Abstract
Studies were undertaken to examine the natural role of ErbB2, ErbB3, and ErbB4 during the development of normal rat mammary epithelial cells (MECs) in vivo and in vitro. Immunohistochemical analysis demonstrated that mammary gland terminal end buds expressed abundant ErbB2 and ErbB4 but limited ErbB3 in pubescent rats, whereas luminal epithelial cells in nulliparous rats expressed ErbB2, ErbB3, and/or ErbB4. During pregnancy, ductal epithelial cells and stromal cells expressed abundant ErbB3 but limited ErbB2. Although ErbB2 and ErbB3 were downregulated throughout lactation, both receptors were re-expressed during involution. In contrast, ErbB4 was downregulated throughout pregnancy, lactation, and involution. Immunoblotting and immunoprecipitation studies confirmed the developmental expression of ErbB2 and ErbB3 in the mammary gland and the co-localization of distinct ErbB receptors in the mammary gland of nulliparous rats. In agreement with our in vivo findings, primary culture studies demonstrated that ErbB2 and ErbB3 were expressed in functionally immature, terminally differentiated and apoptotic MECs, and downregulated in functionally differentiated MECs. ErbB receptor signaling was required for epithelial cell growth, functional differentiation, and morphogenesis of immature MECs, and the survival of terminally differentiated MECs. Finally, ErbB4 expression did not interfere with functional differentiation and apoptosis of normal MECs.
T
The mammary gland undergoes cyclic postnatal development and regression during each estrous cycle and during pregnancy, lactation, and involution (Topper and Freeman 1980; Borellini and Oka 1989). Briefly, the epithelium expands and invades the surrounding fatty stroma by the process of branching morphogenesis during puberty and pregnancy, produces and secretes milk during lactation, and undergoes apoptosis and remodeling during involution. Stromal and epithelial cells in the mammary gland have been shown to express various ErbB ligands including EGF, transforming growth factor-α (TGFα), amphiregulin, and neuregulin-α during distinct developmental stages (Panico et al. 1996; DiAugustine et al. 1997; Herrington et al. 1997), and these ligands regulate a variety of effects in normal and tumor cells within the mammary gland (Oka et al. 1988; Earp et al. 1995; Normanno and Ciardiello 1997). ErbB receptors are mutated and/or overexpressed in a variety of human malignancies, including breast cancer, and these events are often associated with ligand-independence, invasive and metastatic aggressiveness, and drug resistance (Hynes and Stern 1994; Salomon et al. 1995; Carraway et al. 1997).
The roles played by ErbB2, ErbB3, and ErbB4 in normal mammary gland development and the early stages of breast cancer progression are unclear. Unfortunately, transgenic mice lacking ErbB2 (Lee et al. 1995), ErbB3 (Riethmacher et al. 1997), or ErbB4 (Gassmann et al. 1995) die during embryogenesis. Therefore, it is not possible to appreciate the natural role of these receptors in the mammary glands of homozygous null mice. As another approach, ErbB2, ErbB3, and/or ErbB4 expression have been examined in normal rat, human, and/or mouse mammary glands. The data obtained from these studies, however, are contradictory and/or reflective of a limited spectrum of mammary gland development (Press et al. 1990; Lemoine et al. 1992; Dati et al. 1996; Gompel et al. 1996; DiAugustine et al. 1997; Schroeder and Lee 1998; Sebastian et al. 1998; Srinivasan et al. 1998). Differences in the type of detection method, the assay conditions, and/or the species, strain, stage of the estrous or menstrual cycle, and reproductive history of the mammary gland donor probably contributed to this disparity in ErbB receptor expression. Additional studies are therefore required for the expression of distinct ErbB receptors to accurately predict breast cancer risk, prognosis, metastasis, drug resistance, and/or survival.
To test the hypothesis that select ErbB receptors play natural, non-oncogenic roles in regulating growth, differentiation, and/or apoptosis of normal mammary epithelial cells (MECs), immunohistochemistry studies were undertaken to define the cell type-specific and developmental stage-dependent profile of ErbB2, ErbB3, and ErbB4 in serial sections of rat mammary glands during puberty, sexual maturation, pregnancy, lactation, and involution. Immunoblotting and immunoprecipitation studies were used to confirm the specificity of the ErbB-selective antibodies and the developmental expression of ErbB2 and ErbB3. A primary culture model system (Darcy et al. 1995) and an inhibitor of the tyrosine kinase domain of the ErbB receptor family (Fry et al. 1997) were used to examine changes in ErbB receptor expression during the in vitro development of MECs and to identify the effects that were ErbB receptor-dependent.
Materials and Methods
Materials
Progesterone, hydrocortisone, insulin, ascorbic acid, fatty acid-free fraction V bovine serum albumin (BSA), 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl tetrazolium bromide (MTT), phenylmethylsulfonyl fluoride (PMSF), 3,3'-diaminobenzidine (DAB), transferrin, fetal bovine serum (FBS), and phenol red-free F12/DMEM medium were purchased from Sigma (St Louis, MO). Grade II dispase was obtained from Boehringer-Mannheim Biochemicals (Indianapolis, IN). Gentamycin and porcine trypsin were purchased from Gibco BRL Life Technologies (Grand Island, NY). Culture grade human recombinant TGFα was acquired from Collaborative Research (Bedford, MA). Ovine prolactin (NIDDK oPRL-19 and −20) was a gift of the National Hormone and Pituitary Program, National Institute of Diabetes and Digestive and Kidney Diseases. PD158780 was generously provided by Dr. David Fry at Parke-Davis Pharmaceutical Research (Ann Arbor, MI). Sterile tissue culture plastic flasks and plates were purchased from Becton Dickenson Labware (Franklin Lakes, NJ).
Affinity-purified rabbit polyclonal anti-peptide antibodies against EGFR (SC-03), ErbB2 (SC-284), ErbB3 (SC-285), and ErbB4 (SC-283), as well as ErbB receptor-specific peptides (SC-03P for EGFR, SC-284P for ErbB2, SC-285P for ErbB3, and SC-283P for ErB4) were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). A mouse monoclonal antibody against ErbB3 (clone 2F12, UBI 05-390) and a sheep polyclonal antibody against EGFR (UBI 06-129) were acquired from Upstate Biotechnology (Lake Placid, NY). NeoMarkers (Freemont, CA) was the source of rabbit polyclonal antibodies against either ErbB2 (Ab-1; clone 21N, cat # RB-103) or ErbB4 (Ab-2; cat # RB-284), the synthetic Ab-1 ErbB2 peptide (cat # PP-103) as well as mouse monoclonal IgG1 anti-ErbB2 Ab-9 (clone B10; cat # MS-326) or Ab-17 (a mixture of clones e2-4001 + 3B5; cat # MS-730). Chromopure rabbit IgG, chromopure mouse IgG1, and preabsorbed F(ab')2 fragments of affinity-purified donkey anti-rabbit IgG or anti-mouse IgG conjugated to either biotin or horseradish peroxidase were purchased from Jackson ImmunoResearch Laboratories (West Grove, PA). Mouse myeloma IgG2a and the streptavidin-horseradish peroxidase conjugate were purchased from Zymed (San Francisco, CA). ImmunoPure Immobilized Protein G was acquired from Pierce (Rockford, IL). Immobilon P membrane was obtained from Millipore (Bedford, MA) and the enhanced chemiluminescence (ECL) reagent was acquired from Amersham (Arlington Heights, IL). Peroxidase In Situ ApopTag Plus Detection Kit was obtained from Oncor (Gaithersburg, MD).
Animals
Pubescent nulliparous and untimed pregnant female Sprague-Dawley CD rats (Crl:CDBR) were purchased from Charles River (Kingston, NY) or Taconic Laboratories (German-town, NY). Nulliparous rats were used at 50-54 or 83-85 days of age. Pregnant rats were allowed to advance to mid-pregnancy (days 14-16 of gestation), lactation (5-7, 15, or 21 days after birth), or involution (3-4 or 6-10 days after weaning 21-day-old pups). CD2F1 mice, purchased from NCI Frederick Cancer Research Facility, Biological Testing Branch (Frederick, MD), were used to carry the Engelbreth-Holm-Swarm (EHS) sarcoma. Animals were fed rat or mouse chow diets (Teklad, Madison, WI), ad libitum and had free access to water. Animal rooms were temperature-and humidity-controlled. Light cycles were 14 hr on/10 hr off for rats and 12 hr on/12 hr off for mice. The care and use of animals were in accordance with National Institute of Health guidelines and Institute Animal Care and Use Committee regulations.
Isolation of Mammary Epithelial Cells and Fibroblasts
Abdominal and inguinal mammary glands, excised from six to 12 50-54-day-old female Sprague-Dawley rats were mechanically and enzymatically disaggregated. Mammary epithelial organ-like fragments (organoids) were isolated as previously described (Darcy et al. 1995). Briefly, the digest was fractionated by centrifugation to remove floating adipocytes and the suspension sequentially filtered through 530-μm and 60-μm nitex filters to remove large tissue fragments and single cells, respectively. Mammary epithelial organoids (MEOs) retained on the 60-μm filter, were recovered and cultured on tissue culture plastic for 4 hr at 37C to allow the attachment and subsequent removal of residual stromal cells. Cell number within the organoid suspension was estimated from triplicate nuclei counts as previously described (Hahm and Ip 1990).
Mammary fibroblasts were recovered from the 60-μm filtrate by a 10-min, 500 × g centrifugation at 4C, suspended in F12/DMEM medium with 10% (v/v) FBS and 50 μg/ml gentamicin, and cultured on tissue culture plastic for 2 hr at 37C. The nonadherent cells were removed from the flasks and a relatively homogenous population of fibroblasts was expanded and passaged.
Primary Culture Conditions for Mammary Epithelial Organoids
The reconstituted basement membrane (RBM) used for the primary culture studies was extracted from the EHS mouse sarcoma as previously described (Darcy et al. 1995). MEOs were suspended in ice-cold RBM at a density of 1.5 × 106 cells per ml or 4 × 106 cells per ml. Using 24-well tissue culture plates, 200 μl of organoid-RBM suspension containing approximately 3 × 105 cells was plated on top of 200 μl of presolidified RBM in each well. Alternatively, 2.5 ml of organoid-RBM suspension containing approximately 1 × 107 cells was plated on top of 2.5 ml of presolidified RBM in 100 mm tissue culture dishes. MEOs were cultured for up to 21 days with 1 ml per well in 24-well plates or 12.5 ml per 100-mm dish of TGFα medium (F12/DMEM medium with 10 μg/ml insulin, 1 μg/ml prolactin, 1 μg/ml progesterone, 1 μg/ml hydrocortisone, 5 μg/ml apotransferrin, 880 ng/ml ascorbic acid, 1 mg/ml fatty acid-free BSA, 50 μg/ml gentamycin, and 10 ng/ml human recombinant TGFα). Medium was changed every 3.5 days. The pyrido-pyrimidine analogue PD158780 (Rewcastle et al. 1996) was used at 0.5 μM to examine ErbB-dependent effects in cultured MEOs. PD158780 was added fresh to TGFα medium from a 10-mM stock stored at −80C in 100% DMSO. Each treatment condition in the primary culture studies was carried out in at least two and as many as eight separate experiments.
Culture Conditions for Preadipocytes, Fibroblasts, NMU Cells, and RBA Cells
Rat mammary preadipocytes were isolated from the abdominal and inguinal mammary glands from eight to 12 50-54-day-old female Sprague-Dawley rats and retained a fibroblast-like morphology when cultured in F12/DMEM medium with 10% (v/v) FBS and 50 μg/ml gentamicin (Zangani et al. 1999). Rat mammary adenocarcinoma cell lines NMU (derived from primary rat mammary tumors induced with the carcinogen N-methyl-nitrosourea) and RBA (derived from primary rat mammary tumors induced with the carcinogen dimethylbenzanthracene) were obtained from American Type Culture Collection (Rockville, MD). Preadipocytes, fibroblasts, NMU cells, and RBA cells were set up at 1 × 106 cells in Falcon T175 tissue culture flasks in F12/DMEM with 10% (v/v) FBS and 50 μg/ml gentamicin. Medium was changed every 2 days and lysates were prepared as described below when the cells reached ~85% confluency.
Immunolocalization of ErbB2, ErbB3, and ErbB4
ErbB localization was examined in serial sections of formalin-fixed, paraffin-embedded abdominal mammary glands from at least three Sprague-Dawley rats during puberty (5054 days of age), at sexual maturation (83-85 days of age), at days 14-16 of pregnancy, during lactation (5-7 and 15 days after birth), and during involution (3-4 and 6-10 days after weaning), or in MEOs cultured for up to 21 days in 100-mm dishes in two primary culture studies. For the developmental series, one large assay was carried out for the detection of each ErbB receptor to allow intensity levels and localization of that receptor to be compared in the different sections. ErbB4 but not ErbB2 and ErbB3 detection required a PBS microwave antigen retrieval incubation. ErbB2 localization and specificity were evaluated using a 2-hr room temperature (RT) incubation with either rabbit ErbB2 antibody, ErbB2 antibody preincubated with ErbB2 peptide or EGFR peptide, rabbit IgG, or PBS only. ErbB4 localization and specificity were analyzed using a 2-hr RT incubation with either rabbit ErbB4 antibody, ErbB4 antibody preincubated with ErbB4 peptide or ErbB2 peptide, rabbit IgG, or PBS. The two rabbit primary antibodies were used at 1 μg/ml and preincubated overnight at 4C with either 0 or 10 μg/ml of the appropriate peptide before use. ErbB3 localization and specificity were examined using a 2-hr RT incubation with 10 μg/ml of either mouse IgG2a monoclonal ErbB3 antibody or the isotype control, mouse IgG2a. Reactive proteins were visualized using preabsorbed F(ab')2 fragments of affinity-purified donkey anti-rabbit or anti-mouse antibody conjugated with biotin, a streptavidin-horseradish peroxidase conjugate, and DAB, and sections were counterstained with hematoxylin. Color photographs were taken of identical fields in serial sections using a Nikon FX-35A camera and an Olympus BH-2 microscope, photographs were scanned using an AGFA Argus II flatbed scanner, and images were processed using Adobe Photoshop (Adobe Systems; San Jose, CA) and printed with a Kodak DS 8650 PS printer.
Lysate Preparation
Lysates were prepared of intact mammary glands, isolated MEOs, and various types of cultured cells using ice-cold lysis buffer [50 mM Tris, pH 8.0, at 4C, with 150 mM NaCl, 2 mM EDTA, 10 mM Na2HP04, 10 mM Na4P2O7-10H2O, 5 mM Na3VO4, 1% (v/v) Triton X-100, 0.1% (w/v) Nadodecyl sulfate (SDS), 0.5% (w/v) Na deoxycholate, 1 mM phenylmethylsulfonyl fluoride (PMSF), 100 μg/ml soybean trypsin inhibitor, and 20 μg/ml leupeptin]. Rat liver or intact mammary glands excised either from nulliparous adult (85-day-old or 6-month-old) rats or from rats at Day 21 of lactation were harvested in 5 ml lysis buffer per gram of tissue. MEOs isolated from 50-day-old nulliparous rats were harvested in 1 ml of lysis buffer per 2 × 107 MECs. Finally, lysates were prepared of cultured cells, using 1 ml of lysis buffer per T175 flask containing PBS-washed cells and scraping. Each lysate was sonicated three times on ice for 10 sec using a Tekmar sonic disruptor, vortexed for 10 min at 4C, and microfuged for 15 min at ~12,000 × gat 4C. Mammary gland samples were homogenized with three 30-sec bursts of a Polytron on ice before the sonication step. Supernatants were quickly frozen in liquid nitrogen and stored at −20C.
Alternatively, Trizol extracts were prepared of MEOs isolated from the mammary glands of rats during puberty (50-52 days of age), pregnancy (Days 14-16 of gestation), lactation (6-7 days after birth), and involution (7 days after weaning 21-day-old pups) as previously described (Varela and Ip 1996). MEOs isolated from at least three independent sets of rats for each developmental stage were homogenized in 1 ml of Trizol reagent per 107 cells, and total cellular protein recovered as recommended by the manufacturer. Lyophilized proteins were solubilized in 10 M urea containing 50 mM dithiothreitol and mixed with sample buffer as previously described (Varela and Ip 1996).
Immunoblot Detection of EGFR, ErbB2, ErbB3, and ErbB4
Lysates mixed with reducing and denaturing sample buffer were separated on 4-20% polyacrylamide gradient gels and transferred to Immobilon P membranes. Immunoblot analysis was performed on membranes using an overnight 4C incubation either with 0.1 μg/ml of rabbit antibodies against EGFR (SC-03), ErbB2 (SC-284), or ErbB3 (SC-285) or with 4 μg/ml of NeoMarkers' ErbB4 Ab-2. Alternatively, membranes were incubated with 2 μg/ml of either rabbit IgG or rabbit anti-ErbB2 Ab-1, or with 0.5 μg/ml of either mouse IgG1 or mouse anti-ErbB2 Ab-17. Membranes were then incubated for 1 hr at RT with preabsorbed F(ab')2 fragments of affinity-purified donkey anti-rabbit IgG or anti-mouse IgG conjugated to horseradish peroxidase. Immunoreactive bands were visualized on X-ray film using ECL reagent, exposed films were scanned, and images were processed as described above and printed with a Kodak DS 8650 printer.
Immunoprecipitation of EGFR, ErbB2, ErbB3, and ErbB4
Mammary gland lysates (100 μl/reaction) from nulliparous adult (85-day-old) rats or rat liver lysates (100 μl/reaction) were precleared with protein G/agarose and incubated overnight at 4C with 1 μg of rabbit IgG, or affinity-purified rabbit antibodies against EGFR (SC-03), ErbB2 (SC-284), ErbB3 (SC-285), or ErbB4 (SC-283). Alternatively, precleared lysates were incubated overnight with 10 μg of either rabbit IgG or rabbit anti-ErbB2 Ab-1 (clone 21N), or with 2 μg of either mouse IgG1, mouse anti-ErbB2 Ab-9, or mouse anti-ErbB2 Ab-17. Samples were then incubated with protein G/agarose for 90 min at 4C. Immune complexes were precipitated with protein G/agarose, washed four times with ice-cold lysis buffer, and dissociated from the protein G/agarose by boiling in 50 μl of 2 × sample buffer [62.5 mM Tris-HCl, pH 8.8, at RT with 2% (w/v) SDS, 2.5% (v/v) glycerol, 5% (v/v) β-mercaptoethanol, and 0.0125% (w/v) bromophenol blue]. Thirty μl of each sample was separated on replicate 4-20% polyacrylamide gradient gels and transferred to Immobilon P membranes. Individual membranes were incubated overnight at 4C with 0.1 μg/ml of affinity-purified rabbit polyclonal antibody to EGFR (SC-03), ErbB2 (SC-284), ErbB3 (SC-285), or ErbB4 (SC-283) and then for 1 hr at RT with preabsorbed F(ab')2 fragments of affinity-purified donkey anti-rabbit IgG conjugated to horseradish peroxidase. Immunoreactive bands were visualized and processed as indicated above.
MTT Assay to Quantify Viable Cell Numbers
Cell numbers were quantified in triplicate culture wells per treatment for each of the different time points using an MTT assay (Hahm and Ip 1990; Ip et al. 1992). Briefly, cultures were incubated with MTT for 16 hr at 37C, the RBM was digested away from the MEOs using 5 U/ml dispase in F12/ DMEM medium, and the aqueous-insoluble formazan crystals were collected and solubilized in 2-propanol. Sample absorbance was evaluated with a Bio-tek EL-311 automatic plate reader at 570 nm. Production of formazan crystals and absorbance of the solubilized crystals were directly proportional to viable cell number. A standard curve was set up with the newly isolated MEOs for each experiment. A comparison of nuclei counting in cultured MEOs with cell numbers obtained using the MTT assay confirmed that PD158780 did not interfere with MTT metabolism in these cells (Varela et al. 1997).
Evaluation of MEC-specific Functional Differentiation
Casein accumulation, used as an indicator of MEC functional differentiation, was monitored in triplicate culture wells per treatment for each of the different time points. Culture medium was removed and lysates of RBM without or with MEOs were prepared at 4C with ice-cold lysis buffer containing 0.1 mM PMSF, 100 ng/ml soybean trypsin inhibitor, and 20 ng/ml leupeptin. Each lysate was sonicated three times on ice for 10 sec, vortexed, and microfuged for 15 min at ~12,000 × g at 4C. Supernatants were quickly frozen in liquid nitrogen and stored at − 20C. Casein levels were then quantified in these lysates using a previously described noncompetitive ELISA (Ip et al. 1992) with a rat casein polyclonal rabbit antiserum (Hahm et al. 1990).
Microscopic Examination of and Apoptosis Detection in Cultured MEOs
An Olympus CK2 microscope mounted with a Nikon FX-35A camera was used for time-lapse light microscopic examination and photography of individual colonies. Three culture wells of living organoids for each treatment group were repeatedly examined to identify changes in colony number, size, coloring, and/or shape during the course of the 21-day culture period. Cultured MEOs in 100-mm dishes were exposed to TGFα medium with 0 or 0.5 μM PD158780 from Days 18-20 of the study, fixed with formalin, embedded in paraffin, and sectioned at 4–5 μm. Sections were processed using Oncor's ApopTag Peroxidase In Situ Apoptosis detection kit as described by the manufacturer, and then analyzed and photographed using an Olympus BH-2 microscope and a Nikon FX-35A camera. Images were processed using Adobe Photoshop and printed using a Kodak DS 8650 PS printer.
Statistics
Data are presented as mean ± SEM. Statistical significance was evaluated using a one-way analysis of variance (ANOVA) with the Student-Newman-Keuls test for pair-wise multiple comparisons. p<0.05 was judged to be statistically significant.
General immunohistochemical localization of ErbB2, ErbB3, and ErbB4 in the mammary glands of either nulliparous rats or rats during pregnancy, lactation, or involution
aErbB receptor staining is indicated as being positive (P), weak (W), or negative (N).
bRepresentative of mammary glands from at least three nulliparous rats at 50-53 and 83-85 days of age.
cRepresentative of mammary glands from at least three pregnant rats at Days 14-16 of gestation.
dRepresentative of mammary glands from at least three rats at Days 5–7 and 10-15 of lactation.
eRepresentative of mammary glands from at least three rats at Days 3–4 and 6-10 of involution.
Results
Localization of ErbB2, ErbB3, And ErbB4 During Rat Mammary Gland Development
If select ErbB receptors play natural, non-oncogenic roles in regulating growth, differentiation, apoptosis, and/or remodeling in normal mammary glands, then these receptors should be differentially expressed in mammary epithelial and/or stromal cells during stages of extensive growth (puberty and pregnancy), differentiation (pregnancy and lactation), apoptosis (involution), and/or tissue remodeling (puberty and involution). This hypothesis was examined using an immunohistochemical study and the results are summarized in Table 1. During puberty, terminal end buds expressed a high level of ErbB2 (Figure 1A) and ErbB4 (Figure 1C), but limited ErbB3 (Figure 1B). Most fibroblasts and adipocytes surrounding these end buds exhibited ErbB2 (Figure 1A), ErbB3 (Figure 1B), and/or ErbB4 (Figure 1C) staining.
Specificity of ErbB2 reactivity was demonstrated using rabbit IgG (not shown) and by selective competition of the rabbit ErbB2 antibody SC-284 with its immunogenic peptide (Figure 1D) but not the EGFR peptide (not shown). Moreover, rat sweat glands, skeletal muscle, salivary gland, and kidney were positive controls, whereas rat liver was the negative control tissue for the ErbB2 immunolocalization studies (not shown). The absence of brown staining in mammary sections exposed to mouse IgG2a isotype control antibody confirmed the specificity of the ErbB3 monoclonal antibody clone 2F12 (Figure 1E). The immunizing peptide was not available for peptide competition of this antibody. Specificity of ErbB4 staining was determined by selective competition of the rabbit ErbB4 antibody (SC-283) with its immunizing ErbB4 peptide (Figure 1F), and using rabbit IgG (not shown). Infiltrating macrophages were found to nonspecifically react with the ErbB4 antibody (Figure 1F, arrowhead). It should also be noted that rat brain and epidermis were the positive control tissues for the ErbB3 and ErbB4 immunohistochemistry assays (not shown).
During puberty (50-54 days of age; Figures 2A–2I) and at sexual maturation when proliferation has decreased (83-85 days of age; not shown), luminal epithelial cells in mammary gland ducts and alveoli exhibited ErbB2 (Figures 2A and 2B), ErbB3 (Figures 2D–2F), and/or ErbB4 (Figures 2G and 2H) staining either along the apical cell membrane or in both the cytoplasm and along apical and/or basolateral cell membranes. Myoepithelial cells surrounding epithelial ducts and alveoli were positive for ErbB2 (Figure 2B) and/or expressed moderate ErbB3 (Figure 2E) and minimal ErbB4 (Figures 2G and 2H) reactivity. Mammary fibroblasts and adipocytes surrounding ducts and alveoli exhibited ErbB2 (Figures 2A–2C) and ErbB3 (Figures 2D–2F) staining, and either limited or no ErbB4 reactivity (Figures 2G–2I).
Immunolocalization of ErbB2, ErbB3, and ErbB4 in mammary gland terminal end buds. Serial sections of mammary glands from 50-day-old virgin rats were evaluated with the rabbit ErbB2 antibody (SC-284,
During pregnancy (Days 14-16), ductal epithelial cells expressed a low level of cytoplasmic and apical membrane-associated ErbB2 (Figure 3A) and extensive ErbB3 reactivity (Figure 3D). Alveolar epithelial cells did not exhibit ErbB2 (Figure 3B) nor ErbB3 (Figure 3E) staining along the cell membrane. ErbB2 expression in myoepithelial and stromal cells was either limited or below detection limits (Figures 3A–3C). In contrast, most of the remaining fibroblasts (Figure 3D) and adipocytes (Figure 3F) exhibited strong ErbB3 staining, whereas myoepithelial cells either failed to express or expressed moderate ErbB3 reactivity (Figures 3D and 3E). ErbB4 staining was almost completely absent in mammary epithelial and stromal cells at this time point in pregnancy (Figures 3G–3I).
During lactation, ErbB2 staining was downregulated in the functionally differentiated epithelial cells within ducts (Figure 3J) and alveoli (Figure 3K). Modest ErbB3 staining was observed in the apical cytoplasm and on the cell membrane of luminal epithelial cells (Figures 3M and 3N). Weak ErbB4 staining was detected in certain luminal epithelial cells, and the localization was supranuclear (not shown). Although myoepithelial cells, fibroblasts, and adipocytes surrounding functionally differentiated MEC did not exhibit ErbB2 (Figures 3J–3L) or ErbB4 (not shown) reactivity, a number of these cells displayed distinct ErbB3 staining (Figures 3M and 3O). There was no difference in the intensity or localization of ErbB2, ErbB3, or ErbB4 staining within the various cell types in mammary glands from rats at Days 5–7 compared with Day 15 of lactation (not shown).
Immunolocalization of ErbB2, ErbB3, and ErbB4 in mammary glands from nulliparous female rats. Serial sections of mammary glands from 50-day-old virgin rats were evaluated with the rabbit ErbB2 antibody (SC-284,
Immunolocalization of ErbB2, ErbB3, and ErbB4 in mammary glands during pregnancy, lactation, and involution. Serial sections of mammary glands from mid-pregnant rats at Days 14-16 of gestation (
Differential expression of EGFR, ErbB2, ErbB3, and ErbB4 in rat cells and tissues. Triton lysates of cultured rat mammary tumor cell lines RBA (Lane 1) and NMU (Lane 2), cultured rat mammary fibroblasts (MFC, Lane 3) and preadipocytes (MPA, Lane 4), rat liver (Lane 5), and mammary glands excised either from rats at Day 21 of lactation (L21-MG, Lane 6) or from 85-day-old virgin rats (V85-MG, Lane 7) were separated on replicate 4-20% polyacrylamide gradient gels and then transferred to membranes. Immunoblot analysis was carried out using the EGFR antibody SC-03 (
ErbB receptor expression was examined at Days 3–4 (Figures 3P–3V) of postweaning involution, and ductal MECs displayed distinct ErbB2 (Figure 3P) and ErbB3 (Figure 3T) reactivity that was cytoplasmic, supranuclear, and/or cell membrane-associated. Luminal epithelial cells in alveoli (Figures 3Q and 3U) and residual epithelial cells in apoptotic alveoli (Figures 3R and 3V) exhibited grainy ErbB2 (Figures 3Q and 3R) and ErbB3 (Figures 3U and 3V) staining. Limited supranuclear ErbB4 staining was observed in a number of healthy epithelial cells as well as in apoptotic alveoli (not shown). Myoepithelial cells and fibroblasts exhibited ErbB2 (Figures 3P–3R) and/or ErbB3 (Figures 3T–3V) staining but not ErbB4 reactivity (not shown). Mature adipocytes displayed limited staining for ErbB2 (Figure 3S) and prominent ErbB3 reactivity (not shown) but were not ErbB4-reactive (not shown). Mammary epithelial and stromal cells continued to express ErbB2 and ErbB3 through at least Day 10 of involution (not shown).
Differential Expression of Select ErbB Receptors in Rat Mammary Cells and Tissues
The observation that the ErbB2, ErbB3, and ErbB4 antibodies recognized distinct cell types during select developmental stages in rat mammary glands strongly supports the contention that these antibodies did not have overlapping reactivity. However, to confirm the specificity of these ErbB antibodies, lysates prepared from different rat mammary cells and tissues were evaluated by immunoblot analysis (Figure 4). This experiment demonstrated that the EGFR, ErbB2, and ErbB3 antibodies each detected distinct proteins. First, Lanes 1 and 2 in Figures 4A–4C indicate that in RBA and NMU rat mammary tumor cells, the proteins detected by each antibody were of a different molecular weight (165, 185, and 180 kDa for EGFR, ErbB2 and ErbB3, respectively). Second, although EGFR and ErbB2 were detected in mammary fibroblasts (MFCs) and mammary preadipocytes (MPAs), ErbB3 was not, demonstrating that the EGFR and ErbB2 antibodies did not crossreact with ErbB3 (Lanes 3 and 4 in Figures 4A–4C). Third, although EGFR was detected in rat liver, ErbB2 and ErbB3 were not, showing that the EGFR antibody did not crossreact with ErbB2 and ErbB3 (Lane 5 in Figures 4A–4C). Fourth, EGFR, ErbB2 and ErbB3 were all detected in normal rat mammary gland lysates from 85-day-old virgin female rats (V85-MG) and rats at Day 21 of lactation (Lanes 6 and 7 in Figures 4A–4C). Finally, we found that the ErbB4 antibody SC-283 (not shown) and Ab-2 (Figure 4D) were not suitable for direct immunoblot detection. It should be noted that Schroeder and Lee (1998) reported similar results using the SC-283 antibody. Together, these tissue/cell-specific differences in ErbB receptor expression suggest that the individual ErbB receptor antibodies are not reacting nonspecifically with other family members.
Confirmation of ErbB2 Expression in Mammary Glands of Nulliparous Rats
Although Dati and co-workers (1996) used an immunofluorescence assay to demonstrate that ErbB2 was expressed in the mammary glands of virgin and early pregnant rats but not of late pregnant and lactating rats, their immunoblot study indicated that high levels of ErbB2 were detected in mammary gland lysates from rats at late pregnancy and lactation but not from virgin rats. To further bolster our data that ErbB2 is present in the mammary gland of nulliparous rats, immunoblot and immunoprecipitation studies were carried out using different ErbB2-reactive antibodies including Ab-1 (clone 21N) used by Dati and co-workers. Figure 5A demonstrates that full-length ErbB2 was readily detected with the ErbB2 selective antibodies Ab-17 (clones e2-4001+3B5), SC-284, and Ab-1 (clone 21N) in a mammary gland lysate from 85-day-old virgin female rats (MG) but not in the negative control rat liver lysate. In addition, mouse IgG1 (Lanes 1 and 2 in Figure 5A) and rabbit IgG (Lanes 7 and 8 in Figure 5B) did not detect a 185-kD protein in either lysate. Incubation of Ab-1 with its immunogenic peptide competed the 185-kD reactive protein but not the lower molecular weight nonspecific proteins (compare Lanes 1 and 3 in Figure 5B).
Detection of ErbB2 in mammary gland lysates using different ErbB2-reactive antibodies. Triton lysates of mammary glands excised from 85-day-old virgin rats (MG) or rat liver were separated on 4-20% polyacrylamide gradient gels and then transferred to Immobilon P membranes. (
An immunoprecipitation study demonstrated that full-length ErbB2 was detected in the pellet recovered when a mammary gland lysate from 85-day-old female rats was immunoprecipitated with the ErbB2-selective antibodies SC-284, Ab-1 (clone 21N), Ab-9 (clone B10), or Ab-17 (clones e2-4001 + 3B5) but not with rabbit IgG or mouse IgG1 (Figure 6A). Furthermore, ErbB2 was depleted in the supernatant fraction recovered when another aliquot of the same lysate was immunoprecipitated with any of the four ErbB2-selective antibodies but not when immunoprecipitated with rabbit IgG or mouse IgG1 (Figure 6B). Finally, rat liver lysate was used as a negative control to demonstrate that none of the ErbB2-selective antibodies, rabbit IgG, or mouse IgG1 was able to immunoprecipitate a 185-kD protein (Figure 6C).
An additional immunohistochemical study was carried out to determine whether Ab-1 and/or Ab-17 could reliably detect ErbB2 in formalin-fixed, paraffin-embedded rat salivary gland, kidney, and mammary gland, but not in liver. Although Ab-1 showed strong reactivity in salivary gland, kidney, and mammary gland, the staining was not competed when Ab-1 was preincubated with its immunogenic peptide (not shown), presumably because of the nonspecific protein with which this antibody also reacts (Figure 5B). In addition, Ab-17 did not show positive staining in any of the tissues tested (not shown). In this study, SC-284 staining was observed in rat salivary gland and kidney as well as in mammary glands from nulliparous, pregnant, and postlactational rats but not in rat liver or mammary glands from lactating rats.
Relative Epithelial Expression of ErbB2 And ErbB3 During Mammary Gland Development
Our immunohistochemical studies demonstrated that ErbB2 and ErbB3 levels were decreased in the lactating animals compared with the rats at other developmental stages (Figures 1–3). To examine this more definitively, epithelial cell-specific changes in ErbB2 and ErbB3 were examined in lysates of epithelial organoids isolated from the mammary glands of rats during puberty, pregnancy, lactation, and involution. Both mechanical and enzymatic steps were employed to dissociate the epithelium from the surrounding stroma. This resulted in detection of multiple immunoreactive proteins in these lysates (Figure 7). Figure 7A illustrates the detection of three immunoreactive proteins using the ErbB2-selective antibody SC-284 in lysates of isolated mammary epithelial organoids. The 125-and 145-kD proteins are believed to represent proteolytic fragments of full-length ErbB2 (Cassel and Glaser 1982; Yang et al. 1994; McIntyre et al. 1995). All three ErbB2-reactive proteins were competed when the ErbB2-selective antibody SC-284 was preincubated with its immunogenic peptide (Figure 7B). The ErbB3-selective antibody SC-285 detected two immunoreactive proteins in lysates of isolated mammary epithelial organoids (Figure 7C). The 160-kD protein is believed to represent a proteolytic fragment of full-length ErbB3 (Figure 7C). The 180- and 160-kD proteins were competed when SC-285 was preincubated with its immunogenic peptide (Figure 7D). Epithelial expression of ErbB2 (Figure 7A) and ErbB3 (Figure 7C) were found to be high during puberty and pregnancy, low throughout lactation, and then to increase again by Day 7 of involution. Similar results were obtained in at least three independent experiments. Downregulation of ErbB2 and ErbB3 during lactation was not simply an artifact associated with a higher level of casein in these samples because casein levels were previously shown to be similar in other epithelial lysates isolated from pregnant and lactating mammary glands (Varela and Ip 1996).
Immunoprecipitation of ErbB2 using four different ErbB2-reactive antibodies. Triton lysates of mammary glands excised from 85-day-old virgin rats (MG, Lane 1) and rat liver (Lane 2) were used as positive and negative control samples, respectively, for the ErbB2 immunoprecipitation study. The pellet (
Relative expression of ErbB2 and ErbB3 in MEC during different stages of rat mammary gland development. Trizol samples were prepared of mammary epithelial organoids isolated from 50-day-old virgin rats (V-MEO, Lane 1), or from rats at Days 14-16 of pregnancy (P-MEO, Lane 2), Day 7 of lactation (L-MEO, Lane 3), or Day 7 of involution (I-MEO, Lane 4). Samples were separated on 4-20% polyacrylamide gradient gels and transferred to Immobilon P membranes. Immunoblot analysis was carried out using the ErbB2-reactive antibody SC-284 (
Immunoprecipitation of Distinct ErbB Receptor Heterodimers
The coincident localization of the various ErbB receptors in the mammary glands of virgin rats (Figures 1 and 2) suggested the possibility that these receptors formed heterodimers. In addition, the ErbB ligands EGF, TGFα, amphiregulin, and neuregulin-α have previously been shown to be expressed in mammary epithelial and stromal cells (Panico et al. 1996; DiAugustine et al. 1997; Herrington et al. 1997). These proteins induce their effects by binding to select ErbB receptors, inducing the formation of ErbB receptor homo- and heterodimers and activating the tyrosine kinase domain of the ErbB receptors (Earp et al. 1995). An immunoprecipitation study was carried out to examine ErbB receptor dimerization in mammary gland lysates prepared from nulliparous adult rats. Figure 8A demonstrates that EGFR was detected in the pellets recovered when lysates where immunoprecipitated with sheep or rabbit antibodies against EGFR but not with the rabbit IgG or rabbit antibodies against ErbB2, ErbB3, or ErbB4. Furthermore, EGFR was depleted from the recovered supernatant fraction only when the lysate was immunoprecipitated with an EGFR antibody (Figure 8B). ErbB2 was detected in pellets recovered when lysates were immunoprecipitated with sheep or rabbit antibodies against EGFR and with rabbit antibodies against ErbB2, ErbB3, and ErbB4 but not with rabbit IgG (Figure 8C). In addition, ErbB2 was depleted from the recovered supernatant fraction only when the immunoprecipitation reaction was carried out using the antibody against ErbB2 (Figure 8D), which suggests that the antibodies against EGFR, ErbB3 and ErbB4 did not nonspecifically precipitate ErbB2 out of the lysate and that no one dimer pair predominated in the lysate. ErbB3 (Figure 8E) and ErbB4 (Figure 8G) were detected in the pellet recovered when the lysate was immunoprecipitated with rabbit polyclonal antibodies against EGFR, ErbB2, ErbB3, or ErbB4 but not with rabbit IgG. Analysis of the supernatants from these immunoprecipitation reactions confirmed that only the antibody against ErbB3 was able to deplete a 180-kD protein from the mammary gland lysate (Figure 8F) and that SC-283 was not able to directly detect ErbB4 in any of the immunoprecipitation supernatants (Figure 8H).
Co-immunoprecipitation of distinct ErbB receptor heterodimers. A Triton lysate was prepared of homogenized mammary glands from nulliparous adult (85-day-old) rats and immunoprecipitated (IP) with a sheep antibody against EGFR (UBI 06-129, Lane 1), rabbit IgG (Lane 2), or affinity-purified rabbit antibodies against EGFR (SC-03, Lane 3), ErbB2 (SC-284, Lane 4), ErbB3 (SC-285, Lane 5), or ErbB4 (SC-283, Lane 6). The pellet (
Role of ErbB Receptors in Normal MECs
The above data suggested that ErbB receptors may be important in rat mammary gland development during puberty, pregnancy, and involution and at sexual maturation but not during lactation. Because embryonic lethality prevented the study of mammary gland development in ErbB2 (Lee et al. 1995), ErbB3 (Riethmacher et al. 1997), or ErbB4 (Gassmann et al. 1995) null mice, a primary culture model system was used to test the hypothesis that ErbB receptor signaling is required for the growth, differentiation, and/or apoptosis of normal MECs. MECs were isolated as organoids from pubescent female rats and cultured under defined serum-free culture conditions in the absence of mammary fibroblasts and/or adipocytes (Hahm et al. 1990; Darcy et al. 1991, 1995). Changes in ErbB receptor expression were examined in serial sections of organoids cultured for 7, 14, or 21 days, and TGFα-stimulated ErbB receptor-dependent effects were identified using the pyrido-pyrimidine analogue PD158780 (Rewcastle et al. 1996). This drug inhibited EGF- and heregulin-stimulated tyrosine phosphorylation and growth in human breast cancer cells that expressed EGFR, ErbB2, and ErbB3 but not ErbB4, as well as in cells that expressed ErbB2, ErbB3, and ErbB4 but not EGFR (Fry et al. 1997). PD158780 at 0.5 μM inhibited ErbB-dependent tyrosine kinase activity but had no effect on platelet-derived growth factor receptor-or fibroblast growth factor receptor-dependent tyrosine phosphorylation or growth (Fry et al. 1997).
Immunolocalization of ErbB2, ErbB3, and ErbB4 during the in vitro development of mammary epithelial cells. Serial sections of MEO cultured for 7 (
On the first day of culture, immature organoids were small, lobular, and heterogenous in their expression of cytoplasmic and cell membrane-associated ErbB2, ErbB3, and/or ErbB4 (not shown). After 7 days in culture, most MECs developed into complex multilobular alveolar organoids, with the individual lobes being primarily composed of a monolayer of well polarized epithelial cells organized around a distended central lumen (Figures 9A, 9D, and 9G) (Darcy et al. 1991,1995). The majority of these luminal epithelial cells expressed cytoplasmic and cell membrane-associated ErbB2 (Figure 9A), ErbB3 (Figure 9D), and/or ErbB4 (Figure 9G) reactivity. ErbB2 and ErbB4 reactivity were selectively competed with the corresponding immunizing peptide but not with a nonspecific peptide (not shown). In addition, epithelial organoids did not stain with rabbit IgG or mouse IgG2a (not shown).
To evaluate the role of ErbB receptor signaling in the development of functionally immature MECs, newly isolated epithelial organoids were cultured in TGFα-containing medium for 4 or 7 days and exposed to 0.5 μM PD158780 from Days 0–4 or 4–7 of the study, respectively. Exposure of MEOs to 0.5 μM PD158780 during the first week of culture inhibited epithelial cell growth (Figure 10A) and casein accumulation (Figure 10B). In the presence of PD158780, the epithelial organoids at Day 4 or Day 7 of the study resembled organoids at Day 0 of the study (not shown). These findings support the hypothesis that ErbB receptor signaling was required for the induction of epithelial growth, functional differentiation, and organoid morphogenesis and invasion into the surrounding RBM.
Effect of the pyrido-pyrimidine analogue PD158780 on growth, casein accumulation, and apoptosis in cultured MECs. (
After 14 days in culture, epithelial cell numbers continued to increase (Figure 10C) and the multilobular alveolar organoids were primarily composed of functionally differentiated epithelial cells that produced and accumulated extensive casein (Figure 10D) and intracellular lipid (Figures 9B, 9E, and 9H). ErbB2 (Figure 9B) and ErbB3 (Figure 9H) stainings were significantly downregulated in these functionally differentiated epithelial cells. In contrast, distinct ErbB4 staining was noted in the cytoplasm and/or along the cell membrane (Figure 9H). Exposure of MEOs to PD158780 from Days 11-14 of culture did not inhibit epithelial cell growth (Figure 10C), casein accumulation (Figure 10D), or the morphologic appearance of the organoids. This supports the hypothesis that ErbB receptors are not required to maintain functional differentiation of MECs.
In this study, viable cell numbers (Figure 10C) and casein accumulation (Figure 10D) continued to increase during the third week of culture. Exposure of MEOs to 0.5 μM PD158780 from Days 18-21 of culture did not affect casein accumulation (Figure 10D) but cell numbers were decreased (Figure 10C). The decrease occurred as large numbers of cells along the periphery of many of these organoids broke away and died (Figure 10F). Cell death was judged microscopically by the inability of these cells to convert the yellow MTT dye to purple formazan crystals. Furthermore, apoptotic DNA can be readily seen in the lumen of organoids exposed to PD158780 from Day 18-21 (Figures 9C, 9F, and 9I). It should be noted that exposure of MEOs to 0.5 μM PD158780 before Day 18 of culture did not induce morphological evidence of apoptosis. Moreover, MEOs cultured with 0 (not shown) or 0.5 μM (Figures 9C, 9F, and 9I) PD158780 from Days 18-21 expressed a similar level of cytoplasmic and cell membrane-associated ErbB2 (Figure 9C), ErbB3 (Figure 9F), and ErbB4 (Figure 9I). The latter finding confirmed that suppression of ErbB signaling was not required for the re-expression of ErbB2 and ErbB3 in terminally differentiated MECs.
An additional primary culture study was carried out to quantify the apoptotic response observed when MEOs were exposed to 0.5 μM PD158780 starting on Day 18 of culture. PD158780 exposure for 48 but not for 24 hr starting on Day 18 of culture induced 75% of the luminal epithelial cells within individual organoids to undergo apoptosis (Figure 10G). Analysis of serial sections of these organoids confirmed that apparently healthy and apoptotic MECs exhibited extensive ErbB2, ErbB3, and ErbB4 staining (not shown).
Discussion
This report presents novel and distinct localization and expression profiles for ErbB2, ErbB3, and ErbB4 in the rat mammary gland during puberty, pregnancy, lactation, and involution. ErbB2 localization was detected immunohistochemically using antibody SC-284, levels were examined by immunoblot analysis using three different antibodies (SC-284, Ab-1, and Ab-17), and immunoprecipitation was accomplished using four different antibodies (SC-284, Ab-1, Ab-9, and Ab-17). ErbB3 localization was identified using a mouse monoclonal anti-ErbB3 antibody, and antibody SC-285 was used for ErbB3 immunoblotting and immunoprecipitation. SC-283 was the only ErbB4-reactive antibody shown to detect ErbB4 in formalin-fixed, paraffin-embedded rat mammary glands or to immunoprecipitate ErbB4 from rat mammary gland lysates.
Our immunohistochemical studies demonstrate that the naturally proliferative and invasive parenchymal cells within terminal end buds co-express a high level of ErbB2 and ErbB4 but not ErbB3, whereas luminal epithelial cells within ducts and alveoli of nulliparous (pubescent and adult) rats coordinately co-express ErbB2, ErbB3, and/or ErbB4. By Days 14-16 of pregnancy, the epithelium undergoes extensive proliferation, differentiation, and alveolar morphogenesis. At this stage, ductal epithelial cells maintain a high level of ErbB3 but ErbB2 is markedly reduced and ErbB4 expression is lost completely. Concurrently, limited or no ErbB receptors were detected along the cell membrane of alveolar epithelium. Although all three ErbB receptors are downregulated in MECs throughout lactation, high levels of ErbB2 and ErbB3 but not ErbB4 are re-expressed in rat mammary glands during involution. In comparison with the epithelial cells, mammary stromal cells express a high level of ErbB2 during puberty, at sexual maturation, and during involution, but this receptor is downregulated and then lost during pregnancy and lactation, respectively. In contrast, ErbB3 is expressed in mammary fibroblasts and/or adipocytes throughout the four developmental stages, whereas stromal cells surrounding terminal end buds, but not ducts or alveoli, express a high level of ErbB4. These studies lay the groundwork for future studies in which activation and signaling of the ErbB receptors can be investigated in the various mammary cell types during these distinct developmental stages. Our immunoblot and immunoprecipitation studies demonstrate the specificity and/or selectivity of the different ErbB receptor antibodies in various types of rat tissues and cells and confirm the developmental expression of ErbB2 and ErbB3 in rat mammary glands.
Our laboratory recently described selective changes in EGFR (ErbB1) expression and localization during rat mammary gland development (Darcy et al. 1999). EGFR localization was examined in serial sections of the paraffin blocks used for this study. A high level of EGFR was detected in the parenchymal cells throughout terminal end buds during puberty, and in luminal epithelial cells as well as stromal cells in nulliparous rats. Our immunoprecipitation study confirms the immunohistochemical findings suggesting that EGFR, ErbB2, ErbB3, and/or ErbB4 are differentially co-expressed in terminal end buds, ducts, alveoli, and/or stromal cells in functionally immature mammary glands. All six ErbB receptor heterodimers (ErbB2-EGFR, ErbB2-ErbB3, ErbB2-ErbB4, ErbB3-EGFR, ErbB3-ErbB4, and ErbB4-EGFR) were immunoprecipitated from mammary gland lysates prepared from adult nulliparous rats. In addition, epithelial cells and stromal cells expressed a high level of EGFR during pregnancy (Days 14-16 of gestation) and from Days 3-10 of involution, whereas EGFR was downregulated in rat mammary glands throughout lactation.
The individual ErbB receptors have been shown to possess unique promoters, catalytic activities, cellular routing, transmodulation domains, and cytoplasmic binding sites for a particular subset of tyrosine kinase substrates (Earp et al. 1995; Alroy and Yarden 1997; Carraway et al. 1997; Pinkas-Kramarski et al. 1997). Because this receptor family functions in homo- as well as heterodimeric pairs, the differential expression and localization of select ErbB receptors during development allow the activation of a diverse array of signal transduction pathways and the regulation of distinct biological responses in epithelial and/or stromal cells within the same microenvironment. In addition, the selective insertion of specific ErbB receptors into distinct apical and/or basolateral membranes in luminal epithelial cells, or along the entire cell membrane of immature parenchymal cells in terminal end buds, fibroblasts, and/or adipocytes, may contribute to cell type-specific functions of this receptor family. In luminal epithelial cells, basolateral ErbB receptor expression may regulate junctional complex stability as well as cell-extracellular matrix interactions, whereas apical ErbB receptor localization may regulate cytoskeletal organization and/or vectoral secretion. In contrast, general localization of ErbB receptors along the cell membrane may stimulate (a) growth, migration, and invasion of immature parenchymal cells in terminal end buds as well as stromal fibroblasts and preadipocytes, (b) lipogenesis in mature adipocytes, and/or (c) survival of immature and terminally differentiated epithelial cells. Cytoplasmic staining in the various mammary gland cell types is considered to be physiologically relevant and probably represents the synthesis and routing of ErbB2, ErbB3, or ErbB4.
Expression and Localization of ErbB2
Similar to our studies in rat, a high level of ErbB2 was detected in mammary gland lysates from nulliparous mice (Schroeder and Lee 1998; Sebastian et al. 1998). In addition, ErbB2 was shown to be expressed in ductal and alveolar epithelial cells in the mammary glands of virgin and pregnant rats (Dati et al. 1996) and in normal human breast (Press et al. 1990; Gompel et al. 1996). Interestingly, Press and co-workers localized ErbB2 to the basolateral but not the apical cell membrane of ductal and lobular epithelium, whereas we demonstrated ErbB2 localization to apical and/or basolateral cell membranes in luminal epithelial cells. Our immunohistochemical studies in rat mammary glands extend the human studies by demonstrating ErbB2 expression throughout terminal end buds and downregulation in mammary epithelial and stromal cells throughout lactation. In contrast, studies in mouse mammary glands revealed that ErbB2 was selectively expressed in the outer cap cell layer of terminal end buds of 35-day-old mice (DiAugustine et al. 1997) and in functionally differentiated luminal epithelial cells in mouse mammary glands during pregnancy and/or lactation (DiAugustine et al. 1997; Schroeder and Lee 1998). Histiotypic differences in mouse compared with rat mammary glands may explain the discrepancy in ErbB2 expression in terminal end buds between these two species. Mouse terminal end buds contain a highly proliferative and invasive outer layer of pale cells referred to as the cap or “stem” cell layer (Coleman et al. 1988; Russo and Russo 1996), whereas cap-like cells are only rarely seen in rat terminal end buds (Masso-Welch et al. in press). In addition, distinct regulatory machinery probably exists to selectively regulate ErbB2 expression and localization in the mammary glands of rats compared with mice. It should also be noted that our studies are the first to demonstrate ErbB2 expression in both healthy and apoptotic epithelial cells during involution, a stage of extensive apoptosis and tissue remodeling.
Expression and Localization of ErbB3
ErbB3 expression was demonstrated by immunoblot and immunohistochemical analysis in mouse mammary glands during pregnancy, lactation, and involution but not during puberty (Schroeder and Lee 1998). In contrast, our data show a high level of ErbB3 in rat mammary ducts and alveoli during puberty and at sexual maturation. Similar to our results, ErbB3 was preferentially localized in ductal rather than alveolar epithelium in mouse mammary glands during pregnancy (Schroeder and Lee 1998). Although a high level of ErbB3 was detected in mouse mammary glands during lactation (Schroeder and Lee 1998), ErbB3 is downregulated in rat mammary glands at this time. Our ErbB3 immunohistochemical study is the first to demonstrate that a high level of ErbB3 is expressed in ductal and alveolar epithelial cells but not in terminal end buds during puberty, in a majority of mammary fibroblasts and adipocytes during all four developmental stages, and in apparently healthy and apoptotic alveolar epithelial cells during involution. It should also be noted that ErbB3 was expressed in both stromal and epithelial cells in normal breast tissue from nonpregnant women (Lemoine et al. 1992). Finally, our immunoblotting data confirm our immunohistochemical results and demonstrate lower levels of ErbB3 in lysates of mammary gland (Figure 4) and isolated mammary epithelial organoids (Figure 7) from lactating compared with virgin rats, and that a high level of ErbB3 is re-expressed in MEC during involution (Figure 7).
Expression and Localization of ErbB4
Recently, functionally immature epithelial cells and myoepithelial cells in normal human breast tissue were shown to express cytoplasmic and cell membrane-associated ErbB4 (Srinivasan et al. 1998). Our immunohistochemical study extends these findings by demonstrating a high level of ErbB4 in luminal epithelial cells in ducts and alveoli of nulliparous rats during puberty and at sexual maturation, and a downregulation of ErbB4 in rat mammary glands during pregnancy and lactation. In contrast, ErbB4 was preferentially expressed in ductal compared with alveolar epithelium in mice at Day 18 of pregnancy and Day 2 of lactation, but not in the mammary glands of nulliparous mice nor mice during involution (Schroeder and Lee 1998). Our study is the first to show that ErbB4 is expressed throughout the naturally invasive terminal end buds, that stromal cells surrounding terminal end buds, but not ducts or alveoli, express a high level of ErbB4, and that ErbB4 remains downregulated during involution. Taken together, these studies support the hypothesis that the developmental expression and localization of ErbB4 are differentially regulated in rat and human compared with mouse mammary glands.
Implications of the Primary Culture Studies
Our primary culture studies not only examined the relationship between ErbB receptor expression and function but also tested the hypothesis that mammary stromal cells are not required for functionally immature and apoptotic MEC to express ErbB2, ErbB3, and/or ErbB4 and for the downregulation of all three receptors in functionally differentiated MECs. The culture studies demonstrate that EGFR (Darcy et al. 1999), ErbB2, and ErbB3 but not ErbB4 are downregulated in MECs that undergo extensive functional differentiation in a manner reminiscent of that observed during lactation, and then are re-expressed in terminally differentiated and apoptotic MECs similar to that described during involution. In addition, these physiologically relevant changes were observed in the absence of mammary fibroblasts and adipocytes. The in vitro studies also demonstrate that ErbB signaling is required for the development of immature MECs during the first week of culture and later for survival of terminally differentiated MECs, but is not required for maintaining epithelial cell growth and casein accumulation during the second week of culture, when a majority of the cells have already differentiated. Taken together, these findings suggest that this primary culture model may be ideal for regulatory studies that examine EGFR, ErbB2, and ErbB3 expression, localization, and function in normal MECs. However, because ErbB4 was not downregulated in culture as it was in vivo during pregnancy, lactation, and involution, additional studies are required to determine the factors that mediate the downregulation in vivo, and which factors are responsible for maintaining the in vitro expression of ErbB4. Interestingly, the expression of ErbB4 in differentiated MECs did not interfere with functional differentiation or survival of the cultured cells. It should be noted that the current culture studies were not able to determine which ErbB receptor dimers elicited the distinct effects observed in normal MECs, because the presence of the RBM interfered with our ability to immunoprecipitate the receptors.
In summary, the immunohistochemistry study demonstrates high level co-expression of ErbB2 and ErbB4 but not ErbB3 in the proliferative and invasive mammary gland terminal end buds, of ErbB2, ErbB3, and/ or ErbB4 in luminal epithelial cells in mammary glands from nulliparous rats, and of ErbB2 and ErbB3 in mammary epithelial and stromal cells during involution. The immunoblotting data indicate that the ErbB receptors were selectively expressed in distinct rat tissues and cells and confirmed the developmental expression of ErbB2 and ErbB3 in the rat mammary gland. All six ErbB receptor heterodimers could be immunoprecipitated from lysates of functionally immature mammary glands from nulliparous rats. These findings, along with the primary culture data, support the hypothesis that select ErbB receptors regulate growth, differentiation, survival, and/or remodeling in the normal mammary gland. These characteristics would enable breast cancer cells that overexpress these receptors to be highly aggressive and invasive. It appears unlikely that this receptor family plays a major role in regulating milk production, secretion, or transportation in rats during lactation because all the ErbB receptors are downregulated in epithelial cells throughout this developmental stage. Finally, additional studies are required to precisely define the factors that regulate ErbB receptor expression, localization, activation, signaling, and/or function in normal and tumor cells in the mammary gland. It is hoped that this information will allow the optimization of strategies to effectively target breast cancer cells that overexpress this receptor family.
Footnotes
Acknowledgements
Supported by DAMD17-94-J-4159, NIH CA33240, and CA64870, and by the shared resources of the NIH Cancer Center Support grant CA16056.
We thank Lawrence Mead, Wendy Shea-Eaton, Nannette Stangle-Castor, and Dr Patricia Masso-Welch for providing information and discussing issues related to this project, and for helping collect some of the samples for the immunohistochemistry study. Special thanks go to Suzanne M. Shoemaker for her technical expertise in quantifying casein accumulation in the cultured MECs, and to Dr Ping-Ping H. Lee and Dr Patricia Masso-Welch for their critical review of this manuscript.