Up-regulation of ARC, a marker of resistance to apoptosis. Paraffin sections from ovariectomized and E2-treated mice were immunostained with anti-ΑRC IgG. Note that ARC staining in E2-treated Cav-1^−/− mice was dramatically increased, and was not strictly limited to epithelial ductal areas, but was also uniformly increased throughout the mammary fat pad.

Status of proliferative markers in Cav-1^−/− mammary lesions. We also examined the levels of proliferative markers, such as PCNA, phospho-RB, and phospho-ERK-1/2 under conditions of estrogen-stimulation. Note that PCNA and phopho-RB (A and B, respectively) levels are clearly elevated in E2-treated Cav-1^−/− mammary lesions. In contrast, phospho-ERK-1/2 levels (C) were more highly elevated in the mammary stroma of E2-treated Cav-1^−/− mice. This is consistent with E2-dependent activation of the mammary stroma in Cav-1^−/− mice and the development of increased vascularization (stromal angiogenesis).

Cav-1^−/− mammary epithelial cells show the overexpression of PR, an estrogen-responsive target gene. The status of progesterone receptor (PR-A/B) expression was monitored by immunostaining. Note that E2-induced Cav-1^−/− mammary lesions show the dramatic up-regulation of PR-A/B. Since PR-expression is estrogen-dependent and is often used as a “reporter” of activated ER receptor signaling, our results are consistent with the idea that ER signaling is highly activated in Cav-1^−/− mammary epithelial cells. In comparison, wild-type (WT) mice (treated identically) show significantly less staining.

Cav-1^−/− mammary glands overexpress ER co-activator genes CAPER and Foxa1. Cav-1^−/− mice are known to over-express ER-α in the luminal mammary compartment. To determine whether Cav-1^−/− mice express other components of the ER-transcriptional machinery, we immunostained Cav-1^−/− mammary glands with a panel of antibodies directed against distinct ER co-activator genes. Note that two ER co-activator genes, CAPER and Foxa1, are dramatically overexpressed in Cav-1^−/− mammary glands; expression of ER-α is also shown for comparison. Interestingly, CAPER is overexpressed in both the luminal epithelial cells (white arrow) and the mammary stromal compartment (black arrow). Paraffin-sections from ovariectomized wild-type and Cav-1^−/− mice were used to better synchronize the cells within the mammary gland.

Estrogen-treated Cav-1^−/− mice develop mammary stromal angiogenesis, and overexpress VEGF in the mammary stroma. A: Trichrome staining. Note that in E2-treated Cav-1^−/− mice, increased mammary stromal vascularization was not strictly limited to the areas surrounding ducts, but was uniformly increased throughout the mammary fat pad. To highlight this increased stromal vascularity, a boxed area is shown at higher magnification. Inset magnification: 10×. B: VEGF immunostaining. Interestingly, increased mammary stromal vascularlization may be due to increased expression of VEGF, an angiogenic growth factor, in the Cav-1^−/− mammary stroma under conditions of estrogen-stimulation.

Diagram summarizing the role of Cav-1 loss of function in estrogen-induced mammary epithelial cell proliferation. Our current findings on the role of Cav-1 loss of function in mediating estrogen hypersensitivity are illustrated schematically. Briefly, we show here that ablation of a single gene is sufficient to confer estrogen-hypersensitivity and DCIS lesion formation in Cav-1^−/− mammary glands, as well as adjacent stromal angiogenesis. Mechanistically, we demonstrate that genetic ablation of Cav-1 causes the overexpression of both ER-α and ER- co-activator genes (such as CAPER, Foxa1, and Gata3).

DCIS-like lesions in Cav-1^−/− mice have an increased mammary stem/progenitor cell population. A: Localization of SPRR1A, a TEB marker, in normal mammary glands. Paraffin sections from 6-week-old virgin female wild-type mice were immunostained with anti-SPRR1A IgG. Note that SPRR1A is preferentially expressed in TEBs, but not in mammary ducts. This is consistent with the idea that TEBs contain mammary stem/progenitor cells. An inset, at right, shows a higher power view of SPRR1A immunostaining. Inset magnification: 5×. B: Cav-1^−/− mammary ductal lesions show up-regulation of SPRR1A, a TEB Marker. Paraffin sections from ovariectomized and E2-treated mice were immunostained with anti35 SPRR1A IgG. Note that in the Cav-1^−/− mammary epithelia, the levels of SPRR1A appear elevated, especially under conditions of estrogen (E2) supplementation. As SPRR1A is a TEB marker, these results suggest that there is an increase in mammary stem/progenitor cells. C: Cav-1^−/− mammary ductal lesions show the up-regulation of β-catenin. Paraffin sections from ovariectomized and E2-treated mice were immunostained with anti-β-catenin IgG. Note that in the Cav-1^−/− mammary epithelia, the levels of β-catenin appear highly elevated. These results suggest that the Wnt/β-catenin pathway is activated in Cav-1^−/− mammary epithelia, under conditions of estrogen (E2) stimulation.

Estrogen-induced Cav-1^−/− mammary lesions morphologically resemble DCIS, and show adjacent stromal angiogenesis. To understand the nature of these estrogen-induced mammary lesions, wild-type and Cav-1^−/− mammary glands were surgically excised after estrogen treatment, and samples were paraffin-embedded and subjected to standard histochemical stains Interestingly, in estrogen-treated Cav-1^−/− mice, these dysplastic foci clearly resemble human DCIS lesions, with complete luminal filling. Representative images of H&E- (A) and Trichrome-stained (B) sections are shown. Trichrome staining also revealed that these lesions are often encircled or associated with small blood vessels (arrow). Also, note that collagen-staining (light blue) is conspicuously reduced surrounding Cav-1^−/− mammary lesions, suggestive of E2-induced local collagen degradation. Boxed areas are shown at higher magnification to highlight nuclear morphology and nucleolar prominence in Cav-1^−/− mammary lesions. Images were taken with a ×60 objective. Inset magnification: 9×.

Changes in the expression and distribution of CAPER during human breast cancer onset and progression. Paraffin sections from normal DCIS, and well-differentiated invasive ductal carcinoma (IDC-WD) were immunostained with anti-CAPER IgG. Note that CAPER is expressed at relatively low levels in normal breast tissue and assumes a cytoplasmic distribution. In contrast, CAPER is expressed at significantly higher levels in DCIS and IDC samples, where it assumes a predominantly nuclear distribution. Representative high power images are shown. At least 3 to 5 samples were stained for each group and similar results were obtained. Boxed areas are shown to the far right at higher magnification to highlight the nuclear and nucleolar localization of CAPER in breast cancer lesions. Inset magnification: 5×.

Cav-1^−/− mammary lesions express elevated levels of B23/nucleophosmin, a nucleolar marker protein and predictor of tamoxifen resistance. Paraffin sections from ovariectomized and E2-treated mice were immunostained with anti- B23/nucleophosmin IgG (brown color). Then, sections were lightly counterstained with hematoxylin (blue color). B23/nucleophosmin is a well-established nucleolar marker protein that is normally elevated in human breast cancers and is a predictor of recurrence in ER(+) patients treated with tamoxifen. A: A comparison between E2-treated mice (WT versus Cav-1^−/−) is shown. Images were taken with a ×100 objective. Inset magnification: 3.5×. B: A comparison of all four experimental groups is shown. Images were taken with a ×40X objective. In (A) and (B), boxed areas are shown at higher magnification. Note that B23/nucleophosmin is up-regulated in Cav-1^−/− mammary lesions in response to E2 treatment. Conversely, its expression is dramatically down-regulated to undetectable levels in wild-type mice treated with E2. Thus, there is a complete reversal of the normal effects of estrogen-treatment on B23/nucleophosmin expression in Cav-1^−/− mammary glands. Inset magnification: 10×.

Analysis of CAPER expression in human normal and cancerous breast tissue by AQUA. A: We examined the expression of CAPER using a human breast cancer tissue microarray, consisting of 84 IDCs and 20 normal breast tissue samples. Note that CAPER is up-regulated selectively in human breast cancer samples. Representative image are shown. CAPER (red); CK (cytokeratin; green); nuclei (blue) B, C: Qua- ntitative analysis of CAPER expression was achieved by using the PM2000 AQUA automated quantitative analysis system (HistoRx, Inc). Note that CAPER protein expression is significantly elevated in all three tumor grades (well, moderately, and poorly differentiated). However, CAPER is most highly expressed in moderately differentiated breast cancers. 1, well; 2, moderate; and 3, poorly differentiated. Two different representations of CAPER expression are shown. *P < 0.01.

Cav-1^−/− mammary lesions show nuclear changes, nucleolar prominence, mitotic figures, and an abnormal distribution of myoepithelial cells. A: Nuclear and nucleoar abnormalities. A higher power view of an E2-induced Cav-1^−/− mammary lesion is shown (Trichrome staining). Note that the mammary epithelial cells within the lesion show nuclear atypia and nuclear heterogeneity. A central mitotic figure is also present (arrow). Nuclei often appear pale, with prominent nucleolar staining. These nuclear changes are thought to be associated with mammary cell transformation. These nuclear changes were not observed in wild-type animals treated identically. B: Altered myoepithelial distribution. Paraffin-sections from ovariectomized and E2-treated mice were immunostained with anti-α-smooth muscle actin IgG to visualize the distribution of myoepithelial cells. Bound antibodies were detected with an HRP-conjugated secondary antibody (brown color). Samples were lightly counterstained with hematoxylin (blue color). Note that in wild-type ducts, the myoepithelial cells appear normal and form a continuous layer surrounding the ducts (Left). In contrast, in Cav-1^−/− ducts, the myoepithelial cells appear in a discontinuous distribution (arrow), and are often mislocalized in the center of the lesions. These changes are suggestive of a potentially invasive phenotype.

Cav-1^−/− mice develop abnormal mammary lesions (dysplastic foci) in response to estrogen. When the mammary whole mounts were analyzed following treatment of Cav-1^−/− mice with a higher (H) dose of E2 (15.0 mg slowly delivered over a 60-day period), the presence of foci was detected. Right, Representative high power images of Cav-1^−/− mammary lesions are shown stained with Carmine dye. In the boxed area, shown at higher magnification, arrows point at mammary lesions (dysplastic foci). Quantitation of mammary lesion frequency (upper left) and size (lower left) was performed using NIH image J software. Although the wild-type mice show some minor increases in mammary lesion formation, following 17-β-estradiol treatment (E2), Cav-1^−/− mice are markedly hyper-responsive to 17-β-estradiol treatment (E2). Cav-1^−/− mice show a greater than fourfold increase in lesion frequency and an ∼2.5-fold increase in lesion diameter. These differences are highly statistically significant. ^†P < 0.01 when compared with wild-type treated with E2; ^‡P < 0.01 when compared with placebo-treated Cav-1^−/− mice. WT, Cav-1^+/+; KO, Cav-1^−/− (n = 7 to 10 for each group). Inset magnification: 10×.

Cav-1^−/− mammary glands are hyper-responsive to the effects of estrogen. As outlined in Figure 2, wild-type and Cav-1^−/− mice were ovariectomized and subjected to estrogen supplementation. After 60 days of placebo (OVX) or E2-treatment (OVX + E2), carmine dye staining of inguinal (no. 4) mammary gland was performed. Then, quantitation of ductal thickness (A) and branching (B) was performed using NIH image J software. The number of primary, secondary and tertiary branch points in the inguinal whole mounts was assessed. Primary branches are defined as ducts that arise from the nipple and extend throughout the gland, secondary branches are ducts that arise from primary ducts, and the tertiary branches are lateral branches that arise from the secondary ducts. Results with both low (L) and higher (H) dose of E2 are shown. Note that at both lower and higher doses of estrogen, Cav-1^−/− mice show an approximately twofold increase in ductal thickening, as compared with E2-treated wild-type mice. Similarly, Cav-1^−/− mammary glands also show greater susceptibility to estrogen induced branching morphogenesis, with approximately three- to fourfold increases in secondary branching and five to sevenfold increases in tertiary branching. For (A), *P < 0.01 when compared with wild-type treated with placebo; ^†P < 0.01 when compared with wild-type treated with E2; and ^‡P < 0.01 when compared with placebo-treated Cav-1^−/− mice (n = 7 to 10 for each group). For (B), *P < 0.001 when compared with wild-type treated with placebo; ^†P < 0.01 when compared with wild-type treated with E2; ^‡P < 0.01 when compared with placebo-treated Cav-1^−/− mice; and ^§P < 0.01 when compared with E2 (L; low dosage) treated Cav-1^−/− mice (n = 7 to 10 for each group).

Experimental approach for determining the estrogen-dependent phenotypes of Cav-1^−/− mammary glands. A: Schematic diagram. Female wild-type and Cav-1^−/− mice were subjected to bilateral ovariectomy (OVX) at 5 weeks of age and allowed to recover for 2 weeks. Then, the mice were randomly given either estrogen or a placebo for 8 consecutive weeks (60 days) via the implantation of a slow-release pellet. Two dosages of 17-β-estradiol (E2) were evaluated in parallel using 7.5 mg (L; low; 125 μg/day) and 15.0 mg (H; high; 250 μg/day) pellets. At the end of the 60-day treatment period, the mice were sacrificed and the morphology of the mammary gland was assessed, by whole mount analysis and using paraffin-embedded tissue sections. In all cases, 7 to 10 mice were used for each treatment group and for each genotype. WT, Cav-1^+/+; KO, Cav-1^−/−. B: Whole mount analysis. Note that wild-type mice treated with placebo have few branches with narrow ducts. Following the treatment with 17-β-estradiol, the wild-type mammary glands do not seem very different, with the exception of twofold more primary branches. Note that Cav-1^−/− mice treated with placebo resemble the wild-type mammary gland. In striking contrast, Cav-1^−/− mice are markedly more responsive to 17-β-estradiol (E2) treatment, as shown by increases in ductal thickening and secondary and tertiary branching. Representative images of mammary glands from mice treated with low (L) dose E2 are shown. Images were taken with a ×1.5 objective. The boxed area is shown at higher magnification to illustrate the increased branching (see arrowhead) observed in E2-treated Cav-1^−/− mammary glands. Inset magnification: 3×.