC (ASGP-2) levels could be regulated post-transcriptionally in cultured rat mammary epithelial cells by both Matrigel and TGF␤. In Matrigel regulation of the expression of SMC (ASGP-2) is markedly different from that of ␤-casein. Matrigel lowers SMC (ASGP-2) levels while it enhances ␤-casein levels. However, regulation of SMC (ASGP-2) and ␤-casein by TGF␤ is similar. 1) Expression of both is repressed under conditions that do not inhibit total protein synthesis. 2) Both SMC (ASGP-2) and caseins are strongly inhibited by physiological picomolar doses of TGF␤ (4,6). 3) The mechanism of regulation appears to be post-transcriptional for both proteins. These data support a role for TGF␤ as an inhibitor of milk protein synthesis and accumulation in the virgin or pregnant mammary gland.
TGF␤ represses SMC (ASGP-2) levels in mammary epithelial cells whether or not the mammary epithelial cells are cycling. This result suggests that TGF␤-induced cell cycle arrest and TGF␤ repression of SMC (ASGP-2) levels occur by different mechanisms (different signaling pathways). Administration of TGF␤ to the mammary glands of pregnant mice does not influence DNA synthesis of alveolar cells, the cells that produce caseins and SMC (ASGP-2) (milk proteins) (1,26,34). Taken together, these data indicate that the repression of SMC (ASGP-2) levels by TGF␤ is independent of the cell cycle and is not a result of growth inhibition. The repression of SMC (ASGP-2) expression by TGF␤ is not the result of an increase in the production of the soluble, secreted form of SMC (ASGP-2), inhibition of biosynthesis of the SMC precursor, or an increase in SMC (ASGP-2) turnover. Instead, TGF␤ interferes with the processing of SMC precursor into mature ASGP-1/ASGP-2, a novel post-translational effect and mechanism (Fig. 8).
In the mammary gland there are several different posttranscriptional mechanisms for controlling (milk) protein expression, and the specifics of these mechanisms are beginning to be elucidated. For example, SMC (ASGP-2) is regulated by Matrigel by inhibition of its biosynthesis and TGF␤ by disrupting SMC precursor processing. ␤-Casein mRNA is stabilized by the presence of prolactin (35), and its synthesis is inhibited by After a 30-min pulse labeling, the medium was replaced with non-radioactive medium, and cells were harvested at various times, as indicated. Samples were immunoprecipitated, and immunoprecipitates and non-immunoprecipitated whole cell lysate samples were subjected to SDS-PAGE and fluorography. B, plot of SMC precursor processing into ASGP-2. The precursor bands from A were quantified by densitometry, and the results were plotted. C, plot of accumulation of mature ASGP-2. The mature ASGP-2 bands from A were quantified by densitometry, and the results were plotted.
TGF␤ (4,5). Lactoferrin message is induced and stabilized by cell rounding (36). Whey acidic protein has an undefined posttranscriptional regulatory mechanism. When MEC are cultured on plastic or basement membrane, whey acidic protein message is transcribed, but requires formation of a hollow alveolar structure with a closed lumen for its synthesis and secretion (37).
TGF␤ has been implicated in a number of post-transcriptional regulatory mechanisms. TGF␤ can regulate gene expression post-transcriptionally by increasing or decreasing the stability of mRNAs. In osteoblast cell cultures TGF␤ can inhibit collagenase 3 expression by accelerating the decay of its transcript (38). In vascular smooth muscle cells TGF␤ can stabilize lysyl oxidase mRNA (39). Other mechanisms of post-transcriptional regulation by TGF␤ have also been proposed. For example, TGF␤ inhibits cdk4 translation in Mv1Lu lung epithelial cells; the CDK4 5Ј-untranslated region is involved in its translational regulation (40). In human prostate cancer cell lines, TGF␤ induces higher secreted levels of collagenase MMP-2 by increasing the stability of the secreted 72-kDa proenzyme (41). TGF␤ represses SMC (ASGP-2) levels by disrupting SMC precursor processing, suggesting that it actually regulates one of the factors necessary for SMC precursor processing. This effect is rapid and does not require new protein synthesis. Thus, this appears to be a different post-transcriptional regulatory mechanism from others reported for TGF␤. The results in this study, along with another recent study, provide a clearer picture of the regulation of SMC (ASGP-2) in normal developing mammary gland and allow us to update our model. Virgin rat mammary epithelial cells are primed for SMC (ASGP-2) production by the presence of SMC (ASGP-2) transcript, whose expression is regulated by cell differentiation and insulin/insulin-like growth factor. 3 Translation of this transcript is repressed by an inhibition related to cell environment, mimicked by Matrigel. High levels of TGF␤ in the virgin mammary gland further control SMC precursor by regulating its processing. As pregnancy proceeds the cell environment changes, and active TGF␤ levels decrease, 4 allowing for increased translation and processing. Finally, at the onset of lactation TGF␤ levels be-come undetectable, and SMC (ASGP-2) is translated and processed at a higher levels. Isolation of MEC causes disruption of the cell environment and loss of TGF␤ signaling, resulting in an overexpression of SMC (ASGP-2), which can be reversed by Matrigel and TGF␤ addition. Similarly, neoplastic transformation can lead to a loss of cell polarization and basement membrane contact, releasing the inhibition on precursor synthesis. Loss of TGF␤ responsiveness during tumor progression (42) will also release the post-translational processing block and lead to frank overexpression of SMC (ASGP-2), with its potential for deleterious consequences.
These studies raise other questions about regulation of SMC (ASGP-2) expression in normal mammary gland by TGF␤. Two signaling pathways have been implicated in TGF␤ effects: the SMAD pathways (43) and the MAP kinase pathway (44). Preliminary experiments with MAP kinase pathway inhibitors suggest that SMC (ASGP-2) regulation by TGF␤ does not involve the MAP kinase pathway. Whether SMADs are involved is uncertain, and studies are currently under way to investigate this possibility. Another question is whether TGF␤ regulates casein and SMC (ASGP-2) by similar mechanisms. This seems unlikely because of the specificity of the effect on SMC (ASGP-2), occurring at a specific stage of SMC (ASGP-2) processing. One possible explanation for the TGF␤ effect on SMC (ASGP-2) is that it inhibits the enzyme that cleaves SMC (ASGP-2) precursor into ASGP-1 and ASGP-2. MUC2 has the same sequence, N-GDPH-C, at its putative cleavage site (24), suggesting that it may be cleaved (processed) by the same enzyme (or family of enzymes). Thus, if the cleavage enzyme is regulated by TGF␤, this mechanism of regulation may be applicable to other mucins, though probably not to casein. However, the TGF␤ effect could also be due to a post-translational modification, such as glycosylation or phosphorylation, which could affect both SMC (ASGP-2) and casein processing and their subsequent behavior. Additional experiments are in progress to investigate these possibilities.