Doxorubicin treatment was found to augment the expression of the extracellular matrix (ECM) protein fibulin-1 in cultured human breast cancer cell lines and in MDA-MB-361 tumors grown in athymic mice. Doxorubicin was also found to augment tumor expression of the fibulin-1–binding proteins fibronectin and laminin-1. Growth of breast cancer cell lines on Matrigel, an ECM extract containing fibulin-1 and laminin-1, resulted in lower levels of doxorubicin-induced apoptosis as compared with controls. Moreover, tumors formed by injection of athymic mice with MDA-MB-361 cells mixed with Matrigel were significantly more doxorubicin resistant and displayed lower levels of apoptosis compared with those that formed in the absence of Matrigel. Monoclonal antibodies against fibulin-1 reversed Matrigel-dependent doxorubicin resistance. Furthermore, small interfering RNA–mediated suppression of fibulin-1 expression in breast cancer cells resulted in a 10-fold increase in doxorubicin sensitivity as compared with control cells. Together, these findings point to a role for fibulin-1 in breast cancer chemoresistance. [Cancer Res 2007;67(9):4271–7]

Adjuvant chemotherapy and/or hormone therapy has improved the probability of survival of breast cancer patients by 10% to 15%. Nonetheless, the incomplete understanding of the biological mechanisms underlying chemoresistance has hampered further improvement in the treatment and outcome of breast carcinoma. To date, the biological classification of tumors as “responsive” or “nonresponsive” to treatment remains controversial, and even extensively studied conventional pathobiological markers of breast cancer such as hormone receptors, grade, p53, and HER-2 are not useful in predicting tumor responsiveness to chemotherapy (1). This likely reflects the multifactorial origin of chemoresistance in most solid tumors.

Mechanisms involved in drug resistance include the presence of a range of drug transporters (such as P-glycoprotein or multidrug resistance–associated protein); changes in the expression of topoisomerases; alterations in metabolic pathways that influence drug metabolism, DNA repair, or apoptosis; delivery and distribution of anticancer drugs to tumor cells; imbalance of the ratio of bax to bcl-2; high levels of constitutive nuclear factor-κB activity, cyclin D1 overexpression, and the microenvironment (26).

Recent evidence increasingly points to the important role of stromal extracellular matrix (ECM) components in the success of chemotherapy (7). Indeed, the disruption of integrin-mediated cell-ECM interactions can influence cancer cell sensitivity to apoptosis and affect drug resistance (1). Several studies have shown that small-cell lung cancer cells (8), myeloma cells (9), glioma cells (10), chronic myelogenous leukemia cells (11), lymphoma (12), uveal melanoma (13), and ovarian (14), bladder (15), and colon (16) carcinomas were protected from the apoptosis induced by various anticancer drugs when the cells were plated on ECM proteins such as fibronectin, laminin-1, and collagens, which directly bind integrin receptors. However, it has remained unclear whether other ECM proteins including those that do not directly engage integrins also contribute to chemoresistance. Among the latter proteins are the fibulins, a family of six proteins thought to function as bridges in the organization of ECM supramolecular structures (17). Fibulin-1 (FBLN-1), the prototype member of this family, has four splice variants (17) and binds to many ECM proteins, including fibronectin, laminin-1, fibrinogen, nidogen, and the proteoglycans aggrecan and versican (18). Several studies have reported that FBLN-1 is expressed in various human neoplasias and it is implicated in processes such as invasion, motility, and in vivo tumor growth (1922). We recently showed that FBLN-1 is a breast cancer–restricted antigen aberrantly expressed in 35% of 528 breast cancer specimens analyzed (23); it induces both specific B-cell– and T-cell–mediated responses in breast cancer patients (23, 24); and it can be exploited as a tool for early detection of breast cancers (25). Here, we report that FBLN-1 acts to promote breast cancer cell survival during doxorubicin treatment.

Cell lines. Human breast carcinoma cell lines SKBR-3, MCF-7, MDA-MB-157, MDA-MB-231, MDA-MB-361, MDA-MB-468, MDA-MB-453, and MDA-MB-175 were obtained from the American Type Culture Collection (ATCC). The MDA-MB-435 cell line, which shows both breast-specific and melanocyte-specific markers (26), was also obtained from ATCC. Cell lines were routinely maintained in RPMI 1640 (Sigma Chemical Co.) supplemented with 10% (v/v) FCS (Sigma Chemical) and l-glutamine in a humidified chamber (95% air, 5% CO2) at 37°C. Human breast tumor MCF-7-DXR (doxorubicin-resistant variant) and MCF-7B (parental) cell lines were kindly provided by Elena Monti [Department of Structural and Functional Biology, Section of Pharmacology, University of Insubria, Busto Arsizio (VA), Italy; ref. 27].

Antibodies. The mouse monoclonal antibodies (mAb) to human FBLN-1, P5B1, P3B4, P1B2, MEM-2, and 3A11 have previously been described (28, 29). Rabbit polyclonal antibodies to human laminin-1 and human fibronectin and mouse mAbs against α-actin and vinculin were purchased from Sigma Chemical.

Indirect immunofluorescence and fluorescence-activated cell sorting analysis. The cytoplasm-associated form of FBLN-1 was detected with the mAb MEM-2 in doxorubicin-treated and untreated permeabilized tumor cells after fixation and permeabilization with BD Cytofix/Cytoperm (BD Biosciences PharMingen) as described (23).

Immunoprecipitation and Western blotting. Tumor specimens from doxorubicin-treated and untreated athymic mice were minced, lysed for 1 h on ice in lysis buffer [50 mmol/L Tris-HCl (pH 7.2), 150 mmol/L NaCl, 1 mmol/L EDTA, 1% Triton X-100, 0.1% SDS and 0.5% deoxycholic acid] containing protease inhibitors, 2 mmol/L phenylmethylsulfonylfluoride, and 10 μg/mL aprotinin, and specifically processed as described (23). Specific proteins were detected by enhanced chemiluminescence according to the manufacturer's recommendations (GE Healthcare Amersham). Autoradiographic signals were measured using a Bio-Rad scanning densitometer (Bio-Rad, ChemiDoc/XRS). Data were acquired and analyzed using Quantity One v 4.6.1 software. The expression of proteins and transcripts was assessed relative to housekeeping gene expression in the given sample.

In-gel tryptic digestion, matrix-assisted laser desorption/ionization-time-of-flight mass spectrometry, and peptide mass fingerprinting. In-gel digestion was done following the protocol of Shevchenko et al. (30, 31) with modification. Reflector matrix-assisted laser desorption/ionization-time-of-flight (MALDI-TOF) spectra were acquired on a Voyager STR (Applied Biosystems) and used to search nonredundant protein sequence databases with Aldente and ProFound software.5

5

For more information about MALDI-TOF data analysis, visit (http://www.expasy.org/tools/aldente/) and (http://www.expasy.org/tools/genomicsolutions/).

Reverse transcription-PCR analysis. RNA was extracted with RNAzol B isolation solvent (Tel-Test, Inc.). Reverse transcription was carried out at 42°C for 60 min using 2 μg of total RNA, 75 pmol of oligo(dT) primer, 0.5 mmol/L of each deoxynucleotide triphosphate, and 50 units of Moloney murine leukemia virus reverse transcriptase (Invitrogen). PCR was done in a total volume of 50 μL using the following sequence-specific primer pairs: FBLN-1, (+) primer 5′-CCGGAGTGGACGCGGATGTCCTCC-3′ and (−) primer 5′-CCTCCAGCTGGCTGTGGCAGCACT-3′ (expected product 149 bp) and (+) primer 5′-ATATGCTACGGAATCCAAAG-3′ and (−) primer 5′-TTGATGCATGTATGCCCAAT-3′ (expected product 821 bp; upstream and downstream oligonucleotide primers both taken from a region of fibulin cDNA common to the four isoforms). After an initial denaturation step at 95°C for 3 min, PCR was carried out at 95°C for 1 min, 56°C for 2 min and 72°C for 2 min for 30 cycles, followed by final extension at 72°C for 10 min. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) transcripts were amplified in a 20-cycle PCR using (+) primer 5′-TCCACCACCCTGTTGCTGTA-3′ and (−) primer 5′-ACCACAGTCCATGCCATCAC-3′.

Cell treatments. Evaluations of FBLN-1 transcript and protein expression in response to chemotherapy were done on breast tumor cell lines first grown at 70% confluence and then treated with doxorubicin (Pharmacia Upjohn) or vehicle for 72 h. The following growth inhibition concentrations (IC30) of doxorubicin were used: 122 nmol/L for MCF-7, 114 nmol/L for SKBR-3, and 132 nmol/L for MDA-MB-157, as previously calculated (32).

ELISA analysis of ECM-associated FBLN-1. ELISA was used to quantify levels of FBLN-1 associated with ECM. Cell lines were treated as above, the culture medium was removed, and cells were repeatedly washed in PBS. After lysing cells by two incubations in 20 mmol/L NH4OH for 10 min, the plates were extensively washed with PBS and then with distilled water and allowed to air dry, as described (33). The ECM proteins were recovered in collection buffer [10 mmol/L Tris-HCl (pH 7.0), 0.1% SDS, 100 mmol/L β-mercaptoethanol]. After 15 min at 37°C, the fluid was collected with a cell scraper. The recovered ECM-associated proteins were used as targets for ELISA as described (34).

Silencing of FBLN-1 by small interfering RNA transfection. MCF-7 cells (3 × 105 per well) were seeded in six-well plates and grown to 60% to 80% confluence and the cell layers washed with serum-free medium. The cell layers were transfected either with a pool of two small interfering RNA (siRNA) oligonucleotides specific for human FBLN-1 (final concentration, 50 nmol/L; Dharmacon) or a pool of control RNA duplexes (Dharmacon) using transIT-TKO transfection reagent (10 μL/well; Mirus). After overnight incubation at 37°C, culture medium was replaced with fresh complete medium containing 10% FCS. Cells were harvested after 48 h, processed for immunofluorescence staining, and treated with escalating doses of doxorubicin for cytotoxicity assays or solubilized for Western blot analysis of FBLN-1 silencing.

Proliferation assay. To study doxorubicin sensitivity of FBLN-1–silenced tumor cells, 3 × 104 MCF-7 cells per well were seeded in 96-well plates in RPMI supplemented with 10% FCS, transfected with FBLN-1–specific siRNA sequences or with control reagents for 48 h, and treated with escalating doses of doxorubicin (six replicates for each drug dose for each cell treatment). Doxorubicin was removed after 72 h and cell proliferation was assessed as described (35).

In vitro apoptosis assays. Human breast cancer cell lines MCF-7, MDA-MB-361, SKBR-3, MDA-MB-435, MDA-MB-175, MDA-MB-231, MDA-MB-453, and MDA-MB-468 in 2 mL of RPMI 1640 supplemented with 1% FCS were seeded at 2 × 105 per well in six-well plates. Replicate cultures were set up in wells precoated or not with a thin layer of Matrigel (BD Biosciences PharMingen; 2 mL of a 1:5 dilution of Matrigel stock solution). In some experiments, cells were seeded in the presence or absence of the FBLN-1 mAbs P3B4, P1B2, and P5B1, used individually or pooled. After 24 h, doxorubicin (200 nmol/L) was added to replicate wells containing cells with or without Matrigel. Apoptosis was evaluated after 72 h of doxorubicin treatment by staining tumor cells with Annexin V-FITC and propidium iodide as described (36) followed by flow cytometry analysis with FACScalibur flow cytometer (BD Biosciences PharMingen).

In vivo tumorigenesis assay. Six- to eight-week-old BALB/c athymic mice were purchased from Charles River. Care and use of the animals were in accordance with institutional guidelines. Mice were injected s.c. with 1 × 106 MDA-MB-361 cells alone; 100 μL of Matrigel alone; cells and Matrigel; cells and doxorubicin (7 mg/mL), administered i.p. at 72 h following tumor cell injection; or cells and Matrigel and doxorubicin, with six mice in each group. Tumors were calibrated twice weekly for 58 days and tumor volume was calculated as 0.5 × d12 × d2, where d1 and d2 are the larger and smaller diameters, respectively.

Terminal deoxyribonucleotidyl transferase–mediated dUTP nick end labeling assay and immunohistochemistry. Apoptotic cell death was evaluated in paraffin-embedded sections of xenografts surgically removed from athymic mice 10 days after s.c. injection of MDA-MB-361 cells, suspended in 0.1 mL of Matrigel or in PBS. Doxorubicin treatment was done 72 h postinjection of cells. Control mice received no injection. A terminal deoxyribonucleotidyl transferase–mediated dUTP nick end labeling (TUNEL)–based assay with the In situ Cell Death Detection Kit, POD (Roche), was carried out according to the manufacturer's instructions. Sections were analyzed as described (37) by acquiring digital images of 10 areas of each tissue section (at ×400) on a Zeiss Axiovert 100 microscope (Carl Zeiss). Other immunohistochemical stainings were done on cryosections using a peroxidase-streptavidin method as previously described (23).

Statistical methods. The results of ELISA and of in vitro apoptosis assay of cells cultured on Matrigel layer and involving the use of FBLN-1 mAbs were evaluated by ANOVA followed by Student-Newman-Keuls test. Statistical significance was annotated as *, P < 0.05; ***, P < 0.001. The results of in vitro apoptosis assays of breast tumor cell lines and of MDA-MB-361 tumor cells admixed or not with Matrigel and injected in mice treated with doxorubicin were evaluated by two-tailed paired and unpaired t test, respectively. Comparison of FBLN-1 gene expression levels in pre- and post-doxorubicin treated breast cancers was evaluated by Mann-Whitney test. Differences were considered significant at P ≤ 0.05.

Modulation of ECM components on doxorubicin treatment. Analyses using reverse transcription-PCR (RT-PCR; Fig. 1A), fluorescence-activated cell sorting (FACS; Fig. 1B), and ELISA (Fig. 1C) revealed increased FBLN-1 expression at both the transcript and protein levels in SKBR-3, MCF-7, MDA-MB-157, and MCF-7-DXR (doxorubicin-resistant) tumor cells as compared with untreated control cells. Densitometric analysis of the expression data indicated that doxorubicin induced 2-, 3-, and 4-fold increases in the expression of FBLN-1 RNA and protein (Fig. 1A). Because a 2-fold increase in FBLN-1 mRNA levels was also observed after a nontoxic physical stress in the form of thermal shock at 42°C for 4 h (Fig. 1A), the doxorubicin-induced increases in FBLN-1 are likely related to drug-induced stress and not to selection of drug-resistant cell variants. The MCF-7 doxorubicin-resistant variant cell line was found to constitutively overexpress FBLN-1 compared with the parental MCF-7B line (Fig. 1A–C).

Figure 1.

Doxorubicin modulation of FBLN-1 RNA and protein expression in cultured breast cancer cells. A, densitometric analysis of FBLN-1 mRNA levels evaluated by RT-PCR before (−) and after (+) doxorubicin treatment (IC30 doses) for 72 h of human breast cancer cell lines SKBR-3, MCF-7, MDA-MB-157, MDA-MB-157 (thermally shocked at 42°C for 4 h), MCF-7B, and MCF-7 doxorubicin (DXR)-resistant cell variant. B, cytofluorimetric analysis of cytoplasm-associated FBLN-1 expression before (gray empty histograms) and after (black empty histograms) doxorubicin treatment of SKBR-3, MCF-7, MDA-MB-157, MCF-7B (gray empty histogram), and MCF-7 doxorubicin-resistant (black empty histogram) cells using the FBLN-1 MEM-2 mAb. Gray filled histograms, internal negative control data. C, ELISA analysis of ECM-associated FBLN-1 released in vitro by SKBR-3, MCF-7, MDA-MB-157 treated (+) or not (−) with doxorubicin, as indicated above, and MCF-7B and MCF-7 doxorubicin-resistant cell variant using MEM-2 mAb. White and light gray columns, internal negative control (PBS); dark gray and black columns, FBLN-1 expression of each tumor cell line tested before (−) and after (+) doxorubicin treatment, respectively. *, P < 0.05; ***, P < 0.001 (ANOVA followed by Student-Newman-Keuls multiple comparison test).

Figure 1.

Doxorubicin modulation of FBLN-1 RNA and protein expression in cultured breast cancer cells. A, densitometric analysis of FBLN-1 mRNA levels evaluated by RT-PCR before (−) and after (+) doxorubicin treatment (IC30 doses) for 72 h of human breast cancer cell lines SKBR-3, MCF-7, MDA-MB-157, MDA-MB-157 (thermally shocked at 42°C for 4 h), MCF-7B, and MCF-7 doxorubicin (DXR)-resistant cell variant. B, cytofluorimetric analysis of cytoplasm-associated FBLN-1 expression before (gray empty histograms) and after (black empty histograms) doxorubicin treatment of SKBR-3, MCF-7, MDA-MB-157, MCF-7B (gray empty histogram), and MCF-7 doxorubicin-resistant (black empty histogram) cells using the FBLN-1 MEM-2 mAb. Gray filled histograms, internal negative control data. C, ELISA analysis of ECM-associated FBLN-1 released in vitro by SKBR-3, MCF-7, MDA-MB-157 treated (+) or not (−) with doxorubicin, as indicated above, and MCF-7B and MCF-7 doxorubicin-resistant cell variant using MEM-2 mAb. White and light gray columns, internal negative control (PBS); dark gray and black columns, FBLN-1 expression of each tumor cell line tested before (−) and after (+) doxorubicin treatment, respectively. *, P < 0.05; ***, P < 0.001 (ANOVA followed by Student-Newman-Keuls multiple comparison test).

Close modal

To verify the findings from in vitro experiments, FBLN-1 levels were evaluated in xenografts of MDA-MB-361 cells grown in athymic mice treated with doxorubicin or buffer 72 h after tumor cell injection. As shown in Fig. 2, FBLN-1 mRNA and protein were increased by 2-fold in tumors grown in mice treated with doxorubicin compared with controls (Fig. 2A and B). In addition, immunohistochemistry analysis of the above-mentioned samples, done with MEM-2 mAb, revealed higher levels of FBLN-1 in surrounding tumor stroma after doxorubicin treatment compared with control (Fig. 2C).

Figure 2.

Modulation of in vivo expression of FBLN-1, fibronectin, and laminin-1 by doxorubicin treatment in breast tumor xenografts. A, RT-PCR analysis of FBLN-1 mRNA expression in MDA-MB-361 breast tumor xenografts grown in athymic mice treated (+) or not (−) with doxorubicin (administered i.p. after 72 h from tumor cell injection). cDNAs were tested for integrity by amplification of GAPDH transcripts as an internal control. Densitometry reveals a 2-fold increase in FBLN-1 transcript biosynthesis after drug treatment compared with the control. B, Western blot analysis of soluble extracts from MDA-MB-361 tumor biopsies grown in athymic mice treated (+) or not (−) with doxorubicin as above using the FBLN-1 MEM-2 mAb. β-Actin was used to normalize protein loading. Densitometry reveals a 2-fold increase in FBLN-1 protein biosynthesis following drug treatment compared with the control. C, immunohistochemical analysis of FBLN-1 expression in a cryosection from a biopsy from a MDA-MB-361 tumor grown in athymic mice treated (+) or not (−) with doxorubicin. The primary antibody used was FBLN-1 mAb MEM-2. D, Western blot analysis of soluble extracts from MDA-MB-361 xenografts grown in doxorubicin treated (+) and control (−) athymic mice using FBLN-1 MEM-2 mAb and antihuman laminin-1 (LN-1) and fibronectin-1 (FN-1) polyclonal sera. The level of the anti-vinculin polypeptide was used to normalize protein loading. Densitometry reveals a 2-, 2.1-, and 3.6-fold increase in expression of FBLN-1, laminin-1, and fibronectin, respectively, in xenotransplants of drug-treated mice as compared with controls.

Figure 2.

Modulation of in vivo expression of FBLN-1, fibronectin, and laminin-1 by doxorubicin treatment in breast tumor xenografts. A, RT-PCR analysis of FBLN-1 mRNA expression in MDA-MB-361 breast tumor xenografts grown in athymic mice treated (+) or not (−) with doxorubicin (administered i.p. after 72 h from tumor cell injection). cDNAs were tested for integrity by amplification of GAPDH transcripts as an internal control. Densitometry reveals a 2-fold increase in FBLN-1 transcript biosynthesis after drug treatment compared with the control. B, Western blot analysis of soluble extracts from MDA-MB-361 tumor biopsies grown in athymic mice treated (+) or not (−) with doxorubicin as above using the FBLN-1 MEM-2 mAb. β-Actin was used to normalize protein loading. Densitometry reveals a 2-fold increase in FBLN-1 protein biosynthesis following drug treatment compared with the control. C, immunohistochemical analysis of FBLN-1 expression in a cryosection from a biopsy from a MDA-MB-361 tumor grown in athymic mice treated (+) or not (−) with doxorubicin. The primary antibody used was FBLN-1 mAb MEM-2. D, Western blot analysis of soluble extracts from MDA-MB-361 xenografts grown in doxorubicin treated (+) and control (−) athymic mice using FBLN-1 MEM-2 mAb and antihuman laminin-1 (LN-1) and fibronectin-1 (FN-1) polyclonal sera. The level of the anti-vinculin polypeptide was used to normalize protein loading. Densitometry reveals a 2-, 2.1-, and 3.6-fold increase in expression of FBLN-1, laminin-1, and fibronectin, respectively, in xenotransplants of drug-treated mice as compared with controls.

Close modal

To ascertain whether the FBLN-1 observed in tumor specimens was of tumor or host origin, FBLN-1 was immunoaffinity purified from xenotransplants from untreated and doxorubicin-treated mice and subjected to peptide mass fingerprinting based on MALDI-TOF mass spectrometry (Supplementary Fig. S1A, lanes 1 and 2, respectively). Human FBLN-1 was identified with a high score in the anti–FBLN-1 immunoprecipitates (see peptides listed with bold characters, Supplementary Fig. S1B). Copurification of human and mouse FBLN-1 could not be excluded. In fact, due to high sequence similarity, a number of the tryptic peptides found in mass spectra were consistent with FBLN-1 of both species (Supplementary Fig. S1B).

Additional analysis of ex vivo changes in ECM composition on doxorubicin treatment revealed 3.6- and 2-fold increases in levels of fibronectin and laminin-1, two ECM proteins capable of binding to FBLN-1 (Fig. 2D). These findings suggest that tumor cells respond to doxorubicin treatment by remodeling the ECM in their microenvironment.

Role of ECM in sensitivity to doxorubicin. To assess the role of ECM proteins in sensitivity to doxorubicin, eight human breast cancer cell lines were seeded on plastic or on a thin layer of Matrigel, a solubilized basement membrane matrix preparation containing high levels of FBLN-1 (Fig. 3A,, inset, lane 2) and laminin-1, as indicated in the BD Matrigel data sheet. The cells were then exposed to doxorubicin (Fig. 3A). Overall, cells cultured on Matrigel showed a significant reduction in doxorubicin-induced apoptosis (P = 0.0007), with the level of protection ranging from 53% with MDA-MB-453 cells to 95% with MDA-MB-435, MDA-MB-361, MDA-MB-468, and MDA-MB-175 cells.

Figure 3.

Matrigel protects breast cancer cells from doxorubicin toxicity. A, human breast cancer cell lines MDA-MB-361 (♦), MDA-MB-435 (□), MDA-MB-468 (▴), SKBR-3 (▪), MCF-7 (•), MDA-MB-453 (▿), MDA-MB-231 (▵), and MDA-MB-157 (▾) were seeded onto plastic or Matrigel and treated with doxorubicin (200 nmol/L) for 72 h. Apoptosis was evaluated by flow cytometry after staining cell samples with Annexin V and propidium iodide. Statistical significance was evaluated by two-tailed paired t test and globally annotated as P = 0.0007 for apoptosis protection mediated by Matrigel. Inset, Western blot analysis of human FBLN-1 from the SKBR-3 cell extract (lane 1; tested as internal control) and of murine FBLN-1 (lane 2) from Matrigel (used in all experiments) probed with the pool of P5B1, P3B4, and P1B2 mAbs. B, tumor growth rate of MDA-MB-361 cells injected s.c. alone or combined with Matrigel in athymic mice treated or not with doxorubicin (7 mg/mL) administered 72 h from tumor cell injection. Mice were divided into five experimental groups of six animals per group: (a) tumor cells only (▴); (b) Matrigel only (▪); (c) tumor cells and doxorubicin (□); (d) tumor cells and Matrigel (♦); and (e) tumor cells, doxorubicin, and Matrigel (•). Statistical significance was evaluated by two-tailed unpaired t test. The difference between mouse group no. 3 (□) and no. 5 (•) was significant (P = 0.05). C, TUNEL staining of xenografts from athymic mice 10 d postinjection with MDA-MB-361 cells in the presence or absence of Matrigel and doxorubicin treatment. Doxorubicin was administered for 72 h. Surgically removed MDA-MB-361 xenografts were evaluated by TUNEL assay for apoptotic cell death 10 d from tumor cell injection. Magnification, ×400.

Figure 3.

Matrigel protects breast cancer cells from doxorubicin toxicity. A, human breast cancer cell lines MDA-MB-361 (♦), MDA-MB-435 (□), MDA-MB-468 (▴), SKBR-3 (▪), MCF-7 (•), MDA-MB-453 (▿), MDA-MB-231 (▵), and MDA-MB-157 (▾) were seeded onto plastic or Matrigel and treated with doxorubicin (200 nmol/L) for 72 h. Apoptosis was evaluated by flow cytometry after staining cell samples with Annexin V and propidium iodide. Statistical significance was evaluated by two-tailed paired t test and globally annotated as P = 0.0007 for apoptosis protection mediated by Matrigel. Inset, Western blot analysis of human FBLN-1 from the SKBR-3 cell extract (lane 1; tested as internal control) and of murine FBLN-1 (lane 2) from Matrigel (used in all experiments) probed with the pool of P5B1, P3B4, and P1B2 mAbs. B, tumor growth rate of MDA-MB-361 cells injected s.c. alone or combined with Matrigel in athymic mice treated or not with doxorubicin (7 mg/mL) administered 72 h from tumor cell injection. Mice were divided into five experimental groups of six animals per group: (a) tumor cells only (▴); (b) Matrigel only (▪); (c) tumor cells and doxorubicin (□); (d) tumor cells and Matrigel (♦); and (e) tumor cells, doxorubicin, and Matrigel (•). Statistical significance was evaluated by two-tailed unpaired t test. The difference between mouse group no. 3 (□) and no. 5 (•) was significant (P = 0.05). C, TUNEL staining of xenografts from athymic mice 10 d postinjection with MDA-MB-361 cells in the presence or absence of Matrigel and doxorubicin treatment. Doxorubicin was administered for 72 h. Surgically removed MDA-MB-361 xenografts were evaluated by TUNEL assay for apoptotic cell death 10 d from tumor cell injection. Magnification, ×400.

Close modal

Tumor volume measurements were made on the tumors from athymic mice injected with MDA-MB-361 cells in the presence or absence of Matrigel and doxorubicin treatment. The results show that there were no differences in the volume of tumors from mice injected with MDA-MB-361 cells alone or those from mice injected with MDA-MB-361 cells mixed with Matrigel but not treated with doxorubicin. By contrast, tumors from mice injected with MDA-MB-361 cells and Matrigel and treated with doxorubicin were significantly larger (P = 0.05) than tumors from mice injected with MDA-MB-361 cells not combined with Matrigel but treated with doxorubicin (Fig. 3B). Tissue sections from xenografts in mice injected with MDA-MB-361 cells without Matrigel and treated with doxorubicin also displayed TUNEL staining. By contrast, little or no TUNEL staining was observed in tumor sections from mice injected with MDA-MB-361 cells and Matrigel but not treated with doxorubicin (Fig. 3C).

Inhibition of FBLN-1 restores sensitivity to doxorubicin. To further support in vivo data, we next evaluated the effect of FBLN-1 antibodies on the observed protective effects of Matrigel on doxorubicin-induced apoptosis. MDA-MB-361 cells were cultured on a thin Matrigel layer and treated with doxorubicin in the presence or absence of FBLN-1 mAbs P3B4, P1B2, and P5B1 (used individually or pooled together). As shown in Fig. 4, each FBLN-1 mAb produced a significant reduction (P < 0.001) in the protective effect of Matrigel against doxorubicin-induced apoptosis, with the highest level of cytotoxicity achieved with mAb P5B1 and the pooled mAbs.

Figure 4.

FBLN-1 mAbs block the protective effects of Matrigel against doxorubicin-induced apoptosis. MDA-MB-361 cells were cultured on plastic or Matrigel and then treated (dark gray columns) or not (light gray columns) with doxorubicin (200 nmol/L) for 72 h in the presence or absence of FBLN-1 mAbs P3B4, P1B2, and P5B1, alone or pooled. Apoptosis was evaluated by flow cytometry after staining cell samples with Annexin V and propidium iodide. *, P < 0.05; ***, P < 0.001, ANOVA followed by Student-Newman-Keuls test.

Figure 4.

FBLN-1 mAbs block the protective effects of Matrigel against doxorubicin-induced apoptosis. MDA-MB-361 cells were cultured on plastic or Matrigel and then treated (dark gray columns) or not (light gray columns) with doxorubicin (200 nmol/L) for 72 h in the presence or absence of FBLN-1 mAbs P3B4, P1B2, and P5B1, alone or pooled. Apoptosis was evaluated by flow cytometry after staining cell samples with Annexin V and propidium iodide. *, P < 0.05; ***, P < 0.001, ANOVA followed by Student-Newman-Keuls test.

Close modal

In another approach to test the consequence of suppressing FBLN-1 expression on doxorubicin sensitivity, MCF-7 cells were transfected with FBLN-1–specific siRNAs or control siRNA duplexes and treated with doxorubicin for 72 h. Immunoblot analysis revealed a 2-fold decrease in FBLN-1 expression in cells transfected with a pool of FBLN-1 siRNAs (Fig. 5, lane 3) as compared with cells transfected with a pool of control RNA duplexes (Fig. 5, lane 4) or cells subjected to treatment with transfection reagent alone (Fig. 5, lane 2). By comparison to cells treated with medium alone, there was a 2.7-fold decrease in FBLN-1 expression in cells transfected with FBLN-1–specific siRNAs (Fig. 5, lane 1). FACS analysis using the same biological samples confirmed the decrease in FBLN-1–silenced cells, as reflected in decreased anti–FBLN-1 immunofluorescence intensity and decreased percentage of FBLN-1–positive events (Table 1). Dose dependence analysis of FBLN-1–silenced MCF-7 cells and controls incubated in the presence of varying concentrations of doxorubicin (0.002 × 10−6–9 × 10−6 mol/L) revealed an IC50 dose of 10 × 10−6 mol/L for siRNA-silenced cells as compared with 1 × 10−6 mol/L for sham transfected cells or cells incubated in medium, and compared with 0.8 × 10−6 mol/L for MCF-7 cells transfected with scrambled oligonucleotides (Table 1). These data indicate that FBLN-1–silenced breast tumor cells display a 10-fold increase in doxorubicin sensitivity compared with controls.

Figure 5.

SiRNA-mediated suppression of FBLN-1 expression in breast cancer cells. Western blot analysis of extracts from untreated MCF-7 cells (lane 1), sham transfected MCF-7 cells (lane 2), MCF-7 cells transfected with a pool of two siRNA oligonucleotides specific for human FBLN-1 (lane 3), or MCF-7 cells transfected with a pool of unrelated RNA duplexes (lane 4). The blot was probed with MEM-2 mAb. Relative levels of FBLN-1 were measured by densitometry with the vinculin immunoreactive polypeptide bands used to normalize protein loading. Densitometry reveals a 2-fold decrease in FBLN-1 expression in lane 3 compared with lanes 2 and 4, and a 2.7-fold decrease compared with lane 1.

Figure 5.

SiRNA-mediated suppression of FBLN-1 expression in breast cancer cells. Western blot analysis of extracts from untreated MCF-7 cells (lane 1), sham transfected MCF-7 cells (lane 2), MCF-7 cells transfected with a pool of two siRNA oligonucleotides specific for human FBLN-1 (lane 3), or MCF-7 cells transfected with a pool of unrelated RNA duplexes (lane 4). The blot was probed with MEM-2 mAb. Relative levels of FBLN-1 were measured by densitometry with the vinculin immunoreactive polypeptide bands used to normalize protein loading. Densitometry reveals a 2-fold decrease in FBLN-1 expression in lane 3 compared with lanes 2 and 4, and a 2.7-fold decrease compared with lane 1.

Close modal
Table 1.

Inhibition of FBLN-1 biosynthesis leads to increased sensitivity to doxorubicin

Cell treatmentsFACS analysis
Sensitivity to doxorubicin (IC50), × 10−6 mol/L
% FBLN-1–positive eventsFluorescence index
Medium 68 10 1 ± 0.06 
Sham transfection 69 1 ± 0.05 
Transfection with FBLN-1 siRNA pool 41 2.5 0.09 ± 0.23 
Transfection with control siRNA pool 74 0.8 ± 0.13 
Cell treatmentsFACS analysis
Sensitivity to doxorubicin (IC50), × 10−6 mol/L
% FBLN-1–positive eventsFluorescence index
Medium 68 10 1 ± 0.06 
Sham transfection 69 1 ± 0.05 
Transfection with FBLN-1 siRNA pool 41 2.5 0.09 ± 0.23 
Transfection with control siRNA pool 74 0.8 ± 0.13 

Findings presented here show a role for the ECM protein FBLN-1 in promoting resistance of breast cancer cells to the antitumor drug doxorubicin. This conclusion was based on several findings including those showing that FBLN-1 mAbs block the ability of Matrigel to protect tumor cells from doxorubicin-induced apoptosis. Furthermore, inhibition of FBLN-1 expression by FBLN-1 siRNA treatment markedly increased doxorubicin sensitivity of breast cancer cells.

Findings from in vitro and in vivo experiments examining FBLN-1 transcript and protein expression in doxorubicin-treated breast cancer cells suggest that FBLN-1 is part of an early response mechanism against drug-induced cell death. In fact, breast cancer cell lines exhibited a marked increase in FBLN-1 mRNA levels, as well as intracellular and extracellular FBLN-1 protein production, on doxorubicin treatment at 72 h. Moreover, analysis of xenografts showed a 2-fold increase in FBLN-1 expression levels in tumors grown in mice treated with doxorubicin compared with controls. In addition to FBLN-1, doxorubicin treatment also increased expression of the FBLN-1–interacting ECM proteins, fibronectin and laminin-1, suggesting a common mechanism activated by doxorubicin that increases tumor cell transcription of functionally interrelated ECM components. It is possible that doxorubicin induces transcription factors that act on the promoters of these ECM genes. Indeed, the promoters of fibronectin, laminin-1, and FBLN-1 genes harbor consensus binding sites for the transcription factor specificity protein 1 (Sp1; ref. 38), and doxorubicin directly induces Sp1 activation (39). Mechanisms involving activation of p53 following doxorubicin-induced DNA damage (40) seem to be less likely because we observed FBLN-1 up-regulation in three breast cancer cell lines differing in p53 status (wild-type in MCF-7, gain-of-function mutant in SKBR-3, and null in MDA-MB-157 cells).

Previous studies have suggested that long-term exposure of cancer patients to antitumor drugs is associated with up-regulation of ECM genes. Through analysis of a database (compiled by Perou and coworkers)6

6

For more information about gene expression, visit (http://genome-www.stanford.edu/molecularportraits/).

of breast cancer specimens isolated after doxorubicin chemotherapy and pretherapy samples from the same patients (41), we found increased expression of ECM genes in the post drug treatment group and mainly of FBLN-1 (see Supplementary Fig. S2). Similarly, Jazaeri et al. (42) also found statistically significant overrepresentation of ECM genes, including FBLN-1, in ovarian tumors after treatment with platinum-based chemotherapy compared with that of primary tumors from different patients (42). Taken together, these findings indicate that stromal-epithelial interactions or ECM may be involved in acquired chemoresistance of breast and ovarian cancers and that the link between FBLN-1 and drug response is not necessarily exclusive to doxorubicin. As noted by Damiano et al. (9), mechanisms of resistance selected by chronic drug treatment may differ from mechanisms of cell survival after acute exposure to chemotherapeutic drugs, so that a selection process might underlie the higher ECM gene expression in the former studies. Thus, in a heterogeneous tumor population, neoplastic cells with high and constitutive expression of ECM genes that exert a “protective” effect against drug-induced death may survive chemotherapy. However, our findings from short-term treatment (72 h) of neoplastic cells with doxorubicin suggest a direct modulation of FBLN-1 gene expression due to doxorubicin-induced stress and not due to selection of a doxorubicin-resistant cell clone, especially because thermal shock induced an altered expression of the FBLN-1 molecule without affecting viability of breast cancer cells. These two different models may yet be reconciled in light of early results of gene expression analysis in doxorubicin-resistant neoplastic cells, indicating increased expression and DNA-binding activity of the Sp1 transcription factor (43). Such a “doxorubicin-Sp1” pathway might explain the observed modulation of ECM genes after brief drug exposure while providing a mechanism that stably affects gene expression after long-term drug exposure of neoplastic cells.

The mechanism by which FBLN-1 exerts chemoresistance in breast cancer cells remains unclear. Based on the current knowledge of the biological activities of FBLN-1, the chemoresistance effects might be mediated through integrin adhesion receptor signaling. FBLN-1 has been shown to suppress the motility and adhesive-promoting effects of fibronectin (20), which are mediated by integrin α5β1. In fact, fibronectin has been shown to deliver survival signals transduced by focal adhesion kinase, leading to suppression of p53-dependent apoptosis in normal cells (44). Furthermore, triggering of β1 integrins by ECM interaction suppresses chemotherapy-induced apoptosis in small-cell lung cancer cells (8). Moreover, integrin-dependent activation by ECM components of the Akt/mammalian target of rapamycin pathway and cell cycle arrest through p27kip1 up-regulation are important mechanisms in promoting resistance of neoplastic cells to chemotherapy (4547). It is possible that FBLN-1 plays an indirect role in promoting doxorubicin resistance by facilitating the interaction of integrins with other ECM components. FBLN-1 may act as a molecular scaffold assisting in the correct assembly of complex ECMs, thereby favoring the appropriate spatial orientation of ECM components needed for efficient engagement and triggering of integrins. The results described in this study, on FBLN-1 mAb–mediated inhibition of Matrigel protection against doxorubicin-induced apoptosis, are consistent with such model of FBLN-1 function.

In conclusion, our study provides evidence that FBLN-1 promotes chemoresistance in breast cancer cells, adding to the number of ECM components that provide such prosurvival function to neoplastic cells. Further studies are needed to determine the potential usefulness of assessing FBLN-1 expression in breast tumors as a predictive marker of drug response and as a therapeutic target to promote drug sensitivity.

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

Grant support: Associazione Italiana per la Ricerca sul Cancro.

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.

1
Pupa SM, Ménard S, Forti S, Tagliabue E. New insights into the role of extracellular matrix during tumor onset and progression.
J Cell Physiol
2002
;
192
:
259
–67.
2
Rintoul RC, Sethi T. The role of extracellular matrix in small-cell lung cancer.
Lancet Oncol
2001
;
2
:
437
–42.
3
Xu ZW, Friess H, Buchler MW, Solioz M. Overexpression of Bax sensitizes human pancreatic cancer cells to apoptosis induced by chemotherapeutic agents.
Cancer Chemother Pharmacol
2002
;
49
:
504
–10.
4
Minchinton AI, Tannock IF. Drug penetration in solid tumours.
Nat Rev Cancer
2006
;
6
:
583
–92.
5
Schuetz JD, Schuetz EG. Extracellular matrix regulation of multidrug resistance in primary monolayer cultures of adult rat hepatocytes.
Cell Growth Differ
1993
;
4
:
31
–40.
6
Miyamoto H, Murakami T, Tsuchida K, Sugino H, Miyake H, Tashiro S. Tumor-stroma interaction of human pancreatic cancer: acquired resistance to anticancer drugs and proliferation regulation is dependent on extracellular matrix proteins.
Pancreas
2004
;
28
:
38
–44.
7
Morin PJ. Drug resistance and the microenvironment: nature and nurture.
Drug Resist Updat
2003
;
6
:
169
–72.
8
Sethi T, Rintoul RC, Moore SM, et al. Extracellular matrix proteins protect small cell lung cancer cells against apoptosis: a mechanism for small cell lung cancer growth and drug resistance in vivo.
Nat Med
1999
;
5
:
662
–8.
9
Damiano JS, Cress AE, Hazlehurst LA, Shtil AA, Dalton WS. Cell adhesion mediated drug resistance (CAM-DR): role of integrins and resistance to apoptosis in human myeloma cell lines.
Blood
1999
;
93
:
1658
–67.
10
Uhm JH, Dooley NP, Kyritsis AP, Rao JS, Gladson CL. Vitronectin, a glioma-derived extracellular matrix protein, protects tumor cells from apoptotic death.
Clin Cancer Res
1999
;
5
:
1587
–94.
11
Damiano JS, Hazlehurst LA, Dalton WS. Cell adhesion-mediated drug resistance (CAM-DR) protects the K562 chronic myelogenous leukemia cell line from apoptosis induced by BCR/ABL inhibition, cytotoxic drugs, and γ-irradiation.
Leukemia
2001
;
15
:
1232
–9.
12
Hazlehurst LA, Argilagos RF, Emmons M, et al. Cell adhesion to fibronectin (CAM-DR) influences acquired mitoxantrone resistance in U937 cells.
Cancer Res
2006
;
66
:
2338
–45.
13
Berube M, Talbot M, Collin C, et al. Role of the extracellular matrix proteins in the resistance of SP6.5 uveal melanoma cells toward cisplatin.
Int J Oncol
2005
;
26
:
405
–13.
14
Sherman-Baust CA, Weeraratna AT, Rangel LB, et al. Remodeling of the extracellular matrix through overexpression of collagen VI contributes to cisplatin resistance in ovarian cancer cells.
Cancer Cell
2003
;
3
:
377
–86.
15
Hurst RE, Kamat CD, Kyker KD, Green DE, Ihnat MA. A novel multidrug resistance phenotype of bladder tumor cells grown on Matrigel or SIS gel.
Cancer Lett
2005
;
217
:
171
–80.
16
Kouniavsky G, Khaikin M, Zvibel I, et al. Stromal extracellular matrix reduces chemotherapy-induced apoptosis in colon cancer cell lines.
Clin Exp Metastasis
2002
;
19
:
55
–60.
17
Argraves WS, Greene LM, Cooley MA, Gallagher WM. Fibulins: physiological and disease perspectives.
EMBO Rep
2003
;
4
:
1127
–31.
18
Roark EF, Keene DR, Haundenschild CC, Godyna S, Little CD, Argraves WS. The association of human fibulin-1 with elastic fibers: an immunohistological, ultrastructural, and RNA study.
J Histochem Cytochem
1995
;
43
:
401
–11.
19
Qing J, Maher VM, Tran H, Argraves WS, Dunstan RW, McCormick JJ. Suppression of anchorage-independent growth and matrigel invasion and delayed tumor formation by elevated expression of fibulin-1D in human fibrosarcoma-derived cell lines.
Oncogene
1997
;
15
:
2159
–68.
20
Twal WO, Czirok A, Hegedus B, et al. Fibulin-1 suppression of fibronectin-regulated cell adhesion and motility.
J Cell Sci
2001
;
114
:
4587
–98.
21
Moll F, Katsaros D, Lazennec G, et al. Estrogen induction and overexpression of fibulin-1C mRNA in ovarian cancer cells.
Oncogene
2002
;
21
:
1097
–107.
22
Greene LM, Twal WO, McDermott EW, et al. Elevated expression and altered processing of fibulin-1 protein in human breast cancer.
Br J Cancer
2003
;
88
:
871
–8.
23
Pupa SM, Argraves WS, Forti S, et al. Immunological and pathobiological roles of fibulin-1 in breast cancer.
Oncogene
2004
;
23
:
2153
–60.
24
Forti S, Scanlan MJ, Invernizzi AM, et al. Identification of breast cancer-restricted antigens by antibody screening of SKBR3 cDNA library using a preselected patient's serum.
Breast Cancer Res Treat
2002
;
73
:
245
–56.
25
Pupa SM, Forti S, Balsari A, Ménard S. Humoral immune response for early diagnosis of breast carcinoma.
Ann Oncol
2002
;
13
:
483
.
26
Sellappan S, Grijalva R, Zhou X, et al. Lineage infidelity of MDA-MB-435 cells: expression of melanocyte proteins in a breast cancer cell line.
Cancer Res
2004
;
64
:
3479
–85.
27
Monti E, Sinha BK. Antiproliferative effect of genistein and Adriamycin against estrogen-dependent and -independent human breast carcinoma cell lines.
Anticancer Res
1994
;
14
:
1221
–6.
28
Pupa SM, Forti S, Invernizzi AM, et al. Monoclonal antibody to fibulin-1 generated by genetic immunization.
J Cell Biochem
2003
;
89
:
647
–52.
29
Argraves WS, Tran H, Burgess WH, Dickerson K. Fibulin is an extracellular matrix and plasma glycoprotein with repeated domain structure.
J Cell Biol
1990
;
111
:
3155
–64.
30
Shevchenko A, Wilm M, Vorm O, Mann M. Mass spectrometric sequencing of proteins silver-stained polyacrylamide gels.
Anal Chem
1996
;
68
:
850
–8.
31
Gorla L, Cantu M, Micciche F, et al. RET oncoproteins induce tyrosine phosphorylation changes of proteins involved in RNA metabolism.
Cell Signal
2006
;
18
:
2272
–82.
32
Campiglio M, Somenzi G, Olgiati C, et al. Role of proliferation in HER2 status predicted response to doxorubicin.
Int J Cancer
2003
;
105
:
568
–73.
33
Plopper G, Falk-Marzillier J, Glaser S, et al. Changes in expression of monoclonal antibody epitopes on laminin-5r induced by cell contact.
J Cell Sci
1996
;
109
:
1965
–73.
34
Berno V, Porrini D, Castiglioni F, et al. The 67-kDa laminin receptor increases tumor aggressiveness by remodeling laminin-1.
Endocr Relat Cancer
2005
;
12
:
393
–406.
35
Castiglioni F, Tagliabue E, Campiglio M, Pupa SM, Balsari A, Ménard S. Role of exon-16-deleted HER2 in breast carcinomas.
Endocr Relat Cancer
2006
;
13
:
221
–32.
36
Zanon M, Piris A, Bersani I, et al. Apoptosis protease activator protein-1 expression is dispensable for response of human melanoma cells to distinct proapoptotic agents.
Cancer Res
2004
;
64
:
7386
–94.
37
Mortarini R, Piris A, Maurichi A, et al. Lack of terminally differentiated tumor-specific CD8+ T cells at tumor site in spite of antitumor immunity to self-antigens in human metastatic melanoma.
Cancer Res
2003
;
63
:
2535
–45.
38
Castoldi M, Chu ML. Structural and functional characterization of the human and mouse fibulin-1 gene promoters: role of Sp1 and Sp3.
Biochem J
2002
;
362
:
41
–50.
39
Uchida Y, Itoh M, Taguchi Y, et al. Ceramide reduction and transcriptional up-regulation of glucosylceramide synthase through doxorubicin-activated Sp1 in drug-resistant HL-60/ADR cells.
Cancer Res
2004
;
64
:
6271
–9.
40
Lowe SW, Bodis S, McClatchey A, et al. p53 status and the efficacy of cancer therapy in vivo.
Science
1994
;
266
:
807
–10.
41
Perou CM, Sorlie T, Eisen MB, et al. Molecular portraits of human breast tumours.
Nature
2000
;
406
:
747
–52.
42
Jazaeri AA, Awtrey CS, Chandramouli GV, et al. Gene expression profiles associated with response to chemotherapy in epithelial ovarian cancers.
Clin Cancer Res
2005
;
11
:
6300
–10.
43
Borellini F, Aquino A, Josephs SF, Glazer RI. Increased expression and DNA-binding activity of transcription factor Sp1 in doxorubicin-resistant HL-60 leukemia cells.
Mol Cell Biol
1990
;
10
:
5541
–7.
44
Ilic D, Almeida EA, Schlaepfer DD, Dazin P, Aizawa S, Damsky CH. Extracellular matrix survival signals transduced by focal adhesion kinase suppress p53-mediated apoptosis.
J Cell Biol
1998
;
143
:
547
–60.
45
Tsurutani J, West KA, Sayyah J, Gills JJ, Dennis PA. Inhibition of the phosphatidylinositol 3-kinase/Akt/mammalian target of rapamycin pathway but not the MEK/ERK pathway attenuates laminin-mediated small cell lung cancer cellular survival and resistance to imatinib mesylate or chemotherapy.
Cancer Res
2005
;
65
:
8423
–32.
46
Matsunaga T, Takemoto N, Sato T, et al. Interaction between leukemic-cell VLA-4 and stromal fibronectin is a decisive factor for minimal residual disease of acute myelogenous leukemia.
Nat Med
2003
;
9
:
1158
–65.
47
Hazlehurst LA, Damiano JS, Buyuksal I, Pledger WJ, Dalton WS. Adhesion to fibronectin via β1 integrins regulates p27kip1 levels and contributes to cell adhesion mediated drug resistance (CAM-DR).
Oncogene
2000
;
19
:
4319
–27.

Supplementary data