Direct proliferative effects of estrogen (E2) on estrogen receptor–positive tumors are well documented; however, the potential for E2 to mediate effects selective for the host (i.e., angiogenesis, vascular permeability, or stromal effects), which influence tumor growth and/or metastasis, has received less attention. In this study, we examine the capacity for E2 to promote tumor growth and/or metastasis independent of direct effects on tumor cells. In these studies, we distinguish host versus tumor compartment components of E2 action in tumor growth and metastasis by analysis of E2-nonresponsive tumor cells implanted in ovariectomized (OVX) mice that contain s.c. implants of placebo (OVX) or E2-containing slow-release pellets (OVX + E2). We show that the D121 lung carcinoma cell line is E2-nonresponsive, and following s.c. implantation in OVX versus OVX + E2 mice, E2 action on the host compartment leads to an increase in spontaneous metastasis but not primary tumor growth or neovascularization. Similarly, experimental lung metastasis of E2-nonresponsive 4T1 mammary carcinoma cells also leads to increased tumor burden in the lungs of OVX + E2 mice. These results suggest that the E2 status of the host compartment influences late steps in tumor cell metastasis that can provide important insights into the role of E2 in the tumor versus host compartments. (Cancer Res 2006; 66(7): 3667-72)

Estrogen signaling has been associated classically with changes in transcription mediated by translocation of dimerized estrogen receptors (ERα) to the nucleus, whereas transcription-independent pathways of estrogen action involve rapid, transient signaling from the cell membrane. Estrogen action promotes cell proliferation, suppresses apoptosis, and, therefore, has been an important target in the diagnosis and treatment of estrogen-dependent breast tumors. However, estrogen-nonresponsive tumor cells are associated with more advanced breast cancer progression, an important variable in tailoring endocrine treatments in a clinical setting (13). Circulating levels of 17β-estradiol (E2) have been shown to affect tumor growth; however, the role of circulating E2 in modulating mechanisms of metastasis and signaling in the host compartment has received less attention (4, 5). For example, E2 mediates vasodilation, angiogenesis, anti-inflammatory responses and attenuates atherosclerosis (68) as well as having an essential role in the female reproductive tract. E2 also regulates remodeling of tumor-associated blood vessels and can engage in signaling crosstalk with other growth factors and their receptors (912). The vascular bed of the lung is of particular interest in these studies as it is the primary organ site of breast cancer metastases in humans and in experimental animals. Furthermore, pulmonary endothelial cells express ERs (13). These reports suggest that E2 functions in tumor-induced angiogenesis and in the maintenance of the vasculature, supporting a hypothesis that E2 mediates critical vascular responses in the host compartment, independent of E2 action, ER activation, and growth promotion within tumor cells themselves.

In this study, we examine spontaneous tumor metastasis in a mouse model in which E2-nonresponsive tumor cells are injected into ovariectomized (OVX) mice implanted with placebo or slow-release E2 pellets. This strategy facilitates the analysis of E2 action on metastasis through host compartment–specific responses rather than tumor cells themselves. Following an analysis of primary tumors and spontaneous lung metastases, experimental metastasis studies were done to examine the capacity for E2 to influence the survival and extravasation of i.v. injected tumor cells into the lung stroma. These studies reveal that although primary growth of E2-nonresponsive tumor cells is unaffected by changes in the E2 status of the host, both spontaneous and experimental metastases of E2-nonresponsive tumor cells are affected by the E2 status of the host compartment.

Immunoblotting. Cell lysates for immunoblotting were collected in radioimmunoprecipitation assay buffer [100 mmol/L Tris (pH 7.5), 150 mmol/L NaCl, 1 mmol/L EDTA, 1% deoxycholic acid, 1% Triton X-100, 0.1% SDS with freshly added 1 mmol/L orthovanadate, 50 mmol/L NaF, and protease inhibitors cocktail; Roche, Indianapolis, IN]. Protein concentrations were determined by the Pierce Bicinchoninic Acid Protein Assay (Rockford, IL). Twenty-five micrograms of each cell lysate sample were separated on a 10% reducing SDS-polyacrylamide gel, transferred to nitrocellulose membrane, and probed with anti-ERα polyclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA). In addition, blots were probed with an anti-phospho-extracellular signal-regulated kinase (phospho-p44/42 MAPK) monoclonal antibody (Cell Signaling Technology, Beverly, MA). Secondary antibodies were anti-rabbit or anti-mouse conjugated to horseradish peroxidase. Visualization of antibody binding was detected using Supersignal West Pico Chemiluminescent Substrate (Pierce).

Immunofluorescence. Indirect immunofluorescence was done on cryosections (10 μm) of tumor samples using anti-ERα monoclonal antibody (Affinity BioReagents, Golden, CO) or anti-CD31 monoclonal antibody (Becton Dickinson, Franklin Lakes, NJ). Secondary antibodies were from The Jackson Laboratory (West Grove, PA) or Molecular Probes (Eugene, OR). Micrographs were captured on a Fluoview 1000 laser-scanning confocal microscope (Olympus with an Olympus BX61 microscope equipped with a ×20/0.7 dry objective lens and Fluoview acquisition software; Olympus, Tokyo, Japan). Tumor sections labeled with dual fluorescent dyes (fluorescein and rhodamine) were imaged with the confocal microscope, and the two channels merged in the Olympus Fluoview software.

Mice. Bilateral oophorectomy (OVX) was done on 35- to 50-day-old C57/BL6 mice for D121 injection and BALB/c mice for 4T1 mammary fat pad injection. Surgeries were done as previously described (14) under modified aseptic conditions using appropriate anesthesia and analgesia. All procedures were approved by the La Jolla Institute for Molecular Medicine Institutional Animal Care and Use Committee. Steroid (E2, 0.36 mg, 90-day release) or placebo pellets (Innovative Research of America, Sarasota, FL) were inserted s.c. through the surgical incision by tunneling under the skin overlying the flank.

Tumor studies. For primary tumor studies, a D121 tumor cell line, the generous gift of L. Eisenbach (Weizman Institute, Rehovot, Israel), was used for tumor implantations in OVX mice with or without estrogen added back. These tumor cells are a subline of the Lewis lung carcinoma and have been selected based on their capacity to metastasize to the lung and form aggressive lung tumors. A murine, mammary carcinoma cell line 4T1 was used for experimental tumor studies to the lung. To identify estrogen-independent cell lines, D121 and 4T1 cells were cultured in phenol red–free media with charcoal-stripped serum (DMEM and RPMI, respectively), with estrogen (E2) added to cultures at 10−10 to 10−8 concentrations. Cells were harvested at 5 minutes for Erk1/2 activation, at 36 hours for ERα identification, and daily for 3 to 4 days for cell proliferation. For the spontaneous lung metastasis assays, 1 × 106 D121 tumor cells were injected s.c. into the flank of C57BL/6 mice, and 1 × 105 4T1 tumor cells were injected i.v. into BALB/c mice. For D121 tumors, cells were injected s.c. and incubated for 14 days, the primary tumor was resected, the wet weight of the primary tumor was determined, and the mice were incubated for an additional 16 days, when the mice were sacrificed and the lungs examined. For D121 tumors, mice were incubated for 20 days following i.v. injection. Tissue samples for subsequent immunostaining were collected for embedding and cryosectioning as described previously (15). Tumor burden was examined with an Olympus SZX12 stereomicroscope (Olympus, Melville, NY) and measured by total wet lung weights. For the experimental tumor metastasis, YFP-labeled 4T1 cells were injected into the tail vein of OVX + E2 and OVX-E2 mice. Twelve days following injection, the lungs were removed and dissociated, and the cells analyzed by fluorescence-activated cell sorting for the presence of YFP cells.

Characterization of a murine lung carcinoma cell line that is estrogen nonresponsive in vitro and in vivo. To isolate the effects of E2 on the host compartment, E2-nonresponsive (D121 and 4T1) and E2-responsive (MCF-7) cell lines were examined for expression of ERα and E2-mediated signaling. Figure 1A shows that immunoblotting to examine the expression of endogenous ERα revealed that ERα was not detected in D121 Lewis lung carcinoma or 4T1 breast carcinoma cells. Immunoblotting of uterus tissue lysates was done as a positive control for the anti-ERα antibody and showed high levels of ERα expression (Fig. 1A). The activation of Erk1/2, a signaling intermediate associated with a wide range of stimuli, was used to evaluate potential non–ER-mediated signaling. Lysates of D121 tumor cells that had been incubated in phenol red–free media with charcoal-stripped serum were stimulated with E2 and subjected to immunoblotting with a phospho-specific antibody recognizing p-Erk1/2, a mitogen-activated protein kinase associated with ER-mediated proliferation, and non–ER-mediated cellular responses (Fig. 1B). E2 treatment of D121 cells resulted in no significant change in Erk1/2 phosphorylation after 5 minutes (Fig. 1B) or 24 hours (data not shown). In addition to Erk1/2 phosphorylation, potential E2-induced cell proliferation was also examined in D121 tumor cells as well as in MCF-7 cells as a positive control, a cell line known to be ERα positive and E2 responsive. E2-treated D121 cells did not undergo an increase in proliferation (Fig. 1C), whereas E2 stimulation of MCF-7 tumor cells led to increased cell proliferation (Fig. 1D). Similar to the E2-induced proliferation of MCF-7 tumor cells, E2 treatment also induced the phosphorylation of Erk1/2 (Fig. 1D , inset), in contrast to the absence of E2-induced Erk1/2 phosphorylation in D121 tumor cells. In combination, these assays indicate that D121 tumor cells are E2 nonresponsive and provide the basis for further analysis in vivo.

Figure 1.

E2 response status of D121 cells. A, whole tissue lysates of mouse uterus and cultured D121 and 4T1 tumor cells were subjected to SDS-PAGE and immunoblotting with an anti-ERα antibody. The predominant 66-kDa band observed in the uterus as a positive control was absent from the D121 and 4T1 tumor cells. B, D121 tumor cells were incubated in the presence or absence of E2 (10−8 and 10−9 mol/L) for 5 minutes (shown) or 24 hours (data not shown) and subjected to immunoblotting with a phospho-specific anti-p-Erk1/2 antibody, resulting in no change in Erk1/2 phosphorylation. C, the effect of E2 on D121 tumor cell proliferation was determined using an 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay (MTT) on lysates of cells cultured in the presence or absence of 10−9 M E2. The combination of these assays is the basis for the characterization of D121 tumor cells as exhibiting an E2-nonresponsive phenotype. D, the E2-responsive breast carcinoma cell line MCF-7 was subjected to a proliferation assay and immunoblotting with an anti-p-Erk1/2 in the presence or absence of E2 (10−9 mol/L; inset). In contrast to the E2-nonresponsive phenotype of D121 tumor cells, E2 stimulation of MCF-7 tumor cells induces cell proliferation and Erk1/2 phosphorylation.

Figure 1.

E2 response status of D121 cells. A, whole tissue lysates of mouse uterus and cultured D121 and 4T1 tumor cells were subjected to SDS-PAGE and immunoblotting with an anti-ERα antibody. The predominant 66-kDa band observed in the uterus as a positive control was absent from the D121 and 4T1 tumor cells. B, D121 tumor cells were incubated in the presence or absence of E2 (10−8 and 10−9 mol/L) for 5 minutes (shown) or 24 hours (data not shown) and subjected to immunoblotting with a phospho-specific anti-p-Erk1/2 antibody, resulting in no change in Erk1/2 phosphorylation. C, the effect of E2 on D121 tumor cell proliferation was determined using an 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay (MTT) on lysates of cells cultured in the presence or absence of 10−9 M E2. The combination of these assays is the basis for the characterization of D121 tumor cells as exhibiting an E2-nonresponsive phenotype. D, the E2-responsive breast carcinoma cell line MCF-7 was subjected to a proliferation assay and immunoblotting with an anti-p-Erk1/2 in the presence or absence of E2 (10−9 mol/L; inset). In contrast to the E2-nonresponsive phenotype of D121 tumor cells, E2 stimulation of MCF-7 tumor cells induces cell proliferation and Erk1/2 phosphorylation.

Close modal

Characterization of ERα expression in vivo. To determine whether s.c. implantation of ER-nonresponsive D121 cells into mice would influence the expression of ERα in D121 cells in vivo, we did immunohistochemistry with an anti-ERα antibody on cryosections of D121 tumor-bearing skin from the flank of an OVX + E2 mouse. Indirect immunofluorescence revealed that the primary tumor was negative for ERα expression in vivo, whereas the overlying skin (i.e., host compartment) was ERα positive (16). Figure 2A shows the cellular distribution of ERα in the overlying skin of the tumor at low magnification. Figure 2B shows the expression of ERα in the skin versus lower expression of ERα in host-derived infiltrating cells/blood vessels. In combination with our in vitro data, these results indicate that D121 tumor cells are E2 nonresponsive in vitro and in vivo and provide the basis for using D121 cells in the analysis of host compartment–mediated mechanisms of E2 action on tumor growth and metastasis.

Figure 2.

ERα expression in tumor and host compartments. A, the E2-nonresponsive lung carcinoma cells were injected s.c. in the skin of C57Bl/6 mice and incubated for 14 days. Cryosections of tumor tissue were prepared and subjected to immunostaining with an anti-ERα antibody, fluorescent labeling with an Alexa 488–conjugated secondary antibody, and detection with an Olympus laser scanning confocal microscope. Tissue sections were also stained with 4′,6-diamidino-2-phenylindole (DAPI), a nuclear stain. Low-magnification micrographs reveal the expression and localization of ERα protein primarily in the skin overlying the tumor. Bar, 100 μm. B, higher-magnification micrographs of the skin and the tumor core reveal ERα protein expression in the skin and lower levels of ERα protein in the tumor. Bar, 200 μm.

Figure 2.

ERα expression in tumor and host compartments. A, the E2-nonresponsive lung carcinoma cells were injected s.c. in the skin of C57Bl/6 mice and incubated for 14 days. Cryosections of tumor tissue were prepared and subjected to immunostaining with an anti-ERα antibody, fluorescent labeling with an Alexa 488–conjugated secondary antibody, and detection with an Olympus laser scanning confocal microscope. Tissue sections were also stained with 4′,6-diamidino-2-phenylindole (DAPI), a nuclear stain. Low-magnification micrographs reveal the expression and localization of ERα protein primarily in the skin overlying the tumor. Bar, 100 μm. B, higher-magnification micrographs of the skin and the tumor core reveal ERα protein expression in the skin and lower levels of ERα protein in the tumor. Bar, 200 μm.

Close modal

E2levels do not affect primary tumor weights or neovascularization. Previous studies have shown that E2 enhances cell proliferation, migration, and invasion of breast tumor cells that express ERs; effects of E2 on host blood vessels that may influence tumor progression through tumor cell–independent mechanisms are less well studied. Therefore, we examined the primary tumor growth variables of E2-nonresponsive D121 tumor cells in syngeneic (C57BL/6) OVX and OVX + E2 mice. In previous studies, we have found that slow-release E2 pellets yield a plasma concentration of 2 × 10−9 mol/L or 574 pg/mL, as determined by RIA (17). Here, we examined the uteri of OVX and OVX + E2 mice for changes in uterine weight, a physiologic measure of E2 status of the host (Fig. 3A). After a 30-day exposure, uterine weights were >10-fold greater (P < 0.001, n = 10) in OVX + E2 versus OVX mice (Fig. 3A). To determine the effect of the E2 status of the host on tumor progression, D121 E2-nonresponsive lung carcinoma cells were implanted by s.c. injection in OVX or OVX + E2 mice and incubated for 14 days. Primary tumors were then resected under modified aseptic conditions, and the mice were housed for an additional 16 days to monitor spontaneous lung metastasis. Analyses of primary tumors from OVX and OVX + E2 mice revealed that there was no significant difference in primary tumor weight between the two mouse groups (n = 10; Fig. 3B). Tumor-induced angiogenesis was measured by CD31 immunostaining of cryosections of primary tumor tissue and revealed a similar level of neovascularization in D121 tumors from OVX and OVX + E2 hosts (n = 10; Fig. 3C).

Figure 3.

Effect of E2 on uterine weight, primary tumor weight, and angiogenesis of an E2-nonresponsive tumor. A, uteri of OVX and OVX + E2 mice were removed and weighed after a 30-day incubation. OVX mice implanted with a placebo pellet and OVX + E2 mice with a slow release E2 pellet. B, 1 × 106 E2-nonresponsive D121 tumor cells were injected s.c. into syngeneic OVX and OVX + E2 C57Bl/6 mice and incubated for 14 days. Lungs were removed and weighed, revealing no significant change in wet lung weight following E2 treatment. C, analysis of CD31-positive endothelial cells in cryosections of primary D121 tumors from OVX versus OVX + E2 mice reveals no significant difference in blood vessel density. Bar, 100 μm.

Figure 3.

Effect of E2 on uterine weight, primary tumor weight, and angiogenesis of an E2-nonresponsive tumor. A, uteri of OVX and OVX + E2 mice were removed and weighed after a 30-day incubation. OVX mice implanted with a placebo pellet and OVX + E2 mice with a slow release E2 pellet. B, 1 × 106 E2-nonresponsive D121 tumor cells were injected s.c. into syngeneic OVX and OVX + E2 C57Bl/6 mice and incubated for 14 days. Lungs were removed and weighed, revealing no significant change in wet lung weight following E2 treatment. C, analysis of CD31-positive endothelial cells in cryosections of primary D121 tumors from OVX versus OVX + E2 mice reveals no significant difference in blood vessel density. Bar, 100 μm.

Close modal

Estrogen (E2) levels affect spontaneous lung metastases of E2 nonresponsive lung carcinoma tumor cells. The influence of the E2 status of the host during spontaneous metastasis of E2-nonresponsive D121 cells was examined by analyzing the lungs of the mice described above 16 days after primary tumor removal. Tumor-bearing OVX + E2 mice exhibited a significant increase in lung tumor nodes and lung wet weight compared with tumor-bearing OVX mice (P < 0.05, n = 10; Fig. 4). These findings suggest that the E2 status of the host enhances the tumor metastasis but not the primary tumor growth of an E2-nonresponsive lung carcinoma. We have previously shown that a combination of spontaneous and experimental metastasis assays can be used to distinguish potential host-mediated mechanisms of metastasis (i.e., tumor extravasation from the primary site versus tumor intravasation from the circulation; ref. 15). Therefore, we next examined experimental metastasis in OVX versus OVX + E2 mice.

Figure 4.

E2 status of host influences spontaneous lung metastasis. A, following the removal of subcutaneous D121 primary tumors from OVX and OVX + E2 mice after 14 days, mice were incubated for an additional 16 days and the lungs removed for analysis. Stereomicrographs of intact lungs reveal an increase in tumor burden in OVX + E2 versus OVX mice. B, analysis of wet lung weights from OVX and OVX + E2 mice reveals an increase in lung weight corresponding to the increase in tumor nodes observed (P < 0.003, n = 10).

Figure 4.

E2 status of host influences spontaneous lung metastasis. A, following the removal of subcutaneous D121 primary tumors from OVX and OVX + E2 mice after 14 days, mice were incubated for an additional 16 days and the lungs removed for analysis. Stereomicrographs of intact lungs reveal an increase in tumor burden in OVX + E2 versus OVX mice. B, analysis of wet lung weights from OVX and OVX + E2 mice reveals an increase in lung weight corresponding to the increase in tumor nodes observed (P < 0.003, n = 10).

Close modal

Reduction in experimental metastasis of an E2-nonresponsive mammary carcinoma. To isolate the role of E2 in the regulation of tumor arrest and/or extravasation in the lung, the murine mammary carcinoma cell line, 4T1, was selected. The E2 response phenotype of 4T1 cells was determined by proliferation assays and Erk1/2 phosphorylation analysis as described for the D121 cells. We observed that E2 did not induce proliferation (Fig. 5A) or Erk1/2 phosphorylation (Fig. 5B) of 4T1 cells. Therefore, like D121 cells, the 4T1 cells were E2 nonresponsive. For the tumor studies, experimental metastasis assays were done with 4T1 cells by direct injection into the tail vein. To enhance the sensitivity and quantitation of lung tumor burden, we have optimized a flow cytometry–based strategy for the detection of fluorescently labeled tumor cells in mouse lung tissue. 4T1 cells were stably transduced with a lentiviral vector expressing yellow fluorescent protein (4T1-YFP). 4T1 cells are syngeneic with BALB/c mice. Therefore, BALB/c females were subjected to OVX surgery as described above for the C57Bl/6 mice followed by the implantation of placebo or slow-release E2 pellets. The E2 pellets led to an increase in uterine weight similar to the data shown for the C57Bl/6 OVX + E2 mice in Fig. 1. For the 4T1-YFP lung carcinoma experimental metastasis assay, 1 × 105 cells were injected i.v. into OVX and OVX + E2 mice followed by a 20-day incubation. To quantitate 4T1-YFP tumor burden in the OVX and OVX + E2 mice, tumor-bearing lungs from mice were removed, dissociated by limited collagenase digestion and subjected to flow cytometry to detect the YFP fluorescence of the 4T1-YFP tumor cells. Flow cytometric analysis of 4T1-YFP tumor cells in this model revealed a significant increase in lung tumor burden in OVX + E2 mice versus OVX control mice (P < 0.05, n = 8; Fig. 5C and D). The E2-mediated increase in experimental metastasis of E2-nonresponsive tumor cells suggests that the primary mechanism(s) of action of the E2 response is associated with E2-induced changes in the host compartment that affects tumor cell survival, arrest, and/or invasion in the lung.

Figure 5.

Analysis of E2-induced experimental lung metastasis of E2-nonresponsive 4T1 breast carcinoma cells. A, E2-induced proliferation of 4T1 cells was analyzed using an 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay. B, the capacity for E2 to induce Erk1/2 phosphorylation was examined by immunoblotting with a phospho-specific anti-Erk1/2 antibody, and an anti-Erk1/2 antibody recognizing total protein. These studies reveal that E2 treatment did not induce cell proliferation or Erk1/2 phosphorylation. C, experimental metastasis of E2-nonresponsive 4T1 cells was done by i.v. injection of 4T1-YFP cells (105) into OVX and OVX + E2 mice. After a 20-day incubation, lungs were removed and subjected to flow cytometric analysis. D, lung tumor burden was determined by flow cytometry from several mice and analysis showed a significant increase in lung tumor burden in OVX + E2 mice in experimental metastasis (P < 0.05, n = 8).

Figure 5.

Analysis of E2-induced experimental lung metastasis of E2-nonresponsive 4T1 breast carcinoma cells. A, E2-induced proliferation of 4T1 cells was analyzed using an 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay. B, the capacity for E2 to induce Erk1/2 phosphorylation was examined by immunoblotting with a phospho-specific anti-Erk1/2 antibody, and an anti-Erk1/2 antibody recognizing total protein. These studies reveal that E2 treatment did not induce cell proliferation or Erk1/2 phosphorylation. C, experimental metastasis of E2-nonresponsive 4T1 cells was done by i.v. injection of 4T1-YFP cells (105) into OVX and OVX + E2 mice. After a 20-day incubation, lungs were removed and subjected to flow cytometric analysis. D, lung tumor burden was determined by flow cytometry from several mice and analysis showed a significant increase in lung tumor burden in OVX + E2 mice in experimental metastasis (P < 0.05, n = 8).

Close modal

Direct effects of E2 on breast cancer cells have been associated with both ER-dependent and ER-independent signaling mechanisms in tumor cells. Although progression of breast cancer cells to an E2-nonresponsive status correlates with a poor prognosis, we examined whether E2 exerts actions on the tumor microenvironment (i.e., host compartment angiogenic blood vessels and stroma) independent of the ER status of the tumor cells. To determine the role of E2 in the host compartment, independent of direct effects of E2 action on tumor cells, we have examined the primary tumor growth and lung metastasis of E2-nonresponsive tumor cells in OVX versus OVX + E2 mice. We have characterized two murine tumor cell lines (D121 and 4T1) for the expression of ERα by immunoblotting, for E2-induced cell proliferation, and for E2-induced Erk1/2 phosphorylation. Based on the negative E2 responses of each cell line in the variables tested, we defined these cells to be E2 nonresponsive. Analyses of tumor growth following implantation of these E2-nonresponsive tumor cells in OVX versus OVX + E2 mice reveals that although primary tumor growth is unaffected by the presence of estrogen, spontaneous and experimental metastases are increased in OVX + E2 mice versus OVX mice, suggesting a host-mediated role for estrogen in tumor cell metastasis.

The extent of metastasis measured in spontaneous versus experimental metastasis assays can provide insights into mechanisms of host compartment–mediated influence on metastasis (15, 1823). For example, in models where primary tumor growth is unchanged between treatment groups, i.v. administration can distinguish between early (i.e., intravasation) and late (i.e., extravasation, arrest, or survival) steps in the metastatic process. In this study, we observed that primary tumor growth was unchanged in OVX versus OVX + E2 mice, whereas spontaneous lung metastasis was increased in OVX + E2 mice versus OVX mice. Furthermore, direct i.v. administration of tumor cells (i.e., experimental metastasis) resulted in a similar increase in lung tumor burden in OVX + E2 versus OVX mice. These findings suggest that E2 action on the host compartment influences the capacity for tumor cells to arrest and/or extravasate from the lung vasculature, rather than influencing invasiveness or intravasation into the circulation from the primary tumor. Although these studies do not address possible effects of E2 on tumor cell survival or immune responses to metastatic cells, E2 mediates additional vascular responses (i.e., vasodilation and angiogenesis in the uterus) that may influence tumor-associated blood vessels (24).

Although we have observed no direct effect of E2 on the neovascular density of primary tumors in the skin of OVX versus OVX + E2 mice, previous studies have shown that E2 can induce vascular endothelial growth factor (VEGF) expression (25). Others indicate that E2 can negatively regulate soluble VEGF receptor 1, a VEGF receptor thought to function as a decoy receptor that neutralizes circulating VEGF (24), whereas VEGF expression in tumor cells can synergize with E2 and override E2-dependent tumor growth (12). Other growth factors, such as basic fibroblast growth factor, depend on E2 action, whereas the capacity for E2 to mediate the expression of other growth factors and cytokines in otherwise E2-nonresponsive tumor cells is poorly understood. The capacity for E2 to regulate other female sex steroids, such as progestins, is also hypothesized to influence host compartment vascular remodeling but was not evaluated in this study (2527). In fact, there should be little progesterone in these mice following ovariectomy. Furthermore, we propose that E2 action and specific E2 signaling pathways will be distinct in E2-responsive tumor cells and the host vascular bed and stroma.

Clinical studies of hormone replacement therapy (HRT) indicate that whereas tumor growth and metastasis can occur in postmenopausal women in the presence of minimal endogenous E2 production, the levels and timing of estrogen replacement therapy can have contrasting effects on the growth or inhibition of various cancers. For example, in the Women's Health Initiative studies, women receiving estrogen/progestin replacement had a significant reduction in colon cancer. However, of the women receiving HRT that did develop tumors, most were diagnosed with a greater number of positive lymph nodes, suggesting a role for estrogen in metastasis (28). In contrast, studies have shown that women receiving HRT have an increased risk of lung adenocarcinoma (29). Correspondingly, anti-estrogen therapy has shown promise in the treatment of non–small cell lung cancer in women (30). Although the effects of E2 on the proliferation of E2-responsive tumor cells have been well documented, E2 may also have a role in ER-negative cancer cell metastasis through nongenomic estrogen actions on tumor cells or indirectly through influences on the host stroma or vasculature. Our data suggest that E2 enhances metastasis through actions on the host and not the tumor cells.

In combination, our studies suggest that further analysis of the specific role(s) of E2 in mediating host versus tumor compartment responses can provide important insights into the mechanisms of E2-dependent versus E2-independent tumor growth and metastasis. A better understanding of these mechanisms may ultimately provide a molecular basis for the tailoring of patient-specific therapies based on the E2 status of the tumor and the specific contribution of host versus tumor compartment signaling responses.

Grant support: Department of Defense Breast Cancer Research Program, California Breast Cancer Research Program (11IB-0079), and National Heart, Lung, and Blood Institute (ROIHL 73396).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Virginia (Jill) Patch, Denny Le, and Willam Kang for technical assistance and Geoff Owens for the fluorescent labeling of tumor cells.

1
Neven P, Vergote I, Van Ginderachter J, et al. Endocrine treatment and prevention of breast and gynaecological cancers.
Eur J Cancer
2002
;
38
Suppl 6:
S1
–11.
2
Thorpe SM. Monoclonal antibody technique for detection of estrogen receptors in human breast cancer: greater sensitivity and more accurate classification of receptor status than the dextran-coated charcoal method.
Cancer Res
1987
;
47
:
6572
–5.
3
Vergote I, Neven P, van Dam P, et al. The oestrogen receptor and its selective modulators in gynaecological and breast cancer.
Eur J Cancer
2000
;
36
Suppl 4:
S1
–9.
4
Cooper LS, Gillett CE, Patel NK, Barnes DM, Fentiman IS. Survival of premenopausal breast carcinoma patients in relation to menstrual cycle timing of surgery and estrogen receptor/progesterone receptor status of the primary tumor.
Cancer
1999
;
86
:
2053
–8.
5
Wood PA, Bove K, You S, Chambers A, Hrushesky WJ. Cancer growth and spread are saltatory and phase-locked to the reproductive cycle through mediators of angiogenesis.
Mol Cancer Ther
2005
;
4
:
1065
–75.
6
Hisamoto K, Bender JR. Vascular cell signaling by membrane estrogen receptors.
Steroids
2005
;
70
:
382
–7.
7
Simoncini T, Fornari L, Mannella P, et al. Novel non-transcriptional mechanisms for estrogen receptor signaling in the cardiovascular system. Interaction of estrogen receptor alpha with phosphatidylinositol 3-OH kinase.
Steroids
2002
;
67
:
935
–9.
8
Shaul PW. Novel role of estrogen receptors in vascular endothelium.
Semin Perinatol
2000
;
24
:
70
–4.
9
Morales DE, McGowan KA, Grant DS, et al. Estrogen promotes angiogenic activity in human umbilical vein endothelial cells in vitro and in a murine model.
Circulation
1995
;
91
:
755
–63.
10
Soares R, Guo S, Gartner F, Schmitt FC, Russo J. 17 beta-estradiol-mediated vessel assembly and stabilization in tumor angiogenesis requires TGF beta and EGFR crosstalk.
Angiogenesis
2003
;
6
:
271
–81.
11
Bogin L, Degani H. Hormonal regulation of VEGF in orthotopic MCF7 human breast cancer.
Cancer Res
2002
;
62
:
1948
–51.
12
Guo P, Fang Q, Tao HQ, et al. Overexpression of vascular endothelial growth factor by MCF-7 breast cancer cells promotes estrogen-independent tumor growth in vivo.
Cancer Res
2003
;
63
:
4684
–91.
13
MacRitchie AN, Jun SS, Chen Z, et al. Estrogen upregulates endothelial nitric oxide synthase gene expression in fetal pulmonary artery endothelium.
Circ Res
1997
;
81
:
355
–62.
14
Marsh MM, Walker VR, Curtiss LK, Banka CL. Protection against atherosclerosis by estrogen is independent of plasma cholesterol levels in LDL receptor-deficient mice.
J Lipid Res
1999
;
40
:
893
–900.
15
Criscuoli ML, Nguyen M, Eliceiri BP. Tumor metastasis but not tumor growth is dependent on Src-mediated vascular permeability.
Blood
2005
;
105
:
1508
–14.
16
Moverare S, Lindberg MK, Faergemann J, Gustafsson JA, Ohlsson C. Estrogen receptor alpha, but not estrogen receptor beta, is involved in the regulation of the hair follicle cycling as well as the thickness of epidermis in male mice.
J Invest Dermatol
2002
;
119
:
1053
–8.
17
Mayer LP, Dyer CA, Eastgard RL, Hoyer PB, Banka CL. Atherosclerotic lesion development in a novel ovary-intact mouse model of perimenopause.
Arterioscler Thromb Vasc Biol
2005
;
25
:
1910
–6.
18
Chambers AF, Groom AC, MacDonald IC. Dissemination and growth of cancer cells in metastatic sites.
Nat Rev Cancer
2002
;
2
:
563
–72.
19
Fidler IJ. Orthotopic implantation of human colon carcinomas into nude mice provides a valuable model for the biology and therapy of metastasis.
Cancer Metastasis Rev
1991
;
10
:
229
–43.
20
Freije JM, MacDonald NJ, Steeg PS. Nm23 and tumour metastasis: basic and translational advances.
Biochem Soc Symp
1998
;
63
:
261
–71.
21
Paget S. The distribution of secondary growths in cancer of the breast. 1889.
Cancer Metastasis Rev
1989
;
8
:
98
–101.
22
Weiss L. Comments on hematogenous metastatic patterns in humans as revealed by autopsy.
Clin Exp Metastasis
1992
;
10
:
191
–9.
23
Welch DR. Technical considerations for studying cancer metastasis in vivo.
Clin Exp Metastasis
1997
;
15
:
272
–306.
24
Elkin M, Orgel A, Kleinman HK. An angiogenic switch in breast cancer involves estrogen and soluble vascular endothelial growth factor receptor 1.
J Natl Cancer Inst
2004
;
96
:
875
–8.
25
Hyder SM, Huang JC, Nawaz Z, et al. Regulation of vascular endothelial growth factor expression by estrogens and progestins.
Environ Health Perspect
2000
;
108
Suppl 5:
785
–90.
26
Hyder SM, Chiappetta C, Stancel GM. Pharmacological and endogenous progestins induce vascular endothelial growth factor expression in human breast cancer cells.
Int J Cancer
2001
;
92
:
469
–73.
27
Wu J, Richer J, Horwitz KB, Hyder SM. Progestin-dependent induction of vascular endothelial growth factor in human breast cancer cells: preferential regulation by progesterone receptor B.
Cancer Res
2004
;
64
:
2238
–44.
28
Chlebowski RT, Wactawski-Wende J, Ritenbaugh C, et al. Estrogen plus progestin and colorectal cancer in postmenopausal women.
N Engl J Med
2004
;
350
:
991
–1004.
29
Caracta CF. Gender differences in pulmonary disease.
Mt Sinai J Med
2003
;
70
:
215
–24.
30
Chen YM, Perng RP, Yang KY, et al. A phase II trial of tamoxifen, ifosfamide, epirubicin, and cisplatin combination chemotherapy for inoperable non-small-cell lung cancer.
Am J Clin Oncol
2000
;
23
:
13
–7.