| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Breast Cancer and the Environment Research Center, Department of Physiology (A.K., S.Z.H.) and Cell and Molecular Biology Program (M.D.A.), Michigan State University, East Lansing, Michigan 48824
Address all correspondence and requests for reprints to: Sandra Z. Haslam, Ph.D., Department of Physiology, 2201 Biomedical and Physical Sciences Building, Michigan State University, East Lansing, Michigan 48824. E-mail: shaslam{at}msu.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Progesterone action in the breast is mediated by two receptors, progesterone receptors (PR) A and B. PR belong to the subfamily of classical nuclear steroid receptors. Both PRA and PRB are transcribed from the same gene by the use of different promoters in human and rat (9). PRB differs from PRA by 164 additional amino acids at the N terminus (10). Despite structural similarities, PRA and PRB possess different functional activities in vitro and in vivo. The PRB-specific N-terminal region contains an additional transactivation function domain (11) that allows PRB to bind to a different set of coactivators than would bind to PRA (12). From in vitro studies, PRB is considered a stronger transactivator, whereas PRA has been shown to act as a transrepressor and can be an inhibitor of other steroid receptors, such as estrogen receptor-
and PRB (13, 14). In the human breast cancer cell line T47D, PRB function is mostly activated by ligand binding, whereas PRA can act in a ligand-independent manner (15). In addition, the two PR isoforms show different responses to antagonists (16). The action of PRA is inhibited by the P antagonist RU486, whereas RU486 can act as a weak agonist for PRB (17).
Immunohistochemical analysis of the adult, normal human premenopausal breast has shown that PRA and PRB proteins are coexpressed in equimolar amounts in the same cell in 10–20% of epithelial cells (18). This ratio has been observed to change in breast cancer. A higher prevalence of PRA is associated with poor prognosis, whereas PRB prevalence is associated with a lack of HER-2/neu expression and a more differentiated tumor phenotype (18, 19). Therefore, advancing our knowledge about how the ratio of PR isoforms impacts biological responses has potential clinical importance.
We have previously examined PRA and PRB isoform expression by immunohistochemistry in the BALB/c strain mouse mammary gland and found that it differs significantly from that reported for the human breast. In the mouse, PRA and PRB are expressed in a distinct spatial and temporal pattern. PRA is the predominant isoform expressed in the pubertal and mature virgin mammary gland, whereas PRB is the predominant isoform detected only during pregnancy. Furthermore, PRA and PRB rarely colocalize in the same cell in the pregnant mammary gland (20). Therefore, P action in the mouse mammary gland is conveyed mainly by either PRA or PRB alone, whereas in the human breast, both isoforms are expressed and may contribute to the biological outcome of P signaling.
Mouse and rat mammary glands are significantly different with regard to their histoarchitecture in the nulliparous state. In the mouse, the mammary gland develops and maintains a primarily ductal organization before pregnancy, when major lobuloalveolar development takes place. In the rat, ductal and lobular development occur simultaneously in the nulliparous state, and during pregnancy, there is an expansion of the mammary lobules. Human and rat mammary glands share a similar ductolobular histoarchitecture in the nonpregnant state (21). Additionally, the human and rat share similar characteristics of hormone responsiveness in mammary cancers (21, 22). Because of these significant similarities between the rat and human, analysis of PR isoform expression and functions in the rat mammary gland may be highly relevant to understanding PR isoform functions in the normal human breast and breast cancer.
Biochemical analyses of PR levels in whole mammary gland provide limited information. Immunoblots do not address cell type-specific expression, subcellular distribution, or colocalization of the PR isoforms. Immunoblots are also limited in their sensitivity and accuracy of PR quantitation due to a dilution effect by mammary stroma because PR are expressed mainly in the epithelial compartment of the gland (20, 23). This is particularly relevant for quantitation of relative levels of PRA or PRB in mammary tissues that exhibit changes in overall epithelial content, such as pubertal gland vs. adult virgin gland vs. pregnant mammary gland. In the first case, there is a predominance of stroma, whereas in the adult virgin gland and pregnant gland, there are overall increases in the ratio of epithelium to stroma with the pregnant gland containing the highest proportion of epithelium. Thus, the same amounts of protein from mammary gland extracts obtained at different physiological states represent different contributions from mammary epithelium. Therefore, immunoblot analysis does not provide an accurate assessment of changes in PR isoform levels in mammary epithelium. In this report, we used immunohistochemical staining with antibodies that detect only PRA or only PRB to analyze expression of PRA and PRB in various mammary gland structures, their colocalization in the same cells, and the relationships of PRA and PRB to proliferation at various stages of rat mammary gland development and function.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Immunoperoxidase staining
Five-micrometer tissue sections were deparaffinized, incubated in 3% H2O2 in methanol for 10 min, and subjected to antigen retrieval by boiling in citrate buffer for 10 min. The sections were blocked in PBS containing 2% BSA and then incubated with rabbit polyclonal antibody to PR from Santa Cruz Biotechnology (Santa Cruz, CA) (C-19, 1:200), which recognizes only PRA by immunohistochemistry (20) or mouse monoclonal anti-BrdU antibody with nuclease (anti-BrdU detection kit; Amersham, Little Chalfont, Buckinghamshire, UK). Sections were incubated with swine antirabbit or goat antimouse biotinylated antibody (Dako, Glostrup, Denmark), followed by incubation with avidin-biotin-horseradish peroxidase complex (ABC reagent; Vector Laboratories, Burlingame, CA), followed by diaminobenzidine reagent (Pierce, Rockford, IL) and counterstained with hematoxylin. Images were obtained using Nikon Eclipse E400 microscope and Qimaging camera and software from MicroPublisher (Burnaby, British Columbia, Canada).
Immunofluorescent labeling
Tissue sections were prepared and subjected to antigen retrieval as previously described (20). Sections were blocked with 2% BSA in PBS followed by incubation with rabbit polyclonal antimouse PRB (B15, 1:800) or rabbit polyclonal antihuman p27 (Santa Cruz Biotechnology; C-19, 1:200) or goat polyclonal antihuman p21 (Santa Cruz Biotechnology; C-19, 1:200). The anti-PRB B15 antibody was raised against a synthetic peptide corresponding to amino acids specific to the PRB region of mouse PR (Affinity Bioreagents, Golden, CO). Because the B15 antibody produced nonspecific background staining with the immunoperoxidase method, the immunofluorescent labeling method was used to detect PRB. Sections were incubated with corresponding secondary goat antirabbit or donkey antigoat Alexa 488-labeled antibody (Molecular Probes, Eugene, OR).
In the case of double-immunofluorescent labeling, tissue sections were first incubated with rabbit polyclonal antihuman PR antibody, which detects only PRA by immunohistochemistry (20) (Dako, Carpinteria, CA; A0098, 1:75), antimouse PRB (B15, 1:800), antihuman phosphorylated retinoblastoma protein (phospho-Rb) (Cell Signaling, Beverly, MA; 1:150), antihuman cyclin D1 (BioSource, Camarillo, CA; 1:50), antihuman p27 (Santa Cruz Biotechnology; C-19, 1:200), or goat polyclonal antihuman p21 (Santa Cruz Biotechnology; C-19, 1:200) overnight at 4 C followed by incubation with goat antirabbit Alexa 488- or Alexa 546-labeled secondary antibody (Molecular Probes). Sections received a blocking step with goat antirat IgG (EMD Biosciences, La Jolla, CA) followed by normal goat serum (Vector) and were then incubated with a second primary mouse monoclonal antibody, such as antihuman PR (Neomarkers, Fremont, CA; hPRa7, 1:50; recognizes only PRA by immunohistochemistry) (20), anti-smooth muscle actin (Sigma; 1:800), anti-BrdU antibody (Amersham anti-BrdU detection kit), or antihuman cyclin D1 (Cell Signaling; 1:100) overnight at 4 C. This was followed by incubation with goat antimouse Alexa 488- or Alexa 546-labeled secondary antibody.
For triple-immunofluorescent labeling, sections were incubated in the following order: rabbit polyclonal antimouse PRB antibody (B15, 1:800) overnight at 4 C, followed by incubation with goat antirabbit Alexa 488-labeled secondary antibody (Molecular Probes) and then with mouse monoclonal antihuman PR (hPRa7; Neomarkers; 1:100) overnight at 4 C. This was followed by incubation with goat antimouse Alexa 633-labeled secondary antibody (Molecular Probes). To prevent cross-reactivity between biotinylated goat antimouse secondary antibody and mouse monoclonal hPRa7 antibody, tissue sections were blocked as previously described (20). Sections were next incubated with mouse monoclonal anti-BrdU antibody (Amersham anti-BrdU detection kit) for 60 min at room temperature and then incubated with goat antimouse biotin-labeled secondary antibody (1:400; Dako, Carpinteria, CA) and Alexa 546-labeled streptavidin. For single- and double-immunofluorescence staining, nuclei were counterstained with 4'-6-diamidino-2-phenylindole (DAPI) or TOPRO-3 iodide (Molecular Probes). All sections were mounted on glass slides using Airvol solution and analyzed with a Nikon inverted fluorescence microscope TE200-PFS (Mager Scientific, Dexter, MI) and Metamorph software or with a Pascal laser scanning confocal microscope (Zeiss, Thornwood, NY) and Zeiss LSM Image Browser software.
Immunoblot analysis
Rat and mouse uteri were flash frozen in liquid nitrogen and stored at –70 C. Frozen uteri were homogenized in a buffer containing 50 mM potassium phosphate (pH 7.0), 10 mM EGTA, 10 mM sodium molybdate, 12 mM thioglycerol, 10% glycerol (1 ml/uterus, 1 ml/100-mm dish of MCF-7 cells) containing protease inhibitor cocktail (Sigma) using a Polytron homogenizer. MCF-7 cells were harvested and sonicated in the same buffer. Protein samples were resolved on a 4–12% NuPAGE Bis-Tris gel (Invitrogen, Carlsbad, CA) gel and transferred to a nitrocellulose membrane (Schleicher & Schuell, Keene, NH). Molecular markers used to determine molecular weights were from Invitrogen. The membrane was divided to allow testing of different antibodies on the same protein samples. The membrane was blocked in 5% dry, fat-free milk in Tris-buffered saline with 0.5% Tween for 1 h and incubated with primary antibody overnight at 4 C. The primary antibodies used to detect PRB were rabbit polyclonal antimouse PRB B15 (1:1000, 5 µg/ml), rabbit polyclonal antihuman PR (Dako A0098, 1:100) (Carpinteria, CA), or rabbit polyclonal antihuman PR (Santa Cruz Biotechnology; C-19, 1:100, 2 µg/ml). Membranes were incubated with horseradish-peroxidase-labeled antirabbit secondary antibody (Santa Cruz Biotechnology; 1:1000) and developed with Western Lightning chemiluminescence reagent (PerkinElmer, Boston, MA).
Quantitation and statistical analysis
The number of PRA- or BrdU-positive cells was determined by immunoperoxidase staining and quantitated with the aid of a Nikon Eclipse 400 microscope. The number of PRA-, PRB-, BrdU-, and cyclin D1-positive cells or colocalization of these proteins was determined by immunofluorescent labeling and was quantitated from captured images using Metamorph software. A minimum of 1000 cells were counted for each structure and a minimum of two to three tissue sections per rat were analyzed. The number of positive cells is expressed as the percentage of total epithelial cells counted. Three to six animals per developmental stage were analyzed. Results are expressed as mean ± SEM, and differences were considered significant at P < 0.05 by Student’s t test or ANOVA where appropriate.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Representative photomicrographs of immunoperoxidase staining of PRA at different developmental stages are shown in Fig. 1
, A–I. Because the mammary gland contains functionally and morphologically different structures such as end buds, ducts, and lobules, PRA+ cells were quantified separately for each type of structure. Results of quantitative analysis of PRA expression are presented in Fig. 1J.
|
). Because the number of end buds declined significantly after 6 wk of age, PRA expression in end buds was only quantitated at 5 and 6 wk of age. At all developmental stages, PRA staining was localized to the nucleus of mammary epithelial cells, and no staining was observed in surrounding stromal fibroblasts or adipocytes (Fig. 1, C–F).
The percentage of PRA+ cells gradually declined after 6 wk of age. By 14 wk, the mammary gland had completed ductal elongation, and end buds were rarely detected. At this age, the percentage of PRA+ cells was significantly lower than at 6 wk of age for both ducts and lobules (P < 0.01) (Fig. 1, E and F). Whole-mount analysis of the pubertal gland showed ducts and developing lobules to be dilated and enlarged (Fig. 1K). By comparison, in the adult gland, lobules were reduced in size and appeared regressed. In the adult gland, there was also a noticeable change in the appearance of the lobules during the estrous cycle such that the lobules were smaller at estrus than during diestrus (Fig. 1K). However, no difference in the percentage of PRA+ cells was observed at estrus vs. diestrus (data not shown). These results indicate that maturation of the mammary gland is associated with a decrease in the percentage of PRA+ cells.
In the pregnant gland (8–10 d of pregnancy), the percentage of PRA+ cells in lobules was significantly decreased to 8% (P < 0.01) compared with lobules of the 14-wk-old virgin gland. However, PRA expression in ducts was not significantly different from the 14-wk-old virgin gland (Fig. 1J). During lactation, no PRA+ cells were detected (Fig. 1H). After lactational involution, at 4 wk post weaning, when the mammary gland is fully regressed, PRA+ cells were again detected, but the percentage was significantly lower in both ducts (12%) and lobules (11%) than in the age-matched adult virgin gland (P < 0.01 for both ducts and lobules) (Fig. 1, I and J). Thus, the percentage of PRA+ cells was permanently decreased as a result of pregnancy.
Expression of PRB isoform at various stages of mammary gland development
The antibody hPRa6 previously used to detect PRB in human (25) and mouse (20) mammary tissues was generated against the human PR and was not effective for staining in rat mammary tissues. We therefore prepared a rabbit polyclonal antibody generated against the N-terminal region of the mouse PRB protein (antibody B15). We reasoned that this antibody would be specific for the PRB isoform because this region of PRB is absent in PRA and that it would detect rat PRB because according to published NCBI sequences of human (NP_000917), rat (NP_074038), and mouse (NP_032855) PR, the amino acid sequence of the N-terminal region of mouse and rat PRB is 84% homologous compared with 67% homology between mouse or rat with the N terminus of human PRB. The specificity of B15 antibody for PRB was confirmed by immunoblot (Fig. 2). Antibody B15 recognized a band of the same mobility and size (120 kDa) in protein extracts of rat and mouse uterus and MCF-7 human breast cancer cells. The band detected by B15 was the same molecular weight as the PRB band obtained with two commercially available antibodies that detect PRB in immunoblots. Specificity of the B15 antibody for PRB in immunohistochemistry was confirmed by colocalization of B15 antibody in pregnant mouse mammary gland with hPRa6, an antibody that detects only PRB by immunohistochemistry (25). The B15 antibody did not produce staining in the virgin mouse mammary gland or colocalize with hPRa7, an antibody that detects only PRA by immunohistochemistry (data not shown).
|
). Moderately intense PRB staining was observed in the prepubertal 3-wk-old gland in end buds, ducts, and lobules (Fig. 3A). At 6 wk of age, the intensity of PRB staining increased (Fig. 3B), and the percentage of PRB+ cells was 50, 45, and 60% in end buds, ducts, and lobules, respectively (Fig. 3, B–D and J). The percentage of PRB+ cells in each type of structure did not change significantly throughout development with the exception of lactation. Additionally, no difference in the percentage of PRB+ cells was observed at estrus vs. diestrus (data not shown). However, the overall intensity of PRB staining was higher in adult virgin glands than in pubertal glands in both ducts and lobules and increased further in the 7-d pregnant gland (Fig. 3, A–J). In the lactating gland, PRB staining was observed in the cytoplasm of basal myoepithelial cells (Fig. 3H). Overall, at all developmental stages, the percentage of PRB+ cells and the intensity of PRB staining in lobules tended to be higher than in ducts.
|
).
|
and 3H). The percentages of PRB+ myoepithelial cells were not significantly different in ducts and lobules of the pubertal and adult virgin gland or pregnant gland (range, 92 ± 3 to 99 ± 1%) (Fig. 3K).
Colocalization of PRA and PRB during development
Colocalization of PRA and PRB was investigated by dual immunofluorescent staining with antibodies that detect only PRA or only PRB. Because PRA was not expressed in myoepithelial cells, we limited our analysis of PRA and PRB colocalization to luminal epithelial cells. Representative photomicrographs of PRA and PRB colocalization in end buds, ducts, and lobules from pubertal 6-wk-old animals are shown in Fig. 4C, and similar experiments were carried out in the adult, pregnant, and involuted gland. PRA and PRB isoforms were often, but not always, expressed in the same cell. At all stages of development, the majority (86–95%) of PRA+ cells coexpressed PRB; this was the case for all mammary gland structures examined (Table 1). There was a small population of PRA+ cells (<15%) that did not coexpress PRB (Fig. 4C). PRB colocalization with PRA in ducts and lobules varied across developmental stages and ranged from 19% in postlaction involuted ducts to 62% in pubertal ducts (Table 1). Importantly, and in contrast to PRA+ cells, there was a significant percentage of PRB+ cells that expressed only PRB (Fig. 4C). In the pubertal gland, the percentage of cells expressing only PRB was 30% in end buds, 23% in ducts, and 34% in lobules. In the adult gland, the percentage of cells expressing only PRB increased to 39% in ducts and 49% in lobules. In the early pregnant gland, PRB was the predominant isoform in lobular epithelium because the percentage of PRA+ cells was significantly decreased; the percentage of cells expressing only PRB increased to 53%. After postlactational involution, the percentage of cells expressing only PRB was 47% in ducts and 54% in lobules. Thus, there was a significant percentage of cells that expressed only PRB, and this percentage changed across development. In contrast, the vast majority of PRA+ cells coexpressed PRB.
|
. The percentage of proliferating cells was highest between 3 and 8 wk of age. Proliferation occurred in ducts and lobules but was highest in end buds. By 10 wk of age, end buds were reduced in number and proliferation was detected mainly in lobules and ducts. At 14 wk of age, end buds were no longer present and the lowest proliferation index was observed. During pregnancy, the proliferation index in ducts and lobules was higher than at any other developmental stage. In the postlactation involuted mammary gland, almost no proliferation was detected.
|
). In the pubertal rat at 6 wk of age, no significant differences in proliferation were observed at estrus vs. diestrus for end buds, ducts, or lobules (data not shown). However, in the 14-wk-old adult rat, where very little proliferation was observed, the greatest amount of proliferation was observed in lobules at diestrus (P < 0.05) (Fig. 5B).
PR isoform expression and proliferation
One model of P action in the breast proposes an indirect stimulation of proliferation in PR-negative (PR–) cells through production of paracrine factors produced by PR+ cells. This is based in part on the observation that PR does not colocalize with proliferation markers and has led to the conclusion that PR+ cells are proliferatively quiescent (27). In those studies, there was no analysis of PR isoform expression in proliferating vs. nonproliferating cells. We have previously reported that there is a significant degree of colocalization of PRB with the proliferation marker BrdU in the murine pregnant mammary gland (20). Thus, it was of interest to examine the relationship between PR isoform expression and proliferation in the rat mammary gland.
To address this question, we analyzed colocalization of the proliferation marker BrdU with PRA or PRB using dual immunofluorescent staining. We analyzed PR isoform colocalization in the pubertal gland (6 wk of age) and in the pregnant gland (7 d pregnant) that have the highest proliferation rates (Table 2).
|
). The higher percentage of PRB+BrdU+ cells in end buds was not surprising because only PRB was expressed in the actively proliferating cap cell layer, whereas PRA+ cells were detected only in body cells. In ducts and lobules, proliferation rates were substantially lower. Of the BrdU+ cells, 20 and 33% of cells in ducts and lobules, respectively, were BrdU+PRA+; 77% of cells in ducts and lobules were BrdU+PRB+. In all structures of the pubertal gland, PRB colocalized with BrdU about 2.5 times more often than PRA colocalized with BrdU. In the 7 d pregnant gland, the proliferation rates in both ducts (9%) and lobules (15%) were significantly higher than the pubertal gland. The percentage of BrdU+ cells coexpressing PRA was significantly decreased compared with the end buds of the pubertal gland; only 12% of the total BrdU+ cells in ducts and 9% in lobules expressed PRA. The percentage of BrdU+ cells expressing PRB was similar to the pubertal gland; 70 and 74% of BrdU+ cells expressed PRB in ducts and lobules, respectively. In the pregnant gland, PRB colocalized with BrdU at least five times more often than PRA.
Because the majority of PRA+ cells coexpressed PRB, this suggested that the proliferation observed in PRA+ cells represented the proliferation of PRA+PRB+ cells. To determine whether this was the case, triple-immunofluorescence labeling was carried out with anti-PRA, anti-PRB, and anti-BrdU antibodies in the pubertal gland. From the images obtained, we determined that the majority of PRA+PRB+ cells were not BrdU+. PRB-only-expressing cells were more frequently observed to be BrdU+ (Fig. 5C).
Colocalization of PR isoforms with cyclin D1
Nuclear localization of cyclin D1 is required for the entrance to and progression through the G1 phase of the cell cycle. In the mouse mammary gland, cyclin D1 was shown to be essential for lobuloalveolar development during pregnancy (28). Phenotypically, cyclin D1 knockout mice resemble PRB knockout mice, which suggests that cyclin D1 and PRB are in the same or parallel pathways that regulate lobuloalveolar development in the mouse (28, 29). We have previously shown that there is a high degree of cyclin D1 and PRB colocalization and correlation with proliferation in the mouse mammary gland during pregnancy (20). It was of interest to determine the relationship of cyclin D1 expression to proliferation in the rat mammary gland. To address this question, we performed double-immunofluorescent labeling with anti-cyclin D1 and anti-PRA or anti-PRB antibody.
Because the active form of cyclin D1 localizes to the nucleus, we quantitated cyclin D1+ luminal epithelial cells with nuclear staining. Representative images of cyclin D1 staining in end buds, ducts, and lobules and quantitative analysis of colocalization between PR isoforms and nuclear cyclin D1 are shown in Fig. 6. Nuclear cyclin D1 was highly expressed in end buds in pubertal mammary glands in both cap cells and body cells; 30% of body cells were cyclin D1+. Of these cyclin D1+ cells, 53% coexpressed PRA and 71% coexpressed PRB (Fig. 6C). In the pubertal gland, 20% of ductal and 23% of lobular cells were cyclin D1+. PRA was coexpressed in 86 and 85% of cyclin D1+ cells in ducts and lobules, respectively. PRB was coexpressed in 79 and 83% of cyclin D1+ cells in ducts and lobules, respectively (Fig. 6, D and E).
|
). In ducts, there was also a trend toward a decrease in the percentage of cyclin D1+cells, which did not reach statistical significance (Fig. 6D). Overall, 14 and 16% of cells were cyclin D1+ in ducts and lobules, respectively. More than 95% of cyclin D1+ cells coexpressed PRA and more than 85% of cyclin D1+ cells coexpressed PRB; this was true for both ducts and lobules (Fig. 6, D and E).
In early pregnancy, the majority of cyclin D1+ cells exhibited only cytoplasmic staining, and 3–4% of cells in ducts and lobules exhibited nuclear cyclin D1+ staining (Fig. 6, D and E). It should be noted that due to the proliferative expansion of the epithelium during pregnancy, the pregnant gland contains significantly more epithelial cells than the virgin gland. Therefore, the percentage calculation of cyclin D1+ cells present in the pregnant gland underestimates the total number of cyclin D1+ cells. Colocalization of PRA with cyclin D1 was dramatically decreased compared with pubertal and adult virgin glands. Only 28 and 15% of cyclin D1+ cells in ducts and lobules, respectively, coexpressed PRA. In contrast, PRB was colocalized in more than 75 and 90% of cyclin D1+ cells in ducts and lobules, respectively.
Coexpression of cyclin D1 with BrdU and phospho-Rb
The results described above show that in the virgin mammary gland, although PRA+ cells rarely proliferated, they expressed nuclear cyclin D1. Because nuclear cyclin D1 is required for transition to G1 phase and entrance to S phase (30, 31), these apparently conflicting results were surprising. To investigate the reason for the discordance between nuclear cyclin D1 localization and proliferation of PRA+ cells, we analyzed colocalization of cyclin D1 with markers of cell cycle progression, namely BrdU and phospho-Rb.
The observation that PRA rarely colocalized with BrdU but was frequently colocalized with nuclear cyclin D1 suggested that cyclin D1 and BrdU would not be colocalized in the same cells. Representative pictures of dual immunofluorescent labeling with anti-cyclin D1 and anti-BrdU antibody in the pubertal end bud and pubertal, adult, and pregnant lobules are shown on Fig. 7A. Indeed, we found a low percentage of cyclin D1+BrdU+ cells in ducts and lobules of the pubertal gland (5 ± 1% of cyclin D1+ cells were also BrdU+; n = 3 animals). Colocalization with BrdU was not quantitated in the adult gland because so few cells were BrdU+. Significantly more colocalization of cyclin D1 with BrdU was observed in ducts and lobules in the pregnant gland compared with the pubertal gland (P < 0.05; n = 3 animals); 16 ± 6% and 20 ± 3% of cyclin D1+ cells were also BrdU+ in ducts and lobules, respectively. These results suggest that the majority of cyclin D1+ cells in ducts and lobules of the pubertal and adult virgin gland do not proliferate.
|
). In the adult virgin gland, less than 0.1% of cyclin D1+ cells coexpressed phospho-Rb. In the pregnant gland, significantly more cyclin D1+ cells coexpressed nuclear phospho-Rb (68 ± 5%; n = 3 animals) (P < 0.01).
Expression of CDK inhibitors p21 and p27 in the rat mammary gland
The observation that the majority of cyclin D1+ cells did not coexpress phospho-Rb in the virgin gland suggested that these cells were cell cycle arrested. The reason for the cell cycle arrest of cyclin D1+ cells in the virgin gland was unclear. It was possible that these cyclin D1+ cells were arrested due to the concomitant expression of cell cycle inhibitors. The CDK inhibitors p21(Cip) and p27(Kip) have been shown to cause G1 arrest in different cell systems (30, 34). Therefore, we analyzed their expression pattern and level by immunofluorescent labeling with anti-p21 or anti-p27 antibody. In the pubertal gland, there was a very low level of p21 staining in end buds; stronger staining was observed in lobules (Fig. 8A). The strongest p21 staining was detected in the adult virgin gland. The least amount of p21 staining was observed in the pregnant gland (Fig. 8A).
|
). In end buds, there was notable staining in the cap cell layer; however, the staining was localized to the cytoplasm. In lobules of the pubertal gland, some epithelial cells showed very strong nuclear p27 staining, whereas others showed only faint staining. In the adult gland, many luminal cells exhibited intense nuclear p27 staining. In the pregnant gland, the lowest amount of nuclear p27 staining was observed. Thus, the strongest nuclear expression of the cell cycle inhibitors p21 and p27 occurred in the adult virgin gland concomitant with the lowest proliferation rate. The lowest nuclear expression of p21 and p27 was observed during pregnancy when proliferation rates in ducts and lobules were maximal.
As described above, many PRA+ cells coexpressed cyclin D1; the majority of PRA+ cells were also PRB+, and these cells appeared to be proliferatively quiescent because they were not BrdU+. To determine whether these cells were cell cycle arrested, we performed dual immunofluorescent staining of p21 or p27 with PRA and p21 or p27 with cyclin D1 in adult virgin gland. As shown in Fig. 8, C and D, we observed PRA+ cells that coexpressed p21 and p21+ cells that coexpressed cyclin D1. The same results were obtained with dual immunofluorescent staining with p27 and PRA or p27 and cyclin D1 (Fig. 8, E and F). These results indicate that the lack of proliferation of PRA+cyclin D1+ cells is likely due to the concomitant expression of the cell cycle inhibitors p21 and p27.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Temporal pattern of PRA and PRB expression
In the rat mammary gland, both PRA and PRB were expressed in prepubertal, pubertal, adult, pregnant, and postlactation involuted mammary glands. This was true for all mammary gland structures examined: end buds, ducts, and lobules. Neither PRA nor PRB was detected in luminal cells during lactation; however, a low level of PRB was expressed in myoepithelial cells. The percentage of PRA+ cells was highest at puberty, decreased in the adult gland, and reached a nadir during pregnancy. The percentage of PRA+ cells was permanently reduced after pregnancy. In contrast, PRB expression was relatively constant at all stages of mammary gland development with the highest expression of PRB in lobules during pregnancy and after postlactational involution. The vast majority of PRA+ cells coexpressed PRB. However, significantly more cells expressed PRB than PRA at all stages of development, and a significant percentage of PRB+ cells expressed only PRB. The proportion of cells expressing only PRB was highest during pregnancy and after postlactational involution. Thus, PRB expression predominated over PRA expression at all developmental stages in the rat.
The temporal and spatial pattern of PR isoform expression in the rat differed significantly from that reported for the mouse (20). In the mouse, PRA was the predominant isoform expressed in the virgin gland, and PRB was the predominant isoform expressed during pregnancy. Additionally, PRA and PRB were rarely colocalized in the same cells during pregnancy. This pattern of PR isoform expression was the same for the BALB/c and C57BL/6 mouse strains (Aupperlee, M. D., and S. Z. Haslam, unpublished observations). The results obtained in the mouse suggest that P action is mediated mainly by either PRA or PRB alone, whereas in the rat, both PR isoforms have the potential to contribute to the biological outcome of P signaling in PRA+PRB+ cells. This is similar to the situation in the adult premenopausal human breast, for which it has been reported that all PR+ cells coexpress PRA and PRB (18).
In the mouse, a ductal morphology predominates in the pubertal and adult gland before a pregnancy. Extensive lobuloalveolar development in many strains is restricted to pregnancy. Studies of PRA or PRB overexpression or deletion in mice have shown that PRB is critically required for lobuloalveolar development (9, 29, 35, 36). Thus, there is good concordance with the timing of PRB expression and lobuloalveolar development in the mouse. In rat and human mammary glands, duct and lobule development occur concurrently from puberty to sexual maturation of the gland (23, 37, 38). In both rat and human mammary glands, PRB is expressed in the nonpregnant state. Additionally, in the rat, concordant with lobular proliferation and expansion that occurs during pregnancy, there is increased PRB expression in early pregnancy. Thus, there is good temporal agreement between lobule development before pregnancy and PRB expression in nonpregnant rat and human mammary glands and lobular expansion during pregnancy and increased PRB expression in the rat. Taken together, these findings in the rat support the hypothesis that PRB expression is required for lobule development. Our studies suggest that PRB expression is likely required for lobule development in the human.
Despite the differences, there are some interesting similarities between the rat and the mouse. There is a prevalence of cells expressing only PRB during pregnancy in both species. These results suggest that PRB performs additional and/or specific functions when not coexpressed with PRA. Additionally, in both the rat and the mouse, PRA expression is decreased with maturation of the mammary gland and during pregnancy and is permanently decreased after pregnancy and postlactational involution. Whether or not PR isoform expression varies at similar stages of human breast development is not known.
It is well known that pregnancy confers a significant protective effect against mammary cancer in rodents and humans (39). In the rat and the mouse, there is a significant decrease in PRA+ cells. Whether or not this decrease in PRA+ cells is causally related to pregnancy-induced protection is not known. It is also not known whether a similar change in PRA expression occurs in the human breast after pregnancy, due to a lack of information about PR isoform expression in the nulliparous vs. parous breast.
Cell type-specific expression of PR isoforms
There are two lineages of epithelial cells in mammary gland: luminal epithelial cells, which line the internal surface of ducts and lobules, and myoepithelial cells, which surround luminal cells and form the contractile network outside ducts and lobules. Based on recent histological analyses conducted in the mouse, PRA and PRB are expressed only in the luminal epithelium (20). In the current study, we showed that PRA is expressed in only luminal cells, whereas PRB is expressed in both luminal and myoepithelial cells. In the pubertal gland, most of the cells in the cap cell layer coexpress PRB and smooth muscle actin and are believed to be the source of myoepithelial cells. The presence of PRB in myoepithelial cells suggests that P may directly stimulate proliferation and/or terminal differentiation of myoepithelial cells. Whether or not PR is expressed in normal human myoepithelial cells is not known because colocalization of PR with luminal-cell- or myoepithelial-cell-specific markers has not been performed. Myoepithelial cells have been shown to function as natural tumor suppressors. These cells can inhibit angiogenesis, induce apoptosis of luminal cells, and create a natural barrier restraining metastasis (40). Recent studies of solid tumors derived from a human breast cancer cell line (T47D), grown as xenografts in athymic mice, indicate that PR may also be involved in differentiation of the myoepithelial cell lineage in human breast cancer (41). The rat mammary gland provides a novel opportunity to study the role of P and PR in myoepithelial cell differentiation and/or function in the normal gland and may provide additional insights about the role of P in the regulation of myoepithelial cells in mammary cancer.
PR isoforms and proliferation
Multiple studies conducted in the rat and mouse mammary gland and in the human breast show that PR expression is dissociated from proliferation; i.e. PR+ cells do not colocalize with proliferation markers (27, 42). These observations form the basis for the proposed paracrine mechanism of P action in mammary gland. Moreover, some candidate proteins such as receptor activator of nuclear factor-
B ligand are reported to be induced in PR+ cells and are proposed to mediate P action in PR– cells through a paracrine mechanism (29, 43). However, it should be noted that in most studies, no discrimination between PR isoforms was made. We previously showed that in the mouse, PRA is rarely colocalized with the proliferation marker BrdU at various developmental stages (20). Conversely, PRB is predominantly expressed only during pregnancy and is highly colocalized with BrdU. These data led us to propose another mechanism of P action in the mouse mammary gland, i.e. that P directly induces proliferation in the PRB-positive cells. This mechanism would be operative in addition to a paracrine mechanism that is most likely mediated by P signaling via PRA.
Our analysis of the relationship between PR isoform expression and proliferation in the rat mammary gland revealed that expression of PRB was positively correlated with proliferation in end buds, ducts, and lobules at puberty and during the expansion of the lobular epithelium during pregnancy. A significantly higher (3- to 7-fold) percentage of PRB+ cells than PRA+ cells colocalized with BrdU at all stages of development. Because the vast majority of PRA+ cells coexpressed PRB, we inferred that cells that coexpressed PRA and PRB (PRA+PRB+ cells) were not proliferating. Triple-antibody labeling with anti-PRA, anti-PRB, and anti-BrdU antibodies showed that few PRA+PRB+ cells were also BrdU+ and confirmed a low level of proliferation of PRA+PRB+ cells. Conversely, more cells that contained only PRB were also BrdU+. At all stages of development, there were more PRB+ cells than PRA+ cells, and based upon the more extensive colocalization of PRB with BrdU, we conclude that cells that express only PRB constitute the major PR+ population of proliferating cells. Our data also indicate that P may directly induce proliferation in PRB-only-expressing cells in the rat mammary gland. Additionally, our data suggest the possibility that when coexpressed with PRB, PRA may have an inhibitory effect on the proliferative activity of PRB. Thus, it is possible that the balance between PRA and PRB expression in a cell may be one contributing factor that determines whether a PR+ cell proliferates or remains proliferatively quiescent after exposure to P.
To the best of our knowledge, the relationship between PR isoform expression and proliferation in the normal human breast is not known. Because earlier studies showed dissociation between PR expression and proliferation in the human breast and because PRA and PRB are coexpressed in the same cell, it is possible to deduce that PRA+PRB+ cells do not colocalize with proliferation markers. However, in breast cancers, PR+ cells do proliferate (23, 27). The mechanisms responsible for proliferation of PR+ cells in breast cancers are not totally understood.
PR isoforms and nuclear cyclin D1 expression and proliferation
Nuclear localization of cyclin D1 is considered to be a marker of cell entry into the G1 phase of the cell cycle (32). We have previously shown that PRA is rarely colocalized with nuclear cyclin D1 in the virgin and pregnant mouse mammary gland, whereas PRB is highly colocalized with nuclear cyclin D1 and the proliferation marker BrdU during pregnancy (20). Thus, the colocalization pattern of PRA or PRB with cyclin D1 in the mouse is in agreement with the colocalization pattern of PR isoforms with proliferation.
In contrast to the mouse, in the rat, there was a high degree of colocalization of nuclear cyclin D1 and PRA in pubertal and adult virgin glands. Additional analysis showed that some cyclin D1+ cells in the pubertal gland and almost all cyclin D1+ cells in the adult gland appeared to be cell cycle arrested. This conclusion is based on the observations that 1) there was a low percentage (5%) of cyclin D1+ cells that were also BrdU+ and 2) there was a low percentage (15%) of cyclin D1+ cells that colocalized with nuclear phospho-Rb. Presently, we do not know the exact cause of cell cycle arrest in these cells. However, we have shown colocalization of p21 with cyclin D1 and p27 with cyclin D1 as well as colocalization of p21 with PRA and p27 with PRA in the adult virgin gland. Therefore, both inhibitors may contribute to cell cycle arrest of PRA+cyclin D1+ cells in the virgin gland. In the pregnant gland, the percentage of PRA+ cells was significantly reduced and the percentage of PRB-only-expressing cells was increased, indicating that colocalization of cyclin D1 with PRB occurred in cells expressing only PRB. During pregnancy, there were 3- to 4-fold more cyclin D1+ cells that colocalized with BrdU compared with the adult virgin, and nuclear phospho-Rb was coexpressed in 68% of cyclin D1+ cells. Additionally, there was reduced or lack of expression of the cell cycle inhibitors p21 and p27, indicating that the cell cycle block is removed in the pregnant gland and that the cyclin D1+ cells proliferate.
Previous studies carried out in vitro in a T47D breast cancer cell line genetically engineered to express only PRB have shown these cells proliferate in response to an initial treatment with progesterone (44). A second treatment 40 h later with R5020 induces an increase of both p21 and p27 and cell cycle arrest. However, it is not known how p21 and p27 may be regulated in cells that coexpress both PRA and PRB. Additional factors may also contribute to the reduced proliferation of PRA+ cells. It has also previously been shown that TGF-ß1 acts to suppress proliferation of steroid receptor-positive mammary cells in the virgin mouse mammary gland (45). It will be of interest to determine whether TGF-ß1 acts in a similar manner in the rat mammary gland.
In summary, we showed distinct developmental profiles of PRA and PRB expression and different relationships of these isoforms to proliferation in the rat mammary gland. Frequent colocalization of PRA and PRB in the same cell is common to the rat and human but is not seen in the mouse. Additionally, we found several features of PR isoform expression that are similar in the mouse and rat, with similar developmental profile of PRA expression, dissociation of PRA isoform expression with proliferation, and high association of PRB-only-expressing cells with proliferation. The concordance between the timing of lobular development, the presence of PRB-only-expressing cells, and the colocalization of PRB and BrdU in these cells during lobular development in both the rat and mouse supports and extends the current hypothesis that PRB expression is specifically required for alveologenesis and lobular development. We speculate that the differences in the pattern and prevalence of PRB-only-expressing cells before pregnancy are responsible for the ductolobular organization vs. the predominantly ductal organization of the mammary gland in the rat and mouse, respectively. How PR isoform expression is related to ductolobular development in the human breast is currently not known. A novel and intriguing observation of PR isoform expression in the nonpregnant rat mammary gland is the lack of association with proliferation of cells coexpressing PRA, PRB, and cyclin D1. This finding suggests that PRA may have an inhibitory effect on the proliferative activity of PRB when both isoforms are coexpressed in the same cells.
Based on studies in the mouse mammary gland, PR expression and P signaling through PR is required for normal mammary gland development (29). According to in vitro studies in human breast cancer cell lines, the functional outcome of P signaling is determined by the ratio between PRA and PRB expression (46). Based on the single study of the normal human premenopausal breast, both PR isoforms are reported to be equimolar and coexpressed in the same cell (18). Very little is known about PR isoform expression at any other developmental stage of the human breast. Understanding the functional roles of PRA vs. PRB becomes important in the context of breast cancer. That progestins play a role in the etiology of breast cancer is evidenced by increased breast cancer risk in postmenopausal women receiving hormonal therapy with estrogen plus progestin compared with estrogen-alone hormonal therapy (6, 7). Furthermore, a shift from equimolar expression of PRA and PRB to a predominance of either isoform in breast cancers is associated with specific tumor features and prognosis for survival (18, 19, 47). The evolutionary divergence of the mouse and rat with regard to PR isoform expression and function provide a unique opportunity to investigate the functional roles of PRA and PRB in the normal gland when expressed separately in different cells or coexpressed in the same cell and may provide novel insights into P action and PR isoform functions that will have relevance to understanding P action in the normal human breast and in the etiology of breast cancer.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Abbreviations: BrdU, 5-Bromo-2'-deoxyuridine; CDK, cyclin-dependent kinase; DAPI, 4'-6-diamidino-2-phenylindole; P, progesterone; PR, progesterone receptor; phospho-Rb, phosphorylated retinoblastoma protein; SMA, smooth muscle actin.
This work was supported by the Breast Cancer and the Environment Research Centers Grant U01 ES/CA 012800 from the National Institute of Environmental Health Science and the National Cancer Institute, National Institutes of Health, Department of Health and Human Services (to S.Z.H.). Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the National Institute of Environmental Health Science or National Cancer Institute, National Institutes of Health.
Disclosure Statement: The authors have nothing to disclose.
Received November 8, 2006.
Accepted for publication February 20, 2007.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
