Cancer Cell Biology
Downregulation of osteopontin contributes to metastasis suppression by breast cancer metastasis suppressor 1
Article first published online: 9 MAY 2008
Copyright © 2008 Wiley-Liss, Inc.
International Journal of Cancer
Volume 123, Issue 3, pages 526–534, 1 August 2008
How to Cite
Hedley, B. D., Welch, D. R., Allan, A. L., Al-Katib, W., Dales, D. W., Postenka, C. O., Casey, G., MacDonald, I. C. and Chambers, A. F. (2008), Downregulation of osteopontin contributes to metastasis suppression by breast cancer metastasis suppressor 1. Int. J. Cancer, 123: 526–534. doi: 10.1002/ijc.23542
- Issue published online: 28 MAY 2008
- Article first published online: 9 MAY 2008
- Manuscript Accepted: 24 JAN 2008
- Manuscript Received: 28 AUG 2007
- Canadian Institutes of Health Research. Grant Number: 42511
- Ontario Cancer Research Network. Grant Number: 04MAY00089
- U.S P.H.S. Grant Number: CA87728
- U.S. Army Medical Research and Materiel Command. Grant Number: DAMD17-2-01-0541
- breast cancer;
- in vivo
- Top of page
- Material and methods
- Supporting Information
Breast cancer metastasis suppressor 1 (BRMS1) inhibits the ability of multiple human and murine cancer cell lines to metastasize to lymph nodes, bones and lungs. Comparison of mRNA expression in metastatic MDA-MB-435 human carcinoma cells (435) and metastasis-suppressed BRMS1 transfectants (435/BRMS1) showed a marked (>90%) reduction of osteopontin (OPN) mRNA and protein expression in BRMS1-overexpressing cells. OPN expression is associated with disease progression in patients, with higher levels of OPN produced by cancer cells associated with poorer patient survival. Furthermore, OPN has been suggested to promote survival of cancer cells in response to stress, although the mechanisms by which this may occur remain poorly understood. This study tested the hypothesis that re-expression of OPN in metastasis-suppressed 435/BRMS1 cells would reverse metastasis suppression and confer protection from stress-induced apoptosis. A stable pooled population of OPN overexpressing 435/BRMS1 cells was created (435/BRMS1/OPN). OPN re-expression did not affect in vitro cell growth rates; however, increased anchorage independent growth/survival and protection from hypoxia-induced apoptosis was observed (p < 0.05). In vivo, OPN re-expression in BRMS1 transfected cells did not affect in vivo primary tumor growth but did increase the incidence of spontaneous metastasis to lymph nodes and lungs in mice. These novel findings suggest that OPN downregulation by BRMS1 may be responsible, at least in part, for BRMS1-mediated metastasis suppression by sensitizing cancer cells to stress induced apoptosis. These studies clarify one mechanism by which BRMS1 can suppress metastasis. © 2008 Wiley-Liss, Inc.
Breast cancer is the most frequent cancer among women in the United States,1, 2 with more than 175,000 new cases diagnosed each year and 40,000 deaths resulting from the disease. The majority of these breast cancer deaths are due to metastasis and not the primary breast tumor.3 Common sites of metastases in breast cancer are the lymph nodes, brain, liver and lung.4–6 Formation of metastases at any of these sites requires cancer cells to complete a number of sequential steps, and failure at any of these steps results in inhibition of metastasis. Currently, various combinations of adjuvant therapy are used as a strategy to reduce metastatic disease in patients at high risk of developing metastases,7 however these therapies reduce the risk of recurrence by only 20–30%.8, 9 In order to decrease morbidity and mortality from breast cancer, it is necessary to gain a greater understanding of metastasis and the molecular factors that contribute to this process. One such class of factors is a set of molecules called metastasis suppressors.
The role of metastasis suppressors in mediating the metastatic process has been investigated in several different cells lines and experimental models.10–13 These proteins differ from tumor suppressors in that they have little or no effect on primary tumor growth rate, while they do inhibit metastatic ability. The metastasis suppressor BRMS1 (breast cancer metastasis suppressor 1) has been shown to reduce the ability of multiple human and murine cell lines to metastasize to bone (Phadke et al., in press), lymph node and lung in experimental models.14–18In vitro, high BRMS1 expression has been shown to reduce breast cancer cell migration, but has no effect on either adhesion or invasion.16 Other studies have shown links between BRMS1 expression and nuclear factor κ-B activity,19, 20 phosphoinositide signaling,21 and altered connexin expression patterns, including restoration of gap junctional intercellular communication.22 The breadth of BRMS1-associated changes is partially explained in recent work by Meehan et al., who showed that BRMS1 may form part of large histone deacetylase complexes within the nucleus.23 However, the mechanisms by which BRMS1 results in suppression of metastasis remain poorly understood.
Previous work has shown that transfection of highly metastatic MDA-MB-435 cells with BRMS1 can reduce metastasis to the lymph node and lung.16 Interestingly, expression analysis has demonstrated that the protein osteopontin (OPN) is highly downregulated when BRMS1 is overexpressed in MDA-MB-435 cells.20 OPN is a secreted, integrin-binding glycoprotein that is widely expressed by numerous cancer types. OPN has been associated with disease progression in patients,24–26 with higher levels of OPN in plasma and tumor tissue associated with poorer breast cancer patient survival.25–29 Experimentally, OPN-mediated signaling results in altered patterns of gene expression, consistent with multiple aspects of malignancy.30 In experimental models of breast cancer, OPN has been shown to functionally contribute to increased malignancy, disease progression and metastasis.31–34 Interestingly, a recent study by Samant et al. demonstrated that a mechanism by which BRMS1 reduces OPN expression levels in 435/BRMS1 cells is via abrogation of NF-κB activation.20 However, the downstream functional implications of BRMS1-induced OPN downregulation on BRMS1-mediated breast cancer metastasis suppression remains to be elucidated.
In the current study, we therefore investigated the functional relationship between the metastasis-promoting protein OPN and the metastasis suppressor protein BRMS1 in an experimental model of breast cancer metastasis. We hypothesized that restoration of OPN levels in MDA-MB-435 breast carcinoma cells expressing high levels of BRMS1 might restore the metastatic ability of these cells. To test this hypothesis, full-length human OPN cDNA was transfected into BRMS1-expressing MDA-MB-435 cells, and the resulting functional effects on malignant and metastatic behavior in vitro and in vivo were examined.
Material and methods
- Top of page
- Material and methods
- Supporting Information
The MDA-MB-435 cell line (435) was originally isolated from the pleural effusion of a woman with metastatic breast adenocarcinoma.35 (It should be noted that there is debate over the origins of the 435 cell line, whether it was derived from the M14 melanoma cell or is in fact a true breast cancer cell line.36, 37 Whether melanoma or breast in origin, BRMS1 expression has been shown to act as a metastasis suppressor in both melanoma and breast human cell cancer lines.16, 17) The 435/BRMS1 cell line was generated by stable transfection of BRMS1 into the original cell line.16 All cell lines were grown in DMEM/F12 (GIBCO-BRL/Life Technologies, Grand Island NY) supplemented with 5% fetal calf serum (FCS; GIBCO-BRL/Life Technologies). 435/BRMS1 cells were supplemented with 500 μg/ml (active) G418 (GIBCO-BRL/Life Technologies) to maintain stable transgene expression. Cultures were regularly monitored to confirm that they were free of Mycoplasma (Mycosensor PCR Assay Kit; Stratagene, Mannheim, Indianapolis, IN).
Plasmids and transfections
Full-length human OPN33 was cloned into the pcDNA3.1/Hygro+ mammalian expression vector (Invitrogen, San Diego, CA) using the HindIII and BamHI sites to generate the pcDNA3.1/Hygro-OPN plasmid. Correct orientation and integrity of the OPN gene was confirmed with restriction digests and verified by sequencing (Genbank accession no. M83248). 435/BRMS1 cells were stably transfected with the pcDNA3.1/Hygro-OPN vector or a control vector (empty pcDNA3.1/Hygro+) using Lipofectamine 2000 (GIBCO-BRL/Life Technologies) as per the manufacturer's guidelines. Following transfection, cells were maintained in growth media containing either 750 μg/ml (435 cells) or 1,000 μg/ml (435/BRMS1 cells) of hygromycin B, levels previously determined to kill 100% of untransfected cells. Colonies were allowed to develop, and stably transfected pooled populations were generated by pooling 7 individual clones from each transfection. Pooled populations were named 435/BRMS1/OPN and 435/BRMS1 from the pcDNA3.1-OPN or pcDNA3.1+ vector transfections, respectively, and subsequently maintained in growth media containing 500 μg/ml (active) hygromycin B (GIBCO-BRL/Life Technologies).
Northern blot analysis
RNA was isolated using TRIzol reagent (Canadian Life Technologies, Burlington, Ontario) as per the manufacturer's guidelines. Total RNA (10 μg) of each sample was run on a 1% agarose gel with 6.8% formaldehyde, capillary transferred to nylon GeneScreen Plus membranes (NEN Life Science Products, Boston, MA) and probed with denatured, oligolabeled 32P-dCTP cDNA probes using an oligolabeling kit (Amersham Pharmacia Biotech, Baie d'Urfe', Quebec). The BRMS1 probe used was constructed from full-length human BRMS1 cDNA (1.5 kb) isolated using EcoRI and XbaI digestion of the Bluescript plasmid. The OPN probe used was a full-length human OPN cDNA (1,493 bp) generated using EcoRI digestion of the OP-10 plasmid.38 Equivalent loading of lanes was confirmed using a probe against human 18S mRNA. Densitometric analysis was carried out using ImageQuant software (Molecular Dynamics, Sunnyvale, CA).
Enzyme-linked immunosorbent assay
Secretion of OPN protein by stably transfected cells was measured by enzyme-linked immunosorbent assay (ELISA) of conditioned media. Conditioned media were prepared as previously described.39 For each cell line, a normalized volume of conditioned media equivalent to 1 × 105 cells was subjected to protein analysis using a validated in-house ELISA assay as described previously.26 Signal quantification was performed on a BioRad plate reader and samples were compared to a standard curve generated from recombinant human OPN in order to determine the amount of OPN protein present.
Western blot analysis
Cell lysates and conditioned media were prepared as described previously.19, 32 For each sample, 20 μg of protein (cell lysates) or a normalized volume of conditioned media (equivalent to 1 × 105 cells) was subjected to standard protein electrophoresis in 10% SDS-polyacrylamide gels and transferred onto polyvinylidene difluoride membranes (Immobilon™, Millipore; Bedford, MA). After transfer, gels were stained with Coomassie Blue to confirm equal loading and transfer efficiency. The primary antibodies used included specific mouse monoclonal antibodies against human BRMS1 (1:1,000),19 and human OPN (1:2,000) (mAb53, Assay Designs Inc., Ann Arbor, MI).29 The secondary antibody used was an anti-mouse HRP conjugate (Sigma) (1:2,000), and the blocking/dilution reagent used was 5% skim milk in TBST (Tris-buffered saline + 0.05% Tween-20). BRMS1 and OPN proteins were visualized using an enhanced chemiluminescence system (Roche Applied Sciences; Laval, QC).
Flow cytometry analysis
Primary antibodies used for flow cytometry were conjugated mouse anti-human monoclonal antibodies directed against integrin αvβ3, integrin αvβ5, integrin α9β1 and CD44 (Chemicon International, Temecula, CA and BD Biosciences). The negative IgG isotype control was a mouse primary IgG1 antibody (Cedarlane Laboratories, Hornby, ON, Canada). Samples were incubated with primary antibodies (1 μg/106 cells) for 1 hr at 4°C on a rotating plate, washed twice with phosphate-buffered saline (PBS) plus 2% fetal bovine serum, re-suspended in 500 μl of PBS plus 2% serum. Samples were analyzed using a Beckman-Coulter EPICS XL-MCL flow cytometer.
Cell proliferation assays
Cell growth in two-dimensional (anchorage-dependent) culture was determined as described previously.32 The doubling time of each cell population was estimated during the exponential growth phase according to the equation Td = 0.693t/ln(Nt/N0), where t is time (in hour), Nt is the cell number at time t and N0 is the cell number at initial time. To determine cell proliferation kinetics in three-dimensional (anchorage-independent) culture, a soft agar assay (0.3%) was used as described by Cook et al.40
To determine the levels of apoptosis under stress, cells were seeded onto 100 mm2 tissue culture dishes at a density of 1 × 105 cells per plate. The following day, culture medium was removed and replaced with serum-free medium (SFM) and left for 24 hr under normoxic conditions (20% O2).41, 42 SFM was removed, replaced with regular growth media, and cells were subjected to the appropriate O2 condition. “Control” cells were maintained in 20% O2 throughout the experimental protocol. “Hypoxia” cells were exposed to 1% O2, 5% CO2 and balanced with N2 in humidified sealed chambers (Billups-Rothenburg, Del Mar, CA) at 37°C for 24 hr. After 24 hr, cells were detached using 2 mM EDTA, washed in ice cold PBS and stained using propidium iodide and Annexin V-FITC (BD Biosciences) for 15 min at room temperature in the dark. Samples were analyzed using a Beckman-Coulter EPICS XL-MCL flow cytometer.
In vivo assessment of tumorigenicity and metastatic ability
Animal procedures were carried out under a protocol approved by the University of Western Ontario Council of Animal Care. Cell lines (435, 435/BRMS1, 435/BRMS1/OPN) were grown to ∼80% confluency in 150 mm2 dishes. Cell suspensions were prepared in Hank's Buffer Salt Solution, HBSS (Sigma), at a concentration of 1 × 107 cells/ml. Cell suspension of 100 μl was injected into the second left thoracic mammary fat pad of 6- to 7-week-old female athymic nude (nu/nu) mice (Harlan Sprague-Dawley, Indianapolis, IN) as described elsewhere,32, 35, 43 using 12 mice per cell line. Primary tumor growth was evaluated twice weekly by measurement with digital calipers in 2 perpendicular dimensions. Estimation of tumor volume (mm3) was calculated using the formula [volume = 0.52 (width)2 (length)]. Mice were allowed to develop tumors until the mean tumor volume of the group had reached 1,500 mm3. Once this volume was reached tumors were resected as described previously43 and mice were left for a further 8 weeks or until mice became moribund.
Histopathology and immunohistochemistry
Formalin-fixed, paraffin-embedded tissues (primary tumor, lymph nodes and lungs) were processed and subjected to standard Hematoxylin and Eosin (H&E) staining and/or immunohistochemistry (IHC) for BRMS1 and OPN. Immunostaining was performed essentially as described previously29, 44 using a streptavidin-biotin complex approach (DakoCytomation, Mississauga, ON). Primary monoclonal antibodies used included 3a1.21 anti-BRMS1 antibody44 (using the Animal Research Kit; Sigma) and mAb53 (Assay Designs, Ann Arbor, MI)29 for OPN. The chromogen used was amino-ethyl carbazol, and slides were counterstained with Meyer's hematoxylin. Slides were analyzed in a blinded fashion by an experienced pathologist in order to determine pathohistological features, micrometastatic involvement and immunostaining for BRMS1 and OPN. Metastatic tumor burden (tumor area/organ total area) was determined quantitatively, as described previously,45 by analysis of histological slides using ImageJ software from the National Institutes of Health. Color RGB (Red Green Blue) images were taken using a Nikon© microscope and converted to 8-bit images. Thresholding was performed on each picture to obtain a black (tumor cells) and white (background) image. Quantification was done using the ‘analyze particles’ option of the ImageJ software after the scale had been set for the microscope objective used.
All in vitro and in vivo experiments were performed at least in triplicate, and data was compiled from 3 separate experiments. Statistical analysis was performed using Graphpad Prism 4.0 software© (San Diego, CA) using either t-test (for comparison between 2 groups), ANOVA with Tukey post test (for comparison between more than 2 groups), Kruskal-Wallis test (for the in vitro soft agar assays) or Fisher's Exact Test (for comparison of proportions for in vivo studies) to determine significance between groups. In all cases, differences were considered to be significant when the p value was less than 0.05.
- Top of page
- Material and methods
- Supporting Information
Characterization of OPN and BRMS1 expression in transfected cell lines
OPN mRNA and protein levels were characterized in the parental 435 cells (435) and 435 cells transfected to overexpress BRMS1 (435/BRMS1). OPN levels were found to be highly downregulated at the mRNA level (greater than 99%) (Fig. 1a) and at the protein level (greater than 90% by ELISA) (Figs. 1b and 1c) when BRMS1 was expressed in these breast cancer cells.
To determine whether re-expression of OPN in BRMS1-expressing MDA-MB-435 cells could restore metastatic phenotype, 435/BRMS1 cells were stably transfected to overexpress human OPN. It should be noted that parental MDA-MB-435 cells were not transfected with the OPN expression vector, since this cell line already expresses high amounts of endogenous OPN (Fig. 1). Cell lines (435, 435/BRMS1 and 435/BRMS1/OPN) were characterized for BRMS1 and OPN mRNA expression by Northern blot analysis, for OPN secretion by ELISA and Western blot analysis and for BRMS1 protein expression by Western blot analysis. OPN mRNA (Fig. 1a) and protein levels (Figs. 1b and 1c) were found to be elevated after OPN transfection relative to 435/BRMS1 expressing cell lines, whereas BRMS1 mRNA (Fig. 1a) and protein levels (Fig. 1c) remained unchanged. Six clones were pooled to form the 435/BRMS1/OPN cell line, all clones showed maintained BRMS1 expression and higher OPN expression relative to control (Supplemental Figs. 1a and 1b).
Flow cytometry was used to analyze the expression of OPN receptors (including OPN-related integrins and CD44) by MDA-MB-435 (435), MDA-MB-435/BRMS1 (435/BRMS1) and MDA-MB-435/BRMS1/OPN (435/BRMS1/OPN) cells (Supplemental Fig. 2). Cells were incubated with anti-CD44, anti-integrin (αvβ3, αvβ5 and α9β1) (black) or nonspecific isotype control antibodies (clear). Cell surface expression of these receptors was assessed by flow cytometry as described in “Material and methods” section. Surface expression of αvβ3 was present in 435 cells but downregulated to the same level as the negative control in 435/BRMS1 and 435/BRMS1/OPN cells. Integrin αvβ5 was highly expressed in 435 cells but showed decreased expression in both 435/BRMS1 and 435/BRMS1/OPN cells. Surface expression of α9β1 was high in 435 cells, downregulated in 435/BRMS1 cells and absent in 435/BRMS1/OPN cells. Surface expression of CD44 was conserved between cell lines, with CD44 being highly expressed in 435 cells and remaining positive in both other cell lines.
OPN re-expression counteracts BRMS1-mediated inhibition of anchorage-independent growth
To assess the effect of BRMS1 on anchorage-independent growth of 435 cells and the effect of OPN re-expression, in vitro assays were used to compare the cell growth and survival characteristics of 435, 435/BRMS1 and 435/BRMS1/OPN cell lines (Fig. 2). In two-dimensional (anchorage-dependent) culture, doubling times were not statistically different among the cell lines (Fig. 2a). Under anchorage-independent conditions (in soft agar), BRMS1 expression alone was found to significantly decrease the number (Fig. 2b) and median size (Fig. 2c) of colonies (Kruskal-Wallis test, p < 0.001) relative to parental 435 cells. Re-expression of OPN resulted in a significant increase in the number (Fig. 2b) and median size (Fig. 2c) of colonies (Kruskal-Wallis test, p < 0.001) relative to cells transfected with BRMS1 alone, although the numbers are still significantly lower than those observed for the parental 435 cells.
The partial re-expression of OPN level achieved in the 435/BRMS1/OPN cell line (approximately 50% of parental 435 cells, see Fig. 1) almost exactly paralleled the increase in anchorage-independent growth suggesting a possible dose-dependent effect of OPN. To investigate this further, experiments were performed with clonal populations of 435/BRMS1 cells transfected with OPN with varying OPN expression levels. A small but significant dose-dependent effect of OPN on BRMS1-mediated metastasis was observed in anchorage-independent growth and all OPN-expressing clones had increased anchorage-independent growth relative to 435/BRMS1 cells (Supplemental Fig. 1).
OPN re-expression counteracts BRMS1-mediated induction of apoptosis in response to hypoxia
To assess the effect of BRMS1 on stress-induced apoptosis in 435 cells and to determine the effect of OPN re-expression, cells were grown normally or under conditions of hypoxic stress (Fig. 2d). All cells grown under hypoxic conditions demonstrated more apoptosis than the same cell type grown under normal culture conditions (ANOVA, p < 0.05). BRMS1 expression alone significantly increased the levels of apoptosis under hypoxic conditions (ANOVA, p < 0.05), and increased apoptosis mediated by BRMS1 was also seen under normoxic conditions (p < 0.05). Interestingly, restoration of OPN levels resulted in a significant decrease in the level of apoptosis under hypoxic conditions (ANOVA, p < 0.05). However, consistent with the results of the anchorage-independent growth assay, under hypoxic stress the OPN-mediated decrease in apoptosis was not sufficient to return the cells to parental 435 levels. As was the case for three-dimensional growth, OPN re-expression in all BRMS1-expressing cells only partially restored the resistance to stress induced apoptosis seen in 435 cells.
Results of the hypoxia assay with clonal populations revealed similar increases in survival to that of 435/BRMS1/OPN cells, however, in contrast to the anchorage-independent growth assay no dose dependent effect was seen (Supplemental Fig. 3). As was the case for three-dimension growth, OPN re-expression in all BRMS1-expressing cells (pooled and clonal cells) only partially restored the resistance to stress induced apoptosis seen in 435 cells.
OPN re-expression counteracts BRMS1-mediated inhibition of metastatic ability
To assess the in vivo effects of expression of BRMS1 (alone or in combination with OPN) 435, 435/BRMS1 and 435/BRMS1/OPN cells were injected into the mammary fat pad of female nude mice and assessed for differences in tumorigenicity and spontaneous metastasis to the lymph nodes and lung. All cell lines formed primary tumors in mice and grew at similar rates (Fig. 3), with no difference in tumor doubling times (ANOVA, p > 0.05). Once the mean tumor volume of each group reached 1,500 mm3, primary tumors were resected as described previously.32, 43 Histopathological features of the resected primary tumors were assessed by an experienced pathologist and all were classified as invasive carcinomas of the breast of “no special type.” All primary tumors contained similar degrees of necrosis, and all tumors contained lymphovascular invasion (data not shown).
At the time of euthanasia, 7 weeks after resection of the primary tumor, mice injected with 435/BRMS1 cells had a significantly lower incidence of axillary lymph node (LN) metastases relative to mice injected with parental 435 cells (Fig. 4a, Fisher's Exact Test, p < 0.0001). Re-expression of OPN in 435/BRMS1 cells resulted in a significant restoration of metastatic incidence to the LN, although again, not to the same level as parental 435 cells (Fig. 4a). However, no dose dependent effect of OPN expression level on the incidence of LN metastases was seen with the clonal populations (Supplemental Table I, p > 0.05). Because local tumor regrowth following resection of the primary tumor is a possible factor that can increase the incidence of LN metastases, mice with primary tumor recurrence were removed from subsequent analyses (435: 4 of 12 mice, 435/BRMS1: 3 of 12 mice, 435/BRMS1/OPN: 3 of 12 mice). Tumor burden in the LN (i.e. percentage area of LN occupied by tumor) was significantly lower in BRMS1-expressing cells (ANOVA, p < 0.01) and significantly restored (ANOVA, p < 0.05) when OPN was re-expressed (Fig. 4b). Similarly, the mean size of LN metastases, as assessed histologically (Figs. 4c and 4d), was significantly lower in mice injected with 435/BRMS1 cells relative to 435 cells (p < 0.0001), and partially but significantly restored in mice injected with 435/BRMS1/OPN cells (p < 0.001).
Analysis of spontaneous metastasis to the lung revealed that OPN expression by transfection in BRMS1-expressing cells resulted in a trend toward increased incidence of lung metastases relative to cells expressing BRMS1 alone (Fig. 5a, ANOVA, p = 0.08). However, unlike metastatic ability to the LN, metastatic tumor burden in the lung was not significantly different between mice injected with 435/BRMS1 versus 435/BRMS1/OPN cells (Fig. 5b, ANOVA, p = 0.5). In parallel, histological analysis demonstrated that the mean size of lung metastases was significantly decreased in mice injected with 435/BRMS1 expressing cells, and OPN did not significantly increase this (Figs. 5c and 5d).
Immunohistochemical staining for OPN was performed on LN and lung metastases for mice in all experimental groups. OPN levels (cytoplasmic) were highest in metastases of mice injected with 435 cells, and lowest in metastases of mice injected with 435/BRMS1 cells. OPN re-expression in BRMS1-expressing cells resulted in increased OPN levels (Figs. 4d and 5d). BRMS1 immunohistochemical staining (nuclear) was also performed on the same tissues as were used for OPN IHC. BRMS1 expression was lowest in metastases of mice injected with 435 cells, and highest in metastases of mice injected with either 435/BRMS1 or 435/BRMS1/OPN cells (Figs. 4d and 5d).
- Top of page
- Material and methods
- Supporting Information
The majority of breast cancer patient deaths are due to irreversible physiological effects of metastasis on normal organ function rather than from the primary tumor. Current surgical and adjuvant therapies have only limited success in reducing the rate of disease recurrence in breast cancer, suggesting that a better understanding of the metastatic process could lead to new avenues for improved prognosis and treatment.
The metastasis suppressor gene BRMS1 has been shown to reduce the ability of multiple human and murine cell lines to metastasize to LN and lung in experimental models,14–17, 20 including the highly metastatic MDA-MB-435 human breast cancer cell line.16 Similar to many aggressive cancer cell lines, MDA-MB-435 cells have been shown to endogenously express high levels of the metastasis promoting protein OPN. Interestingly, expression analysis suggested that transfection-mediated BRMS1 overexpression in MDA-MB-435 cells leads to a reduction in OPN expression, and that a molecular mechanism through which BRMS1 accomplishes this downregulation is via in part reduction of NFκB activation.20 However, the downstream implications of these mechanistic observations for regulating breast cancer cell behavior both in vitro and in vivo required further elucidation. The results of the current study provide, for the first time, experimental evidence of functional link between the metastasis suppressor protein BRMS1 and the metastasis promoting protein OPN.
Previous studies have shown that increased OPN expression leads to increased malignant behavior of cancer cells.30–34, 40 Interestingly, OPN downregulation by BRMS120 has been shown to have little or no effect on in vitro migration, adhesion, or invasion.40 Our findings show that BRMS1 also had no effect on two-dimensional growth, and BRMS1 levels remained unchanged upon OPN re-expression. However, our results indicate that BRMS1 overexpression in MDA-MB-435 cells can significantly decrease other in vitro malignant properties, including colony formation in an anchorage-independent environment (OPN re-expression partially restored colony number and size) and survival under stressed (hypoxic) conditions (OPN re-expression partially reduced the pro-apoptotic effect of BRMS1). Our results for anchorage-independent growth are consistent with previous results demonstrating that OPN overexpression resulted in increased colony forming ability in 21NT human breast cancer cells,40 OPN has also been shown to mitigate the effects of stress-induced apoptosis in endothelial cells,46 and also by recent work by Graessmann et al. showing that OPN is the main anti-apoptotic factor released by chemotherapy resistant breast cancer cells.47 Our functional in vitro findings therefore demonstrate that while anchorage-independent growth and survival (after a pro-apoptotic stimulus) of MDA-MB-435 cells is significantly suppressed by BRMS1 expression, the metastasis promoting protein OPN can effectively counteract these BRMS1-mediated effects.
In vivo, re-expression of OPN in BRMS1-overexpressing cells did not affect primary tumor growth rate or histological features, but did significantly increase the incidence of metastasis to both the LN and lung. Additionally, OPN significantly increased the metastatic tumor burden within the LN, but not the lung. It should be noted that OPN expression levels were restored in the pooled population of 435/BRMS1/OPN cells to approximately 50% of that observed in parental 435 cells. Strikingly, this partial but significant reexpression of OPN level almost exactly paralleled the partial but significant restoration of malignant and metastatic abilities in BRMS1-expressing cells. The observation that OPN can counteract, at least in part, the metastasis-suppressing effect of BRMS1 on growth and survival of MDA-MB-435 cells suggests a possible functional link between BRMS1 and OPN in mediating breast cancer malignancy.
For successful metastatic growth, the molecular interactions between cancer cells and their microenvironment are critical.32, 48 Taken together with the in vitro results, our in vivo data suggest that OPN may provide breast cancer cells with a selective advantage to survive and grow in certain metastatic sites, even in the context of the BRMS1 metastasis suppressor gene. This is supported by our previous experimental work in other cell lines demonstrating that OPN expression gives rise to increased lymphatic involvement and LN metastases in mouse models of breast cancer, and that OPN interaction with the extracellular matrix via integrin and non-integrin binding receptors on the cancer cells may allow for differential growth and survival in the LN versus lung.30–33, 43
Recent studies have shown that it is the interaction between OPN and its receptors (in particular integrins) that governs OPN-mediated cellular behavior such as growth, survival and metastasis.32, 49, 50 Interestingly, results from the study presented here show that BRMS1 reduces not only the expression of OPN, but also the expression of 3 integrin receptors (αvβ3, αvβ5 and α9β1) that have previously been shown to bind OPN and mediate its biologic effects on cell behavior.33 While re-expression of OPN in BRMS1 expressing cells did in part reverse the metastasis suppressing effects of BRMS1, integrin expression remained low. This suggests that 1 reason why OPN only partially (instead of fully) counteracted BRMS1-mediated inhibition of metastasis is that integrin receptors are not expressed at a high enough level to fully mediate OPN's functional effect in BRMS1 expressing cells. In contrast, the OPN receptor CD44 shown little or no change in expression in response to BRMS1, suggesting that OPN may be exerting some of its effect on BRMS1 expressing cells through this receptor instead of through integrins. CD44 expression may also explain the growth seen in the LN but not in lung, with OPN/CD44 related signaling only sufficient to promote survival and growth within the LN but not in lung. Furthermore, the presence of other CD44 ligands that may stimulate growth (such as hyaluronan) may be found in differing amounts within the LN versus the lung.
Clinical studies have demonstrated the importance of OPN in disease progression and patient survival.24–29 The work presented here and the findings of other experimental studies16, 20 demonstrate that the metastasis suppressor gene BRMS1 can reduce metastasis to LN and lung. However, the data to date regarding the clinical importance of BRMS1 is less clear, with some studies showing a reduction in disease-free survival in the presence of BRMS1 expression,44, 51 whereas others show no reduction.52 Limitations to the study by Kelly et al.52 are stromal contamination within the samples analyzed and that only BRMS1 mRNA, and not protein, was assessed in the primary tumor or LN metastases.
Clinically, the significance of LN positivity in managing treatment for breast cancer patients cannot be understated.53, 54 The presence of even a small amount of metastatic tumor tissue (foci > 2 mm) within the LN has been shown to negatively impact survival of breast cancer patients.32 Furthermore, recent clinical breast cancer studies by our group have demonstrated that OPN levels were significantly higher in the LN than in the primary tumor.32 In light of the findings presented here which demonstrate a functional relationship between BRMS1 and OPN in mediating metastatic behavior (particularly in the LN), it is possible that in clinical cases where BRMS1 expression shows no association with reduced disease free survival, OPN expression by these tumors might be counteracting some of the effects of BRMS1 expression. It thus may be of value to examine BRMS1 and OPN levels clinically in the same tumors. Furthermore, as breast cancer is commonly treated with radiation and chemotherapy (known inducers of apoptosis), examination of the relationship between BRMS1 and OPN on apoptosis induction by ionizing radiation and/or chemotherapy may provide valuable clinical insight.
In summary, the present study indicates that BRMS1-mediated metastasis suppression involves downregulation of OPN and subsequent reduction in survival and growth properties, and demonstrates a functional link between BRMS1 and OPN that was suggested by the mechanistic observations of Samant et al.20 Furthermore, the novel findings presented here suggest that in the presence of high levels of metastasis promoters (such as OPN), expression of metastasis suppressor genes (such as BRMS1) may not be sufficient for suppression of LN and lung metastases, and also suggest that a balance between metastasis promoters and metastasis suppressors are key for regulating the metastatic process. Both ideas have important implications for the development of future therapeutic strategies against breast cancer metastasis. A recent experimental study has shown that in vivo restoration of another metastasis suppressor gene, Nm23-H1, is therapeutically feasible and causes a reduction in number and size of lung metastases in mice.55 These findings hold promise for the idea that therapies designed to elevate the expression of metastasis suppressor genes could provide effective treatment for metastatic disease in patients; however, the findings of the present study suggest that high levels of metastasis promoters such as OPN are able to counteract the dramatic effect of a metastasis suppressor gene. Therefore, a good understanding of the molecular pathways involved in regulating the balance between metastasis-promoting and metastasis-suppressing molecules will have important implications for patient prognosis, treatment and development of future therapies.
- Top of page
- Material and methods
- Supporting Information
We thank Dr. Muzaffer Cicek for technical assistance with some of the studies. Dr. Ann F. Chambers is the recipient of a Canada Research Chair in Oncology. Dr. Benjamin D. Hedley is the recipient of a Translational Breast Cancer Studentship from the London Regional Cancer Program.
- Top of page
- Material and methods
- Supporting Information
- 3Principles and practice of oncology, 7th Bk&Cdr edn. Philadelphia: Lippincott Williams & Wilkins, 2005., , .
- 7Early Breast Cancer Trialists' Collaborative Group (EBCTCG). Effects of radiotherapy and of differences in the extent of surgery for early breast cancer on local recurrence and 15-year survival: an overview of the randomised trials. Lancet 2005; 366: 2087–106., , , , , , , , , , , , et al.
- Top of page
- Material and methods
- Supporting Information
This article contains supplementary material available via the internet at http://www.interscience.wiley.com/jpages/0020-7136/suppmat .
|ijc23542-Sup_Table1.doc||43K||Supporting Information file ijc23542-Sup_Table1.doc|
|ijc23542-SupFig1.eps||20236K||Supporting Information file ijc23542-SupFig1.eps|
|ijc23542-SupFig2.eps||20170K||Supporting Information file ijc23542-SupFig2.eps|
|ijc23542-SupFig3.eps||8416K||Supporting Information file ijc23542-SupFig3.eps|
Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.