Osteopontin (OPN) is a secreted phosphoprotein that has been associated with malignancy of breast and other cancers. OPN binds to several cell surface integrins including αvβ3, αvβ5, and αvβ1. Although the relative contribution of these integrins to breast cancer cell malignancy is uncertain, correlative studies suggest that αvβ3 may be particularly associated with increased tumor aggressiveness. Previously, we reported that tumorigenic, nonmetastatic 21NT mammary carcinoma cells respond to OPN through αvβ5 and αvβ1 but not αvβ3. Here, we determined that 21NT cells lack β3 expression, and we asked whether expression of αvβ3 could enhance the ability of breast cancer cells to respond to the malignancy-promoting effects of OPN both in vitro and in vivo. 21NT cells stably transfected with β3 showed significantly increased adhesion, migration, and invasion to OPN in vitro compared with vector control. To determine if β3 could also enhance the response of breast epithelial cells to OPN in vivo, cells stably transfected with both β3 and OPN (NT/Oβ3) were injected into the mammary fat pad of female nude mice and primary tumor growth was assessed relative to controls. Mice injected with NT/Oβ3 cells demonstrated a significantly increased primary tumor take (75% of mice) compared with controls (0–12.5% of mice) as well as a decreased tumor doubling time and a decreased tumor latency period. These results suggest that increased expression of the αvβ3 integrin during breast cancer progression can make tumor cells more responsive to malignancy-promoting ligands such as OPN and result in increased tumor cell aggressiveness.

Osteopontin (OPN) is a secreted glycoprotein that is overexpressed in a number of different carcinomas, including breast cancer (1, 2). OPN levels have been found to be elevated in primary tumors or blood of patients with breast cancer and in some cases have been reported to be associated with poor prognosis (3–7). OPN may also be a useful clinical marker for ovarian, colon, lung, liver, prostate, and perhaps other cancers (8–12).

Elevated OPN expression has also been associated with increased malignancy and metastasis in experimental models of breast cancer (13–16). Functional studies have shown that breast epithelial cells in culture can adhere to OPN (17), migrate toward OPN (18), and invade through basement membrane in response to OPN (15). Associated with these behavioral changes are OPN-mediated changes in gene expression, including increased expression of urokinase-type plasminogen activator as well as increased expression and activation of the hepatocyte growth factor receptor (Met) and the epidermal growth factor receptor (18–20). Cells of higher in vivo malignancy not only tend to express more OPN but also may be more responsive to OPN (15). OPN binding has been found to involve both integrin and nonintegrin receptors in different cell types (17, 18, 21–27). The primary integrin receptors for OPN are αvβ3, αvβ1, and αvβ5 (26, 27). Correlative evidence has shown that the αvβ3 integrin receptor appears to be preferentially used by more malignant breast epithelial cell lines in binding and migrating toward OPN or vitronectin (18, 28). However, it is not known whether experimentally altered levels of αvβ3 can lead to changes in cellular behavior in response to OPN.

Our previous studies (18) have shown that tumorigenic, nonmetastatic 21NT cells (29) differ from tumorigenic, metastatic MDA-MB-435 cells (30) in terms of the specific integrin receptors involved in their response to OPN. Although both cell lines respond to OPN with increased adhesion, migration, and invasion, blocking antibody studies show that 21NT cells respond to OPN via αvβ5 and αvβ1 integrins, whereas MDA-MB-435 cells preferentially use αvβ3 (18). Our goal in the present study was to determine if αvβ3 expression could confer on 21NT cells a more aggressive phenotype by altering their responsiveness to OPN. We first characterized the integrin expression profiles of the two cell lines by flow cytometry and found that both NT/C (control vector-transfected 21NT cells) and MDA-MB-435 cells express αvβ5 and αvβ1 integrins, but only MDA-MB-435 cells express αvβ3. To determine whether increased αvβ3 expression in 21NT cells can increase OPN responsiveness, we generated stable transfectants of 21NT cells that overexpress the β3 integrin subunit (NT/Cβ3). In vitro, the NT/Cβ3 cells were found to adhere better, migrate better, and invade through basement membrane better in response to OPN than the corresponding vector controls. For the in vivo studies, we stably transfected 21NT cells to express both β3 and OPN (NT/Oβ3). After having verified high levels of β3 and OPN in the cotransfectants, we assessed the ability of these cells to form primary mammary tumors relative to cells that express β3 alone (NT/Cβ3), OPN alone (NT/O), or the vector control (NT/C). Following mammary fat pad injection, mice injected with the NT/Oβ3 cells demonstrated a significantly increased primary tumor take, a decreased tumor doubling time, and a decreased tumor latency period relative to controls. Our findings suggest that expression of the αvβ3 integrin may alter how a tumor cell interacts with its environment, at least in part by increasing sensitivity to malignancy-enhancing ligands such as OPN.

Metastatic MDA-MB-435 and Nonmetastatic 21NT Breast Carcinoma Cells Show Different Integrin Expression Patterns

Flow cytometry was used to test for surface expression of the αvβ5, αvβ1, and αvβ3 integrins in MDA-MB-435 and NT/C cells. Both MDA-MB-435 and NT/C cells expressed αvβ5 and αvβ1 integrins (Fig. 1A). When both cell lines were also tested for β3 integrin expression, MDA-MB-435 cells were found to have high levels of surface β3, while NT/C cells were negative for β3 (Fig. 1B). These results thus explain our previous findings (18) that the migration of 21NT cells to OPN was not reduced by blocking αvβ3 due to their lack of expression of the β3 subunit.

FIGURE 1.

Flow cytometry analysis of the integrin expression profiles of MDA-MB-435, NT/C, and NT/Cβ3 cells. Cells were incubated with anti-integrin antibodies, as indicated, or with nonspecific isotype control primary antibodies (blue), and surface expression of specific integrins was assessed by flow cytometry, as described in “Materials and Methods.” A. Surface expression of αvβ5 (yellow) and αvβ1 (green) by MDA-MB-435 and NT/C cells. B. Surface expression of β3 (red) by MDA-MB-435 and NT/C cells. C. Surface expression of β3 (red) by control vector NT/C cells and NT/Cβ3 transfectants (β3 clone A4, β3 pool A, and β3 pool C).

FIGURE 1.

Flow cytometry analysis of the integrin expression profiles of MDA-MB-435, NT/C, and NT/Cβ3 cells. Cells were incubated with anti-integrin antibodies, as indicated, or with nonspecific isotype control primary antibodies (blue), and surface expression of specific integrins was assessed by flow cytometry, as described in “Materials and Methods.” A. Surface expression of αvβ5 (yellow) and αvβ1 (green) by MDA-MB-435 and NT/C cells. B. Surface expression of β3 (red) by MDA-MB-435 and NT/C cells. C. Surface expression of β3 (red) by control vector NT/C cells and NT/Cβ3 transfectants (β3 clone A4, β3 pool A, and β3 pool C).

Close modal

Transfection and Characterization of 21NT Cells Overexpressing the β3 Integrin Subunit

Following determination that the NT/C cells did not express the β3 integrin subunit, we stably transfected them with full-length human β3 integrin cDNA or with the vector only as a control. Clonal and pooled cell populations were isolated and characterized for their β3 protein and mRNA expression levels using flow cytometry and Northern analysis, respectively. Three NT/Cβ3-transfected cell populations, representing a clone and two pools, were characterized in addition to a pooled population of control vector-transfected NT/C cells (Fig. 1C). Flow cytometry analysis demonstrated that the NT/Cβ3 clone A4 cell population had high surface expression levels of β3 (98% of cells), while NT/Cβ3 pools A and C were more heterogeneous (as expected for pooled populations), with 62% and 80%, respectively, of cells expressing β3 integrin. Consistent with the results shown in Fig. 1B, vector control NT/C cells showed undetectable levels of β3 integrin (Fig. 1C). The transfected cells were also characterized for β3 mRNA levels using Northern analysis (Fig. 2). The NT/Cβ3 clone A4 and pools A and C were found to express high levels of β3 mRNA compared with the NT/C control transfected cells. These three β3-transfected cell populations were used in the subsequent studies to assess the abilities of the cells to respond to OPN relative to NT/C vector control cells.

FIGURE 2.

Northern blot analysis of NT/Cβ3 cells. RNA from control vector NT/C cells and NT/Cβ3 transfectants (β3 clone A4, β3 pool A, and β3 pool C) were assessed by Northern blotting for β3 expression, with 18S assessment as loading control, as described in “Materials and Methods.”

FIGURE 2.

Northern blot analysis of NT/Cβ3 cells. RNA from control vector NT/C cells and NT/Cβ3 transfectants (β3 clone A4, β3 pool A, and β3 pool C) were assessed by Northern blotting for β3 expression, with 18S assessment as loading control, as described in “Materials and Methods.”

Close modal

Expression of β3 Integrin in 21NT Breast Carcinoma Cells Increases Cell Adhesion to OPN

In cell adhesion assays, all of the cells tested adhered better to OPN than to BSA (P < 0.01 for all) and those transfected with β3 adhered significantly better to OPN than did the control vector-transfected cells (P < 0.002 for all). In the presence of an anti-αvβ3 antibody, adhesion of NT/Cβ3 cells to OPN was significantly inhibited (P < 0.002 for all) while adhesion of NT/C cells was not inhibited, confirming that this effect was β3 specific (Fig. 3A). Incubation with a control anti-immunoglobulin G antibody had no effect on adhesion of either NT/C or NT/Cβ3 cells (data not shown). Fig. 3B shows the cell morphology of representative NT/C and NT/Cβ3 cells following adhesion to BSA, OPN, or OPN in the presence of the anti-αvβ3 antibody. The NT/Cβ3 cells showed dense packing of adherent, well-spread cells on the OPN substrate, which was inhibited in the presence of the anti-αvβ3 antibody. In contrast, the NT/C cells spread poorly on OPN and were unaffected by the anti-αvβ3 antibody.

FIGURE 3.

Expression of the β3 integrin increases adhesion of NT/C cells to OPN. A. Adhesion of NT/Cβ3 and NT/C control cells to OPN (10 μg/ml) or BSA. Cells (1 × 104) were added to a 96-well plate coated with either OPN or BSA, and the assay was carried out for 3 h at 37°C. For antibody blocking experiments, cells were preincubated with a mouse monoclonal antibody against human αvβ3 for 15 min at 37°C before plating. Columns, mean number of cells from five counts of triplicate wells; bars, SEM. NT/Cβ3 cells showed significantly increased adhesion to OPN versus NT/C control cells (P < 0.002 for all). Adhesion of NT/Cβ3 cells but not NT/C cells could be significantly blocked by an anti-αvβ3 antibody (P < 0.002 for all). B. Cell morphology of NT/Cβ3 clone A4 and NT/C control cells adhered to BSA, OPN, or OPN in the presence of anti-αvβ3. NT/Cβ3 pools A and C showed similar morphology as the NT/Cβ3 clone A4 cell population (data not shown).

FIGURE 3.

Expression of the β3 integrin increases adhesion of NT/C cells to OPN. A. Adhesion of NT/Cβ3 and NT/C control cells to OPN (10 μg/ml) or BSA. Cells (1 × 104) were added to a 96-well plate coated with either OPN or BSA, and the assay was carried out for 3 h at 37°C. For antibody blocking experiments, cells were preincubated with a mouse monoclonal antibody against human αvβ3 for 15 min at 37°C before plating. Columns, mean number of cells from five counts of triplicate wells; bars, SEM. NT/Cβ3 cells showed significantly increased adhesion to OPN versus NT/C control cells (P < 0.002 for all). Adhesion of NT/Cβ3 cells but not NT/C cells could be significantly blocked by an anti-αvβ3 antibody (P < 0.002 for all). B. Cell morphology of NT/Cβ3 clone A4 and NT/C control cells adhered to BSA, OPN, or OPN in the presence of anti-αvβ3. NT/Cβ3 pools A and C showed similar morphology as the NT/Cβ3 clone A4 cell population (data not shown).

Close modal

Expression of β3 Integrin in 21NT Breast Carcinoma Cells Increases Cell Migration in Response to OPN

In transwell migration assays, all three of the NT/Cβ3 cell populations showed significantly increased migration in response to OPN as compared with control vector-transfected cells (P < 0.008 for all; Fig. 4). For each cell line, migration in the presence of OPN was also significantly greater than that with BSA alone (P < 0.04 for all). We also observed that the NT/Cβ3 cells showed some increase in basal migration even in the absence of exogenously added OPN. The NT/Cβ3 cells not only migrated better to the undersurface of the membrane in response to OPN but also appeared to spread better over the undersurface of the membrane.

FIGURE 4.

Expression of the β3 integrin increases migration of NT/C cells to OPN. Cell migration assay was performed using gelatin-coated transwells (6 μg per well). Wells contained 4 × 104 cells in the upper chambers and 50-μg/ml OPN (diluted in αHE + 0.1% BSA media) or αHE + 0.1% BSA media alone in the lower chambers. The assay was carried out for 4.5 h at 37°C. Columns, mean number of cells from five counts of triplicate wells; bars, SEM. NT/Cβ3 cells showed significantly increased migration to OPN versus NT/C control cells (P < 0.008 for all).

FIGURE 4.

Expression of the β3 integrin increases migration of NT/C cells to OPN. Cell migration assay was performed using gelatin-coated transwells (6 μg per well). Wells contained 4 × 104 cells in the upper chambers and 50-μg/ml OPN (diluted in αHE + 0.1% BSA media) or αHE + 0.1% BSA media alone in the lower chambers. The assay was carried out for 4.5 h at 37°C. Columns, mean number of cells from five counts of triplicate wells; bars, SEM. NT/Cβ3 cells showed significantly increased migration to OPN versus NT/C control cells (P < 0.008 for all).

Close modal

Expression of β3 Integrin in 21NT Breast Carcinoma Cells Increases Cell Invasiveness in Response to OPN

In transwell invasion assays through Matrigel (Becton Dickinson, Bedford, MA), all three of the NT/Cβ3 cell populations showed increased invasiveness toward OPN as compared with the NT/C control transfected cell line (P < 0.001 for all; Fig. 5). For each cell line, invasion in the presence of OPN was significantly greater than that with BSA alone (P < 0.004 for all). In contrast to the migration assays, we did not observe an increase in basal invasion (i.e., in the absence of OPN) in the NT/Cβ3 cell populations relative to vector control.

FIGURE 5.

Expression of the β3 integrin increases invasion of NT/C cells to OPN. Cell invasion assay was performed using Matrigel-coated transwells. Wells contained 4 × 104 cells in the upper chambers and 100-μg/ml OPN (diluted in αHE + 0.1% BSA media) or αHE + 0.1% BSA media alone in the lower chambers. The assay was carried out for 72 h at 37°C. Columns, mean number of cells from five counts of triplicate wells; bars, SEM. NT/Cβ3 cells showed significantly increased cell invasion to OPN versus NT/C control cells (P < 0.001 for all).

FIGURE 5.

Expression of the β3 integrin increases invasion of NT/C cells to OPN. Cell invasion assay was performed using Matrigel-coated transwells. Wells contained 4 × 104 cells in the upper chambers and 100-μg/ml OPN (diluted in αHE + 0.1% BSA media) or αHE + 0.1% BSA media alone in the lower chambers. The assay was carried out for 72 h at 37°C. Columns, mean number of cells from five counts of triplicate wells; bars, SEM. NT/Cβ3 cells showed significantly increased cell invasion to OPN versus NT/C control cells (P < 0.001 for all).

Close modal

Expression of β3 Integrin Enhances the Response of 21NT Breast Carcinoma Cells to OPN in Vivo and Results in Increased Tumorigenicity

To assess the effect of β3 integrin expression on the responsiveness of the 21NT cells to OPN in vivo, it was necessary to provide the β3-transfected cells with a source of OPN in the tumor microenvironment. We therefore generated 21NT stable transfectants, which overexpressed both β3 and OPN (NT/Oβ3), and used these cells to determine if β3 could enhance the response of breast epithelial cells to OPN in vivo relative to cells that express β3 alone (NT/Cβ3 pool A), OPN alone (NT/O), or the vector control (NT/C). The mRNA expression levels of β3 and OPN in pooled populations of transfected cells were first confirmed by Northern blot analysis (Fig. 6). Relative to the NT/C vector control cell population, the NT/Cβ3 and NT/Oβ3 cell populations were found to express increased levels of β3 mRNA (12.5- and 12.2-fold, respectively). In contrast, the NT/O cell population expressed very low levels of β3 (2.7-fold increase relative to NT/C control). The NT/O and NT/Oβ3 cell populations were confirmed to overexpress OPN mRNA (7.2- and 6.6-fold, respectively, relative to the NT/C vector control). The NT/Cβ3 cell population expressed levels of OPN comparable with the NT/C control cells.

FIGURE 6.

Northern blot analysis of NT/Cβ3, NT/O, and NT/Oβ3 cells. RNA from NT/C control, NT/Cβ3 (pool A), NT/O, and NT/Oβ3 cells were assessed by Northern blotting for β3 and OPN expression, with 18S assessment as loading control, as described in “Materials and Methods.”

FIGURE 6.

Northern blot analysis of NT/Cβ3, NT/O, and NT/Oβ3 cells. RNA from NT/C control, NT/Cβ3 (pool A), NT/O, and NT/Oβ3 cells were assessed by Northern blotting for β3 and OPN expression, with 18S assessment as loading control, as described in “Materials and Methods.”

Close modal

These cell populations were then injected into the mammary fat pad of female nude mice, and various aspects of primary tumor growth were assessed (Table 1). Primary mammary tumors were first observed between days 62 and 69 in five of eight mice injected with cells expressing both β3 and OPN (NT/Oβ3), and one additional mouse in this group developed a primary tumor by day 105. These mice (six of eight, 75%) were euthanized between days 107 and 167 based on tumor burden. In contrast, a primary tumor was observed in only one of eight (12.5%) mice in the NT/C group, and none of the mice injected with cells expressing β3 alone (NT/Cβ3) or OPN alone (NT/O) formed palpable primary tumors. Primary tumors, which formed from the NT/Oβ3 cells, also demonstrated a decreased doubling time and a decreased latency period compared with vector control (Table 1). Of the mice in the NT/Oβ3 group that formed primary tumors, 50% demonstrated lymphatic invasion at the site of the primary tumor and 33% demonstrated lymph node metastasis. No lymphatic involvement was observed in the control tumor (Table 1).

Table 1.

Summary of in VivoData

Note: Pooled populations of transfected cells were injected into the mammary fat pad of female nude mice using 1 × 107 cells per mouse and eight mice per treatment group. The end point was set at 1 year postinjection.


Group

Tumor Take (No. of mice)

Tumor Latency Period (days to 500 mm3)

Doubling Time (days)

Lymphatic Invasion (No. of Mice with Lymphatic Invasion/No. of Mice With Primary Tumors)

Lymph Node Metastasis (No. of Mice With Lymphatic Metastasis/No. of Mice With Primary Tumors)
NT/C (vector control) 1/8 233 6.4 0/1 0/1 
NT/Cβ33 alone) 0/8 N/A N/A 0/0 0/0 
NT/O (OPN alone) 0/8 N/A N/A 0/0 0/0 
NT/Oβ33 + OPN)
 
6/8a
a

Tumor take was significantly higher in the NT/Oβ3 group relative to the other groups (P < 0.001 using Fisher's exact test).


 
109.3 ± 11.7b
b

Data are presented as the means ± SE.


 
4.9 ± 0.42b
 
3/6c
c

Lymphatic invasion at the site of the primary tumor and lymph node metastasis was determined using standard H&amp;E staining as assessed by a pathologist (A.B.T.).


 
2/6c
 

Group

Tumor Take (No. of mice)

Tumor Latency Period (days to 500 mm3)

Doubling Time (days)

Lymphatic Invasion (No. of Mice with Lymphatic Invasion/No. of Mice With Primary Tumors)

Lymph Node Metastasis (No. of Mice With Lymphatic Metastasis/No. of Mice With Primary Tumors)
NT/C (vector control) 1/8 233 6.4 0/1 0/1 
NT/Cβ33 alone) 0/8 N/A N/A 0/0 0/0 
NT/O (OPN alone) 0/8 N/A N/A 0/0 0/0 
NT/Oβ33 + OPN)
 
6/8a
a

Tumor take was significantly higher in the NT/Oβ3 group relative to the other groups (P < 0.001 using Fisher's exact test).


 
109.3 ± 11.7b
b

Data are presented as the means ± SE.


 
4.9 ± 0.42b
 
3/6c
c

Lymphatic invasion at the site of the primary tumor and lymph node metastasis was determined using standard H&amp;E staining as assessed by a pathologist (A.B.T.).


 
2/6c
 

Elevated levels of OPN in both tumor tissue and blood have been identified in patients with breast and other cancers and have been shown to be associated with poor prognosis (3–12, reviewed in 2, 31). Recent microarray gene expression studies have also implicated OPN in the progression of various tumor types (10, 12, 32, 33), and several other studies have demonstrated that OPN can contribute functionally to cancer progression (reviewed in 2, 31). Potential mechanisms of OPN-mediated malignancy include induction of expression and activity of proteases, such as urokinase-type plasminogen activator, as well as growth factor receptors, such as hepatocyte growth factor receptor (Met) and epidermal growth factor receptor, which are important in tumor motility, invasion, and metastasis (15, 18–20). Other possible mechanisms include protection from apoptosis or host immune-mediated destruction (34–37). The effects of OPN on cancer cells are thought to involve integrin signaling and cross-talk with growth factor receptors, such as Met, as well as interactions with other nonintegrin surface receptors such as CD44 (reviewed in 2, 21, 31, 38, 39).

OPN can bind to a number of cell surface integrins, including αvβ3, αvβ1, and αvβ5 (26, 27). High levels of the αvβ3 integrin combined with low levels of αvβ1 and αvβ5 have, in some cases, been associated with malignancy and progression in breast cancer, melanoma, and prostate cancer (28, 40–45). However, it has not previously been shown whether experimentally altered levels of cell surface integrin expression can lead to changes in cellular behavior in response to OPN. The current study provides, for the first time, direct experimental evidence that changing the pattern of integrin expression (in this case, expression of the αvβ3 integrin) can contribute functionally to increased malignancy of breast cancer cells in response to OPN both in vitro and in vivo.

We previously showed that tumorigenic, nonmetastatic 21NT cells and metastatic MDA-MB-435 human breast cancer cells both responded to exogenous OPN or transfected, endogenous OPN with increased aggressive behavior in vitro, including cell migration and invasion, indicating that OPN can contribute functionally to the malignant behavior of breast cancer cells (15, 18–20). These effects could be functionally blocked in MDA-MB-435 cells with antibodies to the αvβ3 integrin but not with antibodies to αvβ1 or αvβ5. Conversely, in 21NT cells, anti-αvβ3 antibodies had no effect, but invasion and migration were blocked by antibodies to αvβ1 or αvβ5 (18). These findings suggested that different patterns of integrin expression might determine how cells responded to a ligand such as OPN in the tumor microenvironment.

In the present study, we first characterized the integrin expression profile of 21NT cells and found that they were positive for surface expression of both αvβ5 and αvβ1 but were negative for expression of the αvβ3 integrin subunit both on the cell surface and at the RNA level. This finding provides an explanation for the failure of an anti-β3 blocking antibody to inhibit their migration or invasion in response to OPN, as previously reported (18). In contrast, MDA-MB-435 cells showed cell surface expression of all three integrins, although only anti-αvβ3 and not anti-αvβ5 or anti-αvβ1 antibodies were effective at blocking their migration and invasion in response to OPN (18), indicating that these more aggressive cells preferentially use αvβ3 to respond to OPN. We then experimentally altered the integrin expression pattern of the 21NT cells by transfecting them to overexpress the β3 integrin subunit. Relative to the vector control transfectants, the β3-expressing 21NT cells were found to be more aggressive in terms of in vitro adhesion, migration, and invasion in response to OPN. Furthermore, in vivo studies demonstrated that mammary fat pad injection of 21NT cells, which expressed β3 in combination with OPN (NT/Oβ3), resulted in an increased tumor take, a decreased tumor doubling time, and a decreased tumor latency period relative to controls. Parental 21NT cells typically take up 1 year to develop primary tumors and only in 10–20% of mice (29 and A.B. Tuck, unpublished data). Thus, the observation that the NT/Oβ3 cells can form tumors in 75% of mice as early as 62 days postinjection is a novel and significant finding. In addition, the primary tumors that do form from parental 21NT cells have never been observed to demonstrate either lymphatic invasion at the site of the primary tumor or lymph node metastasis.3

3

A.B. Tuck, unpublished data.

The observation that tumors, which develop from NT/Oβ3 cells, show lymph node invasion in three of eight mice and lymph node metastasis in two of eight mice therefore indicates that, in addition to affecting tumorigenicity, cooperative expression of both β3 and OPN in 21NT cells can also lead to increased tumor aggressiveness in terms of lymphatic involvement.

OPN present in tumor tissue can be produced both by tumor cells and by host infiltrating inflammatory cells (1, 5, 46), although it remains to be elucidated whether the specific functional consequences of OPN on tumor cell malignancy differ depending on the cellular source of OPN. However, it is clear that OPN produced by various cell types in the tumor microenvironment can influence cell behavior in either an autocrine or a paracrine manner. The model system described here indicates that the αvβ3 integrin may be important for tumor cell response to both exogenously produced OPN (as demonstrated by the in vitro studies) and endogenously produced OPN (as demonstrated by the in vivo studies). The present study provides evidence that increased β3 integrin expression may contribute to breast cancer progression by making the cancer cells more responsive to the malignancy-enhancing effects of microenvironmental ligands such as OPN.

Cell Lines and Growth in Tissue Culture

The tumorigenic, nonmetastatic 21NT human mammary carcinoma cell line was used (a gift from Dr. Vimla Band, Dana-Farber Cancer Institute, Boston, MA; 29). 21NT cells were previously stably transfected with an unmodified pcDNA3 mammalian expression vector (NT/C cells) or with a full-length human OPN cDNA contained within the pcDNA3 vector (NT/O cells; 15). Transfected 21NT cells were grown in αHE medium with 10% fetal bovine serum, as previously described (15), with 200 μg/ml active G418 (Life Technologies, Inc., Burlington, Ontario) to maintain expression of the transfected plasmids. The highly metastatic MDA-MB-435 human breast carcinoma cell line (a gift from Dr. Janet Price, Anderson Cancer Center, Houston, TX; 30) was also used and was grown in αMEM medium supplemented with 10% fetal bovine serum.

Flow Cytometry

Primary antibodies used for flow cytometry were mouse antihuman antibodies directed against the αvβ5 integrin (clone P1F6, Life Technologies), the αvβ1 integrin (clone P4C10, Life Technologies), and the αvβ3 integrin (clone PM6/13, Chemicon International, Inc., Temecula, CA). A negative immunoglobulin G isotype primary mouse antibody (Cedarlane Laboratories Ltd., Hornby, Ontario) was used as a control. The secondary antibody used was goat antimouse immunoglobulin G conjugated to R-phycoerythrin (Chemicon International).

Cell lines were grown in T75 tissue culture flasks to ∼85% confluency (log phase of growth). The cells were gently trypsinized and washed twice with ice cold, sterile PBS (0.1-m potassium phosphate, 0.1-m sodium phosphate, 0.15-m sodium chloride, pH 7.4) plus 2% fetal bovine serum (flow buffer). Cells (1 × 106) were resuspended in 100 μl of flow buffer, the primary antibody was applied (1 μg per 106 cells), and the suspension was incubated at 4°C on a rotating plate for 1 h. The cells were centrifuged and washed twice with 1 ml of flow buffer and resuspended in 100 μl of buffer. The secondary antibody was then added (0.5 μg per 106 cells) and incubated at 4°C on a rotating plate for 1 h in the dark. The cells were recentrifuged, washed twice, and resuspended in 1 ml of flow buffer before being scanned on an EPICS XL-MCL flow cytometer (Beckman-Coulter Canada Inc., Mississauga, ON, Canada).

Cloning of the pcHygro-β3 Plasmid

Full-length human β3 integrin cDNA, contained within the pcDNA3 expression vector (Invitrogen, San Diego, CA; pcDNA3-β3), was obtained as a gift from Dr. David Wilcox (Medical College of Wisconsin, Milwaukee, WI). Because the recipient NT/C cells already contained the pcDNA3 vector, the β3 integrin cDNA was cloned into the pcDNA3.1/Hygro+ mammalian expression vector (Invitrogen) using the HindIII and XhoI sites to generate the pcHygro-β3 plasmid. Correct orientation and integrity of the β3 gene was confirmed with restriction digests and verified by sequencing (Genbank accession no. NM 000212).

Transfections

Two sets of transfections (with NT/C cells and NT/O cells) were performed using the pcHygro-β3 and the control vector pcDNA3.1/Hygro+ plasmids in both cases. The Lipofectin reagent (Life Technologies) was used for the transfections, according to the manufacturer's guidelines. Triplicate 60-mm tissue culture dishes (Becton Dickinson) were seeded with 2.5 × 105 cells and incubated for 24 h at 37°C in a CO2 incubator. After 24 h, cells had reached 50% confluency and were transfected as per the manufacturer's guidelines. For each plate, 1 μg of DNA was mixed with 10 μl of Lipofectin reagent and the transfection was carried out overnight at 37°C in a CO2 incubator. In the morning, the transfection medium was removed and replaced with normal growth medium and the cells were allowed to recover and grow for an additional 48 h at 37°C. The cells were then trypsinized, replated, and grown again for 48 h. Normal growth medium containing hygromycin B (500 μg/ml, a concentration previously determined to give 100% death of parental NT/C and NT/O cells) was added to the cells and colonies were allowed to develop. Clonal cell populations were generated from a single colony, whereas pooled cell populations represent a collection of colonies. The specific cell populations isolated were as follows: three β3-transfected NT/C populations were named NT/Cβ3 clone A4, NT/Cβ3 pool A (consisting of 20 colonies), and NT/Cβ3 pool C (consisting of 15 colonies). One control vector-transfected NT/O pooled population was named NT/O (consisting of 5 colonies). One β3-transfected NT/O pooled population was named NT/Oβ3 (consisting of 15 colonies) and one control vector-transfected pooled population was named NT/C (consisting of 8 colonies).

Northern Analysis

RNA was isolated using TRIzol reagent (Life Technologies) from subconfluent cells according to the manufacturer's recommendations. RNA (10 μg per sample) was run on a 1.1% agarose gel with 6.8% formaldehyde and was capillary transferred to a nylon GeneScreen Plus membrane (NEN Life Science Products, Boston, MA). Blots were probed with denatured, oligolabeled 32P-dCTP cDNA probes using an oligolabeling kit (Amersham Pharmacia Biotech, Inc., Baie d'Urfé, Quebec) according to the manufacturer's recommendations. The β3 integrin probe used was a 551 bp, KpnI-digested fragment of the pcHygro-β3 plasmid containing the first 500 bp of the human β3 cDNA. The OPN probe used was a full-length human OPN cDNA (1493 bp) generated using EcoRI digestion of the OP-10 plasmid (15). To confirm even loading of lanes, human 18S rRNA from the p100D9 plasmid (a gift from Dr. David T. Denhardt) was also used. Densitometric analysis was carried out using ImageQuant software (Molecular Dynamics, Sunnyvale, CA).

Cell Adhesion Assay

On the day prior to the experiment, sterile 96-well nontissue culture plates (Titertek, Flow Laboratories, McLean, VA) were treated in triplicate with either 10 μg/ml of human recombinant OPN (17) or 0.2% BSA (Sigma Chemical Co., St. Louis, MO) and incubated overnight at 4°C. The wells were then rinsed and blocked with αHE + 0.2% BSA medium for 2 h at 37°C. The cells were prepared using gentle trypsinization with 1× trypsin/EDTA (Life Technologies) followed by inactivation using soybean trypsin inhibitor (Life Technologies). The cells were resuspended in αHE + 0.2% BSA, and 1 × 104 cells were added per well. For antibody blocking experiments, cells were preincubated on a rotating plate with saturating concentrations (as determined by preliminary titration experiments) of mouse monoclonal antibodies against human αvβ3 (Chemicon International) or a control negative immunoglobulin G isotype primary mouse antibody (Cedarlane Laboratories) for 15 min at 37°C before plating. Plates were incubated for 3 h at 37°C, and the medium was then removed and nonadhered cells were rinsed away with three gentle washes of PBS. Adhered cells were fixed using fresh 2% gluteraldehyde for 20 min then washed with PBS and stained using Harris' hematoxylin for 10 min. The staining was intensified using a 1% solution of ammonia water and the blue-stained cells could then be counted using a light microscope. Five high-power fields were counted for each well, and the mean number of cells per field was calculated. This assay was performed twice and gave similar results.

Cell Migration Assay

Cell migration was assessed as previously described (15). Briefly, on the day prior to the experiment, triplicate 6.5-mm transwells with 8-μm pores (Costar Corp., Cambridge, MA) were coated with 6 μg of gelatin (denatured collagen; Sigma Chemical) per well and dried overnight in a tissue culture hood. On the following morning, the gelatin was reconstituted with αHE medium for 90 min. Subconfluent cells were gently trypsinized and washed twice with αHE + 0.1% BSA. The cells were resuspended in αHE + 0.1% BSA and 4 × 104 cells were added to each upper well. Each lower chamber contained either αHE + 0.1% BSA (negative control) or 50-μg/ml recombinant human OPN diluted in αHE + 0.1% BSA. The plates were then incubated for 4.5 h at 37°C. The upper well was removed and fixed in fresh 2% gluteraldehyde for 20 min then washed with PBS and stained using Harris' hematoxylin for 15 min. The staining was then intensified using a 1% solution of ammonia water. Nonmigrated cells remaining on the upper side of the membrane were removed with a cotton swab. The blue-stained cells on the lower surface were then counted using a light microscope. Five high-power fields of cells were counted for each well, and the mean number of cells per field was calculated. This assay was performed twice and gave similar results.

Cell Invasion Assay

Cell invasion through Matrigel was assessed as previously described (15). Briefly, on the day prior to the experiment, triplicate 6.5-mm transwells with 8-μm pores were coated with 20 μg of Matrigel per well and dried overnight in a tissue culture hood. On the following morning, the Matrigel was reconstituted by adding αHE medium and the plate was agitated for 90 min. Subconfluent cells were prepared by gentle trypsinization followed by two washes with αHE + 0.1% BSA medium. The cells were resuspended in αHE + 0.1% BSA and 4 × 104 cells were added to each upper well. Each lower chamber contained either αHE + 0.1% BSA (negative control) or 100-μg/ml human recombinant OPN diluted in αHE + 0.1% BSA medium. The plates were then incubated for 72 h at 37°C to allow for invasion of the cells through the Matrigel to the underside of the transwell membrane. The wells were then fixed, stained, and counted as described above. This assay was performed thrice and gave similar results.

Mammary Fat Pad Injections

Female athymic NCr nude mice (nu/nu) were housed and cared for in accordance with the recommendations of the Canadian Council on Animal Care under a protocol approved by the University of Western Ontario Council on Animal Care. Cell lines were grown in 150-mm tissue culture dishes to ∼80% confluency (log phase of growth). The cells were gently trypsinized, washed twice with sterile PBS, and resuspended in serum-free αHE media at a concentration of 1 × 107 cells per 100 μl. Pooled populations of 21NT cells stably transfected with the control vector (NT/C), β3 alone (NT/Cβ3 pool A), OPN alone (NT/O), or both OPN and β3 (NT/Oβ3) were injected into the second thoracic mammary fat pad of 8–9-week-old female nude mice, as described elsewhere (30), using 1 × 107 cells per mouse and eight mice per treatment group. Animals were routinely monitored for health and primary tumors were measured every 7–14 days. Animals were euthanized early if the tumor burden became too great or at the end point of the experiment (1 year postinjection). Animals were necropsied, and tissues and organs (including primary mammary tumor and/or mammary fat pad, lymph nodes, lungs, liver, and spleen) were fixed in 10% neutral-buffered formalin before processing. Tissues were embedded in paraffin wax, sectioned (4 μm thick), and subjected to standard H&E staining to observe histopathological characteristics.

Statistical Analysis

For the in vitro studies, statistical differences in cell adhesion, migration, and invasion between groups were determined by Student's t test. For the in vivo studies, statistical differences in tumor take between groups were determined using Fisher's exact test. In all cases, values of P < 0.05 were considered to be significant.

We thank Michael Keeney of the London Health Sciences Centre Flow Cytometry Unit for his help in carrying out and analyzing the flow cytometry studies and Larry Stitt of the London Regional Cancer Centre Clinical Research Unit for his help with statistical analysis of the data.

1
Brown, L. F., Papadopoulos-Sergiou, A., Berse, B., Manseau, E. J., Tognazzi, K., Perruzzi, C. A., Dvorak, H. F., and Senger, D. R. Osteopontin expression and distribution in human carcinomas.
Am. J. Pathol.
,
145
:
610
–623, 
1994
.
2
Furger, K. A., Menon, R. K., Tuck, A. B., Bramwell, V. H., and Chambers, A. F. The functional and clinical roles of osteopontin in cancer and metastasis.
Curr. Mol. Med.
,
1
:
621
–632, 
2001
.
3
Singhal, H., Bautista, D. S., Tonkin, K. S., O'Malley, F. P., Tuck, A. B., Chambers, A. F., and Harris, J. F. Elevated plasma osteopontin in metastatic breast cancer associated with increased tumor burden and decreased survival.
Clin. Cancer Res.
,
3
:
605
–611, 
1997
.
4
Tuck, A. B., O'Malley, F. P., Singhal, H., Tonkin, K. S., Harris, J. F., Bautista, D., and Chambers, A. F. Osteopontin and p53 expression are associated with tumor progression in a case of synchronous, bilateral, invasive mammary carcinomas.
Arch. Pathol. Lab. Med.
,
121
:
578
–584, 
1997
.
5
Tuck, A. B., O'Malley, F. P., Singhal, H., Harris, J. F., Tonkin, K. S., Kerkvliet, N., Saad, Z., Doig, G. S., and Chambers, A. F. Osteopontin expression in a group of lymph node negative breast cancer patients.
Int. J. Cancer
,
79
:
502
–508, 
1998
.
6
Fedarko, N. S., Jain, A., Karadag, A., Van Eman, M. R., and Fisher, L. W. Elevated serum bone sialoprotein and osteopontin in colon, breast, prostate, and lung cancer.
Clin. Cancer Res.
,
7
:
4060
–4066, 
2001
.
7
Rudland, P. S., Platt-Higgins, A., El-Tanani, M., De Silva Rudland, S., Barraclough, R., Winstanley, J. H., Howitt, R., and West, C. R. Prognostic significance of the metastasis-associated protein osteopontin in human breast cancer.
Cancer Res.
,
62
:
3417
–3427, 
2002
.
8
Chambers, A. F., Wilson, S. M., Kerkvliet, N., O'Malley, F. P., Harris, J. F., and Casson, A. G. Osteopontin expression in lung cancer.
Lung Cancer
,
15
:
311
–323, 
1996
.
9
Kim, J. H., Skates, S. J., Uede, T., Wong, K. K., Schorge, J. O., Feltmate, C. M., Berkowtiz, R. S., Cramer, D. W., and Mok, S. C. Osteopontin as a potential diagnostic biomarker for ovarian cancer.
JAMA
,
278
:
1671
–1679, 
2002
.
10
Agrawal, D., Chen, T., Irby, R., Quackenbush, J., Chambers, A. F., Szabo, M., Cantor, A., and Coppola, D. Osteopontin identified as lead marker of colon cancer progression, using pooled sample expression profiling.
J. Natl. Cancer Inst.
,
94
:
513
–521, 
2002
.
11
Hotte, S. J., Winquist, E. W., Stitt, L., Wilson, S. M., and Chambers, A. F. Plasma osteopontin: associations with survival and metastasis to bone in men with hormone-refractory prostate carcinoma.
Cancer
,
95
:
506
–512, 
2002
.
12
Ye, Q. H., Qin, L. X., Forgues, M., He, P., Kim, J. W., Peng, A. C., Simon, R., Li, Y., Robles, A. I., Chen, Y., Ma, Z. C., Wu, Z. Q., Ye, S. L., Liu, Y. K., Tang, Z. Y., and Wang, X. W. Predicting hepatitis B virus-positive metastatic hepatocellular using gene expression profiling and supervised machine learning.
Nat. Med.
,
9
:
416
–423, 
2003
.
13
Oates, A. J., Barraclough, R., and Rudland, P. S. The identification of osteopontin as a metastasis-related gene product in a rodent mammary tumor model.
Oncogene
,
13
:
97
–104, 
1996
.
14
Sung, V., Gilles, C., Murray, A., Clarke, R., Aaron, A. D., Azumi, N., and Thompson, E. W. The LCC15-MB human breast cancer cell line expresses osteopontin and exhibits an invasive and metastatic phenotype.
Exp. Cell Res.
,
241
:
273
–284, 
1998
.
15
Tuck, A. B., Arsenault, D. M., O'Malley, F. P., Hota, C., Ling, M. C., Wilson, S. M., and Chambers, A. F. Osteopontin induces increased invasiveness and plasminogen activator expression of human mammary epithelial cells.
Oncogene
,
18
:
4237
–4246, 
1999
.
16
Urquidi, V., Sloan, D., Kawai, K., Agarwal, D., Woodman, A. C., Tarin, D., and Goodison, S. Contrasting expression of thrombospondin-1 and osteopontin correlates with absence or presence of metastatic phenotype in an isogenic model of spontaneous human breast cancer metastasis.
Clin. Cancer Res.
,
8
:
61
–74, 
2002
.
17
Xuan, J. W., Hota, C., and Chambers, A. F. Recombinant GST-human osteopontin fusion protein is functional in RGD-dependent cell adhesion.
J. Cell. Biochem.
,
54
:
247
–255, 
1994
.
18
Tuck, A. B., Elliott, B. E., Hota, C., Tremblay, E., and Chambers, A. F. Osteopontin-induced, integrin-dependent migration of human mammary epithelial cells involves activation of the hepatocyte growth factor receptor (Met).
J. Cell. Biochem.
,
78
:
465
–475, 
2000
.
19
Tuck, A. B., Hota, C., Wilson, S. M., and Chambers, A. F. Osteopontin-induced migration of human mammary epithelial cells involves activation of EGF receptor and multiple signal transduction pathways.
Oncogene
,
22
:
1198
–1205, 
2003
.
20
Tuck, A. B., Hota, C., and Chambers, A. F. Osteopontin (OPN)-induced increase in human mammary epithelial cell invasiveness is urokinase (uPA)-dependent.
Breast Cancer Res. Treat.
,
70
:
197
–204, 
2001
.
21
Sodek, J., Ganss, B., and McKee, M. D. Osteopontin.
Crit. Rev. Oral Biol. Med.
,
11
:
279
–303, 
2000
.
22
Weber, G. F., Ashkar, S., Glimcher, M. J., and Cantor, H. Receptor-ligand interaction between CD44 and osteopontin (Eta-1).
Science
,
271
:
509
–512, 
1996
.
23
Katagiri, Y. U., Sleeman, J., Fujii, H., Herrlich, P., Hotta, H., Tanaka, K., Chikuma, S., Yagita, H., Okumura, K., Murakami, M., Saiki, I., Chambers, A. F., and Uede, T. CD44 variants but not CD44s cooperate with β1-containing integrins to permit cells to bind to osteopontin independently of arginine-glycine-aspartic acid, thereby stimulating cell motility and chemotaxis.
Cancer Res.
,
59
:
219
–226, 
1999
.
24
van Dijk, S., D'Errico, J. A., Somerman, M. J., Farach-Carson, M. C., and Butler, W. T. Evidence that a non-RGD domain in rat osteopontin is involved in cell attachment.
J. Bone Miner. Res.
,
8
:
1499
–1506, 
1993
.
25
Katagiri, Y. U., Murakami, M., Mori, K., Iizuka, J., Hara, T., Tanaka, K., Jia, W. Y., Chambers, A. F., and Uede, T. Non-RGD domains of osteopontin promote cell adhesion without involving αv integrins.
J. Cell. Biochem.
,
62
:
123
–131, 
1996
.
26
Hu, D. D., Lin, E. C., Kovach, N. L., Hoyer, J. R., and Smith, J. W. A biochemical characterization of the binding of osteopontin to integrins αvβ1 and αvβ5.
J. Biol. Chem.
,
270
:
26232
–26238, 
1995
.
27
Hruska, K. A., Rolnick, F., Huskey, M., Alvarez, U., and Cheresh, D. Engagement of the osteoclast integrin αvβ3 by osteopontin stimulates phosphatidylinositol 3-hydroxyl kinase activity.
Endocrinology
,
136
:
2984
–2992, 
1995
.
28
Wong, N. C., Mueller, B. M., Barbas, C. F., Ruminski, P., Quaranta, V., Lin, E. C., and Smith, J. W. αv integrins mediate adhesion and migration of breast carcinoma cell lines.
Clin. & Exp. Metastasis
,
16
:
50
–61, 
1998
.
29
Band, V., Zajchowski, D., Swisshelm, K., Trask, D., Kulesa, V., Cohen, C., Connolly, J., and Sager, R. Tumor progression in four mammary epithelial cell lines derived from the same patient.
Cancer Res.
,
50
:
7351
–7357, 
1990
.
30
Price, J. E., Polyzos, A., Zhang, R. D., and Daniels, L. M. Tumorigenicity and metastasis of human breast carcinoma cell lines in nude mice.
Cancer Res.
,
50
:
717
–721, 
1990
.
31
Tuck, A. B. and Chambers, A. F. The role of osteopontin in breast cancer: clinical and experimental studies.
J. Mammary Gland Biol. Neoplasia
,
6
:
419
–429, 
2001
.
32
Markert, J. M., Fuller, C. M., Gillespie, G. Y., Bubien, J. K., McLean, L. A., Hong, R. L., Lee, K., Gullans, S. R., Mapstone, T. B., and Benos, D. J. Differential gene expression profiling in human brain tumors.
Physiol. Genomics
,
5
:
21
–33, 
2001
.
33
Euer, N., Schwirzke, M., Evtimova, V., Burtscher, H., Jarsch, M., Tarin, D., and Weidle, U. H. Identification of genes associated with metastasis of mammary carcinoma in metastatic versus non-metastatic cell lines.
Anticancer Res.
,
22
:
733
–740, 
2002
.
34
Denhardt, D. T. and Chambers, A. F. Overcoming obstacles to metastasis—defenses against host defenses: osteopontin (OPN) as a shield against attack by cytotoxic host cells.
J. Cell. Biochem.
,
56
:
48
–51, 
1994
.
35
Feng, B., Rollo, E. E., and Denhardt, D. T. OPN may facilitate metastasis by protecting cells from macrophage NO-mediated cytotoxicity: evidence from cell lines down-regulated for OPN expression by targeted ribozyme.
Clin. & Exp. Metastasis
,
13
:
453
–462, 
1995
.
36
Scatena, M., Almeida, M., Chaisson, M. L., Fausto, N., Nicosia, R. F., and Giachelli, C. M. NF-κB mediates αvβ3 integrin-induced endothelial cell survival.
J. Cell Biol.
,
141
:
1083
–1093, 
1998
.
37
Geissinger, E., Weisser, C., Fischer, P., Schartl, M., and Wellbrock, C. Autocrine stimulation by osteopontin contributes to antiapoptotic signaling of melanocytes in dermal collagen.
Cancer Res.
,
62
:
4820
–4828, 
2002
.
38
Eliceiri, B. P. Integrin and growth factor receptor crosstalk.
Circ. Res.
,
89
:
1104
–1010, 
2001
.
39
Weber, G. F. The metastasis gene osteopontin: a candidate target for cancer therapy.
Biochim. Biophys. Acta
,
1552
:
61
–85, 
2001
.
40
Nesbit, M. and Herlyn, M. Adhesion receptors in human melanoma progression.
Invasion Metastasis
,
14
:
131
–146, 
1994
.
41
Gui, G. P., Wells, C. A., Browne, P. D., Yeomans, P., Jordan, S., Puddefoot, J. R., Vinson, G. P., and Carpenter, R. Integrin expression in primary breast cancer and its relation to axillary nodal status.
Surgery
,
117
:
102
–108, 
1995
.
42
Liapis, H., Flath, A., and Kitazawa, S. Integrin αvβ3 expression by bone-residing breast cancer metastases.
Diagn. Mol. Pathol.
,
5
:
127
–135, 
1996
.
43
Van Belle, P. A., Elenitsas, R., Satyamoorthy, K., Wolfe, J. T., Guerry, D. IV, Schuchter, L., Van Belle, T. J., Albelda, S., Tahin, P., Herlyn, M., and Elder, D. E. Progression-related expression of β3 integrin in melanomas and nevi.
Hum. Pathol.
,
30
:
562
–567, 
1999
.
44
Felding-Habermann, B., O'Toole, T. E., Smith, J. W., Fransvea, E., Ruggeri, Z. M., Ginsberg, M. H., Hughes, P. E., Pampori, N., Shattil, S. J., Saven, A., and Mueller, B. M. Integrin activation controls metastasis in human breast cancer.
Proc. Natl. Acad. Sci. USA
,
98
:
1853
–1858, 
2001
.
45
Cooper, C. R., Chay, C. H., and Pienta, K. J. The role of αvβ3 in prostate cancer progression.
Neoplasia
,
4
:
191
–194, 
2002
.
46
Hirota, S., Ito, A., Nagoshi, J., Takeda, M., Kurata, A., Takatsuka, Y., Kohri, K., Nomura, S., and Kitamura, Y. Expression of bone matrix protein messenger ribonucleic acids in human breast cancers: possible involvement of osteopontin in development of calcifying foci.
Lab. Invest.
,
72
:
64
–69, 
1995
.
1

Grant 12078 from the Canadian Breast Cancer Research Initiative (A.F.C. and A.B.T.) A.L.A. is supported by the H.L. Holmes Postdoctoral Award from the National Research Council of Canada. S.A.V. is supported by a predoctoral award from the U.S. Medical Research and Material Command Breast Cancer Research Program. This work was performed in partial fulfillment of the requirement for the M.Sc. degree by K.A.F. Note: K.A.F. and A.L.A. contributed equally to the preparation of this manuscript.