Tip30 Deletion in MMTV-Neu Mice Leads to Enhanced EGFR Signaling and Development of Estrogen Receptor–Positive and Progesterone Receptor–Negative Mammary Tumors

Despite considerable success in the treatment of estrogen receptor–positive and progesterone receptor–positive (ER+/PR+) breast cancers with therapies directed at targeting estrogen and ERα, a substantial fraction of patients with ER+/PR− tumors do not benefit significantly from these therapies (1, 2). It is estimated that 15% to 25% of all human breast cancers are ER+/PR− with more aggressive malignant characteristics and poorer response to selective estrogen receptor modulators than ER+/PR+ breast cancers (2–4). Moreover, 25% of ER+/PR− tumors are found to have HER2/Neu overexpression; patients with this subtype of ER+/PR− breast cancer have an extremely poor response to endocrine treatment. While several lines of evidence suggest that ER+/PR− tumors can be derived from ER+/PR+ tumors by the loss of PR expression due to anti-hormone therapy, studies indicate that ER+/PR− tumors could arise de novo from other etiologic factors (5). To date, the mechanisms underlying de novo and acquired ER+/PR− breast cancer remain poorly defined. Thus, elucidation of the molecular basis of ER+/PR− breast tumor development has the potential to reveal new therapeutic targets for the treatment, and even prevention of the resistance to anti-estrogen therapy in breast cancer patients.

There are several hypotheses to explain the development of ER+/PR− breast cancers. These include inhibition of PR transcription by aberrant ER cofactors or nonfunctional ER, reduced ER activity due to lower circulating estrogen levels, hypermethylation of PR promoter, or by growth factor signaling pathways (6). Of particular interest are growth factor signaling pathways, in which aberrations are common in many human cancers (7, 8). Among the growth factor receptors, HER2/Neu is the most frequently altered receptor in breast cancers. While most of HER2 positive breast cancers are ER−/PR−, only a small fraction are ER+/PR+ or ER+/PR−, suggesting that HER2 may inhibit ER expression as well as PR expression (6). This hypothesis is supported by the observation that mouse models of breast cancer harboring a HER2/Neu transgene almost exclusively develop ER−/PR− mammary tumors. In addition, when transfected with HER2-expressing vectors, ER+/PR+ breast cancer cells exhibited a significant decrease in ER and PR expression (8). Nevertheless, the mechanism by which activation of HER2/Neu leads to the development of ER−/PR− but not ER⁺ breast cancer remains poorly understood.

TIP30, also known as CC3, is a 30-kDa human cellular protein that was purified as a HIV-1 Tat interacting protein (9) and its expression is altered in human liver, lung, and breast cancers (10–13). Our previous studies showed that Tip30-deficient mice spontaneously develop tumors in several tissues and mammary preneoplastic lesions, suggesting that TIP30 acts as a tumor suppressor (10, 14). Its tumor suppressor activity is probably due to multiple mechanisms. TIP30 functions as a transcription cofactor to repress expression of genes that are involved in proliferation and apoptosis (15, 16) and it can induce apoptosis as an inhibitor of nuclear import (17). In particular, TIP30 can act as a repressor of ERα-mediated c-Myc transcription in mammary glands and breast cancer cells (15). In addition, recent evidence has highlighted that TIP30 controls endocytic downregulation of the EGFR signaling pathway in primary hepatocytes and hepatocellular carcinoma cells (C. Zhang, A. Li, X. Zhang, and H. Xiao, submitted manuscript).

The multiple functions of TIP30 have prompted the speculation that its loss may contribute to mammary tumorigenesis induced by activation of oncogenes. Therefore, we aimed to determine whether Tip30 deletion enhances mammary tumorigenesis in MMTV-Neu mice. We report here that Tip30 deletion cooperates with HER2/Neu activation to promote the development of ER+/PR− mammary tumors, in part, through upregulation of the EGFR signaling pathway. The growth of these ER+/PR− tumors seemed to depend upon both estrogen and progesterone. Thus, the Neu⁺/Tip30^−/− mouse model may help decipher the mechanisms leading to ER+/PR− mammary tumors and identify therapeutic targets for this subgroup of tumors.

Tip30^+/− mice in FBV genetic background were generated by backcrossing Tip30^+/− C57BL/6 mice (14) 7 times with FBV mice. Tip30^+/− mice in FBV background were bred with MMTV-Neu mice (FVB/N-Tg; Jackson Laboratory) to generate Neu⁺/Tip30^−/−, Neu+/Tip3^+/− and Neu⁺/Tip30^+/+ mice. Primary MECs and tumor cells were isolated and cultured as described previously (43). For tumor transplantation assays, all recipient mice were 8-week-old Nu/Nu female nude mice (Charles River). For ovariectomized mice, both ovaries were removed under anesthesia. Placebo (25 mg, 90-day release), 17-estradiol or estrogen (0.1 mg E2, 90-day release), progesterone (35 mg P4, 21-day release), or E2 + P4 pellets (0.1 mg E2 + 32.5 mg P4, 90-day release) were purchased from Innovative Research of America and implanted subcutaneously in the front flanks of each mouse. RU486 was purchased from Calbiochem. Mice were sacrificed at the end of 3 months or when the tumor volume reached 1 cm³. All mice were housed and cared for in the Animal Facility at Michigan State University according to institutional guidelines.

Immunofluorescent staining of mouse mammary tissues was carried out as follow. After deparaffinization and rehydration, tissue sections were autoclaved and then incubated with primary antibody specific for ERα (MC-20, 1:50; Santa Cruz Biotechnology), p-ERα (Santa Cruz Biotechnology), PR-A (hPRa7, 1:50; Labvision), PR-B (hPRa6, 1:50; Labvision), or β-Gal (Promega) at 4°C overnight. After PBS rinse, tissue sections were sequentially incubated for 30 minutes at room temperature with diluted goat anti-rabbit or mouse Alexa-488- or Alexa-594–conjugated secondary antibody (1:200; Molecular Probes). Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI). Immunohistochemical staining of mouse mammary tissues was described previously (44). Immunohistochemical analysis of p-Akt at Ser473 (1:50; Cell Signaling Technologies) and p-ERα (Santa Cruz Biotechnology) were carried out as described previously (44).

Tip30^+/+ and Tip30^−/− primary mammary epithelial cells were isolated from 2-month-old Tip30^+/+ and Tip30^−/− female mice and cultured as previously described (14) and then serum-starved for 3 hours. Cells were incubated with 100 ng/mL of Alexa⁴⁸⁸-EGF (Invitrogen) and 20 μg/mL of cycloheximide on ice for 1 hour and then washed 4 times with cold PBS before being incubated in DMEM with 20 μg/mL of cycloheximide at 37°C for 2 hours. Cells were fixed in 4% paraformaldehyde in PBS for 15 minutes, permeabilized with 0.1% Triton X-100 for 2 minutes, and stained for EGFR. Images were obtained with a Zeiss LSM 510 Meta confocal microscope (Carl Zeiss), using Plan-Apochromat 63×/1.40 Oil objective. Pinhole size was set to 1 airy unit for all channels. All images are representative of single optical sections.

Comparisons among groups were analyzed by 2-sided t test or Fisher's exact test. A difference of P < 0.05 was considered to be statistically significant. All analyses were done with SPSS software, Version 11.5. Data are expressed as mean ± SEM.

To investigate whether Tip30 deletion cooperates with HER2/Neu to promote mammary tumorigenesis, we generated Tip30-knockout mice with overexpression of Neu by crossing the MMTV-Neu transgene from MMTV-Neu mice (FVB/N-Tg; ref. 18) into Tip30^−/− FBV mice. Neu⁺/Tip30^−/− mice appeared similar to Neu⁺/Tip30^+/+ mice in size and reached weaning age at the expected Mendelian frequency. A cohort of Neu⁺/Tip30^−/−, Neu⁺/Tip30^+/−, and Neu⁺/Tip30^+/+ female mice were monitored for 75 weeks. Mammary tumors were noted to appear earlier in Neu⁺/Tip30^−/− mice than in Neu⁺/Tip30^+/+ mice. Kaplan–Meier survival curves (Fig. 1A) were generated on the basis of the time of palpable tumor formation. We observed that 50% of Neu⁺/Tip30^+/+ mice developed mammary tumors with a relatively long median latency of 58 weeks. The relatively longer median latency and lower frequency of tumors arising in Neu⁺/Tip30^+/+ mice than those in MMTV-Neu mice (18) are possibly due to only 1 MMTV-Neu wild-type transgene allele in Neu⁺/Tip30^+/+ mice. In contrast, all Neu⁺/Tip30^−/− mice developed tumors at a median age of 37 weeks and 87% of Neu⁺/Tip30^+/− mice developed tumors at a median age of 45 weeks. Histologic analysis showed that Neu⁺/Tip30^−/− tumors were poorly or moderately differentiated mammary tumors with solid or glandular growth patterns, which are morphologically similar to the mammary tumors arising in MMTV-Neu mice (Fig. 1B1 and B2). Immunohistochemical staining of paraffin-embedded tumor sections revealed that Neu⁺/Tip30^−/− or Neu⁺/Tip30^+/+ tumor cells were mostly Keratin (K8)-positive and α-smooth muscle actin (αSMA)-negative (Fig. 1C), indicating that Neu⁺/Tip30^−/− tumors are of the luminal cell type similar to MMTV-Neu tumors (19). The presence of metastasis in the lung was observed in 4 of 10 Neu⁺/Tip30^−/− mice, whereas metastasis was detected in 1 of 10 Neu⁺/Tip30^+/+ mice. These results suggest that Tip30 loss accelerates the onset of mammary luminal tumors in MMTV-Neu mice and possibly increases metastasis.

It is well known that MMTV-Neu transgenic mice develop mammary tumors composed almost exclusively of ER⁻/PR⁻ luminal epithelial cells. Surprisingly, we found that all Neu⁺/Tip30^−/− tumors (n = 8) and 50% of Neu⁺/Tip30^+/− tumors (n = 6) examined showed an ER+/PR− staining pattern (Fig. 2B) whereas 89% Neu⁺/Tip30^+/+ tumors (n = 9) were ER−/PR− (Fig. 2C), indicating that Neu⁺/Tip30^−/− mice were more likely to develop ER+/PR− mammary tumors than Neu⁺/Tip30^+/+ mice (Table 1, 100% vs. 11%; P = 0.004). These results suggest that Tip30 loss combined with the activation of Neu promotes development of ER⁺/PR⁻ mammary tumors.

To determine whether ER+/PR− mammary tumors arising in Neu⁺/Tip30^−/− were ovarian hormone dependent, we first transplanted small pieces of freshly dissected tumors into ovary-intact (non-OVX) and ovariectomized (OVX) nude mice and then monitored the growth of transplanted tumor tissues. Remarkably, removal of both ovaries from recipient mice drastically reduced growth and progress of transplanted tumors, suggesting that ER+/PR− mammary tumors that developed in Neu⁺/Tip30^−/− mice are ovary dependent (Fig. 3A). Next, we transplanted small pieces of freshly dissected tumors into OVX mice supplemented with placebo, estrogen, progesterone, or estrogen plus progesterone pellets (Fig. 3B and Supplementary Fig. 1). Surprisingly, estrogen plus progesterone strongly promoted tumor growth compared with placebo (P = 0.04) whereas estrogen or progesterone alone only slightly increased tumor growth compared with placebo (P > 0.05). Moreover, the progesterone antagonist RU486 could significantly delay the growth of Neu⁺/Tip30^−/− tumors (Fig. 3C). These results suggest that both estrogen and progesterone are required for promoting growth of ER+/PR− tumors arising in Neu⁺/Tip30^−/− mice.

The effect of progesterone on the growth of ER+/PR− tumors from Neu⁺/Tip30^−/− mice raises the question of whether these tumor cells express any PR proteins. Because previous studies have suggested that active PRs are rapidly degraded in breast cells (20), we speculated that PR was expressed and then degraded rapidly in ER+/PR− tumors. To test this hypothesis, we examined PR-A and PR-B expression in cultured tumor cells derived from ER+/PR− tumors. Indeed, PR-A, but not PR-B, was clearly detected by Western blot analysis after cells were serum-starved and treated for 6 hours with the proteasome inhibitor MG132 (Fig. 3D), implying that PR-A is expressed but rapidly turned over in these tumors. Together, these results suggest that estrogen and progesterone play stimulating roles in the development of ER+/PR− mammary tumors in Neu⁺/Tip30^−/− mice.

The preceding data imply that Tip30 may play a key role in suppressing tumorigenesis in ERα-positive (ER⁺) epithelial cells. To test whether the Tip30 gene promoter is active in ER⁺ epithelial cells and ER+/PR− tumors, we carried out immunofluorescent double staining for ERα and β-galactosidase in the mammary glands and tumors derived from Neu⁺/Tip30^+/− mice harboring a knock-in β-galactosidase (β-Gal) gene at the Tip30 gene locus under the control of Tip30 promoter. The β-Gal protein was predominantly detected in ER⁺ mammary epithelial cells and tumor cells (Fig. 4A), indicating that the Tip30 promoter is activated mainly in ER⁺ mammary epithelial cells (MEC) and tumor cells. Given that ERα is activated through ligand binding and phosphorylation in response to estrogen- and growth factor–induced signaling and that Tip30^−/− mice do not exhibit a significant increase in the number of ERα-positive cells in the mammary gland at the age of 4 months (14), we therefore asked whether the proportion of phospho-ERα (p-ERα)-positive cells is altered in Neu⁺/Tip30^−/− mammary glands. Immunohistochemical analysis was used to examine phosphorylation of ERα at Ser171 (equivalent to Ser167 in human) in mammary glands and tumors from Neu⁺/Tip30^−/− and Neu⁺/Tip30^+/+ mice (Fig. 4B). No significant difference in the numbers of p-ERα–positive cells was detected between 2-month-old mammary glands from Neu⁺/Tip30^−/− and Neu⁺/Tip30^+/+ mice (Fig. 4C). Strikingly, 12-month-old Neu⁺/Tip30^−/− mammary glands and tumors displayed an increase in the number of p-ERα–positive cells compared with Neu⁺/Tip30^+/+ mammary glands and tumors (P < 0.05). These results suggest that Tip30 deletion preferentially increases the number of p-ERα–positive luminal cells in the mammary gland of MMTV-Neu mice.

Akt is one of the most important downstream factors in HER2/Neu and EGFR signaling pathways that can phosphorylate ERα and regulate MEC apoptosis (21, 22). Previous studies have shown that the expression of a constitutively active form of Akt-1 accelerates HER2/Neu-mediated mammary tumor formation (23) whereas disruption of Akt-1 delays HER2/Neu-mediated mammary tumorigenesis (24–26). To examine whether the deletion of Tip30 affects the activation of Akt (phospho-Akt, p-Akt) in preneoplastic mammary glands from MMTV-Neu mice, we carried out the immunohistochemical analysis for p-Akt in preneoplastic mammary glands from Neu⁺/Tip30^−/− and Neu⁺/Tip30^+/+ mice at the age of 2 and 12 months. MECs at the different ages exhibit negative, weak, intermediate, or strong staining for p-Akt (Fig. 5; A1, A2, A3, or A4, respectively). No significant difference in p-Akt expression levels and numbers of p-Akt–positive cells (P = 0.678 or 0.972, respectively) was detected between Neu⁺/Tip30^−/− and Neu⁺/Tip30^+/+ mammary glands at 2 months of age (Fig. 5B). However, at 12 months, the number of MECs having strongly positive p-Akt staining in Neu⁺/Tip30^−/− mammary glands was significantly increased compared with that in Neu⁺/Tip30^+/+ mammary glands (strong staining in Neu⁺/Tip30^−/− mammary gland: 41.4%; strong staining in Neu⁺/Tip30^+/+ mammary gland: 9.9%; P = 0.02; Fig. 5C). However, there was no significant difference in the levels of p-Akt between mammary tumors from Neu⁺/Tip30^−/− and Neu⁺/Tip30^+/+ mice (Fig. 5D). These data indicate that the relatively earlier onset of enhanced Akt activation in the mammary glands due to Tip30 loss may contribute to accelerated mammary tumorigenesis in Neu⁺/Tip30^−/− mice.

Upon binding EGF, EGFR proteins are rapidly internalized and localized in early endosomes, where they are either sent back to the plasma membrane or sorted into late endosomes and lysosomes for destruction (27, 28). Early endosomes serve as a platform for signaling receptors to activate specific downstream signaling until ligand–receptor dissociation occurs due to early endosomal acidification mediated by vacuolar (H⁺)-ATPases (29, 30). Recently, we have shown that TIP30 regulates EGFR signaling by controlling endocytic downregulation of EGFR in primary hepatocytes and liver cancer cells. Tip30 deletion impairs the fusion of Rab5 vesicles carrying vacuolar (H⁺)-ATPases with early endosomes that contain internalized EGF and EGFR, leading to delayed EGFR degradation and sustained EGFR signaling (C. Zhang, A. Li, X. Zhang, and H. Xiao, submitted manuscript). Therefore, we questioned whether the increased phosphorylation of Akt and ERα in Neu⁺/Tip30^−/− mammary gland is also caused by a similar mechanism. First, we measured the protein levels of EGFR in mammary tumors cells isolated from Neu⁺/Tip30^−/− and Neu⁺/Tip30^+/+ mammary tumors in response to EGF treatment at various time points after EGF internalization. We used an experimental approach that eliminates the interference from continuous ligand internalization and nascent protein synthesis to measure endocytic degradation of EGFR. The comparison revealed that endocytic degradation of EGFR was significantly delayed in Neu⁺/Tip30^−/− mammary tumors cells compared with Neu⁺/Tip30^+/+ mammary tumors cells, indicating that Tip30 deletion impairs endocytic degradation of EGFR (Fig. 6A and B).

To determine whether Tip30 deletion can block EGFR trafficking from early endosomes to lysosomes for degradation, we tracked Alexa-488–conjugated EGF (Alexa⁴⁸⁸-EGF) and EGFR in normal primary MECs isolated from Tip30^−/− and Tip30^+/+ mice. The majority of internalized EGFs dissociated from EGFR in wild-type MECs 2 hours after EGF internalization. In contrast, they remained associated with EGFR in Tip30^−/− MECs (EGF-EGFR colocalization in wild-type primary MECs: 11%; EGF–EGFR colocalization in Tip30^−/− primary MECs: 55%; n = 20, P = 0.004; Fig. 6C and D), indicating that Tip30 deletion causes the trapping of EGF–EGFR complex in endosomes and sustained endosomal EGFR signaling. To rule out the possibility that Tip30 deletion increased Neu transgene expression at the level of transcription, we used quantitative reverse transcription-PCR (RT-PCR) to examine the mRNA expression of Neu transgene in 5- to 9-week-old Neu⁺/Tip30^+/+ and Neu⁺/Tip30^−/− mice and found no significant difference (data not shown). Together, these results suggest that Tip30 loss may prolong EGFR signaling, which cooperates with Neu activation to enhance Akt activation and to promote the formation of ER+/PR− tumors.

This study was designed to investigate the relationship between HER2/Neu overexpression and Tip30 deletion in mammary tumorigenesis by using genetically engineered mice containing both Tip30 deletion and an MMTV-Neu transgene. Strikingly, the data show that Tip30 deletion cooperates with Neu overexpression to promote exclusive development of ER+/PR− mammary tumors in mice. In addition, we show that Tip30 loss impairs endocytic trafficking of EGF–EGFR, delays EGFR degradation in primary MECs and tumor cells, and enhances Akt and ERα phosphorylation in the mammary gland. These findings, combined with our recent observation that TIP30 formed a protein complex with ACSL4 and EndoB1 to control EGF–EGFR endocytic trafficking in hepatocytes (C. Zhang, A. Li, X. Zhang, and H. Xiao, submitted manuscript), strongly suggest a novel mechanism by which the loss of Tip30 contributes to the development of ER+/PR− tumors, at least in part, through enhancing EGFR signaling in ER⁺ MECs.

It is not immediately obvious why Tip30 deletion combined with Neu overexpression causes the exclusive development of ER+/PR− mammary tumors. The observation that the promoter of Tip30 was predominantly active in ER⁺ MECs suggest that Tip30 deletion may mainly affect the proliferation of ER⁺ cells by inducing enhanced ERα activities, thereby selecting ER⁺ cells to initiate tumorigenesis. Indeed, ER⁺/PR⁺ mammary tumors developed spontaneously in 22% of aged Tip30-knockout female mice in the BALB/c genetic background (A. Li, C. Zhang, S. Gao, R. Luo, and H. Xiao, unpublished data) and TIP30 could inhibit ERα-mediated transcription (15). The correlation between progressively increased p-Akt and p-ERα–positive cells in Neu⁺/Tip30^−/− mammary glands observed in this study implies that Tip30 deletion may promote the development of ER⁺ mammary tumors by enhancing Akt activation and increasing active ERα-positive cells. Consistent with this scenario, a previous study showed that Akt overexpression could increase the intensity of ERα staining, the number of ERα-positive cells, and the frequency of ER⁺ tumors in DMBA-treated mice (31), although it did not show whether these tumors were PR positive. Moreover, expression and phosphorylation of ERα in ER⁺ human breast cells are enhanced by the activation of Akt (21, 31). It should be noted that enhanced Akt activation alone is insufficient for driving the tumorigenic process in mouse MECs in vivo (32); therefore, other mechanisms such as increased expression of c-Myc and IGF-1 induced by Tip30 deletion may also contribute to the formation of ER+/PR− mammary tumors in MMTV-Neu mouse models (14, 15).

Even thought tumors arising in Neu⁺/Tip30^−/− mice were stained negatively for PRs and did not display significantly more PR-A/PR-B mRNAs according to our quantitative RT-PCR analysis (data not shown), these ER+/PR− tumors were sensitive to progesterone stimulation and RU486 inhibition, and PR-A proteins were detectable in cultured ER+/PR− tumor cells when proteasomes were inhibited. One explanation for these observations is that PR-A is expressed in ER+/PR− tumors but rapidly turns over due to enhanced activation of EGFR and HER2/Neu in vivo. This explanation is supported by previous reports that PR degradation by the 26S proteasome is mediated by MAPK/ERK-induced phosphorylation at Ser294 in cultured breast cancer cells (20). It has been shown that ERK1/2 activation in human cancer cells induces PR-B Ser294 phosphorylation and blocks PR-B sumoylation, which leads to 2 coupled events, hyperactive transcription activity and rapid turnover of PR-B proteins. PR-A was shown to be relatively resistant to these events compared with PR-B (20, 33–35). In agreement with these results, we also observed significantly enhanced ERK1/2 activation at Neu⁺/Tip30^−/− tumor periphery compared with Neu⁺/Tip30^+/+ tumor periphery (Supplementary Fig. 2). Notably, we detected only PR-A after treatment with proteasome inhibitor (Fig. 3D), possibly because PR-A is the dominant form in the mammary glands of mature virgin mice whereas human breast cells express both PR-A and PR-B (36). Our observation that ER+/PR− tumors were responsive to progesterone and RU486 indicates a critical role of progesterone signaling in the growth of ER+/PR− tumors and implicates that intervention of ER+/PR− breast cancers may be achieved, in part, through suppression of PR function.

Our data seem to support the hypothesis that ER⁺ breast cancers arise from ER+ or otherwise estrogen-responsive progenitor cells (37). However, our data do not exclude the possibility that Tip30 deletion may cause ER⁻/PR⁻ cells to reexpress ERα or promote transformed ER⁻/PR⁻ luminal progenitor cells to differentiate to ER+/PR− epithelial cells. Studies on the origin of tumor cells in MMTV-Neu mice have suggested that tumor cells from this model originate from transformed luminal progenitor cells committing to ER⁻/PR⁻ cells (38). Therefore, the cell origin of ER+/PR− tumors arising in Neu⁺/Tip30^−/− mice remains to be determined.

Currently, there remains a profound need for more effective therapies for treating HER2⁺/ER+/PR− breast cancers because of their poor response and development of resistance to existing therapies. Nonetheless, the majority of preclinical studies of ER-positive breast cancer have relied on cultured cell lines or on xenograft tumor models, in which breast tumor development and progression do not accurately represent clinical human breast cancer. Alternatively, the use of genetically engineered mouse models of breast cancer has major advantages for investigating the molecular mechanism of mammary tumorigenesis as well as developing anticancer agents. To date, there have been many mouse mammary cancer models generated by overexpression or deletion of specific genes that are associated with human breast cancer. Unfortunately, most mammary tumors arising in those animal models are ER⁻/PR⁻ and do not morphologically resemble the major subtype of human breast cancer (ER⁺ ductal carcinoma); ER⁺ mammary tumors are observed in only a few genetically engineered mouse models (39–42). To our knowledge, there has been no animal model of HER2⁺/ER+/PR− mammary tumors reported. Therefore, our mouse model of HER2⁺/ER+/PR− breast cancer provides a valuable tool for deciphering the mechanisms of HER2⁺/ER+/PR− breast cancer development and for testing single or combination therapies.

No potential conflicts of interest were disclosed.

We are grateful to Jill Pecha, for backcrossing Tip30 deletion allele into FBV mice, and Ryan Brooks, Adam C. Edmunds, George Chen, for the excellent technical assistance. We thank Ying Qin for histopathologic expertise and Eran Andrechek for critical reading of the manuscript.

The contents of this work are solely the responsibility of the authors and do not necessarily represent the official views of the NIDDK, National Institute of Environment Health Science, or National Cancer Institute, National Institutes of Health.

This work was supported by grants RO1 DK066110-01 and W81XWH-08-1-0377 (to H.X.) from the NIDDK and Department of Defense and grant U01 ES/CA 012800 (to S.Z.H.) from the National Institute of Environment Health Science and the National Cancer Institute, National Institutes of Health, Department of Health and Human Services.

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.