Purpose: An inherent problem in breast cancer treatment is that current therapeutic approaches fail to specifically target the dissemination of breast cancer cells from the primary tumor. Clinical findings show that the loss of Wnt-5a protein expression in the primary breast tumor predicts a faster tumor spread, and in vitro analyses reveal that it does so by inhibiting tumor cell migration. Therefore, we hypothesized that the reconstitution of Wnt-5a signaling could be a novel therapeutic strategy to inhibit breast cancer metastasis.

Experimental Design: We used in vitro techniques to show that 4T1 mouse breast cancer cells responded to the reconstitution of Wnt-5a signaling using our novel Wnt-5a mimicking hexapeptide, Foxy-5, in the same way as human breast cancer cells. Therefore, we could subsequently study its effect in vivo on the metastatic spread of cancer following the inoculation of 4T1 cells into mice.

Results:In vitro analyses revealed that both recombinant Wnt-5a and the Wnt-5a–derived Foxy-5 peptide impaired migration and invasion without affecting apoptosis or proliferation of 4T1 breast cancer cells. The in vivo experiments show that i.p. injections of Foxy-5 inhibited metastasis of inoculated 4T1 breast cancer cells from the mammary fat pad to the lungs and liver by 70% to 90%.

Conclusions: These data provide proof of principle that the reconstitution of Wnt-5a signaling in breast cancer cells is a novel approach to impair breast tumor metastasis by targeting cell motility. In combination with existing therapies, this approach represents a potential novel therapeutic strategy for the treatment of breast cancer patients.

Translational Relevance

The primary tumor is rarely the cause of death for cancer patients. In fact, in the majority of cases cancer-associated mortality is the result of cancer metastasis. The progression of most primary tumors leads to an increase in tumor cell motility. A low-level expression of Wnt-5a protein in primary invasive breast carcinomas is associated with breast cancer cells with increased motility and in shortened recurrence-free survival. Consequently, an attractive antimetastatic approach would be to reconstitute Wnt-5a signaling in breast cancers that have a reduced expression of this protein (50% of all cases). Based on sequence analysis of Wnt-5a, we identified a hexapeptide fragment that after modification (formylation) mimicked the inhibitory effect of Wnt-5a on breast cancer cell migration. In the present study we show that i.p. injection of this peptide, Foxy-5, causes a significant reduction (70-90%) of liver and lung metastases in immunodeficient as well as in immunocompetent mice. These results indicate that the reconstitution of Wnt-5a signaling is a valid approach to impair breast cancer metastasis. It is reasonable to presume that the development and use of a specific antimetastatic drug, based on Foxy-5, will increase the quality of life and the overall survival of breast cancer patients.

Breast cancer is the most common malignancy among women, and it is predicted to afflict 1 in 10 females during their lifetime. The development of improved diagnostics and therapeutics for breast cancer had led to a reduction in overall mortality rates (1). However, tumor metastasis is the predominant cause of death among patients, and the 5-year survival remains less than 30% in patients with distal spread of their tumor (1).

In this context, the nontransforming Wnt-5a protein seems to be a promising therapeutic candidate, as the lack of Wnt-5a protein expression in the primary tumor of breast cancer patients strongly correlates with an increased risk of metastasis and reduced survival (2, 3). This inhibitory effect of Wnt-5a on metastasis may be due to the ability of Wnt-5a to increase adhesion and thereby decrease migration of breast epithelial cells (4, 5), as well as to its ability to counteract activation of NFAT (6), a transcription factor that has been implicated in promoting breast cancer cell invasiveness (7, 8). Indirect support for a suppressive effect of Wnt-5a on breast cancer cell migration comes from a recent and elegant study of mammary gland development (9). In this study the investigators clearly show that Wnt-5a signaling is capable of inhibiting ductal extension and lateral branching in the mammary gland most likely via an increased cell adhesion. A tumor-suppressing effect of Wnt-5a has also been found in B lymphomas (10, 11), thyroid carcinoma (12), and colon cancer (13, 14). In contrast, Wnt-5a has been shown to induce migration and invasion in gastric cancer (15) and melanoma (16, 17).

The Wnt proteins function in an auto or paracrine fashion. This mechanism of action is promoted by the fact that immediately following their release from a cell, they rapidly bind to heparan sulfate proteoglycans on the same or neighboring cell surface, which functions to effectively present Wnt proteins to their respective cell surface receptor (18). Consequently, the heparan sulfate-binding domain on Wnt proteins limits the distribution of a released Wnt protein to the area closest to the cell from which it originated. This implies that administration of a Wnt protein is likely to result in a very restricted dispersal from the site of deposition, thus presenting difficulty in using an intact Wnt protein in vivo.

Wnt proteins bind to G-protein coupled receptors of the Frizzled family (19, 20). The functional effect of a particular Wnt protein is not only determined by the Frizzled receptor it binds to, but also by the presence of low-density lipoprotein receptor-related proteins, LRP5/6, and the tyrosine kinase receptor Ror2 that have been shown to function as coreceptors or complementary receptors (2123). Regarding their function as coreceptors in Wnt-5a signaling, Ror2 has been shown to affect the migratory response to Wnt-5a in human melanoma and mouse fibroblasts (23), whereas LRP5/6 has not been reported to affect the migratory response to Wnt-5a.

In the present study, we have taken advantage of the newly characterized formylated Wnt-5a–derived hexapeptide Foxy-5, which lacks a heparan sulfate-binding domain but mimics the Wnt-5a–induced inhibition of human breast cancer cell migration in vitro (5), to investigate the potential in vivo effect of reconstituting Wnt-5a signaling. Although it is possible that degradation of such a peptide can occur in vivo, small formylated peptides released from bacteria at an infection-induced inflammatory site can potently activate cells of the innate immune system despite the low pH and high protease activities in this milieu (24). Consequently, the fact that the Foxy-5 peptide was designed with a formyl group on its NH2-terminal methionine residue means that it is potentially similarly protected from degradation in vivo.

In this study, we investigated the mechanisms whereby recombinant Wnt-5a (rWnt-5a) and the Foxy-5 peptide affect mouse 4T1 breast tumor cells in vitro as a means to interpret the effects of Foxy-5 administration on metastasis of 4T1 breast tumors in BALB/c mice.

Antibodies and peptides. The following antibodies and peptides were used: β-actin monoclonal AC-15 antibody and formyl-Nle-Leu-Phe-Nle-Tyr-Lys (fcp) peptide (Sigma Aldrich Corporation); Ki67 polyclonal antibody (Affinity BioReagents); phospho-p44/42 MAPK Thr202/Tyr204 polyclonal antibody and total p44/42 MAPK polyclonal antibody (Cell Signaling Technology, Inc.); poly(ADP)ribose polymerase monoclonal antibody (Biomol International); rWnt-5a and Ror2 goat polyclonal antibody (R&D Systems); and formyl-Met-Asp-Gly-Cys-Glu-Leu (Foxy-5), Met-Asp-Gly-Cys-Glu-Leu (nfp), and formyl-Met-Ser-Ala-Asp-Val-Gly (RnF) peptides (Pepscan Systems). The synthesized peptides (>95% pure) were quality-controlled by reversed phase–high performance liquid chromatography and mass spectrometry and synthesized on four different occasions. The polyclonal antibodies toward Wnt-5a, Frizzled-2, and Frizzled-5 were produced in our laboratory, as previously described (4, 5). The polyclonal antibody against amino acids 202-220 in the ectodomain of the Frizzled-2 receptor was used as a control antibody. The polyclonal anti–Wnt-5a antibody was raised against amino acids 275-290 of the Wnt-5a molecule (4), a segment that exhibits 100% homology between human and mouse Wnt-5a.

Cell culture. The human mammary carcinoma cell lines MDA-MB-468 (negative control for Wnt-5a expression) and T47D (positive control for Wnt-5a expression) were grown in DMEM D5796 supplemented to contain 10% fetal bovine serum, 5 U/mL penicillin, 0.5 U/mL streptomycin, and 2 mmol/L glutamine. The mouse mammary epithelial cell line 4T1 was grown in RPMI 1640 supplemented to contain 10% fetal bovine serum, 1.5 g/L sodium bicarbonate, 10 mmol/L HEPES, 1 mmol/L sodium pyruvate, 4.5 g/L glucose, 5 U/mL penicillin, 0.5 U/mL streptomycin, and 2 mmol/L glutamine. The mouse macrophage cell lines J774 and Raw264.7 were grown in DMEM D5796 supplemented to contain 10% fetal bovine serum and 2 mmol/L glutamine, and serum-starved in Macrophage SFM serum-free medium 12065. All cells were incubated at 37°C in 5% CO2.

Western blot. The cells were lysed with a lysis buffer (5, 13) and centrifuged at 15,000 rpm for 8 min at 4°C. The protein content in each sample was determined and adjusted to ensure equal loading of protein in each lane. Thereafter, 50 mmol/L DTT and 5× concentrated Laemmli buffer were added, and the lysates were boiled for 5 min and separated by SDS gel electrophoresis. The proteins were subsequently transferred to polyvinylidene difluoride membranes, blocked in 5% nonfat milk or 3% bovine serum albumin for 45 min, and incubated for 1 h with the indicated primary antibody [1:15,000 for actin; 1:5,000 for Frizzled-5; 1:2,000 for Wnt-5a and Ror2; and 1:1,000 for phospho p44/42 MAPK, p44/42 MAPK, Frizzled-2, and poly(ADP)ribose polymerase] in 1% nonfat milk or bovine serum albumin. After incubating for 1 h with a horseradish peroxidase-conjugated secondary antibody and extensive washing, the antibody-antigen complexes were detected using an enhanced chemiluminescence kit from Amersham Pharmacia Biotech. For reprobing, the membranes were stripped with a Reblot Strong solution from Chemicon International. The Western blots shown are representative of at least five independent experiments.

Cell migration. Cell migration (Fig. 1B and C) was analyzed in a modified Boyden chamber (Transwell; Costar). At the onset of each experiment, the cells were detached with Versene and resuspended as single cells in serum-free RPMI medium supplemented with 0.5% bovine serum albumin. A 0.2-mL aliquot of the cell suspension containing 25,000 4T1 cells was added to the upper transwell chamber together with the indicated peptide or rWnt-5a. In Fig. 1C, 5 μg/mL Frizzled-2 antibody, Frizzled-5 antibody, or rabbit immunoglobulin fraction was added to the upper chamber. The lower chamber was filled with serum-free, bovine serum albumin–containing medium supplemented with the chemoattractant IGF-I (1 ng/mL). The cells were allowed to migrate in the Boyden chamber at 37°C in a humidified atmosphere containing 5% CO2. After discarding the medium and removing the nonmigratory cells from the membrane, the membrane was cut out of the chamber. The cells that had migrated through the membrane and were located on its lower surface were stained for 20 min with 0.5% crystal violet in 20% methanol. The membrane was subsequently washed, and the stained cells were counted (number per membrane). The result of each experiment is presented as the median value from three identical samples done in separate wells. The experiments were independently repeated 5 to 6 times.

Cell invasion. Cell invasion (Fig. 1D) was analyzed in the BD Matrigel invasion chamber (BD Biosciences). At the onset of each experiment, the cells were detached with Versene and resuspended as single cells in serum-free RPMI medium. An aliquot of the cell suspension containing 25,000 4T1 cells was added to the upper transwell chamber, and the lower chamber was filled with serum-containing (10%) medium and supplemented with the indicated peptide or rWnt-5a. In the data presented by the two right-hand columns in Fig. 1D, Foxy-5 or rWnt-5a was added to the upper chamber together with the cells. The cells were allowed to invade at 37°C in a humidified atmosphere of 5% CO2. The experiment was terminated by discarding the medium and fixing the cells for 10 min in 4% paraformaldehyde. After staining with 0.5% crystal violet in 20% methanol and removing the noninvading cells, the membrane was washed and cut out of the chamber, and the stained cells were counted (number per membrane). The experiment was repeated independently seven times.

Tumor metastasis animal model. All animal experiments were carried out at the central animal laboratory of Turku University. In the first series of animal experiments, 25 female athymic BALB/C mice (8 wk old; Harlan) were randomly divided into 3 groups. The 4T1 mouse breast cancer cells (2.5 × 104 in 0.1 mL PBS) were inoculated into the mammary fat pads under isoflurane anesthesia (1.5-3%, air flow 200 mL/min, Univentor 400 anesthesia unit). The first dose of PBS alone (100 μL) or PBS (100 μL) containing 5 μg of formylated control peptide (fcp) or Foxy-5 was injected on the day of 4T1 cell inoculation. Thereafter, the mice were i.p. injected with the same amount of peptide or PBS alone every 4th day, with the last dose administered the day before the experiment was terminated on day 25. The primary tumor diameters (d1 and d2) were measured using a caliper, and the tumor volumes were calculated every 4th day throughout the experiment (Fig. 2A), according to the formula, V = (π/6)(d1×d2)3/2 (25). The mice were also weighed every 4th day throughout the experiment (Fig. 2B). The experiments in Figs. 2 and 3 were done with 9 PBS-treated, 8 fcp-treated, and 8 Foxy-5–treated mice.

In Fig. 4, 26 female normal BALB/c mice (8 wk old; Harlan) were randomly grouped and inoculated with 1 × 104 4T1 mouse breast cancer cells. The cells were suspended in 0.5 mL PBS and injected into the mammary pads. The animals were treated exactly as above with PBS alone, 20 μg fcp or Foxy-5. The primary tumor volume was calculated as previously described, and the mice were weighed every 4th day throughout the experiment and showed no significant differences (Fig. 4A). The experiment was done with 8 PBS-treated, 9 fcp-treated, and 9 Foxy-5–treated mice. The dose-response study of Foxy-5 (Supplementary Fig. S3) included 52 female normal BALB/c mice (8 wk old). All animals were inoculated with 1 × 104 4T1 mouse breast cancer cells, as described above. The animals were treated exactly as above with PBS alone or the indicated concentration of Foxy-5. The data were calculated as percentage of the PBS-treated controls and given as means from 17 mice treated with PBS alone (9 of those obtained from the second series of experiment; Fig. 4), 9 mice treated with 5 μg Foxy-5, 9 mice treated with 20 μg Foxy-5 (obtained from the second series of experiments; Fig. 4), 6 mice treated with 40 μg Foxy-5, and 7 mice treated with 160 μg Foxy-5.

In Fig. 5, 24 female normal BALB/c mice (8 wk old; Harlan) were randomly grouped and injected with 1 × 104 4T1 mouse breast cancer cells, as described above. The first i.p. injection of PBS alone (100 μL) or PBS (100 μL) containing 40 μg of Foxy-5 was done on the day of 4T1 cell inoculation. Thereafter, the mice were i.p. injected with the same amount of peptide or PBS alone every 4th day, with the last dose administered the day before the experiment was terminated on day 25. The primary tumor volume was calculated as previously described, and the mice were weighed every 4th day throughout the experiment (Fig. 5A and B). The data are given as means from 11 PBS-treated and 11 Foxy-5–treated mice.

The animal experiment procedures were reviewed by the local ethics committee on animal experimentation at the University of Turku and approved by the local provincial state office of Western Finland (number 1494/05 decision number LSLH-2005-3322/Ym-23, 4.4.2005).

Histomorphometric examination. The mice were euthanized in a CO2 chamber. The lungs and liver were excised, fixed in 4% buffered formalin for 2 d, and embedded in paraffin. In the first experiment (Fig. 2), the lungs were cut through and 12 horizontal sections (5 μm) were taken at 250-μm intervals for H&E staining. Four horizontal sections (5 μm) were taken from the center area of the liver with a 500-μm interval. In later experiments (Figs. 4 and 5; Supplementary Fig. S3), sections from only six levels of lungs and two levels of liver were taken because in a control evaluation of the first experiment they gave the same result as the double number of sections. The H&E sections were evaluated and photographed with an Olympus BX51 microscope controlled by the IrfanView 3.97 software (Irfan Skiljan, Graduate of Vienna University of Technology). The relative area of metastases (total metastatic area/total area of section) was measured using the ImageJ 1.37o software (NIH). All metastasis analyses (numbers and areas) were done on blinded samples. The numbers and areas of liver metastases throughout this study are calculated per histologic liver or lung section.

Terminal deoxynucleotidyl transferase biotin-dUTP nick end labeling staining. To evaluate the degree of apoptosis in the lung and liver metastases, an In situ Cell Death Detection Fluorescein kit (Roche Diagnostics GmbH) based on the terminal deoxynucleotidyl transferase biotin-dUTP nick end labeling (TUNEL) method was used. Tissue sections were deparaffinized, incubated in the dark with the TUNEL reaction mixture for 1 h at 37°C, and then counterstained with 4′,6-diamidino-2-phenylindole. Samples treated with 30 U/mL DNase I for 10 min (Invitrogen) were used as positive controls, whereas samples incubated without the terminal transferase were used as negative controls. No staining was detected in any of the negative controls (data not shown). All metastases in two lung and two liver sections from each mouse were analyzed.

Ki67-staining. To investigate the amount of proliferation, lung and liver sections were stained with a Ki67 antibody. All metastases in two lung and two liver sections from each mouse were analyzed.

Statistical analysis. Student's t test for paired samples was used to analyze for statistically significant differences in all the in vitro experiments. The Kruskal Wallis test was used to determine if an overall difference was present in the animal experiments; this was complemented with the Mann-Whitney's test to determine between which treatments the difference existed and its corresponding P value (*P < 0.05; **P < 0.01; ***P < 0.001).

The effect of Foxy-5 and rWnt-5a on migration, invasion, apoptosis and proliferation in vitro. To investigate the role of the Wnt-5a–derived peptide Foxy-5 in breast cancer metastasis we chose the 4T1 cancer cells originally derived from a spontaneous breast cancer in a BALB/c mouse (26), because following inoculation into the mammary fat pad, the resultant 4T1 breast tumors give rise to distal metastases in both athymic and normal BALB/c mice (27). 4T1 cells lack endogenous expression of the Wnt-5a protein but still express the proposed Wnt-5a receptors Frizzled-2 and Frizzled-5 (Fig. 1A). Furthermore, similar to human MDA-MB-468 breast cancer cells, murine 4T1 breast cancer cells also lack expression of the coreceptor Ror2 at both the mRNA and protein levels (Supplementary Fig. S1). We therefore believed that this murine cell line was an appropriate model for human breast cancer cells lacking or having a reduced endogenous expression of Wnt-5a (2, 3) and thus suitable for testing the in vivo effects of Wnt-5a reconstitution. We first evaluated the mechanism(s) of action of rWnt-5a and the Foxy-5 peptide in 4T1 breast cancer cells.

Both rWnt-5a and Foxy-5 significantly inhibited the migration of 4T1 cells in a modified Boyden chamber assay (Fig. 1B). A nonformylated version of Foxy-5 (nfp) and a formylated control peptide (fcp) were used as controls, and neither of them had any detectable effect on migration (Fig. 1B). Next, we found that the previously described Frizzled-5 receptor-blocking antibody (16, 28) effectively abolished the effects of rWnt-5a and Foxy-5 on 4T1 cell migration (Fig. 1C). As rWnt-5a was previously suggested to have a chemotactic effect on the human breast cancer cell line MCF-7 in an invasion assay (29), we tested the effect of rWnt-5a and Foxy-5 on 4T1 cells in a similar commercially available invasion assay. We found no chemotactic effect of either rWnt-5a or Foxy-5 on 4T1 cells when added to the lower chamber of the Matrigel invasion system (Fig. 1D). On the contrary, we observed an inhibitory effect of Foxy-5 on 4T1 cell invasion even when added to the lower chamber (Fig. 1D). This effect of Foxy-5 can possibly be explained by its more rapid diffusion across the Matrigel layer compared with that of the much larger rWnt-5a molecule. In support of inhibition of 4T1 cell migration by Wnt-5a signaling, we observed that when either rWnt-5a or Foxy5 was added to 4T1 cells in the upper chamber of the Matrigel invasion system, the invasion of these cells was significantly inhibited (the two columns to the right in Fig. 1D).

As shown in Supplementary Fig. S2, neither rWnt-5a nor Foxy-5 or either of the two control hexapeptides had any effect on apoptosis of 4T1 cells in vitro. In addition, neither rWnt-5a nor Foxy-5 or either of the two control hexapeptides had any effect on the proliferation of 4T1 cells in vitro. Thus, Wnt-5a and Foxy-5 lack an effect on apoptosis and proliferation of 4T1 breast cancer cells, although they readily inhibit migration and invasion in vitro.

The effect of Foxy-5 on breast cancer metastasis in athymic BALB/c mice. We first tested if Foxy-5 could inhibit metastasis of 4T1 cells inoculated into the mammary fat pad of athymic BALB/c mice. Following i.p. administration of Foxy-5 (5 μg) every 4th day, our data revealed that Foxy-5 had no significant effect on the growth of the primary tumors or on the weight gain of the animals when compared with animals injected with a formylated control hexapeptide (chemotactic for neutrophils and macrophages), or saline alone (Fig. 2A and B). However, evaluation of the amount of metastases in histologic tissue sections stained with H&E revealed that treatment with Foxy-5 significantly inhibited the number of liver and lung metastases by 70% compared with the control animals (Fig. 2C). This inhibition was even more pronounced when the total area of tumor metastases was analyzed (Fig. 2D). The effect on lungs was not quite statistically significant (P = 0.055 for lung area and P = 0.074 for lung number), which is readily explained by the heterogeneous sizes of the lung metastasis.

The effect of Foxy-5 on apoptosis and proliferation in vivo. To evaluate the possible effect of Foxy-5 on proliferation and apoptosis in the 4T1 cells in vivo, we did a TUNEL assay and stained lung and liver tissue sections with an antibody directed against the proliferative marker Ki67. TUNEL staining of liver and lung sections revealed that neither the Foxy-5 peptide nor the control peptide had any effect on the number of apoptotic cells in the tumor metastases (Fig. 3A and B). Immunohistochemical analysis of the proliferative marker Ki67 in liver sections showed no effect of Foxy-5, whereas we noted a small inhibitory effect of Foxy-5 on the number of Ki67-positive cells in the lung metastases but not in the liver (Fig. 3C and D). The fact that Foxy-5 has no effect on 4T1 cell proliferation in vitro (Supplementary Fig. S2D and E) even when incubated over a period of 1 week (Supplementary Fig. S2F) and that the Wnt-5a protein expression in primary human breast tumors was unrelated to the expression of the proliferate marker Ki67 (3) suggests that the limited effect of Foxy-5 on 4T1 tumor cell proliferation in the lungs is indirect. These data support the concept that Wnt-5a does not act via inhibition of β-catenin signaling and a subsequent inhibition of proliferation. This is in contrast to the situation in colon cancer where Wnt-5a signaling inhibits β-catenin signaling and subsequently tumor cell proliferation (14). A reasonable explanation for this difference relates to the well-known elevated β-catenin signaling in colon cancer cells not present in breast cancer cells. Dolled-Filhart and coworkers have recently shown in a clinical cohort of 688 ductal breast cancers that nuclear β-catenin staining was detected in less than 10 tumors (30). Furthermore, Wnt-5a does not affect the low level of active β-catenin in human breast cancer cells (5). One therefore has to conclude that the effect of Wnt-5a depends on the receptor, coreceptor, and signaling context of the cell, and that Wnt-5a inhibits breast cancer cell migration via a β-catenin independent signaling pathway possibly relating to its effect on adhesion receptors (4), NFAT activity (6), and cytoskeletal changes (5).

The effect of Foxy-5 on breast cancer metastasis in normal BALB/c mice. Athymic mice are often used in experimental cancer research, but a defective immune response might influence, possibly reduce, tumor progression and thus the results of an antimetastatic substance could be overestimated. Inoculation in athymic BALB/c mice has been shown to lead to the development of less metastases than in normal syngeneic BALB/c mice (31). Therefore, we took advantage of the present mouse model enabling us to test the antimetastatic effect of Foxy-5 in normal BALB/c mice. We initially chose a 4-fold higher dose of Foxy-5 than that used in the immunocompromised BALB/c mice. This dose of Foxy-5 had no, as in Fig. 2, statistically significant effect on the growth of the primary tumor or on the weight gain of the animals (Fig. 4A). We noted that the effects of Foxy-5 on the development of liver and lung metastases were qualitatively similar to that seen in athymic mice but less pronounced and statistically not significant. The Kruskal Wallis analysis revealed that P = 0.473 in Fig. 4B (left), P = 0.376 in Fig. 4B (right), P = 0.541 in Fig. 4C (left), and P = 0.396 in Fig. 4C (right). In accordance with the findings of increased development of tumor metastases in normal syngeneic BALB/c mice (31), these data indicate the need of a higher dose of Foxy-5 to obtain inhibitory effects similar to that found in athymic mice. We therefore initiated a limited dose-response study (0-160 μg) on the effect of Foxy-5 on the metastasis to the lungs. The data revealed that 40 μg Foxy-5 exhibited a near maximal inhibitory effect on 4T1 cancer cell metastasis in normal BALB/c mice (Supplementary Fig. S3).

Interestingly, administration of the control, non–Foxy-5 formylated hexapeptide could not be used as a control peptide in the experiments in normal BALB/c mice, because it led to increased metastasis (Fig. 4B and C). The fact that this “control” peptide had no effect on migration, invasion, proliferation or apoptosis of 4T1 cells in vitro (Fig. 1; Supplementary Fig. S2), suggests that its effect is related to nontumor cells, for example tumor-associated macrophages. This possibility was supported by the findings that both the formylated control hexapeptide and to a lesser extent a formylated random hexapeptide, but not Foxy-5, triggered extracellular signal-regulated kinases activation in two different murine macrophage–like cell lines, J774 and Raw264.7 (Fig. 4D). Extracellular signal-regulated kinases activation was chosen as a representative downstream signal, as it has been shown to be essential for both proliferation and migration of murine macrophages (32, 33).

Based on the above results, we tested how i.p. injection of 40 μg Foxy-5 in normal BALB/c mice affected the progression of 4T1 cancer development and metastasis. Our observation that even the formylated random hexapeptide stimulated murine macrophages and 4T1 cell metastasis led us to use i.p. injection of saline alone as a control to Foxy-5. Clearly, if we had used any of the peptides that stimulated murine macrophages as a control we would incorrectly have overestimated the effect of Foxy-5. The dose of Foxy-5 used in this series of experiments had no statistically significant effect on the growth of the primary tumor or on the weight gain of the animals (Fig. 5A and B). It should be mentioned that the volumes of the primary tumors in this series of experiments (Fig. 5A) were smaller than those in the previous series of experiments (Fig. 4A). This reduced growth of the primary tumor is a plausible explanation for the reduced number of metastases and smaller areas of the metastatic lesions observed in this series of experiments (Fig. 5C and D). Despite this difference, Foxy-5 inhibited the number of liver and lung metastases by 60% to 70% (Fig. 5C). These inhibitions were even more pronounced (70% and 90%) if the total area of tumor metastases were analyzed as an indicator of metastasis (Fig. 5D).

In the present investigation we have shown that the Foxy-5 peptide reduced 4T1 breast cancer metastasis to the lungs and liver. The clinical relevance of this finding is strengthened by the fact that we used an in vivo model in which 4T1 breast cancer cells, derived from a spontaneous breast cancer in BALB/c mouse (26), were inoculated into the mammary fat pad of immunocompromised (nude) or normal BALB/c mice. Orthotopically implanted syngeneic tumors are preferable when evaluating the response of a possible antimetastatic treatment, because the importance of the interaction of tumor cells with the surrounding tissue has become increasingly evident (31). In support of a metastasis-suppressing role of Wnt-5a, the Wnt-5a mimicking peptide Foxy-5 inhibited breast cancer metastasis to both the lungs and the liver in normal BALB/c mice. As could be expected, a higher dose of the Foxy-5 peptide was required in the normal BALB/c mice to attain a similar level of liver metastasis suppression as in the immunocompromised BALB/c mice. Despite the increased amount of Foxy-5 (40 μg, i.e., 1.6 mg/kg body weight) needed to obtain a 70% to 90% inhibition of tumor metastasis in normal mice, this amount of Foxy-5 is still relatively low when compared with similar studies using other types of tumor suppressors (3436). The increased metastasis in normal BALB/c mice is most readily explained by an intact immune system of this mouse strain, as has been previously observed (31). Activation of immune cells in such animals, and in particular tumor-associated macrophages, can promote metastasis formation (37). The increased metastasis to the lungs and liver in the normal BALB/c mice treated with the formylated control peptide relative to those given PBS alone support this idea because this peptide is known for its ability to activate murine macrophages, as also shown in our in vitro experiments. Importantly, the Foxy-5 peptide had no such effect on murine macrophage activation in vitro. Consequently, the finding that Foxy-5 has the ability to impair 4T1 breast cancer metastasis in BALB/c mice with a normal immune response, i.e., animals with a high metastatic spread, ensures that the antimetastatic effect shown by the Foxy-5–induced reconstitution of Wnt-5a signaling is not overestimated as could have been the case if only immunocompromised mice would have been used.

In this study, we used histomorphometrical analyses of lung and liver metastases as an end point to establish the effects of Foxy-5 on breast cancer cell metastasis. Although it was not possible to quantitatively assay tumor cell migration in vivo, the accumulated in vitro and in vivo data suggest that the main mechanism by which Foxy-5 suppresses the metastatic process is by impairing tumor cell migration and invasion. Previous investigations on how Wnt-5a signaling inhibits breast epithelial cell migration have indicated the involvement of several different mechanisms, including increased DDR1-dependent adhesion (4, 38), impaired NFAT activity (6), and an increased amount of actin stress fibers (5, 13). Consequently, to ascertain that all these mechanisms and also yet unidentified ones are activated, it makes more sense to use a receptor agonist than an approach that triggers activation of a specific downstream signaling event in the cascades of Wnt-5a signaling. In future studies it will be interesting to apply in vivo imaging techniques that enable longitudinal follow-up of the metastatic process in live animals (39, 40). The in vivo imaging will probably also enable studies on how Foxy-5 affects the dynamics of cancer cell migration and invasion at the cellular level (41).

In contrast to our previous findings, it has been suggested that Wnt-5a released from tumor-associated macrophages could have a chemotactic effect on breast cancer cells and thereby increase metastasis (29). Because breast cancer cells that exhibit an endogenous expression of Wnt-5a have a lower metastatic potential, this would mean that Wnt-5a must affect the tumor cells differently when released from tumor-associated macrophages than when released from breast cancer cells. However, we did not in the present study find a chemotactic effect of Wnt-5a or Foxy-5 on 4T1 breast cancer cell invasion when using a commercially available invasion assay, similar to that previously used (29). Thus, the correlation between reduced Wnt-5a expression in clinical breast cancer samples with reduced disease-free survival, the present and previous in vitro data, and the present in vivo data strongly support the concept that endogenous Wnt-5a expression in breast cancer cells and the reconstitution of Wnt-5a signaling in breast cancer cells with reduced endogenous Wnt-5a expression act to significantly reduce breast cancer metastasis. In principle, reconstitution of Wnt-5a signaling might have a similar therapeutic potential in other forms of cancer in which Wnt-5a has been shown to have a tumor suppressor function (1214, 42).

In conclusion, reconstitution of Wnt-5a signaling by the administration of Foxy-5 significantly inhibits the metastasis of 4T1 breast cancer cells. Importantly, this Wnt-5a–derived peptide ligand can serve as a model substance for the development of a novel therapeutic approach to specifically target the dissemination process in breast cancer metastasis, in particular for the ∼50% of human breast cancer patients that exhibit a reduced endogenous expression of Wnt-5a in their primary tumor (2).

T. Andersson has filed a patent for the use of the Foxy-5 peptide in anti-metastatic treatment.

Grant support: Swedish Cancer Foundation (T. Andersson, P. Härkönen), the Swedish Research Council (T. Andersson), the Research School in Pharmaceutical Sciences (T. Andersson), the U-MAS Research Foundations (T. Andersson), the Gunnar Nilsson's Cancer Foundation (T. Andersson), Sigrid Jusélius Foundation (P. Härkönen), and the Royal Physiografic Society (A. Säfholm).

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.

Note: Supplementary data for this article are available at Clinical Cancer Research Online (http://clincancerres.aacrjournals.org/).

We thank J. Seppänen for technical assistance with the animal experiments, M. Haglund for technical assistance with the immunohistochemistry staining, and Dr. L.M. Bjursten for statistical advice.

1
Jemal A, Siegel R, Ward E, Murray T, Xu J, Thun MJ. Cancer statistics, 2007.
CA Cancer J Clin
2007
;
57
:
43
–66.
2
Jonsson M, Dejmek J, Bendahl PO, Andersson T. Loss of Wnt-5a protein is associated with early relapse in invasive ductal breast carcinomas.
Cancer Res
2002
;
62
:
409
–16.
3
Dejmek J, Leandersson K, Manjer J, et al. Expression and signaling activity of Wnt-5a/discoidin domain receptor-1 and Syk plays distinct but decisive roles in breast cancer patient survival.
Clin Cancer Res
2005
;
11
:
520
–8.
4
Jonsson M, Andersson T. Repression of Wnt-5a impairs DDR1 phosphorylation and modifies adhesion and migration of mammary cells.
J Cell Sci
2001
;
114
:
2043
–53.
5
Safholm A, Leandersson K, Dejmek J, Nielsen CK, Villoutreix BO, Andersson T. A formylated hexapeptide ligand mimics the ability of Wnt-5a to impair migration of human breast epithelial cells.
J Biol Chem
2006
;
281
:
2740
–9.
6
Dejmek J, Safholm A, Kamp Nielsen C, Andersson T, Leandersson K. Wnt-5a/Ca2+-induced NFAT activity is counteracted by Wnt-5a/Yes-Cdc42-casein kinase 1α signaling in human mammary epithelial cells.
Mol Cell Biol
2006
;
26
:
6024
–36.
7
Jauliac S, Lopez-Rodriguez C, Shaw LM, Brown LF, Rao A, Toker A. The role of NFAT transcription factors in integrin-mediated carcinoma invasion.
Nat Cell Biol
2002
;
4
:
540
–4.
8
Yiu GK, Toker A. NFAT induces breast cancer cell invasion by promoting the induction of cyclooxygenase-2.
J Biol Chem
2006
;
281
:
12210
–7.
9
Roarty K, Serra R. Wnt5a is required for proper mammary gland development and TGF-β-mediated inhibition of ductal growth.
Development
2007
;
134
:
3929
–39.
10
Liang H, Chen Q, Coles AH, et al. Wnt5a inhibits B cell proliferation and functions as a tumor suppressor in hematopoietic tissue.
Cancer Cell
2003
;
4
:
349
–60.
11
Ying J, Li H, Chen YW, Srivastava G, Gao Z, Tao Q. WNT5A is epigenetically silenced in hematologic malignancies and inhibits leukemia cell growth as a tumor suppressor.
Blood
2007
;
110
:
4130
–2.
12
Kremenevskaja N, von Wasielewski R, Rao AS, Schofl C, Andersson T, Brabant G. Wnt-5a has tumor suppressor activity in thyroid carcinoma.
Oncogene
2005
;
24
:
2144
–54.
13
Dejmek J, Dejmek A, Safholm A, Sjolander A, Andersson T. Wnt-5a protein expression in primary dukes B colon cancers identifies a subgroup of patients with good prognosis.
Cancer Res
2005
;
65
:
9142
–6.
14
Ying J, Li H, Yu J, et al. WNT5A exhibits tumor-suppressive activity through antagonizing the Wnt/ -catenin signaling, and is frequently methylated in colorectal cancer.
Clin Cancer Res
2008
;
14
:
55
–61.
15
Kurayoshi M, Oue N, Yamamoto H, et al. Expression of Wnt-5a is correlated with aggressiveness of gastric cancer by stimulating cell migration and invasion.
Cancer Res
2006
;
66
:
10439
–48.
16
Weeraratna AT, Jiang Y, Hostetter G, et al. Wnt5a signaling directly affects cell motility and invasion of metastatic melanoma.
Cancer Cell
2002
;
1
:
279
–88.
17
Dissanayake SK, Wade M, Johnson CE, et al. The Wnt5A/protein kinase C pathway mediates motility in melanoma cells via the inhibition of metastasis suppressors and initiation of an epithelial to mesenchymal transition.
J Biol Chem
2007
;
282
:
17259
–71.
18
Ma P, Beck SL, Raab RW, et al. Heparanase deglycanation of syndecan-1 is required for binding of the epithelial-restricted prosecretory mitogen lacritin.
J Cell Biol
2006
;
174
:
1097
–106.
19
Yang-Snyder J, Miller JR, Brown JD, Lai CJ, Moon RT. A frizzled homolog functions in a vertebrate Wnt signaling pathway.
Curr Biol
1996
;
6
:
1302
–6.
20
Wang Y, Macke JP, Abella BS, et al. A large family of putative transmembrane receptors homologous to the product of the Drosophila tissue polarity gene frizzled.
J Biol Chem
1996
;
271
:
4468
–76.
21
Mikels AJ, Nusse R. Purified Wnt5a protein activates or inhibits β-catenin-TCF signaling depending on receptor context.
PLoS Biol
2006
;
4
:
e115
.
22
Tamai K, Semenov M, Kato Y, et al. LDL-receptor-related proteins in Wnt signal transduction.
Nature
2000
;
407
:
530
–5.
23
Nishita M, Yoo SK, Nomachi A, et al. Filopodia formation mediated by receptor tyrosine kinase Ror2 is required for Wnt5a-induced cell migration.
J Cell Biol
2006
;
175
:
555
–62.
24
Hartt JK, Barish G, Murphy PM, Gao JL. N-formylpeptides induce two distinct concentration optima for mouse neutrophil chemotaxis by differential interaction with two N-formylpeptide receptor (FPR) subtypes. Molecular characterization of FPR2, a second mouse neutrophil FPR.
J Exp Med
1999
;
190
:
741
–7.
25
Warri AM, Huovinen RL, Laine AM, Martikainen PM, Harkonen PL. Apoptosis in toremifene-induced growth inhibition of human breast cancer cells in vivo and in vitro.
J Natl Cancer Inst
1993
;
85
:
1412
–8.
26
Aslakson CJ, Miller FR. Selective events in the metastatic process defined by analysis of the sequential dissemination of subpopulations of a mouse mammary tumor.
Cancer Res
1992
;
52
:
1399
–405.
27
Hiraga T, Williams PJ, Ueda A, Tamura D, Yoneda T. Zoledronic acid inhibits visceral metastases in the 4T1/luc mouse breast cancer model.
Clin Cancer Res
2004
;
10
:
4559
–67.
28
Sen M, Chamorro M, Reifert J, Corr M, Carson DA. Blockade of Wnt-5A/frizzled 5 signaling inhibits rheumatoid synoviocyte activation.
Arthritis Rheum
2001
;
44
:
772
–81.
29
Pukrop T, Klemm F, Hagemann T, et al. Wnt 5a signaling is critical for macrophage-induced invasion of breast cancer cell lines.
Proc Natl Acad Sci U S A
2006
;
103
:
5454
–9.
30
Dolled-Filhart M, McCabe A, Giltnane J, Cregger M, Camp RL, Rimm DL. Quantitative in situ analysis of β-catenin expression in breast cancer shows decreased expression is associated with poor outcome.
Cancer Res
2006
;
66
:
5487
–94.
31
Talmadge JE, Singh RK, Fidler IJ, Raz A. Murine models to evaluate novel and conventional therapeutic strategies for cancer.
Am J Pathol
2007
;
170
:
793
–804.
32
Sanchez-Tillo E, Comalada M, Farrera C, Valledor AF, Lloberas J, Celada A. Macrophage-colony-stimulating factor-induced proliferation and lipopolysaccharide-dependent activation of macrophages requires Raf-1 phosphorylation to induce mitogen kinase phosphatase-1 expression.
J Immunol
2006
;
176
:
6594
–602.
33
Don MJ, Liao JF, Lin LY, Chiou WF. Cryptotanshinone inhibits chemotactic migration in macrophages through negative regulation of the PI3K signaling pathway.
Br J Pharmacol
2007
;
151
:
638
–46.
34
Hiraga T, Ueda A, Tamura D, et al. Effects of oral UFT combined with or without zoledronic acid on bone metastasis in the 4T1/luc mouse breast cancer.
Int J Cancer
2003
;
106
:
973
–9.
35
Huang X, Wong MK, Yi H, et al. Combined therapy of local and metastatic 4T1 breast tumor in mice using SU6668, an inhibitor of angiogenic receptor tyrosine kinases, and the immunostimulator B7.2-IgG fusion protein.
Cancer Res
2002
;
62
:
5727
–35.
36
Ge R, Rajeev V, Ray P, et al. Inhibition of growth and metastasis of mouse mammary carcinoma by selective inhibitor of transforming growth factor-β type I receptor kinase in vivo.
Clin Cancer Res
2006
;
12
:
4315
–30.
37
Coussens LM, Werb Z. Inflammation and cancer.
Nature
2002
;
420
:
860
–7.
38
Dejmek J, Dib K, Jonsson M, Andersson T. Wnt-5a and G-protein signaling are required for collagen-induced DDR1 receptor activation and normal mammary cell adhesion.
Int J Cancer
2003
;
103
:
344
–51.
39
Hoffman RM. Orthotopic metastatic mouse models for anticancer drug discovery and evaluation: a bridge to the clinic.
Invest New Drugs
1999
;
17
:
343
–59.
40
Hoffman RM. The multiple uses of fluorescent proteins to visualize cancer in vivo.
Nat Rev Cancer
2005
;
5
:
796
–806.
41
Yamauchi K, Yang M, Jiang P, et al. Development of real-time subcellular dynamic multicolor imaging of cancer-cell trafficking in live mice with a variable-magnification whole-mouse imaging system.
Cancer Res
2006
;
66
:
4208
–14.
42
Blanc E, Roux GL, Benard J, Raguenez G. Low expression of Wnt-5a gene is associated with high-risk neuroblastoma.
Oncogene
2005
;
24
:
1277
–83.

Supplementary data