Transforming growth factor-β (TGF-β) is a multifunctional cytokine involved in cellular growth, migration, tissue repair and immune regulation.1 Many types of cancer cells are known to secrete TGF-β2 and TGF-β exerts dual actions in carcinogenesis. During the early stages, TGF-β acts as a potent growth inhibitor, but becomes a stimulant of invasion and metastasis at later stages.3 Moreover, TGF-β has been found to influence cancer cell adhesion and migration.4–6 More recently, inhibition of autocrine TGF-β signaling in carcinoma cells was shown to reduce cell invasiveness and tumor metastasis.7
After TGF-β stimulation, signaling is conducted through the activation of heteromeric complexes of 2 transmembrane receptor serine/threonine kinases consisting of a type II ligand binding receptor (TβR II) and a TGF-β type I signaling receptor (TβR I).8 The activation of this membrane complex occurs via ligand-dependent phosphorylation of TβR I by TβR II. Subsequently, TβR I phosphorylates its immediate downstream effectors Smad2 and Smad3, members of the Smad family of intracellular signaling molecules. This phosphorylation induces a conformational change in Smad2 and Smad3 which allows binding to another member of the Smad family, Smad4. This Smad complex then translocates to the nucleus, where it regulates the transcription of various target genes.9, 10
Matrix metalloproteases (MMPs) are involved in extracellular matrix degradation, a process that requires the presence of zinc. Because MMPs can degrade the extracellular matrix, they participate in tumor cell invasion and migration.11 Contemporary research divides the MMP family into 2 categories, including soluble-type MMPs and membrane-type MMPs (MT-MMP).12 One of the latter proteins, MT1-MMP, is frequently produced by invasive cancer and endothelial cells during angiogenesis13, 14 along with its substrates type I, type II and type III collagen; laminin-1 and laminin-5; vitronectin; fibronectin and aggrecan.15 MT1-MMP can also activate other proMMPs, including proMMP-2 and proMMP-13.16, 17 The expression of MT1-MMP on the cell surface can trigger various activation cascades.
CD44, a surface receptor for hyaluronan, serves as an adhesion molecule in cell–substrate and cell–cell interactions, and participates in lymphocyte recruitment to inflammatory sites as well as tumor metastasis.18–20 The CD44 adhesion molecule is known to be involved in the process of cancer metastasis. High CD44 levels on the surface of cancer cells are correlated with high metastatic activity. Furthermore, it has been shown in animal models that injected reagents that interfere with the binding of CD44 to its ligand also inhibit local tumor growth and metastatic spread.21, 22 Tumor cell invasion is comprised of cell adhesion to the extracellular matrix, degradation of extracellular matrix components, tumor cell motility and cell detachment.23 Tumor metastasis involves detachment of cells from the primary tumor mass, penetration of cells into the circulation, binding to the vascular wall, extravasation into the surrounding tissues and finally the seeding and proliferation of the cells with and the formation of metastatic foci.24, 25 The molecular mechanism(s) by which CD44 promotes tumor metastasis are poorly understood. Recent research demonstrates that the extracellular domain of CD44 undergoes cleavage by MT1-MMP on the surface of cells and this cleavage process plays a decisive role during cell migration in a pancreatic tumor cell line.26 This process of inducing CD44 deadhesion through metalloproteinase interaction appears to play a crucial role in cellular migration.
It has been observed that patients with more invasive breast cancer have higher serum TGF-β levels.2 TGF-β signaling is also involved in cancer cell migration and invasion.27 However, the exact molecular mechanism(s) by which TGF-β induces tumor metastasis remain unclear. During metastasis, cancer cells can break down the extracellular matrix to clear the path for cellular migration, and then enter the blood vessels or the lymphatics.15 Munshi et al.28 reported that TGF-β promoted MMP-dependent cell scattering and collagen invasion through increased expression of MMP-2 and MT1-MMP, and enhanced MMP-2 activation in oral squamous cell carcinoma.
In this article, we used an MDA-MB-435s cell line that likely represents a breast epithelial cell line that has undergone lineage infidelity29 and demonstrated that treatment of MDA-MB-435s cells with TGF-β results in de novo MT1-MMP synthesis on the cell membrane, a process that induces CD44 cleavage thereby increasing cell motility. We postulate that 1 function of TGF-β in tumor cells is the induction of MT1-MMP expression, which, at least for certain tumor cell types, is a critical step in the formation of metastatic colonies.
Material and methods
Human breast carcinoma cell lines, MDA-MB-435s, SK-BR-3 and human umbilical vein endothelial cells (HUVECs) were obtained from the Food Industry Research and Development Institute (Hsinchu, Taiwan). MDA-MB-435s cells were grown in Leibovitz's L-15 medium supplemented with 15% fetal bovine serum (FBS; Hyclone, Logan, UT), 10 μg/ml insulin, 100 U/ml penicillin and 100 μg/ml streptomycin at 37°C. SK-BR-3 cells were grown in DMEM supplemented with 10% FBS, 10 μg/ml insulin, 100 U/ml penicillin and 100 μg/ml streptomycin at 37°C. HUVECs were cultured in Medium 200 (Cascade Biologics, Portland, OR) supplemented with 2% FBS, 1 μg/ml hydrocortisone, 10 ng/ml human epidermal growth factor, 3 ng/ml basic fibroblast growth factor, 10 μg/ml heparin, 100 U/ml penicillin and 100 μg/ml streptomycin at 37°C in a humidified atmosphere of 5% CO2.
Mouse monoclonal antibody against human MT1-MMP and rabbit monoclonal antibody against CD44 were purchased from R&D systems (Minneapolis, MN). Mouse monoclonal antibody against CD44 and horseradish peroxidase-conjugated goat anti-mouse secondary antibody were purchased from NeoMarkers (Fremont, CA). Rabbit polyclonal antibody against MT1-MMP and Rhodamine-conjugated goat anti-rabbit secondary antibody were purchased from Chemicon International (San Diego, CA). FITC-conjugated bovine anti-mouse secondary antibody was purchased from Santa Cruz Biotechnology (Santa Cruz, CA).
The 21-nucleotide MT1-MMP dsRNAs were prepared by Ambion (Ambion, Austin, TX). The sequence of the MT1-MMP dsRNAs were prepared using single-stranded complementary RNA molecules with a two-nucleotide tail (UU): AUGCAGAA GUUUUACGGCUUGUU and CAAGCCGUAAAACUUCUG CAUUU.30 For transfection, cells (1 × 106 cells per well) were seeded in six-well plates, and 1.5 μg MT1-MMP siRNA was added and transfected using a Cell Line Optimization Nucleofector™ Kit (Amaxa, Gaithersburg, MD).
Western blot analyses
For immunoblotting on polyvinylidene difluoride (PVDF) membranes (Amersham Life Science, Piscataway, NJ), cells from two 100-mm dishes per treatment group were pooled, rinsed briefly with phosphate-buffered saline (PBS), then lysed on ice for 10 min in 1 ml PBS containing 1% sodium dodecyl sulfate, 0.5 mM phenylmethylsulfonyl fluoride, 10 μg/ml leupeptin, 10 μg/ml aprotinin, 5 μg/ml pepstatin, 10 μg/ml soybean trypsin inhibitor, and 0.5 mM dithiothreitol. After centrifugation for 20 min at 1,300g and 4°C, the supernatant was removed and centrifuged at 6,000g for 1 hr at 4°C. The final supernatant and pellet fractions were used as the cytosolic and membrane fractions, respectively. Equal amounts (50 μg protein) of the membrane fractions or cytosolic fractions were run on 10% polyacrylamide gels and transferred to PVDF membrane in transfer buffer (4 parts 25 mM Tris/200 mM buffer pH 8.0, 1 part methanol). The membranes were then blocked for 1 hr at room temperature with 50 mM Tris HCl, 150 mM NaCl, 0.05% Tween-20 (Tris buffered saline with Tween 20, TBST), pH 7.0, containing 5% nonfat dry milk and then incubated overnight at 4°C with the primary antibody. The membranes were then washed with TBST and exposed to the horseradish peroxidase-conjugated secondary antibody for 1 hr at room temperature. The bound antibody was detected by the enhanced chemiluminescence method (PerkinElmer Life Sciences, Wattham, MA).
To prepare soluble CD44, the media was collected and concentrated by Centricon YM-30 (Millipore, Billerica, MA). The samples were analyzed by SDS-PAGE using a 10% polyacrylamide gel. Separated polypeptides were then transferred onto PVDF membranes. After blocking nonspecific sites with 5% nonfat milk, the PVDF membranes were incubated with each of the specific immunoreagents (e.g., rat anti-CD44 Ab [5 μg/ml]), followed by an incubation with horseradish peroxidase-labeled immunoreagents (e.g., goat anti-rat Ab, goat anti-rabbit Ab or goat anti-mouse Ab). The blots were developed using the ECL system (Amersham Biosciences, Piscataway, NJ).
For immunoprecipitate experiments, cell lysate were collected and either immunoprecipitated with mouse anti-CD44 mAb (2 μg of primary antibody per 1 mg of protein sample; 0.6 μg of primary antibody per 1 ml medium) overnight at 4°C, and then immunoblotted with anti-MT1-MMP, or immunoprecipitated with MT1-MMP antibody before being immunoblotted with anti-CD44. Proteins were resolved by 12% SDS-PAGE, transferred to PVDF and subsequently probed with mAbs as indicated. Immunoreactive polypeptides were visualized using horseradish peroxidase-labeled secondary antibodies and the ECL detection system.
RNA extraction and RT-PCR analysis
Total RNA was extracted from both untreated (control) and treated cells using RNeasy Purification Reagent (Qiagen, Valencia, CA) and then a sample (1 μg) was reverse transcribed with Mmlv reverse transcriptase for 30 min at 42°C in the presence of the oligo-dT primer. PCR was performed using specific primers designed from the published sequence of each cDNA as follows: the oligonucleotides (sense 5′-GTGATGGATGGATACCCAA TGC-3′ and antisense 5′-GAACGCTGGCAGTAAAGCAGTC-3′) corresponding to human MT1-MMP were used for specific amplification of a 786-bp fragment of MT1-MMP mRNA. The initial temperature for the RT-PCR was 95°C (5 min) and then there were 35 cycles of denaturation (95°C, 30 sec), annealing (55°C, 30 sec) and elongation (72°C, 30 sec), with 1 additional 7-min incubation at 72°C after completion of the last cycle. To exclude the possibility of contaminating genomic DNA, the PCRs were also run without reverse transcriptase. The amplified cDNA was separated by electrophoresis through a 2% agarose gel.
Immunofluorescence stain and confocal laser scanning microscopic analysis
Breast cancer cells (1 × 105 cells) were cultured on glass cover slips, serum starved for 3 hr and then treated with TGF-β for various times. After incubation, the cells were fixed in 4% paraformaldehyde for 15 min, washed with PBS and preincubated in blocking solution (5% non-fat milk in PBS) for 15 min. After being washed with PBS, the cells were incubated with diluted anti-MT1-MMP mAb (1:500) or anti-CD44 mAb (1:500) in PBS for 60 min at room temperature. After being washed with PBS, the cells were incubated with diluted FITC-conjugated secondary antibody or rhodamine-conjugated secondary antibody for 60 min at room temperature. Samples were washed with PBS, mounted in Mounting medium (Vector, North Hollywood, CA) and visualized using a confocal microscope (Leica, Allendale, NJ). In the controls where the primary antibodies had been omitted, there was negligible immunofluorescence.
ECIS™ (electric cell-substrate impedance sensing)
To detect the invasive activities of the metastatic cells in vitro, an electric cell-substrate impedance sensing (ECIS) assay was used.31 Electrode arrays were gold film electrodes supplied by Applied Biophysics (Troy, NY) and were prepared by rinsing with PBS and then precoating with collagen for 1 hr. A HUVEC suspension was prepared at 5 × 105 cells/ml, and 200 μl were added to each well, resulting in a final surface concentration of 1.25 × 105 cells/cm2 and the cells were allowed to grow to confluence. Subsequently, 2 × 104 MDA-MB-435s cells were cultured in a 35-mm dish with or without TGF-β as described earlier. They were then prepared in suspension using fresh Medium 200. The cell suspension was then added to the well containing the monolayer of HUVEC. The impedance of the challenged HUVEC was monitored via ECIS for the next 20 hr to determine the invasiveness of the MDA-MB-435s cells. Highly metastatic cells reduce the total resistance across the HUVEC monolayer.
The transendothelial assay used was a transwell, the upper chamber of which consisted of cell culture inserts coated with 20 μl of matrigel (BD, Bedford, MA) for 30 min, then HUVEC cells (1 × 105 cells/well) were added until the cells reached confluence. MDA-MB-435s cells were transfected with/without MT1-MMP siRNA for 24 hr, then treated with TGF-β for 12 or 24 hr. Subsequently, MDA-MB-435s cells were trypsinized and resuspended in cell culture medium, then 2 × 105 cells were added to the upper chamber and incubated for 12 hr at 37°C in a humidified 5% carbon dioxide environment. Cells that had invaded through the matrix and become adherent to the undersurface of the filter were quantified using Hoechst staining and fluorescence microscopy.
The results are expressed as the mean ± the standard deviation (SD) of at least 3 experiments and comparisons were analyzed by one-way ANOVA. A p-value < 0.05 was considered significant.
Effects of TGF-β on MT1-MMP expression on MDA-MB-435s cells
MDA-MB-435s cells were treated with TGF-β (10 ng/ml) for different time periods and the level of MT1-MMP was monitored by Western blot analyses. TGF-β stimulation for 12 hr resulted in a significant increase in the level of MT1-MMP on the cell membrane with the highest level being observed after 24 hr of TGF-β treatment (Fig. 1a). In addition, the expression of MT1-MMP was increased after TGF-β treatment for12 hr and also at 18 and 24 hr in SK-BR-3 cell lines (Fig. 1b).
MT1-MMP expression in MDA-MB-435s cells treated with TGF-β for different time periods was evaluated by RT-PCR. After stimulation for 1, 3 and 6 hr, MT1-MMP expression steadily increased (Fig. 1c). To confirm that TGF-β-induced MT1-MMP mRNA transcription resulted in protein synthesis, the effect of cycloheximide on the MT1-MMP level after treatment with TGF-β was observed by Western blot analyses. Cycloheximide blocked TGF-β-induced MT1-MMP expression (Fig. 1d). Furthermore, transfection of MDA-MB-435s cells with MT1-MMP dsRNAs (siRNA) resulted in an inhibition of TGF-β-induced MT1-MMP expression (Fig. 1e).
These results demonstrate that TGF-β treatment of MDA-MB-435s cells led to an enhanced MT1-MMP expression resulting in an increased level of MT1-MMP on the membrane of MDA-MB-435s cells. Those observations were confirmed by immunocytochemical staining using mouse anti-MT1-MMP mAb (Figs. 2a–c and 2e). When the cells were pretreated with cycloheximide and subsequently treated with TGF-β for 18 and 24 hr, the expression of MT1-MMP decreased (Figs. 2d and 2f). These results further confirmed that TGF-β increases the level of MT1-MMP on MDA-MB-435s cells involving de novo protein synthesis.
Effects of various signal transduction inhibitors on the TGF-β-induced MT1-MMP level on the cell membrane of MDA-MB-435s cells
Although it is known that TGF-β signaling can be transmitted through the Smad pathway after the receptors are activated, only recently has it been discovered that this activated receptor complex can also signal through other pathways such as the mitogen-activated protein kinases (MAPKs)32 or phosphoinositol-3 kinase (PI3K)33 pathways. Therefore, we investigated the involvement of any of those pathways in TGF-β-induced MT1-MMP expression using various inhibitors, including the PI3K inhibitor LY294002, extracellular regulated kinase-1/2 (ERK1/2) inhibitor PD98059 and c-Jun NH2-terminal kinase (JNK) inhibitor SP600125.27, 28, 32 MDA-MB-435s cells treated with either LY294002 (Fig. 3a), PD98059 (Fig. 3b) or SP600125 (Fig. 3c) for 1 hr and then stimulated with TGF-β for different periods of time exhibited a significantly decreased level of MT1-MMP in the cell membrane when compared with MDA-MB-435s cells that were similarly stimulated with TGF-β without being pretreated with any of the inhibitors. These results suggest that TGF-β-induced MT1-MMP expression likely involves the PI3K, ERK1/2 and JNK pathways.
In situ localization of CD44 and MT1-MMP in TGF-β-treated MDA-MB-435s cells
In addition, we investigated the possible link between CD44 and TGF-β-induced MT1-MMP expression. After 24 hr of TGF-β stimulation, MDA-MB-435s cells were treated with mouse anti-CD44 mAb and FITC-conjugated bovine anti-mouse Ab to visualize CD44, and with rabbit anti MT1-MMP polyclonal antibody and rhodamine-conjugated goat anti-rabbit Ab to visualize MT1-MMP using confocal microscopy. Under normal conditions, the CD44 receptor (Fig. 4d) and MT1-MMP (Fig. 4e) are localized on the MDA-MB-435s cell membrane (Fig. 4f). After TGF-β stimulation for 24 hr, the CD44 level decreased (Fig. 4g) with an increase in the level of MT1-MMP (Fig. 4h). Interestingly, TGF-β induced MT1-MMP and CD44 colocalized in the cell membrane (Fig. 4i) which was supported by the histogram analysis. Under normal conditions, MDA-MB-435s cells contain CD44 on the cell membrane while MT1-MMP is predominantly localized in the cytosol (Figs. 4j–4l). After TGF-β stimulation, the level of CD44 on the cell membrane decreases while MT1-MMP and CD44 colocalize on the cell membrane (Figs. 4m–4o).
Effect of TGF-β treatment on the MT1-MMP and CD44 association and soluble CD44 level in MDA-MB-435s cells
To confirm the association between CD44 and MT1-MMP, we coimmunoprecipitated MT1-MMP using anti-CD44 antibodies. Greater amounts of MT1-MMP were observed in immunoprecipitates after TGF-β treatment (Fig. 5a). MDA-MB-435s cells were treated with TGF-β for different periods of time and the membrane fraction proteins were prepared for Western blot analyses to determine whether TGF-β could cleave CD44. After 12 hr of TGF-β treatment, the CD44 level in the cell membrane fraction decreased, a trend which was significant 12 hr later (Fig. 5b). Interestingly, an increased level of soluble CD44 was observed in concentrated medium, 12 hr after TGF-β stimulation (Fig. 5c).
The penetration of cells through the endothelial monolayer has been suggested to be similar to the invasive activities that take place during the metastatic process in vivo. An electric cell-substrate impedance sensing (ECIS™) cell-electrode system was adopted to study the invasive behavior of MDA-MB-435s cells in terms of impedance change. MDA-MB-435s cells with or without TGF-β stimulation were inoculated into the electrode well to challenge the HUVEC cells. Individual electrodes were followed to measure the time-course of any impedance changes from the time of inoculation to 8 hr after inoculation. In Figure 6a, real time impedance changes are shown for several attachment curves, where each curve represents the time-course impedance change measured for 1 electrode. In this experiment, MDA-MB-435s cells pretreated with TGF-β for 12 and 24 hr were able to reduce the intact HUVEC layer resistance. Moreover, there is a substantially greater drop elicited by the TGF-β-treated cells when compared with untreated cells, as well as a quieting of the typical endothelial cell impedance fluctuations. The change in impedance started at 4 hr when the tumor cells first attached to the HUVECs; as they then invaded through the HUVEC monolayer, there was a breakdown in the cell-cell contact of the monolayer. Once the cell-cell contact between HUVEC cells was broken, the resistance dropped fast; the real time invasion happened in this study between ∼ 4 and 8 hr (Fig. 6a). To determine whether TGF-β-mediated tumor cell migration is dependent on MT1-MMP-mediated CD44 cleavage, cells were transfected with MT1-MMP siRNA. Transfection with MT1-MMP siRNA resulted in an inhibition of TGF-β-induced MT1-MMP expression at 12 and 24 hr. The levels of TGF-β induced soluble CD44 at 12 and 24 hr decreased in cells transfected with MT1-MMP siRNA (Fig. 6b), suggesting that MT1-MMP cleavage is involved in enhancing the level of TGF-β-induced soluble CD44. Furthermore, using the transendothelial migration assay, we demonstrate that transendothelial migration of MDA-MB-435s cells was markedly increased upon TGF-β treatment and this effect was blocked in MDA-MB-435s cells transfected with MT1-MMP siRNA (Fig. 6c). These data suggest that TGF-β-induced MT1-MMP production is involved in MDA-MB-435s cell migration through the endothelial cells.
In this study, we investigated whether TGF-β induces CD44 cleavage through the expression of MT1-MMP during the process of migration using a MDA-MB-435s cell line that likely represents a breast epithelial cell line that has undergone lineage infidelity.29 MDA-MB-435s cells treated with TGF-β exhibited a higher level of MT1-MMP in the cell membrane. TGF-β signals through a heteromeric complex of the type I (TβR-I) and type II (TβR-II) transmembrane receptor kinases.34 When it links with a receptor, TβR-I is phosphorylated by the constitutively active TβII kinase; then the signal propagates to the Smad proteins.34 TGF-β can activate the ERK MAPKs and JNK type of the MAPK pathways.35 TGF-β induces aggrecan gene (Agc) expression in chondrogenic cells and this response requires TGF-β-induced activation of the R-Smad2/4 complexes as well as the p38 MAPK and ERK1/2 pathways.36 Our observations suggest that the ERK1/2 and JNK kinase pathways are involved in TGF-β-induced MT1-MMP expression as demonstrated in the inhibitor study. However, no effect on MT1-MMP expression was detected when MDA-MB-435s cells were pretreated with a p38 inhibitor (SB203580). The involvement of multiple signaling pathways has also been shown to be required for TGF-β-mediated metastatic breast tumor progression.37
CD44 is expressed in many types of metastatic tumor cells.19, 20 Soluble CD44 can be shed from cell surfaces through a proteolytic process.38, 39 CD44 cleavage contributes to the regulation of its interaction with hyaluronic acid which is required for the migration process, and consequently promotes CD44-mediated cancer cell migration.40 This proteolytic cleavage of CD44 is involved in tumor invasion and metastasis.38 Mori et al.41 discovered that CD44H appears to link MT1-MMP to the cytoskeleton and regulate its localization. Moreover, MT1-MMP forms a complex with CD44H via a hemopexin-like domain. This interaction is also critical for the shedding of CD44H and the cell migration-promoting activity of MT1-MMP.41 MT1-MMP appears to possess CD44 shedding capabilities and promotes cell migration39 as demonstrated by the colocalization of MT1-MMP and CD44, which is important for cell migration.39 Evidence is accumulating that suggests that MT1-MMP plays a pivotal role in cell migration and invasion.42–44 Indeed, we found that CD44 colocalizes with TGF-β-induced MT1-MMP on the cell membrane which resulted in a marked rise in the soluble form of CD44 after 12 hr of TGF-β stimulation, likely due to CD44 cleavage. We also observed that after TGF-β stimulation, MT1-MMP was associated with CD44v6, however, soluble CD44v6 was not detected in the concentrated medium (data not shown). Because the results presented herein do not allow ruling out that other CD44 isoforms play a role in cell migration and invasion, additional studies are required to further investigate this route.
CD44 cleavage has been demonstrated to play a critical role in CD44-mediated tumor cell migration by providing off-setting changes in adhesive interactions between CD44 and ECM.38, 44 To analyze the relationship between TGF-β stimulation of MT1-MMP expression and CD44 cleavage and tumor cell mobility, we used the ECIS method to observe cell invasion and demonstrated that TGF-β-induced CD44 cleavage through MT1-MMP is related to tumor cell migration.
In addition, MMP-2, MMP-9, MMP-1 and MT1-MMP are known to be involved in invasion, although the MT-MMPs seem to possess a greater capability in promoting cell invasion.45 MT1-MMP is known to degrade the extracellular matrix, activating other MMPs,15 and it acts as a shedding enzyme involving CD44 cleavage and the promotion of cell movement.38, 46, 47 Researchers have found that invasive breast cancer cells have higher levels of CD44 than normal tissue by tissue staining.19, 48 Moreover, TGF-β can enhance the invasiveness of rat mammary tumor cells,49 although the details of this enhancing mechanism are still unclear. Here, we demonstrated the involvement of various different signal-transduction pathways in TGF-β-induced MT1-MMP expression, which then causes CD44 cleavage and thus enhances the invasiveness of the cancer cells. Although additional experiments are required to pinpoint the exact mechanism of MT1-MMP-mediated CD44 cleavage on the cell surface of MDA-MB-435s cells and how this relates to enhanced migration and envasiveness in vivo, this study yields useful information on tumor cell invasion and metastasis.