Estrogen receptor positivity in mammary tumors of Wnt-1 transgenic mice is influenced by collaborating oncogenic mutations

Article metrics

Abstract

The majority (75%) of human breast cancers express estrogen receptor (ER). Although ER-positive tumors usually respond to antiestrogen therapies, 30% of them do not. It is not known what controls the ER status of breast cancers or their responsiveness to antihormone interventions. In this report, we document that transgenic (TG) expression of Wnt-1 in mice induces ER-positive tumors. Loss of Pten or gain of Ras mutations during the evolution of tumors in Wnt-1 TG mice has no effect on the expression of ER, but overexpression of Neu or loss of p53 leads to ER-negative tumors. Thus, our results provide compelling evidence that expression of ER in breast cancer may be influenced by specific genetic changes that promote cancer progression. These findings constitute a first step to explore the molecular mechanisms leading to ER-positive or ER-negative mammary tumors. In addition, we find that ER-positive tumors arising in Wnt-1 TG mice are refractory to both ovariectomy and the ER antagonist tamoxifen, but lose ER expression with tamoxifen, suggesting that antiestrogen selects for ER-negative tumor cells and that the ER-positive cell fraction is dispensable for growth of these tumors. This is a first report of a mouse model of antiestrogen-resistant ER-positive breast cancers, and could provide a powerful tool to study the molecular mechanisms that control antiestrogen resistance.

Introduction

Breast cancer develops and evolves as genetic and epigenetic alterations accumulate in the ductal epithelium, in which approximately 10–15% of the normal cells express the α isoform of estrogen receptor (ER, Clarke et al., 1997). These changes initially give rise to precursor lesions such as usual or atypical ductal hyperplasias, which may or may not progress to in situ and then to invasive breast cancer (Allred et al., 2004). This process of malignant evolution is promoted by the female hormone estrogen. In the end, two distinct types of breast cancer develop – 75% of all breast cancers express ER in a few to nearly 100% of the tumor cells (ER+) and the rest lack ER entirely (ER; Allred et al., 2004). The majority of ER+ tumors respond initially to antiestrogen therapies despite their high or low level of ER expression, but approximately 30% of them do not and have a poorer prognosis (Allred et al., 2004).

It is not known what determines whether a breast cancer expresses ER, and it is poorly understood why approximately 30% of ER+ breast cancers do not respond to antihormone therapies (Clarke et al., 2003; Allred et al., 2004). Most studies attempting to address these issues have been carried out using the few available ER+ human breast cancer cell lines and their xenograft models; however, clinical breast cancer is comprised of heterogeneous cells, including usually both ER+ and ER tumor cells and cells at different stages of differentiation (Dontu et al., 2004). Animal models would be very useful for deciphering the mechanisms leading to ER+ cancers and to antiestrogen resistance. However, only a few ER+ models have been made (Medina et al., 1980, 2002; Matsuzawa, 1986; Nandi et al., 1995; Rose-Hellekant et al., 2003; Tilli et al., 2003; Gattelli et al., 2004; Lin et al., 2004; Torres-Arzayus et al., 2004). To our knowledge, ER+ tumors arising in the few genetically engineered mouse models have not been reported to either respond to or to resist antiestrogen interventions.

Wnts control many developmental processes including mammary morphogenesis and progenitor cell renewal (reviewed by Alonso and Fuchs, 2003; Chu et al., 2004). Made as secreted glycoproteins, Wnts exert their biological effects by binding to their membrane receptors, the frizzled and low-density-lipoprotein receptor-related proteins. As a result, β-catenin is stabilized, translocates to the nucleus and transactivates different genes depending on cellular context. Genes encoding components and transcriptional targets of the Wnt signaling pathway are mutated or deregulated in several types of human tumors including breast cancers (Ugolini et al., 2001; Hatsell et al., 2003; Bafico et al., 2004; Brennan and Brown, 2004; Klopocki et al., 2004; Milovanovic et al., 2004).

Wnt-1 (Int-1) was discovered as a gene frequently activated in mammary tumors arising in mice infected with mouse mammary tumor virus (MMTV, Nusse and Varmus, 1982). The MMTV-Wnt-1 transgenic (TG) model was created by overexpressing Wnt-1 using its endogenous promoter and the MMTV enhancer elements (Tsukamoto et al., 1988). This unique construct leads to the expression of Wnt-1 in mammary buds during embryogenesis (Cunha and Hom, 1996). This transgenes appears to cause expansion of mammary progenitor cells, since there is a significant increase in cells expressing putative progenitor cell markers (such as Sca-1 and keratin 6) and cells effluxing fluorescent Hoechst 33342 dye – the dye-excluding property has been associated with stem cells in the hematopoietic system (Goodell et al., 1996; Li Y et al., 2003; Liu et al., 2004). The resulting tumors in MMTV-Wnt-1 TG mice also seem to arise from progenitor cells because the tumor cells also express Sca-1 and keratin 6, and because they contain heterogeneous tumor cells that share a common genetic mutation in Pten, implying a common progenitor (Cui and Donehower, 2000; Rosner et al., 2002; Li Y et al., 2003; Henry et al., 2004; Liu et al., 2004).

In this report, we present evidence to suggest that there is a causal relationship between aberrant Wnt-1 signaling and ER+ mammary tumors, and that mutant p53 or overexpressed Neu, but not defective Pten or activated Ras, can suppress the expression of ER in mammary tumors arising in Wnt-1 TG mice. In addition, we provide evidence to suggest that tumors arising in Wnt-1 TG mice, although ER+, are resistant to ovariectomy and to tamoxifen, but lose ER+ cells with tamoxifen treatment.

Results

ER and PR are expressed in mammary tumors and pulmonary metastases in Wnt-1 TG mice

Using a commercial rabbit antibody against the C-terminus of ER-α in immunohistochemical staining, we detected positively stained cells in the majority of mammary tumors and their pulmonary metastases from Wnt-1 TG mice (ages 3–9 months, Figure 1a). To verify the specificity of this antibody, we also stained normal mammary glands and two human breast cancer cell lines. As expected, staining was detected in heterogeneous patterns in luminal epithelial cells, but not myoepithelial cells, in normal mouse mammary glands; strong staining was detected in MCF7, an ER+ cell line, but not in CAL51, an ER cell line (data not shown). We further confirmed positive staining in these tumors using a second commercial antibody against the N-terminus of ER-α (data not shown). Using both antibodies in Western blotting, we also detected, in all tumors from Wnt-1 TG mice, a specific band migrating at the same apparent molecular weight (68 kDa) as the stained band of MCF-7-positive control cells (Figure 1b).

Figure 1
figure1

Expression of ER and PR in primary tumors and pulmonary metastases from mice TG for components of the Wnt signaling pathway. (a) Immunohistochemical staining for ER and PR in tumors and metastases of tumors from mice carrying the indicated transgene. A × 20 objective was used. (b) Western blotting for ER in tumor lysates from Wnt-1 TG mice. Lysates from MCF-7 were included as a positive control. Staining for α-tubulin or β-actin was included as a loading control. (c) Distribution of ER+ cells in tumors from Wnt-1 TG mice as determined by counting the percentage of immunohistochemically stained cells. (d) Photomicrograph of coimmunofluorescent staining for ER and keratin 8 (K8) and for ER and α-SMA in mammary tumors from Wnt-1 TG mice. ER is labeled in red, K8 or α-SMA in green. A × 40 objective was used

A faster-migrating band of approximately 46 kDa was also detected when an antibody against the C-terminus was used (Figure 1b, right panel). Truncated variant messages have been reported in mice (Swope et al., 2002), and an N-terminus-truncated ER variant (46 kDa), sometimes detected in human breast cancer cell lines, has been suggested to have a role in membrane functions of ER (Figtree et al., 2003; Li L et al., 2003). It remains to be determined whether this is a true isoform of ER rather than a degradation product, and, if so, whether it has any biological function in these tumors.

Examining 22 tumors from Wnt-1 TG mice, we found that 19 expressed ER in greater than 5% of the tumor cells (Figure 1c) and were therefore defined as ER+. Since tumor cells are heterogeneous in Wnt-1 TG mice, containing both epithelial and myoepithelial populations, we asked whether ER is exclusively located in epithelial tumor cells using coimmunofluorescent staining. Staining for ER was only detected in cells that also stained for keratin 8 (Figure 1d), a marker for ductal epithelial cells in the breast, but not in cells positive for α-SMA (Figure 1d), a marker for myoepithelial cells. Thus, ER is limited to epithelial tumor cells in Wnt-1-induced mammary tumors.

PR is a classical transcriptional target of ER; it is coexpressed in 96% of normal ER+ mammary cells in mammary glands (Clarke et al., 1997). It is also expressed in 67–80% of ER+ human breast cancers, and predicts a favorable prognosis (McGuire et al., 1991). Thus, we asked whether PR is expressed in tumors from Wnt-1 TG mice. Using immunohistochemical staining, we detected PR expression in at least 5% of tumor cells in 82% of tumors from Wnt-1 TG mice (Figure 1a). These data suggest that the transcriptional function of ER is intact in these tumors.

β-Catenin is a central regulator of Wnt signaling. TG expression of a stabilized mutant of β-catenin induces mammary tumors that bear similarities to tumors in Wnt-1 TG mice (Imbert et al., 2001; Michaelson and Leder, 2001). To our surprise, however, neither ER nor PR was detected in five tumors from β-catenin TG mice (Figure 1a; see Discussion for possible explanations). ER was also lacking (Figure 1a) in all five mammary tumors arising in TG mice overexpressing c-Myc, one of the best-characterized targets of the Wnt signaling pathway.

Collaborating genetic alterations in Wnt-1 TG mice modulate the ER status of the resultant tumors

Having found that ER is expressed in the majority of the mammary tumors in Wnt-1 TG mice, we next asked what might modulate the ER status of mammary tumors in Wnt-1 TG mice. Since it takes a mean time of 6 months for the first mammary tumor to appear in one of the 10 mammary glands in Wnt-1 TG mice (Tsukamoto et al., 1988), additional genetic mutations are most likely required to form mammary tumors. Thus, we hypothesized that these additional genetic alterations might affect the ER status of the resulting tumors. We and others have reported several collaborating genetic changes in tumorigenesis in Wnt-1 TG mice, including mutations of H-Ras, loss of Pten or p53, and overexpression of Neu (Donehower et al., 1995; Li et al., 2000, 2001; Podsypanina et al., 2004). When tumors that bore a Ras-activating mutation (G12E, G13V, Q61L and Q61R), determined by sequencing tumor DNA, were compared with those that did not, both ER and PR levels were similar by Western blotting (Figure 2a). Therefore, activation of Ras does not suppress ER expression in mammary tumor cells that already overexpress Wnt-1. However, we were surprised to find that Ras-mutated tumors in Wnt-1 TG mice do not make more activated Erk (a classical downstream factor of Ras) or more activated Akt (another downstream target of Ras) than tumors without Ras mutations (Figure 2a and Podsypanina et al., 2004), since overexpression of mutated H-Ras from the MMTV promoter in TG mice or in cultured cells activates Erk signaling (Sinn et al., 1987; Podsypanina et al., 2004). Similar lack of activation of these two Ras effectors has been reported in tumorigenesis of hematopoietic cells, induced by expressing mutated K-Ras from its endogenous promoter in mice (Braun et al., 2004). Thus, activation of Erk and Akt in vivo may require higher levels of mutated H-Ras that are only achievable by a stronger promoter, such as that of MMTV. Consequently, it is an open question whether activation of Erk or Akt will cause Wnt-1-overexpressing mammary cells to evolve into ER or ER+ tumors.

Figure 2
figure2

Ras status does not affect ER expression in tumors in Wnt-1 TG mice. (a) Protein lysates of mammary tumors with H-Ras status indicated at the top were assayed by Western blotting for the proteins indicated at the right. (b) Quantitative presentation of levels of the four proteins depicted in (a). Normalized signal strength was obtained by dividing the densitometric signal of ER and PR by that of β-actin, and by dividing the signal of P-Akt and P-Erk by that of Akt and Erk, respectively

Pten is mutated in approximately 5% of human breast cancers, and is downregulated in 35–50% of cases (reviewed by Sansal and Sellers, 2004). Loss of Pten is usually associated with ER human breast cancers (Garcia et al., 1999; Perren et al., 1999; Shi et al., 2003). We have previously reported that inactivation of one allele of Pten accelerates mammary tumorigenesis in MMTV-Wnt-1 TG mice, and 70% of the resulting tumors have lost the wild-type allele of Pten (Li et al., 2001). To determine if loss of Pten alters the ER status of tumors in Wnt-1 TG mice, we examined preneoplastic and tumor samples from a cross of Wnt-1 TG mice and Pten heterozygous mice (Podsypanina et al., 1999). Similar to the staining results from mammary glands from adult non-TG and Wnt-1 TG mammary glands, ER and PR were detected in hyperplasias in Wnt-1/Pten+/− mice (Figure 4i and j). Full-blown tumors that had not undergone LOH at the Pten locus also retained ER and PR (data not shown). Furthermore, ER and PR were still expressed in four of the five tumors that had lost the wild-type Pten allele (Figure 3a and b). These results suggest that loss of Pten does not prevent ER expression in mammary cancers, consistent with another study examining ER expression in mammary lesions in Pten+/− mice that do not carry a transgene (Shi et al., 2003).

Figure 4
figure4

ER and PR are expressed in mammary ductal hyperplasias in TG mice. Paraffin sections of mammary glands from adult mice (12 weeks old) with the genotypes shown at the left were stained by immunohistochemical staining for the protein indicated at the top. A × 20 objective was used

Figure 3
figure3

Neu overexpression or loss of p53, but not loss of Pten, leads to ER-/PR- mammary tumors in Wnt-1 TG mice. Paraffin sections of mammary tumors from mice with the genotypes shown at the left were stained by immunohistochemical staining for the protein indicated at the top. Staining of mammary tumors from MMTV-Neu was included as a control. A × 20 objective was used

Neu (HER2, ErbB2) is a member of the epidermal growth factor receptor (EGFR) family of transmembrane tyrosine kinases (Olayioye et al., 2000). Its gene is amplified in approximately 25% of human breast cancers (Slamon et al., 1987), and is inversely correlated with ER and PR (Konecny et al., 2003), but the mechanism underlining this correlation is not known. We have recently reported that the Neu protooncogene collaborates with Wnt-1 in mammary tumorigenesis (Podsypanina et al., 2004), and that it remains wild type in tumors arising in animals bi-TG for Wnt-1 and Neu (Podsypanina et al., 2004), although it is somatically mutated to a more activated form in approximately 70% of tumors arising in MMTV-Neu TG mice that are otherwise wild type (Siegel et al., 1994). Therefore, we asked whether ER is retained in preneoplastic and tumor lesions in bi-TG mice expressing both Wnt-1 and the Neu protooncogene. Similar to what was observed in Wnt-1 or Neu TG mice (Figure 4c–f, and reference Wu et al., 2002), both ER and PR were present in hyperplastic ducts, albeit heterogeneously, in nontumor-bearing mammary glands from Wnt-1/Neu bi-TG mice (Figure 4g and h). However, similar to tumors from Neu TG mice (Figure 3c and d and reference Wu et al., 2002), but different from tumors from Wnt-1 TG mice, neither ER nor PR was detected in any of the four tumors from these bi-TG mice (Figure 3e and f). Therefore, Neu prevents Wnt-1-expressing mammary tissues from evolving into ER+ tumors.

P53 is mutated in approximately 50% of human breast cancers, and the pathway is probably disabled in even higher percentages of cases (Gasco et al., 2002). Loss of p53 does not inhibit the expression of ER in normal mammary cells, since ER is expressed in p53-null mammary epithelia (data not shown and Medina et al., 2003). Loss of p53 does not have a significant correlation with the ER status of human breast tumors, and 20% of tumors arising in mammary epithelia from p53-knockout mice express ER (Barbareschi et al., 1996; Medina et al., 2002). To determine if germline p53 loss might modulate the expression of ER in Wnt-1 TG mice, we examined both preneoplastic and tumor sections from mice that were Wnt-1 TG/p53−/−. While retained in intraductal hyperplasias in these mice (Figure 4k and l), both ER and PR were undetectable in all five tumors we examined by immunohistochemical staining (Figure 3g and h) and Western blotting (data not shown). These results suggest that loss of p53, like activated Neu signaling, inhibits Wnt-1-expressing mammary tissues from evolving into ER+ tumors.

Wnt-1-induced mammary tumors continue to grow when estrogen signaling is inhibited

Approximately 70% of ER+ human breast cancers are growth inhibited by antihormone therapies, but the rest are not (Allred et al., 2004). The mechanisms leading to antihormone responsiveness are poorly understood (Clarke et al., 2003). Although activated Wnt signaling might be associated with a subset of ER+ human breast cancers, based on the correlation between reduced expression of secreted frizzled-related protein 1 (an inhibitor of Wnt signaling) and expression of ER (Ugolini et al., 2001), it is not known whether these ER+ cancers are susceptible to antiestrogens. It has been reported that Wnt-1 can induce mammary tumors in mice that lack ER (Bocchinfuso et al., 1999); however, it is not known whether ER signaling is necessary for survival of the ER+ tumors that develop in ER-intact Wnt-1 TG mice. Therefore, we transplanted three ER+ mammary tumors from Wnt-1 TG mice into 15 nude mice each. When the tumors reached 0.7 cm in diameter, we divided the nude mice in each group into three sets and administered ovariectomy, daily tamoxifen or daily vehicle. We measured tumor size weekly and euthanized all recipient mice 4 weeks later. Both ovariectomy and tamoxifen failed to prevent or inhibit tumor growth (Figure 5a), although the same dose of tamoxifen successfully inhibited the growth of control ER+ tumors in these mice resulting from transplantation with MCF-7 cells (data not shown). ER and PR were detected in the transplanted tumors before and after treatment with the vehicle control (Figure 5b and c), eliminating the trivial explanation that the transplants do not have functional ER due to the lower circulating estrogen levels sometimes found in nude mice. Thus, we conclude that estrogen signaling is dispensable for growth of MMTV-Wnt-1-induced ER+ tumors.

Figure 5
figure5

Ovariectomy and tamoxifen do not suppress growth of transplanted mammary tumors from MMTV-Wnt-1 TG mice, but do reduce the frequency of ER+ cells. (a) Tamoxifen or ovariectomy does not suppress the growth of mammary tumors from Wnt-1 TG mice. Each of three independent MMTV-Wnt-1-induced tumors was transplanted into 15 nude mice (2 mm3 each recipient). When the transplanted tumors reached 0.7 cm in diameter, the mice in each group were divided into three subgroups, which were treated with ovariectomy (ovex), tamoxifen (tam), or vehicle for 4 weeks. (b) ER and PR are suppressed by antiestrogenic therapies. Transplanted tumors were collected after 4 weeks of treatment and stained for the proteins indicated at the left. Photomicrographs are representative of three independent antiestrogenic treatment experiments. (c) Percent cells positive for ER in donors and in transplanted groups after 4 weeks of treatment. Quantitative data were generated by counting positively stained cells in nine tam-or vehicle-treated tumors (three transplants from each of the three donors). Five independent × 40 microscopic fields were counted per section per tumor. P-value is shown

One of the mechanisms leading to estrogen-independent growth is that the transcriptional activity of ER no longer depends on its ligand, due to phosphorylation of ER (Le Goff et al., 1994; Kato et al., 1995; Pietras et al., 1995; Shang and Brown, 2002; Michalides et al., 2004) or to stabilization of its interaction with coactivators (Trowbridge et al., 1997; Zwijsen et al., 1997). However, we found that both ER and its transcriptional target PR were expressed in far fewer cells in tumors from mice treated with tamoxifen than in control tumors or in donor tumors prior to transplantation (Figure 5b and c), suggesting a different mechanism of resistance. The diminished ER expression must be due either to downregulation of ER expression in ER+ tumor cells or to selection against the ER+ subset of tumor cells. With regard to the first possibility, we found that tamoxifen did not downregulate ER in the adjacent normal ducts (data not shown), suggesting that the loss of ER in treated tumors is not caused by downregulation of ER expression by tamoxifen, consistent with reports in cultured breast cancer cell lines (Giamarchi et al., 2002; Cheng et al., 2004). To confirm this in the tumor cells themselves, we examined ER expression in three sets of transplanted tumors that had been treated with tamoxifen for only 5 days, when the ER+ cell population would not yet be selected out. As expected, the percentage of ER+ cells did not change in that period (data not shown). Thus, we conclude that the antiestrogen resistance of ER+ mammary tumors in Wnt-1 TG mice probably does not result either from ligand-independent transactivation by ER or from suppression of ER expression in ER+ tumor cells, but rather from selection against the ER+ population, leaving the ER cells actively growing.

Discussion

In this study, we demonstrate that TG expression of Wnt-1 in the mammary gland induces largely ER+ mammary tumors, while loss of p53 or overexpression of Neu, but not activating mutations of Ras or loss of Pten, lead to ER tumors in these Wnt-1 TG mice. In addition, we show that tumors arising in Wnt-1 TG mice continue to grow in spite of ovariectomy or tamoxifen, possibly through selection against ER+ tumor cells.

While rare in most tumors induced by other oncogenes or defective tumor suppressor genes, ER+ tumors predominate in Wnt-1 TG mice. ER signaling might facilitate Wnt signaling in transforming mammary cells – in fact, ectopic activation of estrogen signaling has been reported to enhance the transcriptional activities of the Wnt signaling intermediate β-catenin in colon cancer cell lines (Kouzmenko et al., 2004). However, it appears that ER signaling is not absolutely required for Wnt-1 to induce mammary tumors. First, mammary tumors have been observed in male Wnt-1 TG mice (Tsukamoto et al., 1988) and in females that have lost exon 1 of ER (but retained the expression of a 61-kDa truncated ER isoform), although these tumors occur much later (Bocchinfuso et al., 1999). Second, a small percentage of mammary tumors arising in Wnt-1 TG mice are actually ER (Figure 1c). Last, there are a great many ER cells even in ER+ hyperplasias and tumors in Wnt-1 TG mice.

Then, what accounts for the predominance of ER+ tumors in Wnt-1 TG mice? We and others have reported evidence that activated Wnt signaling might induce tumors from progenitor cells. If Wnt signaling, which is known to regulate progenitor cell renewal and proliferation, does cause mammary tumors from progenitor cells, the transformed progenitor cells in tumors might still retain the ability to both self-renew and differentiate into multiple cells including both ER and ER+ cells, therefore resulting in ER+ tumors. Indeed, Wnt-1-induced tumors contain a small subset of cells that bear similarities to normal progenitor cells, based on staining for putative progenitor cell markers and detection of cells that exclude the DNA-binding dye Hoechst 33342 (Li Y et al., 2003; Liu et al., 2004). The rest of the tumor cells appear to be differentiated epithelial and myoepithelial cells. The presence of a small subset of tumor cells (tumor progenitor cells or cancer stem cells) that can self-renew and differentiate into cells of multiple lineages has been reported for several tissue types, including the breast (Pardal et al., 2003; Singh et al., 2004), although the cellular origin of these cells or their hosting tumors has not been revealed.

Wnt signaling activates several downstream pathways. The best studied, the so-called ‘canonical’ pathway, leads to stabilization of β-catenin, which in turn forms heterodimers with members of the LEF/TCF family of DNA binding proteins and transactivates a number of transcriptional targets, including c-Myc and Cyclin D1 (Hatsell et al., 2003; Nusse, 2003; Brennan and Brown, 2004). However, to our surprise, neither ER nor PR was detected in tumors from mice TG for β-catenin or c-Myc (Figure 1a). The Wnt-1 induction of ER+ tumors, possibly through regulating progenitor cell self-renewal and multilineage differentiation, may require paracrine Wnt signaling or β-catenin-independent noncanonical Wnt signaling through Rho, JNK and PKC (Hatsell et al., 2003). However, we cannot exclude the possibility that the tumors in β-catenin TG mice arose from a slightly different (probably more differentiated) population of mammary cells due to the use of a different TG promoter (Li Y et al., 2003).

The loss of Pten does not abolish the expression of ER in tumors in Wnt-1 TG mice (Figure 3a). It has been reported that ER is retained in breast lesions arising in Pten heterozygous mice that are otherwise wild type (Shi et al., 2003), although the authors did not indicate whether the wild-type allele of Pten was lost in the epithelial cells in their lesions. Jointly, these data suggest that Pten loss may not suppress ER expression in mammary tumors or prevent transformed progenitor cells from differentiating into ER+ cells. Interestingly, Akt – which is usually phosphorylated and activated in other cell types when Pten is suppressed, and has been reported to suppress ER expression in cultured breast cancer cells (Faridi et al., 2003) – is not highly phosphorylated in mouse tumors that are Wnt-1 TG/Pten+/− with LOH (Li et al., 2001), similar to what has been reported in some human breast cancers that have downregulated Pten (Shi et al., 2003). Therefore, our data do not suggest that activation of Akt during the evolution to tumors in Wnt-1 TG mice leads to ER+ tumors.

Although H-Ras is mutated in approximately 50% of tumors arising in Wnt-1 TG mice (Podsypanina et al., 2004), these H-Ras mutants do not have detectable effects on either the expression of ER or its function in inducing PR in these tumors (Figure 2). Thus, H-Ras mutants most likely do not suppress ER expression in breast cancer or prevent transformed progenitor cells from differentiating into ER+ cells. This in vivo study is consistent with an earlier cell culture study, in which mutant Ras was found not to inhibit the expression of ER in MCF-7 cells (Sukumar et al., 1988). Interestingly, ER is not expressed in mammary tumors induced by TG overexpression of an H-Ras mutant with no pre-existing aberration of Wnt signaling (data not shown). In addition to possible differences in cellular origin between ER+ Ras-mutated tumors in Wnt-1 TG and ER tumors in H-Ras TG mice (Li Y et al., 2003), there are other important dissimilarities between these two tumors, including levels of H-Ras expression and levels of activation of the Erk and Akt effector pathways (Podsypanina et al., 2004). Thus, it remains to be determined whether ER tumors will form in Wnt-1 TG mice if H-Ras mutants are expressed from a stronger exogenous promoter during the evolution to mammary tumors.

It is striking that either overexpression of Neu or loss of p53 causes exclusively ER tumors in Wnt-1 TG mice. Several mechanisms may lead to altered ER status, including the cellular origin of cancer, direct modulation of ER expression by these secondary genetic factors and alteration of differentiation of transformed progenitor cells. It is plausible that additional alteration in Neu signaling and p53 functions may select different cells in the mammary gland to form cancer. However, the ER tumors in Wnt-1 TG mice that either overexpress Neu or lose p53 seem to arise from the same cells as the ER+ tumors in Wnt-1 TG mice, that is, the progenitor cells (Li Y et al., 2003). First, they contain both myoepithelial tumor cells and epithelial tumor cells, implying an origin from a bipotential cell (Li Y et al., 2003; Podsypanina et al., 2004; and K Podsypanina, unpublished). Second, they harbor tumor cells expressing the putative progenitor cell markers keratin 6 and Sca-1 (Li Y et al., 2003; Podsypanina et al., 2004). Last, the microarray gene expression profiles are remarkably similar between tumors from Wnt-1 TG/p53−/− mice and tumors from Wnt-1 TG mice that are otherwise wild type (S Huang, unpublished). Collectively, these results argue against different cellular origins for these tumors.

It is also unlikely that lack of ER expression in tumors in Wnt-1 TG mice that either overexpress Neu or lose p53 is the consequence of direct suppression of ER by inactivation of p53 or overexpression of Neu. As stated above, ER-expressing cells are present in preneoplastic lesions in mammary glands in both Wnt-1 TG/p53−/− and Wnt-1/Neu bi-TG mice (Figure 4), suggesting that these lesions do not directly suppress ER in Wnt-1-overexpressing cells. However, we cannot exclude the possibility that they may be able to suppress ER after these cells have evolved into cancer, or that other genetic events (such as ErbB3 activation; Siegel et al., 1999) that are preferentially selected for during tumorigenesis in these compound mutant cells may inhibit ER in cancer cells.

Since Wnt-1 appears to induce mammary tumors from progenitor cells and the transformed progenitor cells can both self-renew and differentiate into both ER and ER+ tumor cells, it is possible that acquisition of another genetic alteration, such as overexpression of Neu or ablation of p53, might impair the potential of these tumor progenitor cells to differentiate into ER+ cells, thus leading to ER tumors (even though, in precancerous stages, these genetic defects do not seem to totally prevent cancer precursor cells from differentiating into ER+ cells). Consistent with this hypothesis, tumors arising in Wnt-1 TG/p53−/− mice have been reported to exhibit less differentiation than tumors arising in Wnt-1 TG/p53+/+ mice (Donehower et al., 1995); the signaling network activated by Neu has been reported to have a role in regulating cell differentiation (Dai and Holland, 2003). To test whether Neu or defective p53 indeed suppresses the differentiation of cancer progenitor cells in these models, we plan to inhibit Neu signaling or restore p53 in these ER tumors to test whether removing these genetic events can permit the ER tumor progenitor cells to differentiate into ER+ cells.

The likely presence of transformed tumor progenitors in Wnt-1 TG mice might also explain why these ER+ tumors are resistant to antiestrogens. If Wnt-1 transforms ER progenitor cells, the differentiated ER+ progeny cells might be dispensable for tumor growth. A similar hypothesis has been proposed in a recent review article (Dontu et al., 2004) to explain human tumor resistance to antiestrogen therapies. We are currently testing this hypothesis in several mouse models, and additionally asking whether antiestrogen resistance in human ER+ cancers might also result from lack of ER expression in the cancer progenitor cells.

In conclusion, activation of Wnt-1 signaling induces ER+ mammary tumors, but additional genetic mutations can alter the ER status of the resultant tumors, possibly through restricting differentiation of tumor progenitor cells. These ER+ tumors are resistant to antiestrogens, probably because the ER fraction of tumor cells can maintain tumor growth in the absence of estrogen signaling.

Materials and methods

Mice

TG mice expressing Wnt-1 (Tsukamoto et al., 1988), the Neu protooncogene (Guy et al., 1992), an N-terminus-deleted, stabilized mutant of β-catenin (Imbert et al., 2001) and c-Myc (Stewart et al., 1984) have been described, as have mice harboring targeted inactivating mutations of Pten (Podsypanina et al., 1999) and p53 (Jacks et al., 1994). All TG mice were maintained on the FVB background, but Pten and p53 mutant mice were on mixed backgrounds of 129, FVB and C57/BL6. All mice were housed in specific-pathogen-free animal facilities, on a standard diet. Genotyping was carried out using PCR of tail DNAs (Tsukamoto et al., 1988).

Antibodies

The following antibodies were used: purified rabbit antibodies against the ER C-terminus (Santa Cruz, #SC542) or the N-terminus (Novocastra, #NCL-ER-6F11), progesterone receptor (PR, DAKO, #A0098 for immunohistochemical staining; Santa Cruz, #SC-7208 for Western blotting), phospho-AKT (ser473, Cell Signaling, #9102), Akt (Cell Signaling, #9272), phospho-Erk (Cell Signaling, #9101), Erk (Cell Signaling, #9102) and β-actin (Sigma, #5441); purified mouse monoclonal antibodies against α-smooth muscle actin (α-SMA, DAKO) and bromodeoxyuridine (BrdU, Becton-Dickinson, #347580); and partially purified rat antibodies against keratin 8 (Kemler et al., 1981), purchased from the Developmental Studies Hybridoma Bank organized under the auspices of the NICHD and maintained by the University of Iowa, Department of Biological Sciences, Iowa City, IA, USA.

Immunoassays

Tissues were fixed in 10% neutral formalin and processed as described previously (Li et al., 2001) to obtain paraffin sections of 4 μm in thickness. For immunohistochemistry, the sections were boiled for 15 min in citrate buffer, pH 6.0 (to unmask antigen epitopes). Endogenous peroxidase activity was inactivated by a 10 min incubation in 3% hydrogen peroxide, and subsequent steps were performed using Vector ABC and MOM kits and the Novo-Red substrate (Vector Laboratories) following the manufacturer's recommendations. To label cells in S phase of the cell cycle, BrdU (Sigma, #B-5002) at 100 μg/g of mouse body weight in saline was injected intraperitoneally 1 h prior to euthanasia.

For Western blotting, tumors were ground to powder in liquid nitrogen and lysed in the M-PER tissue lysis solution (Pierce) with gentle shaking overnight at 4°C. The resulting supernatant (25 μg protein) was denatured using 2-mercaptoethanol, resolved on 10% polyacrylamide minigels containing 10% SDS and transferred onto nitrocellulose membranes, which were then incubated with primary antibodies and peroxidase-conjugated secondary antibodies (Jackson Laboratory) in tris-borate buffer/0.05% Tween-20/5% nonfat dried milk. The proteins recognized by the specific antibodies were visualized by a chemiluminescent substrate (Supersignal, Pierce).

Analysis of H-Ras mutations in tumors from Wnt-1 TG mice

The method has been described (Podsypanina et al., 2004). Briefly, DNA was extracted from tumors from Wnt-1 TG mice and used for PCR amplification of exons 1 and 2 of H-Ras using two sets of oligos (mouse H-Ras exon 1: HRAS.F1A – 5′-CCTTGGCTAAGTGTGCTTC-3′, HRAS.B1A – 5′-CCACCTCTGGCAGGTAG-3′; mouse H-Ras exon 2: HRAS.F2A – 5′-GGATTCTCTGGTCTGAGG-3′, HRAS.2B2 – 5′-GGATATGAGCCAGCTAGC-3′). The resulting PCR product was purified and sequenced to identify mutations in these two exons. We have reported earlier that only exons 1 and 2 of H-Ras are mutated in Wnt-1-induced mammary tumors; other exons or family members of Ras are normal in these tumors (Podsypanina et al., 2004).

Tumor transplantation and antiestrogen treatment

Tumors of 1.5 cm in diameter were excised from Wnt-1 TG mice, minced into cubes of 2 mm in size and injected subcutaneously using a trocar into female nude mice of 8 weeks of age. In all, 15 nude mice were transplanted for each donor tumor from Wnt-1 TG mice. When transplanted tumors reached 0.7 cm in diameter, the recipient mice were divided into three groups and given ovariectomy, tamoxifen or vehicle. Tamoxifen (Sigma), dissolved in peanut oil at a concentration of 10 mg/ml, was injected subcutaneously five times a week (0.5 mg/mouse). Tumor size were measured weekly. At 4 weeks after the initiation of treatments, all mice were killed, and tumors were collected for immunohistochemical staining for ER and PR.

References

  1. Allred DC, Brown P and Medina D . (2004). Breast Cancer Res., 6, 240–245.

  2. Alonso L and Fuchs E . (2003). Genes Dev., 17, 1189–1200.

  3. Bafico A, Liu G, Goldin L, Harris V and Aaronson SA . (2004). Cancer Cell, 6, 497–506.

  4. Barbareschi M, Caffo O, Veronese S, Leek RD, Fina P, Fox S, Bonzanini M, Girlando S, Morelli L, Eccher C, Pezzella F, Doglioni C, Dalla Palma P and Harris A . (1996). Hum. Pathol., 27, 1149–1155.

  5. Bocchinfuso WP, Hively WP, Couse JF, Varmus HE and Korach KS . (1999). Cancer Res., 59, 1869–1876.

  6. Braun BS, Tuveson DA, Kong N, Le DT, Kogan SC, Rozmus J, Le Beau MM, Jacks TE and Shannon KM . (2004). Proc. Natl. Acad. Sci. USA, 101, 597–602.

  7. Brennan KR and Brown AM . (2004). J. Mammary Gland Biol. Neoplasia, 9, 119–131.

  8. Cheng G, Weihua Z, Warner M and Gustafsson JA . (2004). Proc. Natl. Acad. Sci. USA, 101, 3739–3746.

  9. Chu EY, Hens J, Andl T, Kairo A, Yamaguchi TP, Brisken C, Glick A, Wysolmerski JJ and Millar SE . (2004). Development, 131, 4819–4829.

  10. Clarke RB, Howell A, Potten CS and Anderson E . (1997). Cancer Res., 57, 4987–4991.

  11. Clarke R, Liu MC, Bouker KB, Gu Z, Lee RY, Zhu Y, Skaar TC, Gomez B, O’Brien K, Wang Y and Hilakivi-Clarke LA . (2003). Oncogene, 22, 7316–7339.

  12. Cui XS and Donehower LA . (2000). Oncogene, 19, 5988–5996.

  13. Cunha GR and Hom YK . (1996). J. Mammary Gland Biol. Neoplasia., 1, 21–37.

  14. Dai C and Holland EC . (2003). Cancer J., 9, 72–81.

  15. Donehower LA, Godley LA, Aldaz CM, Pyle R, Shi YP, Pinkel D, Gray J, Bradley A, Medina D and Varmus HE . (1995). Genes Dev., 9, 882–895.

  16. Dontu G, El-Ashry D and Wicha MS . (2004). Trends Endocrinol. Metab., 15, 193–197.

  17. Faridi J, Wang L, Endemann G and Roth RA . (2003). Clin. Cancer Res., 9, 2933–2939.

  18. Figtree GA, McDonald D, Watkins H and Channon KM . (2003). Circulation, 107, 120–126.

  19. Garcia JM, Silva JM, Dominguez G, Gonzalez R, Navarro A, Carretero L, Provencio M, Espana P and Bonilla F . (1999). Breast Cancer Res. Treat., 57, 237–243.

  20. Gasco M, Shami S and Crook T . (2002). Breast Cancer Res., 4, 70–76.

  21. Gattelli A, Cirio MC, Quaglino A, Schere-Levy C, Martinez N, Binaghi M, Meiss RP, Castilla LH and Kordon EC . (2004). Cancer Res., 64, 5193–5199.

  22. Giamarchi C, Chailleux C, Callige M, Rochaix P, Trouche D and Richard-Foy H . (2002). Biochim. Biophys. Acta., 1578, 12–20.

  23. Goodell MA, Brose K, Paradis G, Conner AS and Mulligan RC . (1996). J. Exp. Med., 183, 1797–1806.

  24. Guy CT, Webster MA, Schaller M, Parsons TJ, Cardiff RD and Muller WJ . (1992). Proc. Natl. Acad. Sci. USA, 89, 10578–10582.

  25. Hatsell S, Rowlands T, Hiremath M and Cowin P . (2003). J. Mammary Gland Biol. Neoplasia, 8, 145–158.

  26. Henry MD, Triplett AA, Oh KB, Smith GH and Wagner KU . (2004). Oncogene, 23, 6980–6985.

  27. Imbert A, Eelkema R, Jordan S, Feiner H and Cowin P . (2001). J. Cell. Biol., 153, 555–568.

  28. Jacks T, Remington L, Williams BO, Schmitt EM, Halachmi S, Bronson RT and Weinberg RA . (1994). Curr. Biol., 4, 1–7.

  29. Kato S, Endoh H, Masuhiro Y, Kitamoto T, Uchiyama S, Sasaki H, Masushige S, Gotoh Y, Nishida E, Kawashima H, Metzger D and Chambon P . (1995). Science, 270, 1491–1494.

  30. Kemler R, Brulet P, Schnebelen MT, Gaillard J and Jacob F . (1981). J. Embryol. Exp. Morphol., 64, 45–60.

  31. Klopocki E, Kristiansen G, Wild PJ, Klaman I, Castanos-Velez E, Singer G, Stohr R, Simon R, Sauter G, Leibiger H, Essers L, Weber B, Hermann K, Rosenthal A, Hartmann A and Dahl E . (2004). Int. J. Oncol., 25, 641–649.

  32. Konecny G, Pauletti G, Pegram M, Untch M, Dandekar S, Aguilar Z, Wilson C, Rong HM, Bauerfeind I, Felber M, Wang HJ, Beryt M, Seshadri R, Hepp H and Slamon DJ . (2003). J. Natl. Cancer Inst., 95, 142–153.

  33. Kouzmenko AP, Takeyama KI, Ito S, Furutani T, Sawatsubashi S, Maki A, Suzuki E, Kawasaki Y, Akiyama T, Tabata T and Kato S . (2004). J. Biol. Chem., 279, 40255–40258.

  34. Le Goff P, Montano MM, Schodin DJ and Katzenellenbogen BS . (1994). J. Biol. Chem., 269, 4458–4466.

  35. Li L, Haynes MP and Bender JR . (2003). Proc. Natl. Acad. Sci. USA, 100, 4807–4812.

  36. Lin SC, Lee KF, Nikitin AY, Hilsenbeck SG, Cardiff RD, Li A, Kang KW, Frank SA, Lee WH and Lee EY . (2004). Cancer Res., 64, 3525–3532.

  37. Liu BY, McDermott SP, Khwaja SS and Alexander CM . (2004). Proc. Natl. Acad. Sci. USA, 101, 4158–4163.

  38. Li Y, Hively WP and Varmus HE . (2000). Oncogene, 19, 1002–1009.

  39. Li Y, Podsypanina K, Liu X, Crane A, Tan LK, Parsons R and Varmus HE . (2001). BioMedCentral: Mol. Biol., 2, 2.

  40. Li Y, Welm B, Podsypanina K, Huang S, Chamorro M, Zhang X, Rowlands T, Egeblad M, Cowin P, Werb Z, Tan LK, Rosen JM and Varmus HE . (2003). Proc. Natl. Acad. Sci. USA, 100, 15853–15858.

  41. Matsuzawa A . (1986). Int. Rev. Cytol., 103, 303–340.

  42. McGuire WL, Chamness GC and Fuqua SA . (1991). Mol. Endocrinol., 5, 1571–1577.

  43. Medina D, Butel JS, Socher SH and Miller FL . (1980). Cancer Res., 40, 368–373.

  44. Medina D, Kittrell FS, Shepard A, Contreras A, Rosen JM and Lydon J . (2003). Cancer Res., 63, 1067–1072.

  45. Medina D, Kittrell FS, Shepard A, Stephens LC, Jiang C, Lu J, Allred DC, McCarthy M and Ullrich RL . (2002). FASEB J., 16, 881–883.

  46. Michaelson JS and Leder P . (2001). Oncogene, 20, 5093–5099.

  47. Michalides R, Griekspoor A, Balkenende A, Verwoerd D, Janssen L, Jalink K, Floore A, Velds A, van’t Veer L and Neefjes J . (2004). Cancer Cell, 5, 597–605.

  48. Milovanovic T, Planutis K, Nguyen A, Marsh JL, Lin F, Hope C and Holcombe RF . (2004). Int. J. Oncol., 25, 1337–1342.

  49. Nandi S, Guzman RC and Yang J . (1995). Proc. Natl. Acad. Sci. USA, 92, 3650–3657.

  50. Nusse R . (2003). Development, 130, 5297–5305.

  51. Nusse R and Varmus HE . (1982). Cell, 31, 99–109.

  52. Olayioye MA, Neve RM, Lane HA and Hynes NE . (2000). EMBO J., 19, 3159–3167.

  53. Pardal R, Clarke MF and Morrison SJ . (2003). Nat. Rev. Cancer, 3, 895–902.

  54. Perren A, Weng LP, Boag AH, Ziebold U, Thakore K, Dahia PL, Komminoth P, Lees JA, Mulligan LM, Mutter GL and Eng C . (1999). Am. J. Pathol., 155, 1253–1260.

  55. Pietras RJ, Arboleda J, Reese DM, Wongvipat N, Pegram MD, Ramos L, Gorman CM, Parker MG, Sliwkowski MX and Slamon DJ . (1995). Oncogene, 10, 2435–2446.

  56. Podsypanina K, Ellenson LH, Nemes A, Gu J, Tamura M, Yamada KM, Cordon-Cardo C, Catoretti G, Fisher PE and Parsons R . (1999). Proc. Natl. Acad. Sci. USA, 96, 1563–1568.

  57. Podsypanina K, Li Y and Varmus H . (2004). BMC Med., 2, 24.

  58. Rose-Hellekant TA, Arendt LM, Schroeder MD, Gilchrist K, Sandgren EP and Schuler LA . (2003). Oncogene, 22, 4664–4674.

  59. Rosner A, Miyoshi K, Landesman-Bollag E, Xu X, Seldin DC, Moser AR, MacLeod CL, Shyamala G, Gillgrass AE and Cardiff RD . (2002). Am. J. Pathol., 161, 1087–1097.

  60. Sansal I and Sellers WR . (2004). J. Clin. Oncol., 22, 2954–2963.

  61. Shang Y and Brown M . (2002). Science, 295, 2465–2468.

  62. Shi W, Zhang X, Pintilie M, Ma N, Miller N, Banerjee D, Tsao MS, Mak T, Fyles A and Liu FF . (2003). Int. J. Cancer, 104, 195–203.

  63. Siegel PM, Dankort DL, Hardy WR and Muller WJ . (1994). Mol. Cell. Biol., 14, 7068–7077.

  64. Siegel PM, Ryan ED, Cardiff RD and Muller WJ . (1999). EMBO J., 18, 2149–2164.

  65. Singh SK, Hawkins C, Clarke ID, Squire JA, Bayani J, Hide T, Henkelman RM, Cusimano MD and Dirks PB . (2004). Nature, 432, 396–401.

  66. Sinn E, Muller W, Pattengale P, Tepler I, Wallace R and Leder P . (1987). Cell, 49, 465–475.

  67. Slamon DJ, Clark GM, Wong SG, Levin WJ, Ullrich A and McGuire WL . (1987). Science, 235, 177–182.

  68. Stewart TA, Pattengale PK and Leder P . (1984). Cell, 38, 627–637.

  69. Sukumar S, Carney WP and Barbacid M . (1988). Science, 240, 524–526.

  70. Swope D, Harrell JC, Mahato D and Korach KS . (2002). Gene, 294, 239–247.

  71. Tilli MT, Frech MS, Steed ME, Hruska KS, Johnson MD, Flaws JA and Furth PA . (2003). Am. J. Pathol., 163, 1713–1719.

  72. Torres-Arzayus MI, De Mora JF, Yuan J, Vazquez F, Bronson R, Rue M, Sellers WR and Brown M . (2004). Cancer Cell, 6, 263–274.

  73. Trowbridge JM, Rogatsky I and Garabedian MJ . (1997). Proc. Natl. Acad. Sci. USA, 94, 10132–10137.

  74. Tsukamoto AS, Grosschedl R, Guzman RC, Parslow T and Varmus HE . (1988). Cell, 55, 619–625.

  75. Ugolini F, Charafe-Jauffret E, Bardou VJ, Geneix J, Adelaide J, Labat-Moleur F, Penault-Llorca F, Longy M, Jacquemier J, Birnbaum D and Pebusque MJ . (2001). Oncogene, 20, 5810–5817.

  76. Wu K, Zhang Y, Xu XC, Hill J, Celestino J, Kim HT, Mohsin SK, Hilsenbeck SG, Lamph WW, Bissonette R and Brown PH . (2002). Cancer Res., 62, 6376–6380.

  77. Zwijsen RM, Wientjens E, Klompmaker R, van der Sman J, Bernards R and Michalides RJ . (1997). Cell, 88, 405–415.

Download references

Acknowledgements

We thank Drs Jeff Rosen, Dan Medina, Kent Osborne, Adrian Lee, Steffi Oesterreich, Suzanne Fuqua, Craig Allred, Mike Lewis and Harold Varmus for stimulating discussions and/or critical review of this paper. In addition, we thank the Pathology Core Facility at the Breast Center for tissue processing and the Transgenic Mouse Facility at Baylor College of Medicine for animal husbandry. This work was supported in part by funds from Department of Defense (USAMRMC) BC030500 (to YL), SPORE (a developmental grant to YL) and National Institutes of Health GM47429 (to PC). KP was supported by a Cancer Research Institute fellowship award and by funds from National Institutes of Health P01 CA94060-02 and from the Martell Foundation awarded to her supervisor, Dr Harold Varmus. Some of the tumor samples used in this study were generated by YL when he was still a postdoctoral fellow in the laboratory of Dr Harold Varmus.

Author information

Correspondence to Yi Li.

Rights and permissions

Reprints and Permissions

About this article

Keywords

  • estrogen receptor
  • breast cancer
  • Wnt
  • mouse model

Further reading