hMena (ENAH), an actin regulatory protein involved in the control of cell motility and adhesion, is modulated during human breast carcinogenesis. In fact, whereas undetectable in normal mammary epithelium, hMena becomes overexpressed in high-risk benign lesions and primary and metastatic tumors. In vivo, hMena overexpression correlates with the HER-2+/ER−/Ki67+ unfavorable prognostic phenotype. In vitro, neuregulin-1 up-regulates whereas Herceptin treatment down-modulates hMena expression, suggesting that it may couple tyrosine kinase receptor signaling to the actin cytoskeleton. Herein, we report the cloning of hMena and of a splice variant, hMena+11a, which contains an additional exon corresponding to 21 amino acids located in the EVH2 domain, from a breast carcinoma cell line of epithelial phenotype. Whereas hMena overexpression consistently characterizes the transformed phenotype of tumor cells of different lineages, hMena+11a isoform is concomitantly present only in epithelial tumor cell lines. In breast cancer cell lines, epidermal growth factor (EGF) treatment promotes concomitant up-regulation of hMena and hMena+11a, resulting in an increase of the fraction of phosphorylated hMena+11a isoform only. hMena+11a overexpression and phosphorylation leads to increased p42/44 mitogen-activated protein kinase (MAPK) activation and cell proliferation as evidenced in hMena+11a–transfected breast cancer cell lines. On the contrary, hMena knockdown induces reduction of p42/44 MAPK phosphorylation and of the proliferative response to EGF. The present data provide new insight into the relevance of actin cytoskeleton regulatory proteins and, in particular, of hMena isoforms in coupling multiple signaling pathways involved in breast cancer. [Cancer Res 2007;67(6):2657–65]
- Actin cytoskeleton
- Splice variant
- Breast cancer
Increasing evidence from the analysis of animal models and human tumors highlights the importance of actin cytoskeleton regulatory proteins in early ( 1) as well as advanced ( 2) stages of breast cancer. In this context, we have recently identified by SEREX (serologic analysis of cDNA expression libraries) technology the hMena protein ( 3), the human orthologue of murine Mena, which is overexpressed in benign breast lesions with high risk of transformation and in >70% of primary breast cancers ( 4).
Mena belongs to the Ena/VASP protein family, which includes key regulatory molecules controlling cell shape ( 5, 6) and movement ( 7) by protecting actin filaments from capping proteins at their barbed ends ( 8). Ena/VASP proteins are constituents of the adherens junctions, which are necessary to seal membranes in the epithelial sheet and control actin organization on cadherin adhesion contact ( 9), frequently perturbed processes following transformation ( 10).
Ena/VASP proteins contain specific domains including the NH2-terminal EVH1 domain, which plays a role in intracellular protein localization ( 11). The central proline-rich domain mediates the interaction with proteins containing the SH3 and WW domains and with the small actin monomer binding protein profilin ( 12). The EVH2 COOH-terminal domain is responsible for tetramerization and for the binding to G- and F-actin ( 13, 14); its interaction with the growing ends of the actin filaments is required for targeting the Ena/VASP to lamellipodia and filopodia ( 14). Alternative splicing is frequently responsible for the generation of protein variants associated with the transformation process ( 15). Mena is alternately spliced to give rise to multiple isoforms: a neuronal-specific Mena-140 ( 12) and Mena-S, an isoform ( 16) found in mouse spleen. Thus, tissue-specific alternative splicing is responsible for the generation of Mena isoforms, possessing distinct functions.
Cyclic nucleotide-dependent kinases protein kinase A and protein kinase G regulate Mena function. Murine Mena phosphorylation occurs at two residues, Ser236 and Ser376, playing roles in different processes such as cell motility ( 13), filopodia protrusion in growth cones ( 17), and anticapping activity ( 8).
Ena/VASP regulates actin filament dynamics following the activation of a variety of cell membrane receptors including netrin-1 receptor ( 17), Sema6D ( 18), FcγR ( 19), T-cell receptor ( 20), and c-Met receptor ( 21); thus, they are capable of integrating extracellular signals into appropriate changes in cytoskeleton architecture that are instrumental for cell activity.
Activation of epidermal growth factor (EGF) receptor (EGFR) and other members of the EGFR family plays a central role in breast carcinogenesis and tumor progression ( 22). In a subpopulation of invasive breast tumor cells collected with an in vivo invasion assay in response to EGF, Mena is overespressed among a set of genes coding for the minimum motility machine regulating β-actin polymerization ( 2). Of interest, in invasive primary breast tumors, hMena overexpression correlates with tumor size, proliferation index, and HER-2 overexpression, which are well-established markers of tumor aggressiveness. Furthermore, hMena expression, whereas up-regulated by neuregulin-1, is down-regulated by Herceptin treatment ( 4) in breast cancer cell lines, thus suggesting that hMena couples tyrosine kinase receptor signaling to the actin cytoskeleton.
To acquire further information about the hMena structure and modulation, in the present study, we cloned and characterized hMena from a breast cancer cell line of epithelial phenotype. Data presented show that, differently from the murine counterpart, hMena possesses a splice variant termed hMena+11a, which is characterized by 21 extra amino acids. Overexpression of hMena and hMena+11a occurs in transformed cells of various cell lineages, whereas the hMena+11a isoform is epithelial specific and is phosphorylated after mitogenic stimuli, such as EGF. Of interest, this up-regulation and phosphorylation is accompanied by p42/44 mitogen-activated protein kinase (MAPK) activation and an increased proliferation rate in breast cancer cell lines.
Materials and Methods
Cell lines. The following cell lines were purchased from the American Type Culture Collection (Rockville, MD): MDAMB361, T47D, SKBr3, MCF7, BT474 (breast cancer), A427, Calu3 (lung cancer), SiHa, CaSki (cervical carcinoma), T98G, U87MG, and U373 (glioma). Other cell lines used were glioblastoma U251 ( 23); SBT and DAL, developed in our laboratory from the ascitic fluid of two breast cancer patients ( 24); normal human keratinocytes (NHK) and ADF astrocytoma, kindly provided by Drs. A. Venuti and G. Zupi (Regina Elena Cancer Institute, Rome, Italy), respectively; melanoma ME10538, provided by Dr. A. Anichini (Human Tumor Immunobiology Unit, Istituto Nazional per lo Studio e la Cura dei Tumori, Milan, Italy), and MAS melanoma, developed in our laboratory from a primary melanoma lesion.
MCF7/hMena+11a and MCF7/pcDNA3 stable transfectants were obtained by selecting MCF7 cells transfected with hMena+11a-pcDNA3.1 and empty vector, respectively, using 500 μg/mL G418 (Invitrogen, Paisley, United Kingdom) in complete culture medium.
Cloning and sequences of hMena and hMena+11a. Two micrograms of total RNA extracted from the SBT cell line using Trizol reagent (Life Technologies, Inc., Rockville, MD) were used to obtain the relative cDNA by first strand cDNA synthesis kit (Amersham Pharmacia Biotech, Little Chalfont, United Kingdom). cDNA was amplified using Platinum Pfx DNA polymerase (Invitrogen) in PCR reactions consisting of 30 cycles at a denaturation temperature of 94°C (30 s/cycle), an annealing temperature of 55°C (1 min/cycle), and an extension temperature of 68°C (3 min/cycle). The primers used (Invitrogen) contained the putative ATG and stop codon of hMena, P1-ATG (5′-CACCATGAGTGAACAGAGTATC-3′), and P8-stop (5′-CTGTTCCTCTATGCAGTATTTGAC-3′). PCR products were analyzed on a 1% agarose gel, excised from the gel, and purified using a gel extraction kit (Qiagen, Crawley, United Kingdom). Samples were incubated with 1 unit of AmpliTaq polymerase and 1 μL of 10 mmol/L dATP (both from Applied Biosystems, Branchburg, NJ) to add 3′ adenines and then cloned in pCR4-TOPO plasmid (Invitrogen). Plasmid DNA was isolated by Wizard Plus minipreps DNA purification system (Promega Corp., Madison, WI) and sequenced initially with T7 and T3 primers. Once the initial sequences were obtained, additional primers were synthesized to sequence into the insert (P4-forward, 5′-GAGCGACTGGAACAAGAACAGCTG-3′; P5-forward, 5′-GAGAGCGCAGAATATCAAGTGCTG-3′; P6-reverse, 5′-GGCGATTGTCTTCTGACATGG-3′; P7-forward, 5′-GAATTGCTGAAAAGGGATC-3′; P7-reverse, 5′-GATCCCTTTTCAGCAATTC-3′). DNA sequencing was done by the Nucleic Acid Facility Service (Istituto Dermopatico dell'Immacolata, Rome, Italy) with the use of an ABI PRISM 377-96 automated sequencer (Applied Biosystem). Using EcoR1 restriction sites, the right sequences were then subcloned into pcDNA3.1 vector (Invitrogen).
Reverse transcription-PCR. hMena splice variants were detected by reverse transcription-PCR (RT-PCR) using MTC1f (5′-GCTGGAATGGGAGAGAGAGCGCAGAATATC-3′) and MTC4r (GTCAAGTCCTTCCGTCTGGACTCCATTGGC-3′) primers. PCR reactions consisted of 30 cycles at a denaturation temperature of 94°C (30 s/cycle), an annealing, and an extension temperature of 68°C (2 min/cycle). PCR products were analyzed on a 1% agarose gel electrophoresis and stained with ethidium bromide.
In vitro transcription-coupled translation. The in vitro translation of the hMena and hMena+11a cDNA (inserted into the pcDNA3.1 vector) was examined in an in vitro transcription-coupled translation system following the manufacturer's instruction (TNT, rabbit reticulocyte lysate system, Promega).
Cell treatments. Cells were grown in six-well plates to 50% confluence in RPMI supplemented with 10% fetal bovine serum. The medium was replaced with fresh medium containing 0.5% serum for 18 h and the cells were treated with different amounts of rhEGF, ranging from 10 to 75 ng/mL (Promega) for different periods. AG1478 (500 nmol/L; Calbiochem, La Jolla, CA) was added 1 h before EGF treatment.
Western blot analysis. Cells were lysed as reported ( 4). Normal tissue extracts (Protein Medley) were purchased from Clontech Laboratories (Palo Alto, CA). Lysates (30 or 50 μg) were resolved on 10% polyacrylamide gel and transferred to nitrocellulose membrane (Amersham Bioscience) as described ( 4). Blots were probed with the following antibodies: 10 μg/mL anti-hMena rabbit CKLK1 antibody ( 3), anti–phospho-p44/42 MAPK (Thr202/Tyr204) mouse monoclonal antibody (Cell Signaling, Technology, Beverly, MA), and p44/42 MAPK rabbit antibody (Cell Signaling Technology), in 3% skimmed milk/TBST overnight at 4°C. For actin signal, blots were reprobed with 1 μg/mL monoclonal anti-actin, mouse ascites Fluid clone AC-40 (Sigma-Aldrich, Poole, United Kingdom).
Western blot analysis was also done on 5 μL of in vitro translated hMena and hMena+11a.
Immunofluorescence analysis. T47D cells, grown on glass coverslips to semiconfluence, were fixed in 2% paraformaldehyde (EMS, Fort Washington, PA) in PBS for 6 min and permeabilized with 0.01% Triton X-100 in PBS. Following incubation with 2% (w/v) bovine serum albumin (Sigma, St. Louis, MO), cells were incubated for 1 h with mouse anti-Mena (BD Biosciences, San Jose, CA). Following washing, the coverslips were incubated for 1 h in antimouse immunoglobulin labeled with Alexa 594 (Life Technologies/Invitrogen), washed with PBS, and mounted with H-1200 Vectashield mounting medium containing 4′,6-diamidino-2-phenylindole (DAPI; Vector Laboratories, Inc., Burlingame, CA). The images are representative of two independent experiments done in duplicate.
Two-dimensional electrophoresis. Cells were washed, lyophilized, and proteins solubilized with two-dimensional electrophoresis buffer [9 mol/L urea, 10 mmol/L Tris, 4% CHAPS, 65 mmol/L DTT, 2% IPG buffer ampholine (pH 3–10), protease inhibitor cocktail]. Protein samples (250 μg) were applied to 7-cm IPG strips pH 3–10 nonlinear (Amersham Biosciences) and isoelectrofocusing was done with an IPGphor system (Amersham Biosciences) following a standard protocol as described ( 25). Strips were equilibrated in 50 mmol/L Tris-HCl buffer (pH 8.8) containing 6 mol/L urea, 30% glycerol, 2% SDS, and 2% DTT, followed by an incubation in the same buffer replacing DTT with 2.5% iodoacetamide. The strips were loaded on top of 10% acrylamide SDS-PAGE gels for the second dimension separation. Proteins were electrontransferred onto nitrocellulose membranes and Western blot was done as described above. Images were acquired at high resolution and two-dimensional immunoreactivity patterns analyzed using Progenesis PG240 v2005 software (Nonlinear Dynamics, Newcastle, United Kingdom). Relative molecular mass (Mr) was estimated by comparison with Mr reference markers (Precision, Bio-Rad, Hercules, CA) and isoelectric point (pI) values assigned to detected spots by calibration as described in the Amersham Biosciences guidelines.
Phosphatase treatment. Lambda Protein Phosphatase (λ-PPase; New England BioLabs, Ipswich, MA) treatment was done as described ( 26) with modifications. Pelleted cells (25 × 106) were lyophilized and resuspended in lysis buffer [1% w/v NP40, 1% w/v SDS, 50 mmol/L Tris (pH 7.6), 150 mmol/L NaCl, protease inhibitor cocktail]. Sixty microliters of lysate, corresponding to 600 μg of protein, were brought to a final volume of 600 μL with deionized water, followed by the addition of 20 μL of 20 mmol/L MnCl2 solution and 20 μL of λ-PPase buffer. The mixture was divided into two aliquots and 300 units of λ-PPase were added to one of the aliquots. After mixing, aliquots were incubated for 6 h at 30°C. Proteins were acetone precipitated at −20°C and used for two-dimensional electrophoresis analysis.
Transfections. MCF7 and SKBr3 cells in exponential growth phase were plated in six-well plates at a density of 3 × 105 per well or, for proliferation assay, in 96-well plates at 1 × 104 per well. After 24 h, cells were transfected with 1.5 μg/mL hMena+11a cDNA or with vector alone (pcDNA3). Transfection using LipofectAMINE 2000 reagent (Invitrogen) was done according to the protocol of the manufacturer. After culturing for 24 h, cells were serum deprived for 18 h and then treated with the growth factor for additional 24 h for Western blot analysis or [3H]thymidine incorporation assays.
Small interfering RNA treatment. BT474 cells in exponential growth phase were plated in 48-well plates at a density of 5 × 104 per well. After 24 h, cells were transfected with 100 nmol/L hMena-specific pooled small interfering RNA (siRNA) duplexes (siENA SMART pool) or control nonspecific siRNA (Dharmacon, Lafayette, CO) using LipofectAMINE 2000 reagent (Invitrogen). After culturing for 24 h, cells were serum deprived for 18 h and then treated with 75 ng/mL EGF for additional 24 h for Western blot analysis or [3H]thymidine incorporation assays.
[3H]Thymidine incorporation assay. [3H]Thymidine (Perkin-Elmer Life and Analytical Sciences, Boston, MA) was added at 5 μCi/mL for 4 h on the last day of treatment. After the medium was removed, cells were washed twice with cold PBS, treated with 10% trichloroacetic acid for 30 min at 37°C, solubilized with 0.4 N NaOH, and counted for incorporation of 3H on beta liquid scintillation counter in 5 mL of scintillation fluid. Each experiment was done in sextuplicate and results were expressed as the means of at least three separate experiments.
Statistical analysis. All experiments were repeated at least thrice. Data collected from [3H]thymidine incorporation assay were expressed as mean ±SD. The data presented in some figures are from a representative experiment, which was qualitatively similar with the replicate experiments. Statistical significance was determined by Student's t test (two tailed) comparison between two groups of data. Asterisks indicate significant differences of experimental groups compared with the corresponding control condition (P < 0.05; see figure legends). Statistical analysis was done using GraphPad Prism 4, V4.03 software (GraphPad, Inc., San Diego, CA). Change in the phosphorylation status was evaluated, using Progenesis v.2004 software (Nonlinear Dynamics), by absorbance indicated as normalized spot volume. Normalization was done by multiplying the total spot volume by the constant factor 100, which produces spot percentage volume.
Molecular cloning and characterization of hMena coding sequences. To obtain the full-length cDNA of the human orthologue of mouse Mena (hMena), we analyzed the similarities of our incomplete sequence isolated by the SEREX approach on a primary breast cancer (cDNA clone RMNY-BR-55, SEREX database ID 1385) with the human Expressed Sequence Tags National Center for Biotechnology Information (NCBI) database. The in silico analysis predicted a sequence of 14 exons and 1,713 bp. The coding nucleotide sequence of human Mena was published with ENAH Symbol ( 27), which corresponds to our predicted sequence. This sequence was verified by RT-PCR on SBT breast cancer cell line expressing high levels of RMNY-BR-55 RNA ( 3), and DNA sequencing revealed two different cDNA variants ( Fig. 1 ). The first, hMena, corresponds to the ENAH gene located on chromosome 1q42 (accession no. AY345143); the second, hMena+11a, represents a splice variant including an additional exon of 63 nucleotides between exons 11 and 12, which we named exon 11a (GenBank accession no. AF519769). To distinguish between the two splice variants, we selected two oligonucleotides located in exons 5 and 12, and PCR experiments were done using these two primers, MTC1f and MTC4r. Results reported in Fig. 1B show that the difference in length between hMena and hMena+11a amplified products corresponds to the relative exon 11a.
The hMena cDNA encodes a protein of 570 amino acids that, on alignment with the 541-amino-acid murine Mena (NCBI accession no. U72520), displays an identity of 87% as recently reported ( 27). The two serine phosphorylation sites of murine Mena are both present in the hMena sequence, whereas the tyrosine residue site of phosphorylation in mouse Mena ( 16) is not conserved in the human sequence but substituted by a glutamine residue.
The additional exon 11a present in hMena+11a encodes a peptide of 21 amino acids located in the EVH2 domain of the protein. As predicted by NetPhos 2.0 proteomic tool, 9 the 21-amino-acid sequence contains two putative serine phosphorylation sites in positions 3 and 18 and one potential tyrosine phosphorylation site in position 16, with the highest prediction score for serine at position 3 ( Fig. 1D). Three different tools present on web-sites 10 predict a protein kinase C phosphorylation site on serine at position 18.
hMena and hMena+11a isoform characterization and expression in tumor cell lines. In Fig. 2A , the Western blot analysis of hMena and hMena+11a in vitro translated proteins by the use of an anti-hMena antibody ( 3) recognizing both isoforms is reported. The proteins migrate with an apparent molecular weight of 88 kDa (hMena) and 90 kDa (hMena+11a), whereas the predicted molecular weight is 64 and 66 kDa, respectively. This slower mobility has been reported for mouse Mena and attributed to the proline-rich motif present in the sequence ( 12). The electrophoretic mobility of the two in vitro translated hMena isoforms corresponds to the hMena isoforms evident in SBT tumor cell protein extracts.
The hMena and hMena+11a isoforms are not clearly distinguishable by Western blot because they comigrate (88–90 kDa). The analysis on a large panel of tumor cell lines of different histotypes ( Fig. 2B) shows that hMena overexpression characterizes the transformed phenotype of various cell lineages.
An additional protein form of 80 kDa, which is evident in astrocytoma, neuroblastoma, and glioma cells, is currently under investigation.
RT-PCR experiments conducted on SBT and MCF7 breast cancer cell lines identified two PCR products corresponding to the hMena and hMena+11a splice variants ( Fig. 2C). Two melanoma and the glioma cell lines did not express hMena+11a but rather hMena and an additional splice variant lacking one or more exons. This splice variant might encode a protein with a putative size compatible with that of hMena protein isoform of ∼80 kDa molecular weight observed in the Western blot experiments and expressed in tumor cell lines of neuroectodermal origin.
A two-dimensional Western blot analysis was conducted on protein extracts of breast tumor (SBT, MCF7, and T47D), colon cancer (LS180), and glioma (T98G) cell lines, as shown in Fig. 2D, in which hMena shows a characteristic pattern. In the breast carcinoma cells as well as in the colon cancer cells, two distinct sets of spots with slightly different molecular mass and pI ranging between 5.4 to 6 (lower protein spots) and 5.8 to 6.2 (upper protein spots) were revealed by anti-hMena serum ( Fig. 2D, SBT, MCF7, and LS180). Experiments done with cells transiently transfected with hMena+11a showed that the upper protein spots correspond to this protein isoform (see below).
A different pattern was observed in T98G cells. In fact, out of sharing with breast and colon carcinoma cells the reactivity for the 5.4 to 6 pI set of spots, an extra set of protein spot with lower molecular weight and more basic pI (range, 5.9–6.7) was present ( Fig. 2D).
All these results suggest that whereas hMena is always present, hMena+11a is selectively expressed in epithelial tumor cell lines (breast, lung, cervical, and colon cancer).
EGF treatment up-regulates hMena expression in epithelial breast cancer cell lines. In view of previous findings that hMena expression is up-regulated by neuregulin-1 treatment ( 4), we analyzed whether EGF treatment may affect hMena expression in epithelial breast tumor cell lines. EGF treatment of MDAMB361 and T47D breast cancer cell lines was followed by an increase of hMena protein levels as detected by Western blot. In particular, as shown in Fig. 3 , the treatment with different doses of EGF showed that the highest level of hMena expression occurs at a concentration of 75 ng/mL. This effect is maintained till 72 h in the T47D breast cancer cell line, and the removal of EGF for 1 or 2 days resulted in the restoration of the expression level of the untreated cells. Pretreatment of T47D cells for 30 min with 500 nmol/L of the EGFR inhibitor AG1478 reduced the effect of EGF on hMena ( Fig. 3C), indicating that increased levels are, in part, dependent on EGFR signaling. In agreement with the Western blot data, immunostaining of T47D cells showed that hMena protein, whereas expressed by a small proportion of the cells in the untreated population, is up-regulated following EGF treatment, and this increase is reduced in the presence of AG1478 ( Fig. 3D–F). The addition of AG1478 in the control cells does not affect hMena expression (data not shown).
Densitometric analysis of the two-dimensional electrophoresis images obtained revealed that expression of both isoforms was increased with an EGF/control ratio of 2.13 and 1.28 for hMena and hMena+11a, respectively ( Fig. 4A ).
hMena+11a is phosphorylated on cell treatment with EGF. Interestingly, two-dimensional electrophoresis analysis of the T47D epithelial breast carcinoma cells treated with EGF showed that some of the spots of the upper set corresponding to hMena+11a shifted toward acidic pH ( Fig. 4A and B, top). This change in pI of the protein spots indicates changes in posttranslational modifications affecting protein charge. Phosphorylation is one of the most frequent posttranslational modifications able to change protein pI because phosphate groups that carry negative charges confer acidic features to the protein. We thus incubated cell lysate with λ-PPase to remove phosphates from all phosphorylated amino acid residues. This treatment resulted in a shift of the hMena+11a spots toward less acidic pH ( Fig. 4B, arrows) showing that the less acid spots of hMena+11a are phosphorylated on EGF treatment. This was further confirmed by the partial block of spots shifting toward acidic pH observed when cells were pretreated with the EGFR inhibitor AG1478 ( Fig. 4B, bottom). Because, under resting conditions, hMena+11a already showed acidic spots, we investigated by λ-PPase treatment whether they were phosphorylated. As shown in Fig. 4C, after phosphatase treatment, there was an increase in the relative abundance of the more basic spots, indicating that a fraction of hMena+11a is already phosphorylated in untreated cells and that this fraction is increased on EGF treatment.
hMena+11a overexpression and phosphorylation, increased proliferation, and p42/44 MAPK activation in breast cancer cell lines. In view of previous findings that hMena overexpression correlates with Ki67 proliferation index in 163 primary breast tumors ( 4), to gain insights on the functional role of hMena+11a, we investigated whether this isoform could be involved in the control of breast tumor cell proliferation. To this end, hMena+11a was stably overexpressed in MCF7 and transiently overexpressed in SKBr3 cancer cell line. As shown by the two-dimensional electrophoresis analysis ( Fig. 5A and B ), hMena+11a was constitutively phosphorylated both in MCF7 and SKBr3 cells with higher level of phosphorylation in SKBr3 cells (63.7% versus 90.6%). In MCF7 cells that exhibit lower levels of basal phosphorylation, the exogenous expression of hMena+11a was accompanied by increased amount of phosphorylated hMena+11a with respect to control cells (63.7% up to 77%), as indicated by a shift of the spots toward more acidic pH ( Fig. 5A).
In addition, both transfected cell lines ( Fig. 5C and D) exhibited a significant growth rate over control cells, as evaluated by [3H]thymidine proliferation assay. Furthermore, parallel analysis of protein extracts showed higher levels of p42/44 MAPK phosphorylation ( Fig. 5E and F) both in MCF7 (1.7-fold) and SKBr3 (1.8-fold) hMena+11a–transfected cells as compared with control cells, suggesting that the hMena+11a increased phosphorylation and p42/44 MAPK activation observed on transfection could be associated with breast cancer cell proliferation. Because EGF treatment clearly enhanced the phosphorylation of hMena+11a ( Figs. 4 and 5B), we then evaluated whether EGF-induced proliferation of hMena+11a transiently transfected SKBr3 cells was also associated with increased phosphorylation of hMena+11a and p42/44 MAPK activation. Figure 5B shows a further increase in the rate of hMena+11a phosphorylation in SKBr3/hMena+11a cells treated with EGF as indicated by the shift toward more acidic pH of the spots, which was hampered by pretreatment of the cell lysate with λ-PPase (data not shown). Figure 5D indicates that SKBr3/hMena+11a cells displayed a higher proliferative response to EGF than the control cells (SKBr3/pcDNA3), which was associated with concomitant enhanced p42/44 MAPK activation ( Fig. 5F). Conversely, the knockdown of the endogenous hMena expression with specific siRNA, by decreasing the expression of both protein isoforms ( Fig. 6A ), significantly reduces the EGF-mediated p42/44 MAPK phosphorylation and cell proliferation in BT474 breast cancer cell line ( Fig. 6A and B).
Collectively, these results show a link between hMena+11a phosphorylation, proliferation rate, and MAPK activity in breast cancer cell lines.
Malignant transformation and tumor progression are sustained by a profound rearrangement of actin cytoskeleton resulting into altered cell to cell interactions, high proliferation rate, increased survival, and migratory and invasive behavior. Because cytoskeletal proteins such as microtubules are targets of currently used antitumor drugs (i.e., taxanes), the possibility of interfering with tumor-associated actin alterations may be exploited for chemoprevention and chemotherapeutic drug development. On the basis of our present knowledge, the upheaval of actin remodeling is recognized as a general effector event induced by the unrestrained signaling of a wide spectrum of oncogenes and growth factors ( 28). Abnormal expression of actin binding proteins has been described through proteomic analysis of breast cancer ( 1) even at the stage of carcinoma in situ, suggesting that these modifications represent early signatures of transformation. Through the analysis of the antibody repertoire specific for autologous breast cancer, we have recently identified in human breast tumors changes in the expression of the protein hMena ( 3). By in vivo studies, we have shown that hMena, whereas undetectable in normal breast epithelium, is overexpressed in benign breast lesions with an increased risk of transformation and in tumors with the HER-2+ ER(−), PgR(−), Ki67(+) aggressive phenotype ( 4), which requires tailored therapy. In view of the above and with the aim of gaining further insights into hMena sequence and functions in breast tumors, we cloned and characterized its full-length sequence from the breast tumor cell line SBT with an epithelial phenotype. The results of this analysis confirmed our in silico results, the hMena sequence being identical to that recently described in human brain ( 27). The sequence of this protein displays 87% homology with the murine counterpart but is longer with the majority of the additional amino acids located in the Arg/Leu/Glu–rich region (LERER), a motif peculiar to murine Mena within members of the Ena/VASP protein family. Because the LERER motif is believed to function as a protein-protein binding interface ( 7), one can hypothesize that, in the presence of additional amino acids, the murine and human sequences may form different complexes. Of interest, we isolated an hMena splice variant (hMena+11a) that contains an additional exon corresponding to 21 amino acids located in the EVH2 domain. This domain, which characterizes the Ena/VASP family members, mediates protein tetramerization and F-actin and G-actin binding ( 29). Although tissue-specific Mena isoforms have been described in mouse ( 12), no hMena+11a equivalent has been reported. However, an equivalent splicing event has been reported for murine EVL, a member of Ena/VASP family ( 30). The presence of additional sequences through intermolecular interactions within the coiled-coil domain of Mena developing tetramers ( 29) may also generate other functional forms, as suggested for other proteins such as the epimorphin, a key regulator of mammary morphogenesis ( 31).
hMena overexpression characterizes the transformed phenotype of tumor cells of different lineages. Furthermore, we establish that hMena+11a is concomitantly overexpressed only in epithelial tumor cell lines.
Inappropriate growth factor stimulation of tyrosine kinase receptors of the EGF family induced by autocrine and stroma-mediated paracrine mechanisms ( 32) or receptor overexpression ( 33) characterizes breast tumors. We have recently described that hMena expression, whereas up-regulated by neuregulin, is down-modulated by Herceptin ( 4), thus suggesting that hMena may participate in actin remodeling by coupling tyrosine kinase signaling to actin cytoskeleton. Interestingly, a recent tyrosine phosphoproteome analysis of human breast tumors has identified cytoskeleton proteins as the most abundant tyrosine phosphoproteins in breast cancer ( 34). In view of the extensive cross-talk among the members of the EGF receptor family, which represent relevant therapeutic targets, we analyzed whether EGF can differently modulate the expression and the phosphorylation status of hMena and hMena+11a isoforms.
Our results by two-dimensional electrophoresis analysis show that EGF treatment of different breast tumor cell lines, whereas up-regulating both isoforms, induced an increase in the hMena+11a phosphorylation only. Although this event may be related to hMena+11a overexpression, the detection of an increase in the fraction of phosphorylated versus unphosphorylated hMena+11a spots favors the hypothesis that these changes are under the control of an independent regulatory event. This is consistent with the results that EGF-induced hMena+11a phosphorylation is prevented by the specific EGFR inhibitor AG1478 and that both constitutive and exogenously expressed hMena+11a exhibit high levels of basal phosphorylation in SKBr3 cell line overexpressing EGFR, again suggesting that this isoform is coupled to EGFR signaling. Cell signaling downstream of EGFR composes an interconnected network of pathways involving several protein kinases ( 35) and different PKC isoforms implicated in breast cancer development and progression ( 36). Prediction analysis indicates three putative phosphorylation sites in the additional 11a peptide of the hMena+11a isoform, the highest score being attributed to a serine and a consensus binding site for PKC. However, the inclusion of the additional 21 amino acids is likely to reveal cryptic motifs for kinase binding not located in the 11a peptide of hMena+11a protein, and only the mapping of the precise phosphorylation site(s) of hMena+11a will clarify which hMena+11a amino acid residues become phosphorylated and which kinases are implicated. Differently from the murine Mena sequence, which has been described to be phosphorylated ( 14, 16), we never found hMena to be affected by dephosphorylation treatment despite the presence of several phosphorylation sites.
Interestingly, phosphorylation of VASP in the EVH2 domain reduces its ability to bind F-actin and abolishes its anticapping activity ( 8). Because the 11a peptide is in the EVH2 domain, it is possible that this isoform is subject to additional negative regulation.
The finding that hMena+11a overexpression and phosphorylation in breast cancer cell lines resulted in a growth advantage and into activation of p42/44 MAPK, key mediator of mitogenic signaling including those EGF-mediated, strongly suggests a functional role of this isoform in modulating breast cancer cell proliferative activity. This is in line with our recent in vivo immunohistochemical results showing that hMena overexpression correlates with the proliferation index in primary breast tumors ( 4). The present results further underscore the emergent role that actin cytoskeleton regulatory proteins and, in particular, hMena isoforms may exert in coordinating multiple signaling pathways leading to breast cancer development and suggest that the activation status of these proteins may be relevant when these pathways are targets of therapy.
Grant support: Italian Ministry of Health; Associazione Italiana per la Ricerca sul Cancro; Lega Italiana per la Lotta Contro i Tumori; Ministero dell'Istruzione, dell'Università e della Ricerca-Fondo per gli Investimenti della Ricerca di Base; Fondazione Piemontese per gli Studi e le Ricerche sulle Ustioni; Fondazione Cariplo; and NIH grant CA100324 (J.S. Condeelis).
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.
We thank Dr. G. Zupi for her continuous support and helpful discussion; A. Conidi and D. Del Bello for their support and suggestions; and P.I. Franke and M.V. Sarcone for the final revision of the manuscript and for the secretarial assistance.
Note: F. Di Modugno and L. DeMonte contributed equally to this work.
↵10 NetPhosK 1.0, http://www.cbs.dtu.dk/services/NetPhosK; KinasePhos, http://kinasephos.mbc.nctu.edu.tw; and Group-based Phosphorylation Scoring Method, http://973-proteinweb.ustc.edu.cn/gps/gps_web.
- Received May 31, 2006.
- Revision received January 11, 2007.
- Accepted January 16, 2007.
- ©2007 American Association for Cancer Research.