D4-GDI, a Rho GTPase Regulator, Promotes Breast Cancer Cell Invasiveness

Introduction

D4-GDI, also known as Ly-GDI, belongs to a family of Rho GDP dissociation inhibitors (RhoGDI) that include RhoGDI, D4-GDI, and RhoGDI-3 ( 1, 2). Those proteins are pivotal regulators of Rho GTPase function. They control Rho GTPase cellular distribution and interactions with regulatory guanine nucleotide exchange factors, GTPase-activating proteins, and effector targets ( 3, 4). The Rho family of small GTPases, including Rac1, Cdc42, and RhoC, has been implicated in the regulation of many aspects of cell motility and invasion, including cell polarity, cytoskeletal organization, and transduction of signals from the extracellular environment (see reviews in refs. 5– 8). For instance, Rac1 is a key regulator of Tiam1-induced membrane ruffling and invasion ( 9). Activation of Rac1 and Cdc42 induces integrin-mediated motility and invasiveness of mammary epithelial cells ( 10). When exogenously introduced into cells (by transfection or microinjection), the RhoGDIs induce disruption of Rho-dependent cellular activities including the actin cytoskeleton and cell motility ( 11, 12).

In contrast to RhoGDI, which is expressed ubiquitously, D4-GDI was shown to be expressed only in hematopoietic tissues, predominantly in B and T cells ( 13, 14). Thus, much of our knowledge on D4-GDI function derives from studies on lymphocytes and the immune system, demonstrating a role of D4-GDI in the regulation of lymphocyte activation, survival, and responsiveness ( 15, 16). However, recent studies showed that D4-GDI is also expressed in cells of nonhematopoietic neoplasms, including ovarian ( 17) and bladder cancer cells ( 18), suggesting that D4-GDI may play a role in progression of human tumors.

In studies on the role of Rho GTPases in the regulation of apoptosis ( 19– 21), we discovered that D4-GDI is expressed in human breast cancer cells. Among the different human breast cancer cell lines that were examined, cell lines with high invasive activities, such as MDA-MB-231, expressed higher D4-GDI than did weakly invasive and noninvasive cell lines. In addition, we found that targeted disruption of D4-GDI by stably expressing small interfering RNA (siRNA) against D4-GDI transcripts effectively blocks the motility and invasive potential of MDA-MB-231 cells in vitro. Interestingly, cells lacking D4-GDI revert to a normal breast epithelial phenotype characterized by the formation of cysts with a central lumen when grown on Matrigel. Further analysis revealed that D4-GDI depletion is associated with a down-regulation of the cell surface protein β1-integrin and a loss of cell-matrix adhesion, suggesting that D4-GDI induces cell invasion by controlling β1-integrin expression. Restoration of D4-GDI expression to the levels found in the parental MDA-MB-231 cells restores the expression of β1-integrin and invasive activities. These results suggest that D4-GDI may play a critical role in the regulation of breast cancer cell invasiveness.

Materials and Methods

Cell lines and antibodies. Human breast cancer cell lines were obtained from the American Type Culture Collection (Manassas, VA). MDA-MB-231 cells were grown in DMEM/F-12 (1:1 mix) medium (Mediatech, Herndon, VA) supplemented with 5% fetal bovine serum (FBS), 4 mmol/L glutamine, 50 μmol/L β-mercaptoethanol, and 1 mmol/L sodium pyruvate at 37°C and 5% CO₂ in air. Other cell lines were cultured according to the instructions of the manufacturer. Antibodies specific to human D4-GDI, RhoGDI, β1-integrin, Rac1, Cdc42, and phospho-Akt (Ser^472/473) were purchased from BD Transduction Laboratories (Lexington, KY). Antibodies against human RhoA and RhoC were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). A rabbit polyclonal antibody against phospho-PAK1 (Thr⁴²³) and phospho-PAK2 (Thr⁴⁰²) was from Cell Signaling Technology (Beverly, MA). Anti-β1-integrin and an IgG1 negative control, both conjugated to FITC, were from Serotec, Inc. (Raleigh, NC). FITC-labeled anti-ZO-1 was from Invitrogen (Carlsbad, CA).

Plasmid constructs and transfection. The pRNA-U6.1/hygromycin vector expressing siRNA against D4-GDI was custom constructed by GeneScript Corp. (Scotch Plains, NJ). The vector contains a DNA template for the synthesis of siRNA under the control of the U6 promoter. The 19-nucleotide siRNA sequences were selected by the siDesign program of Dharmacon, ¹ and correspond to nucleotides 305 to 324 (designated as siD4-GDI-I) and 373 to 391 (siD4-GDI-II) of human D4-GDI, respectively. The construction of the plasmids was as described previously ( 21). A plasmid that expresses siRNA against firefly luciferase was used as a control (pRNA-U6.1/siLuc). For studies to rescue D4-GDI gene expression, an expression vector for D4-GDI (Guthrie Research Institute, Sayre, PA) was mutated in the encoding region (³¹⁰GGTTCTGAA to ³¹⁰GGCTCTGAG) to prevent destruction of the expressed mRNA by the siRNA already being expressed in the cells. The amino acid sequence of the expressed protein was unchanged. Mutations were introduced using the Quick Change site-directed mutagenesis kit (Stratagene, La Jolla, CA) according to the protocol of the manufacturer. The resulting D4-GDI mutant was excised and subcloned into a pEGFP-C3/kanamycin vector (kindly provided by Dr. Mark R. Philips, New York University School of Medicine, New York, NY), fusing the enhanced green fluorescent protein (EGFP) to the NH₂ terminus of the D4-GDI construct. All constructs were verified by DNA sequencing.

For stable transfection of siRNA expression vectors, MDA-MB-231 cells at 50-60% confluency were grown in six-well plates, and either 1.6 μg pRNA-U6.1/siD4GDI or a control plasmid (pRNA-U6.1/siLuciferase) was introduced using LipofectAMINE (Invitrogen) according to the instructions of the manufacturer. After 72 hours, 0.45 mg/mL hygromycin (Clontech) was added to the cultures to select for hygromicin-resistant clones. Two weeks later, independent colonies were picked using cloning cylinders (Labcor Products, Inc., Frederick, MD), subcultured and tested for expression of D4-GDI by immunoblot analysis with antibodies against human D4-GDI as described below. Selected stable clones with decreased levels of D4-GDI (designated as MDA-MB-231−/−D4-GDI) were maintained in complete culture medium containing hygromycin (0.45 mg/mL). In all experiments, different stable transfectants were used to avoid potential artifacts associated with the selection and propagation of individual clones from single transfected cells. Similarly, BT549 cells were transfected with the siRNA plasmids and the stably expressing pools were used for further analysis. For rescue transfections, cells (∼90% confluent) were grown onto 25 cm² flasks and transfected with 8.0 μg pEGFP-C3/D4-GDI-re. After 48 hours, the EGFP-positive cells (∼20%) were isolated by fluorescence-activated cell sorting (FACS) and maintained in complete medium supplemented with 1.5 mg/mL kanamycin.

Matrigel culture. Cells (2 × 10⁴) were plated onto a thick layer (1 mm) of Matrigel in eight-well chamber slides (Nunc, Naperville, IL). Solidified Matrigel was covered with complete growth medium and incubated at 37°C and 10% CO₂ in air. At the indicated times, cells were processed for analysis of immunostaining or lumen formation. Cell viability was quantified using a 3-(4,5-dimethylthiazol-2yl)-2,5-diphenyltetrazolium bromide (MTT) colorimetric assay ( 21).

Immunofluorescence and microscopy. Immunofluorescence staining of cells cultured on glass chamber slides or Matrigel was done as described ( 22, 23). All steps were done at room temperature. Cells were fixed in 3% paraformaldehyde in PBS (pH 7.4) for 30 minutes, and quenched with 100 mmol/L glycine for 20 minutes. Cells were then washed and permeabilized with 0.2% Triton X-100 in PBS for 20 minutes. After incubation in a blocking buffer containing 10% goat serum and 1% bovine serum albumin (BSA) for 20 minutes, cells were incubated with FITC-conjugated anti-ZO-1 (1:100 dilution) for 2 hours and washed thrice with 1% BSA in PBS. Nuclei were counterstained with propidium iodide (20 μg/mL) for 10 minutes. Finally, the slides were detached from the medium chamber and mounted with antifade reagent (Invitrogen). For the lumen formation study, cells were stained with CellTracker Red for 30 minutes and washed. Confocal images were taken with a Zeiss LSM 5 PASCAL confocal laser scanning microscope.

Cell motility and invasion. Motility and invasion across the 8.0 micron pore transwells coated with extracellular matrix proteins were assayed using BioCoat Matrigel Invasion Chambers (BD Bioscience, Bedford, MA). Motility was assessed in the absence of serum. Invasion across a 1 mm Matrigel layer was assessed on serum-starved cells after adding 20% FBS (chemoattractant) to the lower chamber of the transwells. Control chambers contained the same pore size membrane, but without the Matrigel coating. The chambers were incubated for 20 hours at 37°C in a 5% CO₂ atmosphere. Cells remaining on the upper surface of the membranes were removed by scraping with a cotton swab, and those on the lower surface of the membranes were counted after staining with toluidine blue (Fluka, Buchs, Switzerland). The percent invasion is expressed as the ratio of the mean cell number in the invasion chamber to the mean cell number in control chambers. Statistical analyzes were done using an unpaired, two-tailed t test.

Cell-matrix attachment. The relative attachment of cells to immobilized laminin I and basement membrane complex (BMC) was measured using InnoCyte ECM Cell Adhesion kits (Calbiochem) following the protocol of the vendor. Briefly, 100 μL cell suspension (10⁴ cells per well) was incubated in wells (96-well plate) coated with laminin I or BMC for 1 hour at 37°C. After washing, 100 μL Calcein-AM working solution was added to each well and incubated for another hour. The fluorescence intensities were measured at an excitation of 485 nm and an emission of 520 nm in a Molecular Device plate reader.

Reverse transcription-PCR. Total RNA was isolated from different cells using Trizol reagent (Invitrogen). SuperScript One-Step reverse transcription-PCR (RT-PCR) kit (Invitrogen) was used to quantify mRNA levels as per instructions of the manufacturer. Primers used for human D4-GDI were as follows: 5′-GGCACCCCGGACAGAGACGTG and 5′-TTCTTCCAGGTGGCAAGGGTG. β-actin was used as an internal control. The RT-PCR products were analyzed in 1% agarose gels.

Western blot analysis. Cells (1 × 10⁶) were lysed in SDS lysis buffer containing 50 mmol/L Tris-HCl (pH 7.0), 2% SDS, and 10% glycerol, and incubated for 20 minutes at 95°C. Protein concentrations were estimated using the BCA protein assay (Pierce, Rockford, IL). Equal amounts of cell lysates (20 μg per lane) were resolved by electrophoresis using a 4% to 12% NuPAGE Bis-Tris gel (Invitrogen) and transferred to nitrocellulose membranes (Millipore, Bedford, MA) for immunoblot analysis with an appropriate dilution of antibodies (1:1,000 to 1:2,000). When necessary, the membranes were stripped by Restore Western Blot Stripping Buffer (Pierce) and reprobed with appropriate antibodies. Immunocomplexes were visualized by chemiluminescence using ECL (Santa Cruz Biotechnology) or SuperSignal reagent (Pierce).

Results

D4-GDI is expressed in human breast cancer cells. The expression of D4-GDI in human breast cancer cell lines was tested by Western blot immunoassay ( Fig. 1A ). HeLa cells, which are known not to express D4-GDI, were used as a negative control. As expected, D4-GDI was poorly expressed or undetectable in human mammary fibroblast or benign-derived epithelial cells (578Bst and MCF-12A, respectively). However, it was strongly expressed in several human breast carcinoma cell lines (MDA-MB-231, MDA-MB-468, T47D, BT549, SKBR3, and BT474). In addition to the Western blot results, D4-GDI mRNA was detected in these cells by RT-PCR ( Fig. 1B), thus demonstrating that D4-GDI is expressed in human breast cancer cells.

D4-GDI protein expression seems to correlate with the typical features of a malignant phenotype in these cells ( 24, 25), including increased invasion and motility, except for the MCF-7 cells. MDA-MB-231, a highly invasive breast cancer cell line, showed the most abundant D4-GDI expression. RhoGDI was expressed at essentially the same level in all tested cell lines. These observations suggest that expression of D4-GDI may be associated with the progression to a malignant phenotype in breast cancer cells.

Generation of cell lines stably expressing D4-GDI siRNA vector. To explore the function of D4-GDI in the development of breast cancer cell invasiveness, we suppressed D4-GDI expression by vector-based stable transfection of specific small-interfering RNA (siRNA) to D4-GDI in MDA-MB-231 cells. This cell line has been widely used for studying the molecular basis for human breast cancer invasion and metastasis ( 24). Two weeks after transfection, cells were selected with hygromycin and screened for the expression of D4-GDI. Two target sequences were examined, corresponding to nucleotides 305 to 324 (siD4-GDI-I) and 373 to 391 (siD4-GDI-II), respectively. As expected, pRNA-U6.1/siLuc had no effect on D4-GDI expression (data not shown); however, pRNA-U6.1/siD4-GDI-I greatly diminished its expression. In ∼20% hygromycin-resistant clones, D4-GDI expression was reduced by at least 90% ( Fig. 2A ). The siRNA-mediated suppression of D4-GDI expression was stable over long periods of time ( Fig. 2B). RT-PCR confirmed the disruption of D4-GDI mRNA in the transfected clones (designated −/−D4-GDI; Fig. 2C). Clones that showed no detectable D4-GDI expression were used for further studies. In the D4-GDI knockdown cells, RhoGDI (70% homology to D4-GDI) expression was not altered, indicating the specificity of the siRNA silencing for D4-GDI. In contrast to pRNA-U6.1/siD4-GDI-I, transfection of pRNA-U6.1/siD4-GDI-II was unable to suppress D4-GDI expression in MDA-MB-231 cells (data not shown). Interestingly, both siD4-GDI sequences were effective in silencing D4-GDI in BT549 cells (see below). These results are in agreement with our previous observation that DNA vector–based siRNA effect is cell type dependent ( 21).

Knockdown of D4-GDI blocks breast cancer cell invasive activities. D4-GDI knockdown cells displayed typical epithelial cell morphology similar to that of wild-type 231 cells in monolayer culture. The proliferation rate as determined by MTT assays was slightly slower than that of wild-type cells or cells expressing siLuc (data not shown). When cultured on Matrigel, however, the diminished D4-GDI expression was associated with strikingly altered cell morphology. As expected, the benign-derived mammary epithelial MCF-12A cells developed into individual, acinus-like cysts in three-dimensional culture ( Fig. 3A ; ref. 22). Confocal microscopy revealed lumen formation within each cyst ( Fig. 3D). By contrast, MDA-MB-231 cells formed large, unpolarized, and branching networks without lumen formation in three-dimensional culture, typical of an invasive morphology. Strikingly, D4-GDI knockdown cells lost the invasive phenotype and formed cavitary structures similar to that seen in MCF-12A cells ( Fig. 3A and D). In addition, the viability of MCF-12A and −/−D4-GDI cells differed from MDA-MB-231 cells ( Fig. 3B), peaking around day 10 and decreasing thereafter ( Fig. 3C). This growth pattern is consistent with the view that the lumen formation results from programmed death of cells in the center of the cavitary structures ( 26).

Knockdown of D4-GDI did not affect the tight junction structures in MDA-MB-231 cells. We found that the tight junction is poorly formed in MDA-MB-231 cells as well as MCF-12A cells in monolayer culture. Immunostaining of ZO-1, a tight junction marker, revealed fragmented or diffuse cytoplasmic staining pattern at apical junctions (see Supplementary Data I). A similar ZO-1 staining pattern was seen in D4-GDI knockdown cells. This is in contrast to the well-defined tight junction structure seen in MDCKII or MCF7 cells, in which tight junction maker proteins display smooth, continuous staining ( 23). In addition, staining of ZO-1 in the cavitary structures formed by −/−D4-GDI cells on Matrigel showed no apparent apicobasal polarity.

MDA-MB-231−/−D4-GDI cells lost their ability to migrate through a barrier of Matrigel in the absence of serum ( Fig. 4A ), as well as the ability to invade through Matrigel toward a serum gradient ( Fig. 4B). In contrast, siLuc had no effect on both activities. Similar observation was made with BT549 cells ( Fig. 4C). Together, these results show that D4-GDI is a novel modulator of breast cancer cell invasiveness.

D4-GDI controls β1-integrin expression together with invasion. To investigate how D4-GDI controls cell invasion, we examined the effect of D4-GDI knockdown on protein expression by immunoblot analysis. We focused on the Rho proteins and their effectors, matrix metalloproteases (MMP), and adhesion proteins. The results revealed no apparent change in the expression levels of Rho GTPases (see Supplementary Data II). The activation status of Akt and p21-activated protein kinases (PAK1/PAK2) was also not altered, as shown by immunoblot analysis using antibodies against the phosphorylated/activated species of these enzymes. In addition, zymography analysis showed that D4-GDI knockdown had no effect on the secretion of active MMP-2, although it slightly increased the MMP-9 activity in serum-free conditional medium (see Supplementary Data III).

Consistent with a previous observation ( 10), the motility of MDA-MB-231 cells was mediated by integrin; it was inhibited by an antibody against β1-integrin ( Fig. 5A ). The protein expression levels of β1-integrin, as shown by immunoblot analysis, were significantly diminished in MDA-MB-231−/−D4-GDI cells as well as BT549−/−D4-GDI cells when compared with the parental cells ( Fig. 5B). This was confirmed by FACS analysis using FITC-conjugated anti-β1-integrin ( Fig. 5C). As a result, MDA-MB-231 cells lacking D4-GDI were no longer able to attach to BMC or laminin ( Fig. 5D), a ligand of β1-integrin that mediates cell-matrix adhesion. The results also showed that the laminin alone was sufficient to support cell attachment, supporting the notion that D4-GDI-induced motility and invasion is mediated by β1-integrin.

siRNA inhibition of cell invasiveness and β1-integrin expression is D4-GDI specific. We confirmed that the effects of D4-GDI siRNA can be reversed by the simultaneous expression of a D4-GDI rescue cDNA encoding GFP-tagged D4-GDI (and called GFP-D4-GDI-re). This was accomplished by introducing an exogenous D4-GDI expression vector to the siRNA-expressing cells. The rescue expression plasmid was mutated in the sequences of D4-GDI gene complementary to the siRNA to prevent destruction of the exogenous mRNA by the existing siRNA, but without changing the amino acid sequence of the final protein product ( Fig. 6A ). The level of expression of GFP-D4-GDI-re was comparable with that of the endogenous D4-GDI in parental cells ( Fig. 6B). Importantly, the ectopic expression of D4-GDI-re restored β1-integrin expression ( Fig. 6B) and the invasive activities ( Fig. 6C), demonstrating not only the specificity of the RNA interference inhibition but also the activity of D4-GDI in modulating β1-integrin and breast cancer cell invasion.

Discussion

The acquisition of a motile and invasive phenotype is a prerequisite for the metastatic spread of tumor cells, which constitutes a major cause of poor prognosis for cancer patients. However, the underlying mechanisms controlling cell invasiveness remain to be fully understood. In the present studies, we show that D4-GDI promotes breast cancer cell invasion. We show that D4-GDI is strongly expressed in a variety of human breast cancer cell lines. MDA-MB-231 cells as well as BT549 cells with forced D4-GDI depletion are no longer able to invade through a barrier of extracellular matrix in vitro. The depletion of D4-GDI switches the response of MDA-MB-231 cells to extracellular matrix from a motile, invasive phenotype to a growth-arrested, a more organized structure reminiscent of mammary gland acini; however, it does not seem to induce structural polarization. Importantly, when D4-GDI protein levels are restored, invasiveness is also restored, thus demonstrating that the observed effect on motility and invasion is a result of D4-GDI-specific activity. In addition, we provide evidence that D4-GDI-induced invasion may be mediated by β1-integrin.

Accumulating evidence now shows that D4-GDI is expressed not only in hematopoietic tissues also in nonhematopoietic neoplasms. Results of cDNA microarray analyzes revealed that D4-GDI is up-regulated in ovarian ( 17) and is down-regulated in bladder carcinomas ( 27). However, in contrast to our finding, Theodorescu et al. ( 18) found that D4-GDI (also called RhoGDI2) protein expression in bladder tumors is reduced as a function of bladder tumor progression. This has led to the hypothesis that D4-GDI is a metastasis suppressor gene in models of bladder cancer. In sharp contrast, our results show that D4-GDI expression promotes cell invasiveness in breast cancer cells. These results suggest that D4-GDI may have an opposite role in the progression of different tumor types. Differences in experimental approach may also explain the discrepant findings. In most human breast cancer cells, there is a direct correlation between their in vivo invasive phenotype and in vitro invasion activities ( 24, 25). Thus, it is reasonable to propose that D4-GDI may be involved more generally in the invasive phenotype of human breast cancer.

The mechanism of action of D4-GDI in promoting cell motility and invasion remains to be fully established. One possibility, shown here, is that D4-GDI modulates cell invasion through its effect on the expression of β1-integrin, a cell surface receptor that mediates cell-matrix adhesion through the laminin component of basement membrane. Weaver et al. ( 28) showed that inhibition of β1-integrin activities in T4-2 human breast cancer cells by an anti-β1-integrin antibody restored the formation of polarized acinus-like structures in three-dimensional culture, resulting a reversion similar to the normal phenotype of the S-1 cells. Blockage of β1-integrin was also shown to inhibit Cdc42- or Rac1-induced motility of T47D cells ( 10). Consistent with these reports, we found that the motility of MDA-MB-231 cells was effectively inhibited by an antibody against β1-integrin ( Fig. 5A). Furthermore, we showed that knockdown of D4-GDI in MDA-MB-231 cells was associated with a loss of both β1-integrin protein expression and cell attachment to immobilized laminin ( Fig. 5B-D). Restoration of D4-GDI expression restored both activities ( Fig. 6). Numerous studies have shown that Rho GTPases are key components of integrin-regulated signaling events implicated in breast cancer progression ( 29). Integrins control the activation of Rho proteins by regulating the translocation of activated Rac1-GTP and Cdc42-GTP to the plasma membrane ( 30, 31). In turn, intracellular Rac1 activity induces the assembly of extracellular laminin, which is essential for epithelial apical polarity ( 32). The observed regulation of β1-integrin expression by D4-GDI suggests that alteration of Rho GTPase activities may provide a feedback signal to the upstream signaling components in the pathway.

Like other RhoGDIs, D4-GDI was postulated to bind and inhibit Rho GTPases. However, the specificity of D4-GDI remains much less characterized ( 1, 33, 34). Although recombinant D4-GDI binds to purified Rac1, Cdc42, and RhoA, there is no evidence showing that they can form stable complexes in vivo. We attempted to identify the GTPase(s) complexed with D4-GDI in breast cancer cells, but failed to copurify any of the above mentioned Rho proteins from MDA-MB-231 cell lysates in a controlled coimmunoprecipitation assay using anti-D4-GDI antibodies (data not shown). Consistent with this, the activation status of Rac1, Cdc42, and RhoA was not altered as a result of D4-GDI depletion (data not shown). It is therefore likely that D4-GDI regulates the breast cancer cell motility and invasion through binding to an alternative GTPase.

In summary, we have shown that D4-GDI is a promoter of breast cancer cell invasive activities. If confirmed in primary breast cancers, D4-GDI could serve both as a novel prognostic biomarker for breast cancer and as a new potential target for treatment of metastatic breast cancer.