Leptin Signaling Promotes the Growth of Mammary Tumors and Increases the Expression of Vascular Endothelial Growth Factor (VEGF) and Its Receptor Type Two (VEGF-R2)*

To gain insight into the mechanism(s) by which leptin contributes to mammary tumor (MT) development we investigated the effects of leptin, kinase inhibitors, and/or leptin receptor antagonists (LPrA2) on 4T1 mouse mammary cancer cells in vitro and LPrA2 on 4T1-MT development in vivo. Leptin increases the expression of vascular endothelial growth factor (VEGF), its receptor (VEGF-R2), and cyclin D1 through phosphoinositide 3-kinase, Janus kinase 2/signal transducer and activator of transcription 3, and/or extracellular signal-activated kinase 1/2 signaling pathways. In contrast to leptin-induced levels of cyclin D1 the changes in VEGF or VEGF-R2 were more dependent on specific signaling pathways. Incubation of 4T1 cells with anti-VEGF-R2 antibody increased leptin-mediated VEGF expression suggesting an autocrine/paracrine loop. Pretreatment of syngeneic mice with LPrA2 prior to inoculation with 4T1 cells delayed the development and slowed the growth of MT (up to 90%) compared with controls. Serum VEGF levels and VEGF/VEGF-R2 expression in MT were significantly lower in mice treated with LPrA2. Interestingly, LPrA2-induced effects were more pronounced in vivo than in vitro suggesting paracrine actions in stromal, endothelial, and/or inflammatory cells that may impact the growth of MT. Although all the mechanism(s) by which leptin contributes to tumor development are unknown, it appears leptin stimulates an increase in cell numbers, and the expression of VEGF/VEGF-R2. Together, these results provide further evidence suggesting leptin is a MT growth-promoting factor. The inhibition of leptin signaling could serve as a potential adjuvant therapy for treatment of breast cancer and/or provide a new target for the designing strategies to prevent MT development.

Elevated serum leptin levels are often correlated with obesity and associated with the development of breast cancer in postmenopausal women (1-4) reinforcing a long recognized causal link between obesity and increased risk for some types of cancer. The levels of leptin and its receptor (OB-R) 3 are correlated with distant metastasis, reoccurrence, and survival outcome. Furthermore, expression of leptin and OB-R levels are increased within human breast cancer tissue (2,5). Similar to obese women, mice and rats that have high leptin levels are more likely to develop mammary tumors (6,7). Leptin-deficient mice (ob Ϫ /ob Ϫ ) or mice lacking a functional OB-R (db Ϫ /db Ϫ ) do not develop mammary tumors (MT). Moreover, genetically engineered mouse models of mammary neoplasia exhibit a decreased propensity to develop MT when they are crossed with leptin/OB-R-deficient mice (8,9).
The sequence of leptin is highly conserved among species. However, OB-R has several isoforms with diverse signaling capabilities. These OB-R isoforms have the same extracellular sequence but differ in the length of the cytoplasmic tail. It is believed these variants are derived by alternate splicing (10). The full-length isoform (OB-RL, cytoplasmic tail, 303 amino acid residues) is mainly expressed in the hypothalamus. A short isoform (OB-Ra, cytoplasmic tail, 34 amino acids residues) is found in higher concentrations in peripheral tissues (11). In endometrial cancer cells an increase of the relative concentrations of OB-Ra to OB-RL has been reported (5). However, the exact implication of these findings in leptin-mediated cancer events is unknown.
With the exception of the soluble OB-Re isoform, which lacks transmembrane and intracellular sequence, all other OB-R isoforms contain a highly conserved, proline-rich sequence (Box 1) in their cytoplasmic tail that is required for JAK2 activation. A short amino acid sequence within a second motif, Box 2, is also essential for JAK2 tyrosine phosphorylation (14). Similar to that observed in other species, leptin stimulation of murine OB-RL involves phosphorylation of two specific tyrosine residues (Tyr 985 and Tyr 1138 ) in the C-terminal domain. Binding of Src homology 2 domain containing protein (i.e. STAT3) to Tyr 1138 activates the main leptinsignaling pathway in the hypothalamus (15). Several different mechanisms regulate OB-R activity. Dimerization and translocation of STAT3 to the nucleus promotes gene expression and increases levels of Src homology 2 domain containing the protein, suppressor of cytokine signaling-3 (SOCS-3). The SOCS-3 protein binds to Tyr 985 of OB-R, mediating the negative regulation of the STAT pathway (15). The protein-tyrosine phosphatase SHP2 has also been found to be a regulator of OB-R/STAT3 (16,17) and OB-R/ERK activities (18).
Despite mounting evidence suggesting a role for leptin in the genesis and/or progression of breast cancer the specific role leptin plays in these processes is still unknown but it is likely attributed to its reported proliferative and angiogenic effects (19 -22). The exact mechanism(s) that leptin promotes cellular proliferation is not clear, however, leptin can increase the levels of cell cycle regulators. Leptin/OB-R signaling mediates the expression of cyclin-dependent kinase 2 and cyclin D1 in human MCF-7 mammary cancer cells (20). Similarly leptininduced cell proliferation of ZR-75-1 breast cancer cells is associated with up-regulation of cyclin D1 and c-Myc (22). Leptin also induces aromatase activity in these cells and could consequently increase intracellular estradiol levels (23). Furthermore, in MCF-7 cells leptin transactivates estrogen receptor (ER) ␣ and therefore it has the capacity to amplify ER-dependent proliferation of estradiol-responsive cancer cells (24).
Ligand-dependent activation of OB-R plays an important role in controlling the proliferation, survival, and migration of epithelial cells. In this respect several recent studies suggest the relevance of leptin/OB-R signaling to tumorigenesis (25,26).
Leptin signaling can increase several factors implicated in epithelial cell proliferation, adhesion, inflammation, or angiogenesis, i.e. ␤3 integrin (27,28), leukemia inhibitory factor (LIF) leukemia inhibitory factor receptor (LIF-R), interleukin-1 (IL-1), IL-1 receptor (IL-1R tI), IL-1 receptor antagonist (IL-1Ra) (29), vascular endothelial growth factor (VEGF), and VEGF receptor (VEGF-R2) (30). Thus uncontrolled leptin-induced signaling could contribute to the proliferation of malignant mammary tissue and/or promote angiogenesis required for the growth of mammary cancer and metastasis (7,8). Leptin signaling-mediated effects on mouse 4T1 mammary cancer cells were evaluated in vitro using several signaling inhibitors. To further explore the role of leptin in tumorigenesis, a highly physiological mouse model that closely resembles breast cancer in humans was used. Immunocompetent mice were inoculated with syngeneic mouse 4T1 cells and treated with inhibitors of leptin signaling (29 -31). Collectively, our results suggest that leptin induces an increase in VEGF, VEGF-R2, and cyclin D1 expression that contribute to an increase in cell numbers in vitro. Furthermore, blockade of leptin receptor signaling markedly slows MT growth and leptin-mediated VEGF/VEGF-R2 and cyclin D1 expression in vivo. Specific antibodies IR-Dye-700 nm and IR-Dye-800 nm were from Rockland (Gilbertsville, PA). Mouse, goat, and rabbit secondary antibodies were obtained from Biomedia (Foster City, CA), normal horse serum, biotinylated antibodies, and vectastain ABC kits were obtained from Vector Laboratories (Burlingame, CA). ALI 3409 biotinylated antibody was obtained from BioSource International (Camarillo, CA). Immobilized recombinant Protein A was obtained from Pierce Biotechnology, Inc., (Rockford, IL). 4Ј,6-Diamidino-2-phenylindole was obtained from Molecular Probes Inc. (Eugene, OR). Fetal bovine serum was obtained from Gemini Bioproducts (Woodland, CA), and Dulbecco's modified Eagle's medium/F-12 and antibiotic/antimycotic mixtures were from Invitrogen. siRNA for mouse STAT3 (SMARTpool siRNA) was purchased from Dharmacon, Inc. (Lafayette, CO). SiLentFect TM Lipid reagent was obtained from Bio-Rad. Rabbit antibodies for AKT1 and phosphorylated AKT1 (Ser 473 ), and wortmannin were purchased from Upstate, Inc. (Charlottesville, VA). Tyrphostin AG 490 and other chemicals were obtained from Sigma.

Materials
Signaling Inhibitors-A leptin peptide receptor antagonist, LPrA2, and a scrambled peptide (LPrA-Sc) for a negative control were synthesized, purified, and validated as described previously (29,31). Polyethylene glycol derivatives (PEG, 20 kDa) of LPrA2 and LPrA-Sc were also used. The peptides were dissolved in a sterile filtered vehicle solution (0.0025% dimethyl sulfoxide (Me 2 SO), phosphate-buffered saline (PBS)). AG 490 and mouse siRNA-STAT3 were used to inhibit leptin-induced phosphorylation of STAT3. PD98059 and wortmannin were used to inhibit ERK1/2 and AKT1 phosphorylation, respectively.

Effects of Leptin on VEGF/VEGF-R2 and Cyclin D1 Levels in
Cultured Cells-4T1 cells (mouse mammary cancer cells, ATCC, ER positive) were cultured until they were ϳ70% confluent, washed with PBS, and cultured for 48 h in medium without fetal bovine serum (basal medium). Cells were washed again and cultured for 12-48 h in basal medium containing leptin (0, 0.62, 6.25, or 62.5 nM) in the presence or absence of unconjugated or PEG-LPrA2 or -LPrA-Sc (300 nM) and inhibitors of JAK2/STAT3 (siRNA for STAT3, 5 nM, and AG490, 30 M), MEK-1/MAPK (PD98059, 30 M), and PI3K/AKT1 (wortmannin, 20 M) pathways for 24 h. At the end of the treatment period the supernatants were harvested and kept at Ϫ80°C for VEGF determination. Cells were lysed in lysis buffer (10 mM Tris, 100 mM NaCl, 1 mM EGTA, 1 mM EDTA, 1% Triton, 0.5% Nonidet P-40 containing protease, phosphatase, and kinase inhibitors) and used for Western blot analysis of VEGF-R2, cyclin D1, and signaling pathways as described below. Duplicate wells were analyzed for each treatment and the experiments were repeated at least three times with different cell preparations.
Leptin Induction of VEGF Promoter Activity and VEGF-mediated Cell Proliferation-4T1 cells were grown to ϳ70% confluence and transiently transfected with 400 ng of mouse VEGF reporter plasmid (32) (generously supplied by Dr. P. D'Amore, Harvard Medical School, Boston, MA), and 40 ng of Renilla luciferase control plasmid (phRL-TK; Promega Corp., Madison WI), using 1.0 l of Lipofectamine 2000 reagent (Invitrogen). The cells were treated with IGF-1 (6.6 nM) or leptin (62.5 nM) and harvested 24 h after transfection and lysed in 100 l of passive lysis buffer (Promega Corp.) to monitor luciferase activity (Dual Luciferase Reporter Assay System; Promega Corp.) with a Packard Lumicount Luminescent Plate Reader (Global Medical Instrumentation Inc., Ramey, MN). Reporter activity was calculated as percentage of the ratio of firefly luciferase activity and expressed as relative luciferase units for effects of IGF-1 or leptin. The data were analyzed by Student's t test.
Western Blot Analysis-Protein concentrations in lysates from cell cultures and MT were determined by the Bradford method (Bio-Rad). Thirty g of protein were analyzed by Western blot (30) either directly or after immunoprecipitation with Protein A-agarose beads following the manufacturer's instructions (Pierce). Specific detection of antigens was carried out with antibodies for phosphorylated STAT3, ERK1/2, and AKT1, and VEGF-R2, cyclin D1, and SOCS-3. Antibodies against ␤-actin and non-phosphorylated proteins were used as loading controls. Simultaneous detection of the above antigens with antibodies linked to IR-Dye-800 nm (0.2 g/ml; IRDye800 TM ) and IR-Dye-700 nm (0.2 g/ml; Cy5.5) for the Odyssey InfraRed Imaging System (LI-COR, Inc., Lincoln, NE) or ECL-chemiluminescent assays (Amersham Biosciences) were performed. Nonspecific mouse, rabbit, and goat IgGs were used as negative controls for Western blot analysis. As a control measure for the anti-VEGF-R2 and SOCS-3, the appropriate blocking peptides (20 g/ml) were added to verify specificity. To quantitatively assess the effects of cytokines and inhibitors of leptin/OB-R signaling on antigen expression, the blots were scanned and analyzed by Li-COR Odyssey Software and the x-ray films were analyzed using the NIH Image program.
In Vivo Studies-To explore if leptin signaling is directly involved in the growth of MT we chose a syngeneic mouse model. Eight-week-old BALB/c female mice (Charles River Laboratories, Wilmington, MA) with full reproductive potential were housed in the animal facilities at the Boston Biomedical Research Institute in accordance with National Institutes of Health standards for the care and use of experimental animals. The rooms were on a 14-h light/10-h dark cycle and maintained at the temperature range of 22-24°C. Mice were provided water and food ad libitum and their body weights (BW) were measured at the beginning and end of the experiments. Mouse 4T1 cells suspended in PBS, 0.1% bovine serum albumin (2.5 ϫ 10 5 cells/100 l/mammary gland/mouse) were injected subcutaneously in the region of the mammary pad. To test the impact of blocking leptin signaling on MT growth, the treatments were initiated the same day of cell inoculation and continued daily for 3 weeks. Preliminary experiments showed that local injection of 30 M LPrA2 into the mammary fat pads produced a modest but positive reduction of MT grown. In addition, preliminary pharmacokinetic studies showed that PEG-LPrA2 has an enhanced half-life (18 h) compared with unconjugated LPrA2 (1 h half-life). Therefore, further experiments were carried out with 60 and 30 M unconjugated and pegylated LPrA2, respectively. Mice received 100 l of subcutaneous injections of PBS, 0.0025% Me 2 SO (vehicle control), LPrA-Sc (scrambled peptide receptor antagonist), LPrA2 (receptor antagonist), or PEG-LPrA2 twice daily into the immediate area surrounding the nipple of the mammary gland where malignant cells were to be implanted. In a separate series of experiments the treatments were initiated 1 week before cell inoculation to test the potential for LPrA to prevent tumor formation. Mice received 100 l of intraperitoneal injections of the treatments: PBS, 0.0025% Me 2 SO, PEG-LPrA-Sc, or PEG-LPrA2 twice daily as described previously. A group of mice did not receive treatment (untreated control). The mice were monitored daily for evidence of tumor development. After detection, the size of the MT was determined weekly ( ⌸ /6 ϫ width 2 ϫ length) (33). MT weights were determined after euthanasia. VEGF-R2 and cyclin D1 in MT lysates were determined by Western blot as described before.
ELISA Determination of VEGF and Leptin Levels-Conditioned medium from 4T1 cell cultures, serum from mice inoculated with 4T1 cells and from control (normal healthy mice), and lysates from MT biopsies were used to determine leptin and VEGF concentrations by ELISA (Quantikine, R & D Systems, and Assay Designs, Inc., MI, respectively). VEGF and leptin concentrations were within the dynamic range of the ELISA standard curves and calculated as picograms/ml/mg of protein, picograms/ml/g of BW, and picograms/ml/g of MT. To compare treatments the means from replicate determinations (n ϭ 3) were expressed as percentage of control (basal culture conditions and mice treated with PBS). Standards, controls, and samples were assayed in duplicate. The intra-and interassay coefficients of variation for VEGF and leptin were between 3-11 and 6 -14%, respectively. According to the manufacturer, the performance characteristics of the VEGF-ELISA were as follows: sensitivity, 3 pg/ml, and 100% specificity for mouse VEGF-A (164 and 120 amino acid residue forms) and no significant cross-reactivity with any of several tested cytokines or growth factors, including other VEGF isoforms and their recep-tors. The performance characteristics of leptin-ELISA were as follows: sensitivity, 1.74 pg/ml, 100% specificity for mouse and 43.6% for rat leptin, and no significant cross-reactivity for human leptin, mouse MIP-2, mouse growth-related oncogene-␣ (GRO-␣ or KC) or mouse IL-1␤.
Immunohistochemical Analysis of VEGF-The expression of VEGF was determined in paraffin sections derived from MT after treatments. Negative controls were also included in which the primary antibody was omitted.
Data Analysis-Unless otherwise stated a one-way analysis of variance test with Dunnett error protection and confidence interval of 95% was used from the Analyze-it for Microsoft Excel for data analysis of leptin/antagonist treatments. The experiments were repeated at least three times and all samples were analyzed in duplicate. The data were expressed as mean Ϯ S.D. Values for p Ͻ 0.05 were considered statistically significant. The model included the main effects of treatments and replicates.

Leptin Activates Multiple Signaling Pathways in 4T1 Cells-
Treatment of mouse 4T1 cells with leptin activated MAPK/ERK 1/2, PI-3K/AKT1, and JAK2/STAT3 pathways (Fig. 1, A-C). The leptin receptor antagonist (LPrA2) completely inhibited the leptin-induced increase in pSTAT3, pERK1/2, and pAKT1. In contrast, the negative control, LPrA-Sc had no effect. Based on the observation that leptin activated the JAK/STAT3 pathway we investigated whether the natural regulator of cytokine signaling, SOCS-3, was also regulated by leptin in 4T1 cells. Treatment with leptin significantly increased SOCS-3 and this increase was abrogated by co-incubation of cells with LPrA2 (Fig. 1D).
The Addition of Leptin Increased Cell Number Compared with the Vehicle-treated Control-No detectable concentrations of leptin were found after 24 or 48 h of culture of 4T1 cells in basal medium in either the presence or absence of phenol red. However, the addition of leptin increased the number of cells in culture (1.74-fold over controls) and this effect was inhibited (p Ͻ 0.05) by treatment with unconjugated or pegylated LPrA2. The negative control LPrA-Sc had no effect (data not shown).
Leptin Treatment of 4T1 Cells Increased Cyclin D1 Expression-Leptin stimulated an increase in cyclin D1 in 4T1 cells in vitro within 12 h. Treatment of 4T1 cells with LPrA2 abrogated this increase, whereas vehicle or the scrambled peptide had no effect ( Fig. 2A). Inhibiting phosphorylation of STAT3, AKT1, and ERK1/2 served to markedly disrupt leptininduced cyclin D1 (Fig. 2B). Reduced levels of cyclin D1 were observed after siRNA STAT3 treatment (Fig. 2B). Treatment of the 4T1 cells with the MEK-1 inhibitor significantly reduced the levels of cyclin D1 below that observed in unstimulated cells (basal levels) (Fig. 2B). Similar results were found for leptin at a lower dose (data not included).  SEPTEMBER 8, 2006 • VOLUME 281 • NUMBER 36

Involvement of Leptin in Growth of Mammary Tumors
Leptin Regulates Levels of VEGF, VEGF-R2, and VEGF Promoter Activity-Leptin significantly increased VEGF secretion by 4T1 cells (Fig. 3A) in a dose-dependent manner at high leptin concentrations (6.2 and 62.5 nM). The leptin-induced increase in VEGF was negated by LPrA2 or PEG-LPrA2 (Fig. 3B). To test whether leptin stimulated VEGF promoter activity, 4T1 cells were transiently transfected with a mouse promoter/luciferase reporter constructed for VEGF (32) and treated with IGF-1 or leptin. At 6.6 nM IGF-1 had no significant effect on luciferase activity. However, leptin significantly stimulated (3-fold) luciferase activity (Fig. 3C). Leptin also increased VEGF-R2 levels (Ͼ1.5-fold) in 4T1 cells (Fig. 3D). This increase was unaffected by LPrA-Sc but LPrA2 and PEG-LPrA2 significantly inhibited a leptin-induced increase in VEGF-R2.
Leptin Mediates VEGF and VEGF-R2 Expression-To determine the signaling pathways that are required for leptin-induced expression of VEGF and VEGF-R2, their levels were evaluated in 4T1 cell conditioned medium and cell lysates in the presence or absence of receptor antibodies or inhibitors of specific signaling pathways (Fig. 4). Leptin treatment of 4T1 cells in the presence of  anti-VEGF receptor (KDR/Flk-1) antibody resulted in a further increase in leptin-induced VEGF concentrations in the medium (Fig. 4A). The addition of the siRNA for STAT3 also augmented the leptin-induced increase levels in VEGF. In contrast, treatment of 4T1 cells with leptin in the presence of an inhibitor like JAK2 (AG490) or the PI3K inhibitor (wortmannin) inhibited the induced increase of leptin in VEGF. Inhibition of MEK-1 with PD98059 had little effect on the induction of leptin in VEGF (Fig.  4A). The leptin-induced increase in VEGF-R2 expression was unaffected by the siRNA targeted against STAT3, whereas AG490, wortmannin, and PD98059 attenuated the leptin-induced increase in VEGF-R2 (Fig. 4B).
In Vivo Inhibition of Leptin Signaling Slows MT Growth and Decreases Tumor Levels of VEGF/VEGF-R2-Inhibition of leptin activation of OB-R with LPrA2 or PEG-LPrA2 significantly slowed (50 and 90%, respectively) the growth of MT compared with the control groups (Fig. 5, A and B). This was true when the treatments were initiated before or simultaneously to cancer cell inoculation. Higher concentrations of PEG-LPrA2 (50 M) significantly delayed development of measurable MT (10 days) compared with mice treated with PEG-LPrA-Sc, vehicle, or untreated (Fig. 5C). However, MT in mice treated with PEG-LPrA2 were evident after 2 weeks (Fig. 5D). Histopathological analysis of MT samples derived from the mice treated with LPrA2 or PEG-LPrA2 showed a higher degree of centralized necrosis, which comprised a majority of the tumor bulk, and showed little to no evidence of VEGF-positive cells when compared with MT derived from control mice (Fig. 6A). ELISA analysis of MT lysates showed reduced levels of VEGF in MT from mice receiving LPrA2 and PEG-LPrA2 in comparison to those mice receiving a scrambled peptide or vehicle solution (Fig. 6B). Serum levels of VEGF from mice hosting MT and treated with scrambled peptide or vehicle were higher (Ͼ2-fold) compared with gender-matched mice with no tumor (normal/no cancer cell inoculation) (Fig. 6C). However, LPrA2 treatment decreased serum VEGF levels only in mice hosting MT. Moreover, PEG-LPrA2 treatment of mice inoculated with 4T1 cells normalized the serum level of VEGF to values comparable with healthy controls (see Fig. 6C). There was no difference in the serum levels of VEGF in normal healthy mice treated with vehicle, LPrA-Sc, or LPrA2 (data not shown). Western blot analysis of MT lysates showed that inhibition of leptin signaling with LPrA2 or PEG-LPrA2 negatively affected the level of VEGF-R2 (Fig. 6D).
Tumor and Serum Leptin Levels-Leptin concentrations in MT lysates showed no difference between mice treated with LPrA2 (6.4 Ϯ 1.3 pg/ml/mg of protein) and controls (6.4 Ϯ 1.5; 7.2 Ϯ 2.3 pg/ml/mg of protein, vehicle and LPrA-Sc, respectively). However, increased levels of immunoreactive leptin were found in the MT lysates from mice treated with PEG-LPrA2. Similarly, leptin levels in serum from mice hosting MT and treated with pegylated LPrA2 were significantly higher (351 pg/ml) than those found in mice hosting MT and treated with vehicle or scrambled LPrA2 (183 and 104.4 pg/ml, respectively). However, leptin levels in serum from gender-matched (normal/no tumor/no cancer cell inoculation) control mice were significantly higher (1034 pg/ml). These differences in serum leptin levels did not change after BW correction (data not shown).

DISCUSSION
To date, epidemiological studies, in vitro data, and animal models have implicated leptin as a risk factor for breast cancer Only tissues deemed to be non-necrotic were assessed for VEGF in LPrA2-treated mice. Negative control (C(Ϫ)) using nonspecific species-matched IgG instead the specific anti-VEGF antibody shows no staining. B, VEGF levels in MT lysates determined by ELISA (R&D Systems). Concentrations were expressed as a percent of VEGF in MT from mice treated with PBS.C, serum levels of VEGF were determined at the end of experiment. Mice bearing no tumors were treated with PBS and blood was collected to determine no tumor control (NT) for VEGF values. D, the panels represent qualitative and quantitative analysis of VEGF-R2 in MT lysates after treatments. The VEGF-R2 values were normalized to ␤-actin as a percent of control (n ϭ 3). The data are expressed as the mean Ϯ S.D. Letters (a and b) denote p Ͻ 0.05 and 0.01, respectively, significant differences with respect to control. development. Here we took advantage of the availability of a small peptide leptin receptor antagonist (LPrA2) to define how leptin contributes to mammary tumor establishment or growth. Because the peptide normally would have a relatively short biological half-life, 4 we pegylated it as a means to increase its bioavailability and potentiate its effects (34). Also, rather than using immunocompromised mice to host human tumors, we used syngeneic mice as hosts for tumors developed from mouse 4T1 breast cancer cells. Because leptin is a cytokine, using an immunocompetent mouse allowed for tumor growth more comparable with what would be seen in a naturally occurring breast cancer where one would expect some contribution from the immune system of the host. Treatment with either LPrA2 or PEG-LPrA2 was able to slow 4T1 tumor cell growth in syngeneic mice and reduce overall tumor burden as suggested by increased areas of central necrosis in tumors. Using higher doses of PEG-LPrA2 prior to inoculating mice with cancer cells resulted in markedly delayed MT development although tumors did eventually develop. The delay may be due to resistance to the peptide or the aggressive nature of the 4T1 cells.
Leptin has both angiogenic and mitogenic effects on cells depending on cell type or tissue of origin (1,19,20,26). Promoting either angiogenesis or proliferation could confer an advantage to growth of MT. Mice bearing MT had higher serum levels of VEGF than mice without tumors and these levels were reduced in mice bearing MT that had been treated with the peptide receptor antagonists. In tumors from animals that had been treated with the PEG-LPrA2, there was little to no evidence of immunostaining for VEGF in any area of the tumor that still had viable cells. Furthermore, there were reduced levels of VEGF-R2 detected in lysates of MT from animals treated with the leptin receptor antagonist and in the tumors themselves as detected by immunohistochemistry (data not shown). Whereas tumors from mice that had been treated with the peptide antagonist showed evidence of reduced VEGF, and serum from these mice had reduced levels, it may not have been a direct effect of leptin on the MT cells. In vitro, many types of cancer cells can secrete VEGF. However, it is conceivable that in vivo, the source of VEGF would not be limited to cancer cells but would include inflammatory, stromal, and/or endothelial cells. Therefore, it is probable that the peptide antagonist inhibited production of VEGF produced by such cells and not solely by the MT cells. This concept was indirectly tested using leptin, the peptide antagonists, and the MT cells in vitro.
When 4T1 cells were grown in vitro and treated with leptin, we noted a significant increase in VEGF production and VEGF-R2 expression. Furthermore, using the VEGF promoter linked to a luciferase reporter, we noted an increased promoter activity in response to leptin. All three effects were inhibited by the LPrA2 peptide. These data demonstrate that leptin increases the angiogenic potential of the MT directly and suggests that the decreased serum VEGF we observed in vivo may be to some extent due to direct inhibition of production of the angiogenic factor by the tumor cells. However, this data does not exclude the possibility that other cells (inflammatory, stromal, and endothelial cells) were also targeted by the peptide antagonists. Interestingly, when 4T1 cells were treated with leptin plus a VEGF receptor antibody (KDR/Flk-1), we observed potentiation of the effect of leptin, suggesting a feedback loop between VEGF expression and its receptor. It would seem that inhibition of VEGF receptor activation with the antibody induces more VEGF production. This is an interesting observation because there is currently a trend toward inhibiting angiogenesis as a means to decrease or prevent tumor growth. Our data suggest that attempts to inhibit angiogenesis by inhibiting receptor activity might result in even more VEGF production.
Inhibition of leptin signaling with the receptor antagonist dramatically slowed tumor growth. Leptin is mitogenic in several tissues and normal or cancer cell lines (1,19,20,26,35). In the present studies, leptin increased cell number in cultures and the peptide receptor antagonist abrogated this effect. Consistent with the increase in cell number, there is an increased expression of cyclin D1 induced by leptin, which, like the proliferation, is inhibited by treatment with the receptor antagonist.
Leptin activated several divergent signaling pathways in 4T1 cells in vitro and these pathways in turn seem to play different roles in mediating the effects of leptin on cyclin D1 expression or VEGF/ VEGF-R2 expression. JAK normally interacts with members of the STAT pathway (11). Both OB-Ra and OB-RL, the most abundant isoforms of leptin receptor, can activate JAK2. However, only OB-RL has the capability to activate STAT3 (14). The present data suggest that JAK is a critical component of the leptin-induced increase in VEGF or VEGF-R2 because the JAK inhibitor negated the inductin of leptin of both VEGF and VEGF-R2. In contrast, inhibition of STAT3 with siRNA did not negatively affect the inductin of leptin by VEGF-R2 or VEGF. Inhibiting STAT3 with siRNA resulted in an even greater induction of VEGF than that observed in response to leptin alone. Leptin signaling through JAK/STAT3 is down-regulated by SOCS-3, which is up-regulated by STAT3 (36). In the present study, leptin-induced SOCS-3 was abrogated by LPrA2. Thus in the absence of STAT3 (as a result of STAT3 siRNA action), lower levels of SOCS-3 were produced but leptin-induced VEGF was not down-regulated. These results suggest that VEGF secretion by 4T1 is independent or not entirely dependent on STAT3 phosphorylation. Whether this is due to SOCS-3 is not known, however, as evidenced herein, SOCS-3 was not produced, due to the absence of STAT3 4 and therefore leptin-induced signaling is not down-regulated. Overall, the specific mechanisms for the unanticipated divergent impact of the inhibition of JAK2 and STAT3 phosphorylation on VEGF/VEGF-R2 levels in 4T1 cells are unknown. However, it might be due to leptin/OB-R/JAK2mediated activation of other STAT proteins than STAT3 (i.e. STAT1, STAT5) (37) and/or to a yet unveiled cross-talk between JAK2 and signaling pathway activation that require Box 1 motif (i.e. SHP2/ERK and PI-3K/AKT1) (38).
Both the PI3K and MEK-1/MAPK signaling pathways were also found to be involved in the leptin-mediated increase in VEGF and VEGF-R2. Inhibition of PI3K in 4T1 cells with wortmannin abrogated the leptin induction of VEGF secretion and VEGF-R2 expression suggesting these effects are AKT1 dependent. On the other hand, inhibition of the MAPK pathway completely pre-vented any increase in VEGF-R2 induced by leptin, whereas it did not completely abrogate the effect of leptin on VEGF itself.
In addition to the variable effects the different signaling pathways had on the induced changes of leptin in VEGF or VEGF-R2, each of the different signaling pathways inhibited the induced increase of leptin in cyclin D1. Leptin activated the PI3K pathway in 4T1 cells and this could lead to activation of nuclear factor B (NFB). NFB in turn can activate the cyclin D1 promoter (39). JAK/STAT3 signaling also seems to be involved in leptin induction of cyclin D1 because both the siRNA to STAT3 and the JAK inhibitor both result in inhibition of the increase of leptin induced in cyclin D1.
Whereas we show here that activation of MAPK/ERK1/2 may be involved in the effects of leptin on proliferation and/or cell survival activation of this pathway may have additional roles. Leptin can induce aromatase activity in human breast cancer (24), transactivate the ER␣ through the MAPK pathway (23), and transactivate erbB2 (HER) in human mammary cancer cells (40). Therefore, its pro-estrogenic effects or activation of erbB2 could further enhance and support the growth of estrogen-responsive tumors (i.e. 4T1 cell derived MT). This data suggests that the pleiotropic effects of leptin are likely attributed to its activation of multiple signaling pathways. The contributions of the different pathways become more evident with the use of specific inhibitors.
In the present studies, we observed no overt toxic effects of the leptin receptor antagonist in mice. Moreover, the pegylated receptor antagonist (PEG-LPrA2) does not cross the blood-brain barrier and has no apparent effect on energy balance and food intake. 4 Our data provide evidence to suggest that disruption of leptin signaling could slow MT growth, making this an attractive area to pursue in designing new therapeutic strategies targeting breast cancer. This might also have particular importance in postmenopausal and obese women considered at a higher risk for breast cancer. Perhaps part of that risk is due to higher circulating leptin levels seen in these conditions. Inhibition of leptin signaling in such instances might serve as a preventative measure.