Insulin-Sensitizing Therapy Attenuates Type 2 Diabetes–Mediated Mammary Tumor Progression

All mice were on the FVB/N background. The generation and characterization of MKR^+/+ mice (5,25) as well as mouse mammary tumor virus (MMTV)–PyVmT^+/− mice (27) was described previously. The mice were housed in a clean mouse facility, had free access to a standard mouse chow (Picolab rodent diet 5053; LabDiet, St. Louis, MO) and fresh water ad libitum and were kept on a 12-h light/dark cycle. Animal care and maintenance were provided through the Mount Sinai School of Medicine Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC) accredited animal facility. All procedures were approved by the Animal Care and Use Committee of Mount Sinai School of Medicine.

Body weight and food intake were measured twice a week. Food intake was normalized to body weight and expressed as grams of food per grams of body weight^0.75 per day. Blood glucose levels were measured weekly in the nonfasting state between 9:00 a.m. and 12:00 p.m. with an automated glucometer (Elite; Bayer, Mishawaka, IN). Plasma and serum were obtained in the nonfasting state and collected in heparinized and nonheparinized capillary tubes, respectively. Plasma insulin levels were measured by a radioimmunoassay according to the manufacturer's instructions (Linco, St. Charles, MO). Serum leptin, adiponectin, tumor necrosis factor-α (TNF-α), and interleukin (IL)-6 levels were measured by ELISA (Millipore, Billerica, MA [leptin] and R&D systems, Minneapolis, MN [adiponectin, TNF-α, and IL-6]). Serum free fatty acids (FFAs) and serum triglycerides were measured by a colorimetric assay (Roche Applied Science, Indianapolis, IN [FFAs] and BioVision, Mountain View, CA [triglycerides]). Body composition was determined in nonanesthetized mice using an EchoMRI 3-in-1 NMR system (Echo Medical Systems, Houston, TX).

PyVmT^+/− male mice were interbred with MKR^+/+ or wild-type female mice to generate cohorts of PyVmT^+/− and PyVmT^+/−/MKR^+/+ female mice. CL-316243 (Sigma Aldrich, St. Louis, MO) was dissolved in sterile saline and administered daily intraperitoneally at a dose of 1 mg/kg body wt (30) from 3 to 6 weeks of age. Control mice received an equal amount of vehicle (sterile saline). After euthanasia, inguinal mammary glands (no. 4) were subjected to whole-mount analysis or immediately snap-frozen in liquid nitrogen for further studies.

Met-1 and MCNeuA mouse carcinoma cells were derived from MMTV-PyVmT (FVB/N) and MMTV-Neu (FVB/N) transgenic mice, respectively (28,29). The cells were allowed to grow until confluence in Dulbecco's modified Eagle's medium supplemented with 10% FBS and were detached by a nonenzymatic cell dissociation solution, and Met-1 cells (0.5 × 10⁶) or MCNeuA cells (10⁶) were injected into the left inguinal mammary fat pad (no. 4) of 8-week-old female MKR^+/+ and wild-type mice. One week after tumor cell inoculation, CL-316243 (1 mg · kg body wt⁻¹ · day⁻¹ i.p.) was administered for 21 days. Tumor growth was monitored by palpation and tumor volume was measured in a three-coordinate system using calipers. Tumor volume was calculated by the formula: 4/3 × π × r₁ × r₂ × r₃ (r = radius).

Tissues were lysed in buffer (pH 7.4) containing 50 mmol/l Tris, 150 mmol/l NaCl, 1 mmol/l EDTA, 1.25% CHAPS (Roche Applied Science), 1 mmol/l sodium orthovanadate, 2 mmol/l sodium fluoride, 10 mmol/l sodium pyrophosphate (Sigma Aldrich), 8 mmol/l β-glycerophosphate (VWR, West Chester, PA), and Complete Protease Inhibitor Cocktail (Roche). After denaturation, the proteins were subjected to SDS-PAGE (8% Tris-glycine gel; Invitrogen, Carlsbad, CA) and transferred to a nitrocellulose membrane (Bio-Rad, Hercules, CA). The membrane was sequentially blocked and probed with primary and secondary antibodies and then analyzed by direct infrared fluorescence detection using an Odyssey Infrared Imaging System (Li-cor, Lincoln, NE). Densitometric analysis was performed using MacBAS V2.52 software (Fuji PhotoFilm, Valhalla, NY). The antibodies were purchased from the following sources: Phospho-IRβ^Tyr1150/51/IGF-IRβ^Tyr1135/36 (Cell Signaling Technology, Danvers, MA) and IRβ (Santa Cruz Biotechnology, Santa Cruz, CA).

The no. 4 inguinal mammary glands were carefully excised, spread out on a glass slide, and fixed for 2–4 h in Carnoy fixative (60% ethanol, 30% chloroform, 10% glacial acetic acid). The fixed glands were hydrated in decreasing concentrations of ethanol (100, 95, 70, 50, and 30% for 15 min each), rinsed in double-distilled water, and stained overnight in carmine alum staining. After dehydration in increasing ethanol concentrations (30, 50, 70, 95, and 100% for 15 min each) and clearing in xylene overnight (Fisher Scientific, Pittsburgh, PA), the glands were covered by Mount-Quick mounting medium (Daido Sangyo, Tokyo, Japan) and photographic documentation was performed using a stereomicroscope and MicroSuite FIVE imaging software (Olympus, Center Valley, PA). Quantification of the hyperplastic mammary lesion as a ratio of the total glandular area was performed using ImageJ software (National Institutes of Health, Bethesda, MD).

RNA was extracted from Met-1, MCNeuA, and mammary epithelial cells (MECs) derived from FVB/N mice using the NucleoSpin RNA II kit (Clontech Laboratories, Mountain View, CA). After RT, the resulting cDNA was amplified by PCR using the following primers (Operon, Huntsville, AL): β₃-AR: forward: 5′ ATGGCTCCGTGGCCTCAC 3′ and reverse: 5′ CTGGCTCATGATGGGCGC 3′ (31); 18S rRNA: forward: 5′ TTGACGGAAGGGCACCACCAG 3′ and reverse: 5′ GCACCACCACCCACGGAATCG 3′.

Results are expressed as means ± SEM. Statistical analyses were conducted using ANOVA followed by a Fisher test, with P ≤ 0.05 considered significant. All analyses were performed using STATVIEW version 5.0 (SAS Institute, Cary, NC).

CL-316243 has potent antiobesity and antidiabetic effects in various rodent models (30,32,–34). The action of CL-316243 is highly specific to β₃-AR–expressing cells, which are predominantly the white and brown adipocytes (35,36). Chronic activation of the β₃-AR increases fatty acid oxidation and energy expenditure, thereby reducing adiposity. Subsequently, chronic β₃-AR stimulation leads to improved insulin sensitivity and reduced insulin levels (30,33,34).

To address whether CL-316243 has direct effects on the mammary epithelium or tumor cells, we determined the expression levels of the β₃-AR. As shown in Fig. 1 A, RT-PCR revealed undetectable levels of β₃-AR transcripts in normal mouse MECs and in the two mammary tumor cell lines: Met-1 and MCNeuA.

Chronic β₃-AR activation with CL-316243 leads to an increased formation of brown adipose tissue in the mammary gland (37). Thus, we assessed whether treatment with CL-316243 affects mammary gland development in vivo through alteration of the mammary fat pad. Four-week-old female wild-type mice were subjected to intraperitoneal injections of CL-316243 (1 mg · kg body wt⁻¹ · day⁻¹) or an equal volume of vehicle for 3 weeks. Whole-mount analyses of mammary glands obtained from these animals at the age of 7 weeks demonstrated no significant changes in mammary ductal outgrowth and side branching (Fig. 1 B).

These data suggest that CL-316243 has a minimal effect on mammary gland development and that a direct effect of CL-316243 on mammary tumor cells via activation of the β₃-AR is unlikely.

We have previously shown that chronic treatment with CL-316243 effectively reverses the diabetic phenotype in male MKR^+/+ mice, which, in contrast to female MKR^+/+ mice, develop overt type 2 diabetes (30). To test whether the predominantly hyperinsulinemic phenotype in female MKR^+/+ mice would respond in a similar manner, we treated 8-week-old female MKR^+/+ and wild-type mice chronically with CL-316243 (1 mg · kg body wt⁻¹ · day⁻¹) or vehicle for 3 weeks. Female MKR^+/+ mice exhibit reduced body weight and body adiposity compared with wild-type mice (5). As observed in our previous study in male wild-type and MKR^+/+ mice (30), we found that CL-316243 treatment did not affect body weight (Fig. 2,A), but increased food intake (Fig. 2,B) and decreased body adiposity (Fig. 2,C), predominantly affecting the gonadal fat pads (supplemental Fig. 1, available in an online appendix at http://diabetes.diabetesjournals.org/cgi/content/full/db09-1291/DC1). Additionally, we found that after 1 week of CL-316243 treatment, the moderately elevated blood glucose levels (∼20%) in MKR^+/+ mice were significantly lowered (Fig. 2,D). Importantly, we show that female MKR^+/+ mice display severely elevated plasma insulin levels, which were significantly reduced upon CL-316243 treatment, whereas no effect of the drug on insulin levels was seen in wild-type mice (wild type, vehicle: 0.42 ± 0.01 vs. wild type, CL-316243: 0.43 ± 0.02; MKR^+/+, vehicle: 13.12 ± 3.74 vs. MKR^+/+, CL-316243: 3.28 ± 0.85 ng/ml) (Fig. 2 E).

As chronic CL-316243 treatment lowers body adiposity in both wild-type and MKR^+/+ mice, serum levels of circulating lipids, adipokines, and proinflammatory cytokines were determined (Table 1). Serum levels of FFAs were unchanged in female MKR^+/+ mice compared with wild-type mice, whereas there was a moderate (∼30%) increase in triglyceride levels. Upon chronic treatment with CL-316243, FFAs and triglycerides were significantly reduced to a similar level in both MKR^+/+ and wild-type mice. Serum leptin levels were lower in MKR^+/+ mice compared with wild-type mice, and CL-316243 treatment significantly reduced leptin levels in both MKR^+/+ and wild-type mice. The levels of adiponectin and the proinflammatory cytokines TNF-α and IL-6 were not affected by CL-316243 treatment.

Taken together, these data show that chronic treatment with CL-316243 effectively reduces insulin levels in female MKR^+/+ mice alone while having a comparable effect on food intake, relative body adiposity, circulating lipids, and leptin levels in both wild-type and MKR^+/+ mice. This allows us to study the effect of insulin reduction on mammary tumor progression in MKR^+/+ mice.

To investigate the effect of insulin-sensitizing therapy on early stages of mammary tumor development, we used a PyVmT-induced transgenic tumor model (27). PyVmT^+/− transgenic mice express the PyVmT oncogene exclusively in the mammary epithelium under the control of the MMTV promoter. PyVmT-induced mammary tumors recapitulate many morphologic and pathophysiological processes found in human breast cancer, and their development begins with the formation of a single hyperplastic focus in the lateral part of the mammary gland at 3–4 weeks of age (38).

To assess the effect of antidiabetic therapy on mammary tumor development, we treated PyVmT^+/−/MKR^+/+ and PyVmT^+/− mice with CL-316243 (1 mg · kg body wt⁻¹ · day⁻¹) or vehicle from 3 to 6 weeks of age. The metabolic effects of CL-316243 treatment were similar to those observed in adult mice (supplementary Fig. 2). The inguinal mammary glands (no. 4) were dissected after treatment at 6 weeks of age. Whole-mount analyses of vehicle-treated PyVmT^+/−/MKR^+/+ mice revealed a larger area of hyperplastic mammary lesions (increase of 57%) compared with the PyVmT^+/− mice (Fig. 3,A and B) due to accelerated formation of hyperplasia in a hyperinsulinemic milieu. In contrast, pharmacological correction of hyperinsulinemia in PyVmT^+/−/MKR^+/+ mice resulted in a tumor phenotype similar to nondiabetic PyVmT^+/− mice (Fig. 3,A and B). Note that the CL-316243–treated nondiabetic PyVmT^+/− mice show no changes in mammary transformation. To determine whether the attenuation in tumor progression in the CL-316243–treated mice resulted from an inhibition of insulin signaling, we analyzed the phosphorylation of the IR/IGF-IR in the hyperplastic mammary tissue of vehicle-and CL-316243–treated mice (Fig. 3 C and D and supplementary Fig. 3). We found that hyperinsulinemia led to an increased activation of the IR/IGF-IR in transformed mammary tissue of PyVmT^+/−/MKR^+/+ mice, whereas pharmacological correction of insulin levels by CL-316243 treatment significantly reduced this effect. Taken together, we demonstrate that lowering insulin levels in type 2 diabetic PyVmT^+/−/MKR^+/+ mice reduces IR/IGF-IR activation and abrogates the accelerated formation of hyperplastic mammary lesions.

To investigate whether CL-316243 treatment would also affect growth of fully transformed tumor cells, we injected estrogen receptor–negative PyVmT-transformed mammary tumor cells (Met-1 cells [28]) orthotopically into the no. 4 mammary fat pad (0.5 × 10⁶ cells) of 8-week-old female MKR^+/+ or wild-type mice. Starting 7 days after tumor cell inoculation, when tumors were palpable, the animals were treated with CL-316243 (1 mg · kg body wt⁻¹ · day⁻¹) for 3 weeks. As shown in Fig. 4,A, tumor volume increased in vehicle-treated MKR^+/+ mice, whereas MKR^+/+ mice treated with CL-316243 displayed a significant reduction in tumor growth and were similar in size to wild-type controls. Importantly, CL-316243 treatment had no effect on tumor growth in the wild-type mice. These findings were confirmed by measuring wet tumor weight at the end of the study (Fig. 4 B).

To ascertain that our findings were not limited to PyVmT-induced mammary carcinogenesis, we also studied estrogen receptor–negative Neu-transformed mammary tumor cells (MCNeuA cells [29]) in a similar manner as the Met-1 cells. Neu is the rodent analog of the ERBB2 gene, which is amplified and/or overexpressed in 30% of human breast carcinomas (39). MCNeuA cells were inoculated orthotopically into the no. 4 mammary fat pad (10⁶ cells) and the treatment protocol and follow-up of tumor growth were performed as described above. As shown in Fig. 5 A and B, tumor volume and weight were increased in vehicle-treated MKR^+/+ mice compared with wild-type controls, however, after CL-316243 treatment and subsequent decreases in insulin levels, we observed a significantly reduced tumor size in MKR^+/+ mice. These data suggest that insulin-sensitizing therapy can abrogate insulin-mediated tumor progression independent of the tumor-inducing oncogene.

There is an increasing interest in the effect of insulin-sensitizing therapy on breast cancer risk and mortality. The current study demonstrates that treatment with the insulin-sensitizing drug CL-316243, a highly selective β₃-AR agonist, reverses diabetes-induced mammary tumor progression in hyperinsulinemic, type 2 diabetic MKR^+/+ mice. Our findings corroborate a major role of insulin in type 2 diabetes–mediated mammary tumor progression, which is in concert with our previous findings (5).

High concentrations of insulin are known to activate the IGF-IR (40) and our previous findings demonstrate that the IR, and to lesser extent the IGF-IR, are activated in hyperplastic mammary tumor tissue extracted from type 2 diabetic mice (5). Moreover, pharmacological blockade of the IR/IGF-IR using BMS-536924, a small-molecule tyrosine kinase inhibitor, led to abrogation of diabetes-induced mammary tumor progression in MKR^+/+ mice; however, it also worsened insulin resistance and hyperinsulinemia (5). Here we demonstrate that insulin-sensitizing treatment has a comparable effect on tumor progression but simultaneously exerts antidiabetic activity. Our results imply that elevated insulin levels should be targeted pharmacologically at early stages of the disease to lower breast cancer risk and progression in patients with type 2 diabetes.

The effect of antidiabetic medications on cancer development is poorly understood and no prior experimental studies have thoroughly investigated whether pharmacological reduction of insulin levels reduces mammary tumor growth in a type 2 diabetic organism. The reason for this is that the most widely used insulin-sensitizing drugs (metformin and TZDs) are known to exert direct, predominantly antineoplastic, effects on cancer cells, both in vitro and in vivo in nondiabetic animals (15,–17,19,,,–23,41). Thus, we decided to use CL-316243, a highly selective β₃-AR agonist with no known direct effect on mammary tumors, to study the consequence of lowering insulin levels on diabetes-mediated mammary tumor progression. However, it should be noted that β₃-AR agonists (unlike metformin and TZDs) have not yet been developed for clinical use in humans due to various problems including a lack of selectivity and poor pharmacokinetics (42).

Algire et al. (18) have reported that metformin attenuates the effect of high-energy diet–induced insulin resistance on growth of Lewis lung LLC1 carcinoma cells. Both, the insulin-sensitizing action as well as the direct antineoplastic activity of metformin may be involved in this finding. However, in the setting of excess energy intake and obesity, other factors besides hyperinsulinemia can promote tumor growth, such as adipokines, proinflammatory cytokines, adipose tissue–derived sex steroids, or altered levels of circulating carbohydrates and lipids (43), some of which may be affected by metformin treatment (44,45).

Thus, as an experimental model for insulin resistance independent of obesity, we used a lean, hyperinsulinemic mouse model of type 2 diabetes, the female MKR^+/+ mice. These mice develop severe insulin resistance and hyperinsulinemia but have only mild dysglycemia and thus recapitulate the early stages of type 2 diabetes. As these mice display reduced body adiposity and show no increase in serum FFAs, leptin, or the proinflammatory cytokines TNF-α or IL-6 and only a moderate elevation in serum triglycerides, they serve as an ideal model to study mammary tumor progression uncoupled from obesity-related tumor-promoting factors.

Apart from reducing insulin levels, chronic treatment with CL-316243 also decreases blood glucose levels in MKR^+/+ mice. Hyperglycemia has been proposed to promote tumor growth through the increased flux of glucose, which fuels tumor cells (46). However, a major pathophysiological role of elevated blood glucose levels in the tumor-promoting action of type 2 diabetes in MKR^+/+ mice is unlikely because of the following: 1) female MKR^+/+ mice display only mild dysglycemia (∼20% increase in blood glucose levels compared with wild-type mice), 2) in animals with type 1 diabetes, which are insulin deficient because of immune destruction of the pancreatic β-cells, tumors regress despite severe hyperglycemia (47,48), 3) although some epidemiologic studies show a correlation between elevated blood glucose levels and cancer risk, the studies may be confounded by preexisting hyperinsulinemia (49). Chronic CL-316243 treatment reduces serum FFA, triglyceride, and leptin levels in both MKR^+/+ and wild-type mice. As this decrease is observed in both genotypes and the antineoplastic effect of CL-316243 is specific to MKR^+/+ mice, these factors are most likely not major players in the tumor-reducing effect of CL-316243. Nevertheless, we cannot fully exclude that changes in lipid or leptin levels may, at least in part, have an impact on breast tumor progression in MKR^+/+ mice. Taken together, our experimental approach allowed us to test the effect of insulin-sensitizing therapy on tumor progression in a setting in which obesity-associated confounding factors are minimal.

Our findings demonstrate that hyperinsulinemia in mice with type 2 diabetes significantly increases IR/IGF-IR activation in transformed mammary tissue, whereas pharmacological reduction of insulin levels attenuates this effect. Law et al.(50) reported that phosphorylation of the IR/IGF-IR was present in nearly 50% of all analyzed human primary breast tumors and was predictive for a poor outcome. However, the authors did not obtain information on the insulin levels in the patients studied. It remains to be evaluated whether the extent of IR/IGF-IR activation in breast cancer specimen correlates with underlying type 2 diabetes and hyperinsulinemia in breast cancer patients. If this is the case, then insulin lowering therapy could reduce breast cancer risk and mortality.

In conclusion, we demonstrate that the administration of insulin-sensitizing therapy using a β₃-AR agonist abrogates the accelerated mammary tumor progression in a nonobese mouse model of type 2 diabetes. This effect is accompanied by a reduction of insulin levels leading to a decrease in the phosphorylation of the IR/IGF-IR in transformed mammary tissue. These results demonstrate that insulin-lowering therapy is an important modifier of breast cancer progression. Our findings thus provide a rationale for early administration of insulin-sensitizing therapy in patients with hyperinsulinemia and/or type 2 diabetes, as such treatment may have a significant impact on breast cancer risk, morbidity, and mortality.

This work was funded by the National Cancer Institute (1RO1CA128799-O1A1). Y.F. received grants from the Swiss National Science Foundation (PBBSB-120851 and PBBSB3-120851) as well as from the Novartis Foundation. D.L. and Y.F. received a grant from the Roche Research Foundation.

No potential conflicts of interest relevant to this article were reported.

This study was presented in abstract form at the 69th Scientific Sessions of the American Diabetes Association, New Orleans, Louisiana, 5–9 June 2009.

We thank W.J. Muller (McGill University, Montreal, Canada) for donating MMTV-PyVmT transgenic mice; S.D. Hurstings (Department of Nutritional Sciences, University of Texas, Austin, TX, and Department of Carcinogenesis, University of Texas–M.D. Anderson Cancer Center, Smithville, TX) and N.P. Nunez (Department of Nutritional Sciences, University of Texas, Austin, TX) for donating Met-1 cells; and M.J. Campbell and J.F. Youngren (University of California San Francisco, San Francisco, CA) for providing MCNeuA cells.