Hyperinsulinemia Promotes Metastasis to the Lung in a Mouse Model of Her2-mediated Breast Cancer

Abstract

The Her2 oncogene is expressed in approximately 25% of human breast cancers and is associated with metastatic progression and poor outcome. Epidemiological studies report that breast cancer incidence and mortality rates are higher in women with type 2 diabetes. Here we use a mouse model of Her2-mediated breast cancer on a background of hyperinsulinemia to determine how elevated circulating insulin levels affect Her2-mediated primary tumor growth and lung metastasis. Hyperinsulinemic (MKR^+/+) mice were crossed with doxycycline-inducible NeuNT (MTB/TAN) mice to produce the MTB/TAN/MKR^+/+ mouse model. Both MTB/TAN and MTB/TAN/MKR^+/+ mice were administered doxycycline in drinking water to induce NeuNT mammary tumor formation. In tumor tissues removed at two, four and six weeks of Neu-NT overexpression, we observed increased tumor mass and higher phosphorylation of the insulin receptor (IR)/insulin-like growth factor receptor 1 (IGF-1R), suggesting that activation of these receptors in conditions of hyperinsulinemia could contribute to the increased growth of mammary tumors. After 12 weeks on doxycycline, although no significant further increase in tumor weight was observed in MTB/TAN/MKR^+/+ compared to MTB/TAN mice, the number of lung metastases was significantly higher in MTB/TAN/MKR^+/+ mice compared to controls (MTB/TAN/MKR^+/+ 16.41 ± 4.18 vs. MTB/TAN 5.36 ± 2.72). In tumors at the six week time-point, we observed an increase in vimentin, a cytoskeletal protein and marker of mesenchymal cells, associated with epithelial-to-mesenchymal transition and cancer associated fibroblasts. We conclude that hyperinsulinemia in MTB/TAN/MKR^+/+ mice resulted in larger primary tumors, with more mesenchymal cells and therefore, more aggressive tumors with more numerous pulmonary metastases.

Introduction

Over the last three decades the relationship between type 2 diabetes and breast cancer has been evaluated by numerous epidemiological studies. Most of these support a positive association between diabetes and breast cancer, with several recent meta-analyses suggesting that women with type 2 diabetes are at significantly greater risk of developing, relapsing with and dying from breast cancer than women who do not have diabetes (1);(2);(3). I ASSUME THIS IS NOT THE FINAL WAY THE REF WILL APPEAR????The relationship between insulin, specifically, and breast cancer has been assessed by several epidemiological studies. Independent of all other confounding factors, hyperinsulinemia was reported to be a significant factor for incident breast cancer (4), (5) whilst another study reported an association between raised insulin levels and breast cancer metastasis and mortality (6). Raised C-peptide (a marker of secreted insulin levels) and insulin resistance have both been recently reported to be linked to breast cancer specific death (7, 8). High circulating insulin levels have also been associated with an risk of breast cancer recurrence (9). After breast tumor surgery, elevated insulin levels in the circulation have been reported to be associated with an adverse prognosis during the first five years after diagnosis (10). While epidemiological studies only suggest a link between hyperinsulinemia and accelerated tumor growth, some proof-of-concept experimental studies indicate that hyperinsulinemia indeed promotes tumor development {Novosyadlyy, 2010 #113}{Fierz, 2010 #78}.

In our previous studies we have focused specifically on the impact of hyperinsulinemia on breast cancer progression using the MKR mouse model. Female MKR mice are insulin resistant and hyperinsulinemic whilst remaining non-obese and only mildly hyperglycemic. Using this hyperinsulinemic model we have observed significantly increased growth of transgenic (PyVmT) and orthotopically-induced mammary tumors and have demonstrated a mechanistic link between hyperinsulinemia and increased cancer progression through increased activation of the insulin receptor (IR)/insulin-like growth factor-I receptor (IGF-IR) and the phosphatidylinositol 3-kinase (PI3K)/Akt/mTOR pathway (11).

In the clinical setting, the differential influence of type 2 diabetes and its complications on subtype-specific breast cancer warrants further investigation. Around 25% of breast cancers belong to the subtype characterized by amplification of the gene encoding human epidermal growth factor receptor 2 (Her2), which is associated with high risk of metastasis and poor outcome (12) (13). Recently, a retrospective study has evaluated the benefit of the antidiabetic treatments metformin and thiazolidinediones on women with type 2 diabetes and Her2+ breast cancer. Initial analysis of the study cohort revealed that type 2 diabetes was a significant predictor of reduced overall survival in women with stage 2 or higher Her2+ breast cancer, independent of other confounding factors such as age, estrogen receptor (ER)/progesterone receptor (PR) expression and BMI. Furthermore, the antidiabetic therapies significantly increased overall survival and significantly reduced the risk of Her2+ breast-cancer specific mortality (14). Although insulin levels in these patients were not reported, metformin and thiazalidenidiones are both insulin-sensitizers, suggesting that improvements in insulin resistance in type 2 diabetics could have an impact on Her2+ breast cancer progression.

Her2 (also known as ErbB2 or Neu in rodents) belongs to the epidermal growth factor receptor family of receptor tyrosine kinases (RTK), which includes Her1 (also known as EGFR or ErbB1), Her3 (or ErbB3) and Her4 (or ErbB4). Unlike the other family members, no specific ligand has been identified for Her2 and its activity is dependent on its dimerization with either ligand-activated EGFR or Her3. At high expression levels resulting from ERBB2 amplification, homodimers are also activated (15); (16). Despite the lack of kinase activity of Her3, dimers of Her2/Her3 constitute the most potent signaling combination of all EGFR family dimers (17) (18), and a prevention of Her2/Her3 heterodimerization provides a significant clinical benefit in patients with Her2+ breast cancer (19). Activation of Her2/Her3 dimers leads to upregulation of multiple downstream pathways including the canonical PI3-K/Akt/mTOR signaling cascade and the mitogen-activated protein kinase (MAPK) pathway (20). Importantly, these two pathways are also the principal signaling pathways involved in the growth promoting effects of the activated IR/IGF-IR in tumorogenesis. Indeed, Her2 also dimerizes with the IGF-IR leading to the emergence of resistance to Her2 pharmacotherapies (21).

Transgenic mouse models of Her2 (Neu)-mediated mammary carcinogenesis include those with constitutive activation of wild-type Neu (c-Neu) or oncogenic (activated) Neu (Neu-NT) under control of the MMTV promoter. Both c-Neu and Neu-NT overexpression result in invasive mammary carcinomas with latency periods of around seven and three months, respectively (22). A conditionally activated model of mammary-specific Neu-NT has also been engineered by crossing MMTV-reverse tetracycline transactivator (rtTA) (MTB) transgenic mice with mice bearing the TetO-Neu-NT transgene (TAN) to generate MTB/TAN offspring (23). Tumor latency in this model is short, with development of multiple mammary tumors with 100% penetrance within a few weeks and further progression to spontaneous lung metastasis (23).

Given that in mammary tumors from Neu-NT transgenic mice the PI3-K/Akt/mTOR signaling pathway should already be active, an interesting question that arises is whether the induction of systemic hyperinsulinemia in these mice could enhance Neu-NT-mediated tumor growth via IR/IGF-IR activation. To address this issue we crossed homozygous MKR^+/+ mice with MTB/TAN mice to yield MTB/TAN/MKR^+/+ offspring, thus generating hyperinsulinemic mice expressing a doxycycline-inducible Neu-NT transgene. When induced with doxycycline, MTB/TAN/MKR^+/+ mice develop early Neu-NT- mediated mammary gland hyperplastic changes more rapidly than controls and go on to develop larger mammary tumors. Furthermore, MTB/TAN/MKR^+/+ mice exhibit higher numbers of lung macrometastases suggesting that chronic hyperinsulinemia can augment Neu-NT-mediated primary tumor growth as well as the progression to lung metastasis.

Materials and Methods

Animal studies

Animal care and maintenance were provided by the Mount Sinai School of Medicine AAALAC Accredited Animal Facility. All procedures were approved by the Institutional Animal Care and Use Committee of the Mount Sinai School of Medicine according to the National Institute of Health Guidelines. All mice used in this study were on Friend Virus B (National Institue of health) (FVB/N) genetic background. Mice were housed four per cage in a clean mouse facility and fed a standard mouse chow (PicoLab Rodent Diet 20, 5053; LabDiet, Brentwood, MO, USA) ad libitum on a 12-hour light/dark cycle. Plasma insulin levels were measured by the sensitive rat insulin radioimmunoassay (RIA) kit (Millipore, St. Charles, MO, USA). An insulin tolerance test was performed on animals previously fasted for four hours. Insulin (0.75 Units/kg of body weight) was injected intraperitoneally and blood glucose values were measured immediately before and 15, 30 and 60 min after insulin injection. For induction of Neu-NT, mice were administered 1.5mg/ml doxycycline (Sigma-Aldrich, St Louis, MO, USA) in drinking water for the duration of the period of tumor growth from two weeks up to twelve weeks. Doxycycline water was changed twice per week. Mice were followed on a daily basis and body score conditions were recorded. To determine tumor mass, each animal was euthanized and mammary tumors from all thoracic and inguinal glands were carefully dissected and weighed. For analysis of pulmonary metastases, mice were sacrificed and lungs were inflated via the trachea with 10% formalin, removed and examined for macrometastatic lesions.

Mammary gland whole mount analysis

Inguinal mammary glands were removed, placed on a glass slide and fixed for four hours in Carnoy’s fixative (60% ethanol (100%), 30% chloroform and 10% glacial acetic acid). Glands were serially hydrated in 100%, 95%, 70%, 50% and 30% ethanol for 15 minutes each, rinsed in water for 5 minutes and stained overnight with carmine alum. Glands were then serially dehydrated in 30%, 50%, 70%, 95%, and 100% ethanol for 15 minutes each and cleared overnight in xylene. Glands were then covered by Mount-Quick® mounting medium (Daido Sangyo, Japan), before a glass cover slip was placed on top. Photographs were carried out using a stereomicroscope (Zeiss, Thornwood, NY, USA) at 4X magnification. Quantification of the relative area of end buds was performed using Image J by measuring the mean end bud area of four random images at 4X objective taken from each whole mount.

Histology and immunofluoresence

Lungs were fixed in 10% formalin before being embedded in paraffin and sectioned and stained using haematoxylin & eosin (H & E). For immunofluorescence studies, mammary tumors were cut in cross section at time of sacrifice, fixed in 10% formalin before being embedded in paraffin and sectioned. Five-micron sections were deparaffinized, rehydrated and subjected to antigen retrieval. Primary antibodies used were a rabbit polyclonal antibody to Vimentin (Cell Signaling Cell Signaling Technologies, Danvers, MA, USA) and a mouse monoclonal antibody to Neu (Abcam, Cambridge, MA, USA). Secondary antibodies used were AlexaFluor-568-conjugated goat anti-rabbit IgG and Alexa-Fluor-488-conjugated goat anti-mouse IgG (Invitrogen, Molecular Probes, Eugene, OR, USA). Nuclei were counterstained with 0.2 μg/ml 4′,6-diamidino-2-phenylindole (DAPI), (Sigma-Aldrich, St Louis, MO, USA).

Western Blotting

Tumor tissues were lysed in chilled lysis buffer (pH 7.4) containing 50mM Tris, 150mM NaCl, 1mM EDTA, 1.25% CHAPS, 1mM sodium orthovanadate, 10mM sodium pyrophosphate, 8mM B-glycerophosphate and Complete Protease Inhibitor Cocktail tablet. Protein concentration of samples was measured using the BCA protein assay kit (Thermo Scientific, Rockford, IL, USA). Protein samples were resuspended in 3 x loading buffer containing DTT (Cell Signaling Technologies, Danvers, MA, USA) and denatured by boiling for 5 minutes at 96°C. Samples were then subjected to SDS polyacrylamide gel electrophoresis (8% or 8–16% Tris-glycine gel; Life Technologies, Grand Island, NY, USA) and transferred to a nitrocellulose membrane. Membranes were probed with the appropriate primary antibodies: anti-phospho IR-β^(Y1150/51)/IGF-IRβ^(Y1135/36), anti-phospho Akt^Ser473, anti-total Akt, anti-vimentin (Cell Signaling Technologies, Danvers, MA, USA), and then re-blotted with B-actin (Sigma-Aldrich, St. Louis, MO, USA) or anti-IRβ (Santa Cruz Biotechnology, Santa Cruz, CA, USA) before being incubated with secondary antibodies (LI-COR Biosciences, Lincoln, NE, USA) and being exposed to the LI-COR infrared detection system (LI-COR Biosciences, Lincoln, NE, USA).

Statistical Analysis

Statistical analyses were conducted using the Student’s t-test. Results are expressed as means ± SEM.

Results

MTB/TAN/MKR^+/+ mice express the same metabolic phenotype as the parental MKR^+/+ mouse strain

Transgenic MMTV-rtTA (MTB) and TetO-Neu-NT (TAN) mice were previously crossed to yield bitransgenic MTB/TAN offspring demonstrating doxycycline-dependent Neu-NT expression in luminal mammary epithelial cells and subsequent development of gross mammary tumors and pulmonary macrometastases (23). We crossed MTB/TAN with homozygous MKR mice to yield MTB/TAN/MKR^+/+ mice. Female MTB/TAN/MKR^+/+ mice showed the same metabolic abnormalities as the homozygous female MKR^+/+ mice (11) namely lowered body weight, mild hyperglycemia, increased systemic insulin and severe insulin resistance (Figure 1A–D).

Figure 1. Metabolic characterization of the MTB/TAN/MKR^+/+ mouse model.

(A) Body weight (MTB/TAN, n=14; MTB/TAN/MKR^+/+, n=10), (B) blood glucose (MTB/TAN, n=30; MTB/TAN/MKR^+/+, n=22), and (C) serum insulin (MTB/TAN, n=13; MTB/TAN/MKR^+/+, n=20) of MTB/TAN and MTB/TAN/MKR^+/+ mice at 8 weeks of age. (D) An insulin tolerance test was performed on fasted 8 week-old MTB/TAN/MKR^+/+ (n=5) and MTB/TAN mice (n=5) after intraperitoneal injection of insulin (0.75 Units/kg). Blood samples were obtained from the tail vein and glucose concentrations were determined at the indicated time points. Graphs represent mean values of each group, error bars represent the SEM. *P value <0.05.

MTB/TAN/MKR^+/+ mice show augmented Neu-NT-induced abnormalities of the mammary gland at eight weeks compared to MTB/TAN controls

Others have shown that doxycycline-induced Neu-NT expression results in noticeable hyperplastic abnormalities in the mouse mammary gland such as the presence of cellular masses along the length of ducts, as well as terminal end bud enlargement due to the presence of acinar-like structures (23). We investigated whether or not hyperinsulinemia would have an additive effect on these early Neu-NT-mediated hyperplastic abnormalities by induced Neu-NT expression for three days in eight week-old MTB/TAN/MKR^+/+ and MTB/TAN mice, euthanizing the animals and comparing mammary gland morphology by whole mount analysis. Since MKR^+/+ mice have precocious mammary gland growth and differentiation compared to control mice at both three and fifteen weeks (24), we also performed whole mount analysis and H&E staining of mammary glands from eight week-old MTB/TAN/MKR^+/+ and MTB/TAN mice which had not been administered doxycycline (Fig 2A and 2B, upper panels). In MTB/TAN/MKR^+/+ mice, we found Neu-NT-mediated abnormalities to be enhanced compared to MTB/TAN controls, with larger cellular masses growing on ducts, greater terminal end bud enlargement and precocious lobular development (Fig 2A, lower panels and Fig 2C). This finding was further confirmed by histological evaluation of H&E stained sections of the mammary gland (Fig 2B, lower panels) where we observed advanced ductal hyperplasia in MTB/TAN/MKR^+/+ mice compared to MTB/TAN controls.

Figure 2. Effect of hyperinsulinemia on Neu-NT-mediated mammary gland hyperplasia.

(A) Whole mount analysis of mammary glands obtained from eight week-old female MTB/TAN and MTB/TAN/MKR^+/+ mice after three days of doxycycline treatment. (B) Histological analysis (H & E staining) of mammary glands after three days of doxycycline treatment. (C) Quantification of percentage area of mammary gland whole mounts composed of end buds (MTB/TAN, n=3; MTB/TAN/MKR, n=3. Arrows, End buds (A) and ductal hyperplasia (C) At least five animals per group were analyzed and the representative images are shown. Original magnification, × 4 (A) and × 40 (B). Graph represents mean for each group, error bars represent SEM. * P value <0.05

Neu-NT-induced mammary tumor burden is greater in MTB/TAN/MKR^+/+ mice than MTB/TAN mice, and is regulated through IR/IGF-IR signaling pathways

When administered doxycycline for several weeks, MTB/TAN mice developed multiple invasive mammary adenocarcinomas with tumors arising in all inguinal and thoracic glands (23). We administered doxycycline to eight-week-old MTB/TAN and MTB/TAN/MKR^+/+ mice for two, four, six or twelve weeks. Mice were euthanized at each time-point and the combined tumor weight from all inguinal and thoracic glands from each mouse was recorded. As shown in Table 1 and Figure 3A, MTB/TAN/MKR^+/+ mice exhibited significantly higher total tumor mass than MTB/TAN controls at all time-points with the exception of the twelve week time-point, suggesting that hyperinsulinemia augments Neu-NT-driven mammary tumor growth. We extracted proteins from tumor tissues to examine whether IR/IGF-IR activation was upregulated in MTB/TAN/MKR^+/+ mice compared to MTB/TAN controls. As shown in Figure 3B and 3C, mammary tumor tissues from MTB/TAN/MKR^+/+ mice demonstrated higher levels of phosphorylated insulin receptor/insulin-like growth factor receptor (IR-β/IGF-IRβ) after two, four and six weeks of Neu-NT upregulation, suggesting that these receptors may be involved in mediating the additional tumor growth present in MTB/TAN/MKR^+/+ mice. In MTB/TAN/MKR^+/+ and MTB/TAN mice which had been administered doxycycline for a period of 12 weeks, we observed an increase in IR/IGF-IR phosphorylation in mammary tissues of MTB/TAN/MKR^+/+ mice which was not statistically significant (Fig 3B and 3C). There was no significant difference in tumor weights after 12 weeks of doxycycline administration, possibly due to tumors reaching their physiologically maximal size after this time period (Figure 3A). We also compared activation of Akt, which lies downstream of both the IR/IGF-IR and Her2 and found that at 2 weeks there was a significant upregulation of phosphorylated Akt in tumors from MTB/TAN/MKR^+/+ mice compared to MTB/TAN mice (Figure 3D and 3E). At later stages of tumor development, levels of activation of Akt in the tumors appeared to be similar in the two groups.

Figure 3. Tumor growth, IRβ/IGF-IRβ and Akt signaling are enhanced in tumors from MTB/TAN/MKR^+/+ mice compared to MTB/TAN controls.

(A) MTB/TAN and MTB/TAN/MKR^+/+ mice were maintained on doxycycline for two, four, six or twelve weeks. Mice were euthanized and the combined tumor mass from all mammary glands was recorded. For the two week time-point, MTB/TAN, n=11; MTB/TAN/MKR^+/+, n=4. For the four week time-point, MTB/TAN, n=11; MTB/TAN/MKR^+/+, n=7. For the six week time-point, MTB/TAN, n=16; MTB/TAN/MKR^+/+, n=6. For the twelve week time-point, MTB/TAN, n=12; MTB/TAN/MKR^+/+, n=11. (B) Tumor tissue was subjected to Western blot analysis and probed with antibodies to phosphorylated IR β^(Y1150/51)/IGF-IRβ^(Y1135/36) and phosphorylated Akt^Ser473 (D) and then reblotted with antibodies to IRβ and Akt. Using densitometry, protein expression was quantified (C, E). Error bars represent SEM *, P<0.05. For the two week time-point, Western blot is representative of MTB/TAN, n=9; MTB/TAN/MKR^+/+, n=4 samples. For four and six week time-points, MTB/TAN, n=7; MTB/TAN/MKR^+/+, n=7 samples. For twelve week time-point MTB/TAN, n=10; MTB/TAN/MKR^+/+, n=10.

Neu-NT-induced spontaneous lung macrometastases is enhanced in MTB/TAN/MKR^+/+ mice after 12 weeks of tumor growth

MTB/TAN mice were reported to harbor Neu-NT-dependent spontaneous lung macrometastases (23). We removed the lungs of euthanized MTB/TAN/MKR^+/+ and MTB/TAN mice after two, four, six or 12 weeks of doxycycline induced Neu-NT expression in mammary epithelium and recorded the numbers of visible macrometastases. We also examined H&E stained sections of lung at all time-points for both micrometastases and macrometastases. After two, four or six weeks of Neu expression and mammary tumor growth, no visible macrometastases were observed in either MTB/TAN/MKR^+/+ or MTB/TAN mice. However, H&E staining revealed the presence of micrometastases in lungs of both groups of mice after six weeks of Neu expression (data not shown). Although there was a trend for the number of micrometastases to be increased in MTB/TAN/MKR^+/+ mice (0.3 micromets/lung section in MTB/TAN vs 0.8 micromets/lung section in MTB/TAN/MKR^+/+) this did not reach statistical significance. In contrast, after 12 weeks of Neu expression in mammary tissue, macrometastases could be clearly observed in the lungs, and these were significantly increased in MTB/TAN/MKR^+/+ compared to MTB/TAN mice (Fig 4A and 4B) suggesting that hyperinsulinemia increases the metastatic potential of Neu-driven mammary carcinogenesis.

Figure 4. Metastatic progression is enhanced in MTB/TAN/MKR^+/+ mice.

After 12 weeks on doxycycline, MTB/TAN and MTB/TAN/MKR^+/+ mice were euthanized, lungs removed and inflated and (A) number of lung macrometastases were recorded. (B) Lungs were paraffin-embedded, sectioned and stained with H&E to reveal macrometastases. Original magnification 4X. Arrows indicate metastatic cells.

Neu-NT-induced mammary tumors in MTB/TAN/MKR^+/+ mice express higher levels of vimentin

To determine whether or not hyperinsulinemia enhances the progression of Neu-NT-mediated primary tumors to lung metastases, we analyzed tumor tissue for the expression of vimentin protein, a marker of mesenchymal cells. As shown in Figure 5A and B, western blot analysis of tumor lysates demonstrated increased vimentin expression in tumors from 6 week old MTB/TAN/MKR^+/+ mice, compared to MTB/TAN mice. We additionally, analyzed vimentin expression in six-week tumor tissues by immunofluorescent staining of paraffin-embedded sections. As shown in Figure 5B and 5C, vimentin levels as measured by immunofluoresence were significantly elevated, indicating that more mesenchymal cells were present in MTB/TAN/MKR^+/+ mice at this timepoint. The vimentin positive cells generally did not stain positive for Neu, indicating that these were either cells that had undergone EMT and lost Neu expression, or were cancer associated fibroblasts. At four and 12 weeks of tumor development we did not observe significant differences in vimentin expression (data not shown).

Figure 5. Vimentin expression is enhanced in MTB/TAN/MKR^+/+ mice.

(A). After six weeks Neu-NT upregulation, mammary tumors were removed from MTB/TAN and MTB/TAN/MKR^+/+ mice, tumor lysates were subjected to western blot analysis and probed with anti-vimentin antibodies. Using densitometry, protein expression was quantified (B). Tumor tissue was paraffin embedded, sectioned and analyzed by immunofluorescence microscopy for the expression of vimentin and Neu proteins (C). Vimentin expression was analyzed using ‘Image J’ program to quantify expression of vimentin protein in six-week tumors from MTB/TAN and MTB/TAN/MKR^+/+ mice. Photographs are representative of at least four mice from each group. Five images were taken at the same position on every slide and each image was quantified. Original magnification 40X. Graphs represent the mean for each group, error bars represent SEM. * P value <0.05. WHAT ABOUT D, DO WE NEED IT WHATS IT MEASURING AND WHATS THE AXIS?

Discussion

Epidemiological evidence suggests a causal relationship between type 2 diabetes and the incidence, recurrence and mortality from breast cancer. Hyperinsulinemia has been identified as a specific factor that may drive breast cancer cell proliferation, but data in human patients which focuses on how specific subtypes of breast cancer respond to elevated circulating insulin are limited. Overexpression of ErbB2/Her2 is responsible for around 25% of all human breast cancers and results in aggressive primary tumors, which commonly metastasize to the lungs. In this study we have employed the MKR mouse model with a conditionally activated Neu-NT to represent this subtype of breast cancer. We have demonstrated for the first time in an animal model that hyperinsulinemia significantly affects the rate of Neu-NT-mediated mammary tumor development as well as the progression of breast cancer metastasis to the lungs.

Previously we have employed the MKR mouse model to demonstrate that PyVmT-driven mammary gland hyperplasia at 3 and 15 weeks’ postnatal development is more advanced under conditions of hyperinsulinemia (11). Hyperplasia of the mammary gland has previously been shown to occur rapidly (3–4 days) after conditionally-activated Neu-NT-expression (23). In MTB/TAN/MKR^+/+ mice, we observed that in the presence of elevated insulin, mammary gland hyperplasia was more advanced after 3 days, suggesting that hyperinsulinemia can enhance the mitogenic effect of Neu oncogenic transformation on terminal end buds during mammary growth and development. The MMTV-c-Neu mouse model has been previously used to determine the effect of a high fat diet (HFD) on mammary gland hyperplasia. Although this is a different model from ours, having a much longer tumor latency, it is interesting to note that either 5 or 10 weeks’ high fat diet treatment in FVB/N mice had no effect upon mammary gland hyperplasia in MMTV-c-Neu mice compared to controls suggesting that at very early stages of tumor development, insulin may have a more potent effect than dietary fat in promoting terminal end bud hyperplasia. In the same model it was shown that a HFD based on corn oil resulted in a shorter tumor latency than a HFD based on fish oils suggesting that at later stages of tumor development, fats may play a role in tumor development but individual dietary components have different effects (25).

We found that chronic (2–6 weeks) elevations of Neu-NT led to significantly larger mammary tumors in MTB/TAN/MKR^+/+ mice compared to controls, suggesting that hyperinsulinemia enhances Neu-NT-mediated tumor formation and growth. This result is in agreement with our previous finding that orthotopic inoculation of three different mammary tumor cell lines (one overexpressing Neu) each resulted in larger tumor formation in MKR mice compared to controls (11) (26) and demonstrates that Neu-NT-mediated mammary tumor growth specifically can be enhanced by hyperinsulinemia. The mechanism by which insulin could increase tumor growth in the Neu-NT model may be complex. Our results demonstrate that MTB/TAN/MKR^+/+ mice demonstrate higher levels of phosphorylated IR/IGF-IR than MTB/TAN control mice, suggesting that insulin is acting through its cognate receptor and/or the highly homologous IGF-IR to promote mammary tumor growth. In human breast cancer tissues phosphorylated IR/IGF-IR has been reported to be a prognostic marker of poor outcome for breast cancer, regardless of subtype (27). The ratio of IR isoform A (IR-A) to isoform B (IR-B) is also important in breast cancer, and recently a higher IR-A/IR-B ratio has been observed in the luminal B subtype of breast cancer (28). Breast cancer cells in culture have also been reported to express higher levels of the IR (29) and proliferate directly in response to insulin (30). The signaling pathways activated by the Neu-NT tyrosine kinase significantly overlap with those of the IR/IGF-IR, and include the canonical PI3K/Akt/mTOR and MAPK signaling pathways. It is thus possible that activation of the IR/IGF-IR causes amplification of the canonical Neu signaling pathways, which could possibly lead to increased tumor cell proliferation and/or survival.

Under conditions of hyperinsulinemia, IGF binding proteins (IGFBP)-1 and -2 are repressed (31) which may lead to an increase in circulating “free” IGF-1. In the MTB/TAN/MKR^+/+ mouse model, increased IGF-1 levels at the level of the target tissue could lead to increased activation of the IGF-IR, thus directly increasing tumor growth. The Neu receptor tyrosine kinase has no known ligand, depending for its activation upon dimerization with either ErbB3 or ErbB1 (32, 33). Interestingly, Neu also dimerizes with the IGF-IR, an occurrence which provides a significant source of resistance to Her2-mediated therapies in humans (Lu et al, 2001, 2004). Thus, in our MTB/TAN/MKR^+/+ mouse model, it is possible that elevations of either insulin or “free” IGF-1 could promote tumor growth through increased activation of Neu/IGF-IR hybrids (34).

Our finding of greater numbers of lung metastases in MTB/TAN/MKR^+/+ mice suggests that as well as promoting primary tumor growth, insulin may also enhance primary tumor progression and/or circulating tumor cell survival in the lung. Our previous work has demonstrated that murine mammary tumor cell line Mvt1 is able to form more lung metastases in MKR mice whether as a result of an orthotopic cell inoculation or intravenous injection (26). Evidence that the IR specifically is involved in either homing to, or survival in, the lung has come from a report by Zhang et al, who showed that shRNA knockdown of the IR in breast cancer cell line LCC6 resulted in reduced ability of these cells to form lung metastases following intravenous injection into nude mice (35). The epithelial to mesenchymal transition (EMT) permits tumor cells to gain the plasticity required to extravasate from the primary tumor site. During extravasation cells become increasingly mesenchymal in nature due to numerous changes in cytoskeletal scaffolding protein structures. Vimentin belongs to the intermediate filament (IF) family of proteins and has recently been shown to be an important marker of the EMT in epithelial cells which normally express only cytokeratin-type IFs (36). (Zeisberg JCI 2009). Vimentin is not specific for EMT, and is also a marker of cancer associated fibroblasts (Sugimoto H 2006). Additionally, recent studies have demonstrated that cancer associated fibroblasts induce EMT in breast cancer cells (Soon ERC 2012). In breast cancer cells, vimentin expression has been shown to be correlated with increased migration and invasion (37) (38) and in human breast cancer specimens, several studies have reported overexpression of vimentin as a marker of poor prognosis (39). In the present study, we observed elevated expression of vimentin as early as six weeks of Neu-NT upregulation. The majority of the vimentin positive cells did not co-stain with Neu, suggesting that either these are tumor cells that have undergone EMT and have lost Neu expression, or that the MTB/TAN/MKR^+/+ mice have a greater desmoplastic reaction to the tumors, with more numerous cancer associated fibroblasts. Our finding of hyperinsulinemia leading to greater lung metastasis could thus be due to insulin-mediated increases in dissemination of tumor cells from the primary tumor, due to a direct effect of insulin on the tumors, inducing EMT, or through an increase in the number of cancer associated fibroblasts. It is also possible that hyperinsulinemia may lead to greater cell survival, intravasation and proliferation in lung tissue.

In summary, we have shown that hyperinsulinemia promotes advanced mammary gland hyperplasia, primary tumor growth and lung metastasis in a Her2/Neu model of breast cancer. Further studies are required to determine whether, in promoting lung metastases, hyperinsulinemia enhances cell dissemination from the primary tumor, cell survival/adhesion in the lung, or both of these processes.