Estrogen Induces Vav1 Expression in Human Breast Cancer Cells

Vav1, a guanine nucleotide exchange factor (GEF) for Rho family GTPases, is a hematopoietic protein involved in a variety of cellular events. In recent years, aberrant expression of Vav1 has been reported in non-hematopoietic cancers including human breast cancer. It remains to be answered how Vav1 is expressed and what Vav1 does in its non-resident tissues. In this study, we aimed to explore the mechanism for Vav1 expression in breast cancer cells in correlation with estrogen-ER pathway. We not only verified the ectopic expression of Vav1 in human breast cancer cell lines, but also observed that Vav1 expression was induced by 17β-estradiol (E2), a typical estrogen receptor (ER) ligand, in ER-positive cell lines. On the other hand, Tamoxifen, a selective estrogen receptor modulator (SERM), and ICI 182,780, an ER antagonist, suppressed the expression of Vav1. The estrogen receptor modulating Vav1 expression was identified to be α form, not β. Furthermore, treatment of E2 increased the transcription of vav1 gene by enhancing the promoter activity, though there was no recognizable estrogen response element (ERE). Nevertheless, two regions at the vav1 gene promoter were defined to be responsible for E2-induced activation of vav1 promoter. Chromatin immunoprecipitation (ChIP) and co-immunoprecipitation (Co-IP) analyses suggested that ERα might access to the vav1 promoter via interacting with transcription factors, c-Myb and ELF-1. Consequently, the enhanced expression of Vav1 led to the elevation of Cyclin D1 and the progression of cell cycle. The present study implies that estrogen-ER modulates the transcription and expression of Vav1, which may contribute to the proliferation of cancerous cells.


Introduction
Breast cancer is the most common death-causing cancer in females [1]. The persistent exposure to estrogen has been observed to closely correlate with the development of breast cancer [2][3][4][5]. The expression and responsiveness of estrogen receptor (ER) has been applied as one of the most important markers for the breast cancer classification and prognosis [6]. Two forms of estrogen receptors, ERa and ERb, have been identified [7]. ERa is the dominant form in the breast and uterus, whereas ERb has a wider distribution profile that expands in tissues such as prostate, ovary, lung, and spleen [8]. As a ligand, estrogen binds to ER, and induces its conformational change to activate it. The activated ER associates with the estrogen response element (ERE) at the promoter regions of a variety of genes [9], or complexes with other transcription factors, such as AP1 [10], SP1 [11], or E2F1 [12], modulating the expression of target genes that are involved in cell cycle checkpoint [12,13], cell proliferation [14,15], and apoptosis [16].
Vav1 is first identified as a proto-oncogene in hematopoietic cells [17], with the renowned character as a guanine nucleotide exchange factor (GEF) for RhoGTPases. A plethora of studies revealed that Vav1 is a multidomain protein which not only activates RhoGTPases for cytoskeleton reorganization during lymphocytes activation [18], but also plays a GEF-independent role in diverse cellular processes including calcium mobilization in T cells [19]. The oncogenic form, lacking the N-terminal Calponin homology (CH) domain, is obtained by its transforming effect on NIH3T3 fibroblast cells [20]. Meanwhile, evidence unveils the non-hematopoietic expression profile of Vav1, which associates with several human tumor malignancies, such as neuroblastoma [21], lung cancer [22], and pancreatic ductal adenocarcinomas [23]. Knocking down of vav1 gene in lung cancer and pancreatic cells leads to the decreased cell proliferation and reduces tumor size in nude mice [22,23]. In addition, patients with Vav1-positive pancreatic tumors exhibit poorer prognosis and lower survival rate than those with Vav1-negative tumors [23], suggesting that the ectopic expression of Vav1 plays an innegligible role in tumor development and progression.
Recently, the aberrant expression of Vav1 has been reported and its correlation with estrogen receptor has been addressed in human breast cancer [24][25][26][27]. Herein we aim to investigate the modulation of Vav1 expression in breast cancer cells, and the effect of Vav1 on breast cancer cell proliferation. By confirming the increased vav1 mRNA and protein in several ER-positive cell lines, we found that the transcription and expression of Vav1 was significantly enhanced by E 2 treatment in a time-and dosedependent manner via ERa. We further addressed that E 2 -induced vav1 transcription involved the complex of ERa with other transcription factors. Finally, we showed that the amount of Vav1 expression correlated with the expression of Cyclin D1 and influenced the cell cycle progression in breast cancer cells. Our data suggested that estrogen may promote breast cancer cell growth partially by triggering the aberrant expression of Vav1.

Cell lines and culture
Human breast cancer cell lines (ER positive: MCF7 and T47D; ER negative: MDA-MB-231 and MDA-MB-157) were originally from American Type Culture Collection (ATCC) and maintained in phenol free RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS) (GIBCO, NY, USA). For MCF7 and T47D cells, 0.01 mg/ml human recombinant insulin was added to the medium. The immortalized breast epithelial line 184A1 was maintained in MEGM (Cambrex, NJ, USA) that consisted of modified MCDB 170 medium supplemented with ,52 mg/mL bovine pituitary extract, 10 ng/mL epidermal growth factor, 0.5 mg/mL hydrocortisone, 5 mg/mL insulin, 50 mg/mL gentamicin sulfate, and 50 ng/mL amphotericin B. Jurkat T leukemia cells and vav1-null Jurkat T cells (J.Vav1) were carried in lab as described previously [28] and were grown in RPMI 1640 medium containing 10% FBS. HEK293T cell line was obtained from ATCC, and grown in DMEM supplemented with 10% FBS.

RNA isolation and reverse transcription
MCF7 and T47D cells were cultured in phenol free RPMI 1640 medium for 24 h to deprive estrogen, and then treated with 10 27 mol/L of E 2 for 24 h. The mRNA was extracted from the cells using PolyATract System 1000 (Promega, WI, USA) and reverse-transcribed with Reverse Transcription System (Promega, WI, USA).

Luciferase reporter assay
The vav1 promoter constructs were obtained as described [30]. MCF7 cells were transfected with 2 mg of total plasmid DNA containing Renilla luciferase vector pRL-TK (Promega, WI, USA) and vav1 reporter construct, pVav1-Luc, per 1610 5 cells by Lipofectamine 2000 (Invitrogen, CA, USA). Cells were cultured in RPMI 1640 medium for 24 h followed by medium containing E 2 (10 27 mol/L) for 48 h, then lysed for luciferase activity analyses with Dual Luciferase Assay kit (Promega, WI, USA) and TD20/20 luminometer (Turner Designs Inc, CA, USA). The promoter activity was presented as the ratio of the firefly luciferase activity to Renilla luciferase activity. To determine the effect of ERs on the promoter activity of vav1, MCF7 cells transfected with pVav1-Luc were pre-treated with Tamoxifen or ICI 182,780 for 30 min in prior to E 2 treatment or treated with ER type-specific agonists, PPT or DPN. The cells were then harvested for luciferase assay as described above.

Knockdown of Vav1 expression by lentivirus-based transduction
The lentiviral plasmids were constructed as described [32]. The shRNA sequence targeting Vav1 or control RNA of scrambled sequence were cloned into pLKO.1-TRC vector (Addgene, http:/ www.addgene.org/), respectively. The lentivirus particles were generated according to the standard protocol [33]. Briefly, HEK293T were co-transfected with the vectors containing shRNAs together with vectors pCMV-VSV-G, pMDLg/pRRE, and pRSV-REV. At 48 h post-transfection, the supernatants were harvested, and the viral particles were collected to infect T47D cells at 37uC for 18 h. The transduced cells were selected by 0.5 mg/mL puromycin for 7 days.

Cell cycle analysis
Cells were synchronized to G0/G1 phase by cultured in serumfree medium in the presence of Tamoxifen [34]. After 36 h of treatment with E 2 (10 27 mol/L) or DMSO as control, cells were collected and fixed with ice-cold 70% ethanol, then incubated with 100 mg/mL RNase A (Transgene Biotech, Beijing, China) for 30 min. Cells were then stained with 50 mg/mL propidium iodide (PI) (Sigma, MO, USA) in the dark for 30 min, and subjected to flow cytometer analysis (Calibur, NJ, USA). The DNA contents and cell numbers were plotted using Cell Quest software (Becton Dickinson, NJ, USA).

Co-immunoprecipitation (Co-IP)
T47D cells were cultured in RPMI 1640 medium for 24 h before adding E 2 (10 27 mol/L) or DMSO for 4 h. Then cells were harvested and lysed in RIPA buffer. 1.5 mg of lysate was precleared with protein A/G-agarose beads and subsequently added in 4 mg of indicated antibodies overnight at 4uC. The immunocomplexes were precipitated by protein A/G-agarose, washed 3 times with RIPA buffer, subjected to SDS-PAGE, and analyzed by Western Blot.

Statistics analysis
Graphical data values are presented as mean values of triplicate experiments 6 standard deviations. Each experiment was carried out independently for at least 3 times, and unpaired student T tests were performed. The statistical significance was set at P,0.05 (marked with *) and P,0.01 (marked with **).

The aberrant expression of Vav1 in human breast cancer tissue and cell lines
It was reported previously that Vav1 was detected in ERpositive breast cancer tissue by immunohistochemistry [27]. Here we examined the expression of Vav1 in human breast cancer cell lines ( Fig. 1

Estrogen enhances Vav1 expression through ERs in MCF7 and T47D cell lines
A correlation has been observed between the progression of breast cancer and the exposure to estrogen, which modulates the transcription of many genes by binding and activating ERs [2][3][4][5]. From the results of tissue immunohistochemistry [27] and cell lines Western blot (Fig. 1), higher Vav1 expression was visualized in ER positive breast tumors or cells than that in ER negative samples or cells. We speculated that estrogen-ER was involved in the control of vav1 gene expression. Two ER-positive cell lines, MCF7 and T47D, were tested for Vav1 expression in the presence or absence of 17b-estradiol (E 2 ). After treatment with 10 27 mol/L of E 2 or DMSO as control, the mRNA transcript of vav1 was measured by qRT-PCR. As shown in Figure 2A, E 2 induced an increase in vav1 mRNA expression by 3.15-fold in MCF7 and 2.86-fold in T47D in reference to the DMSO control (P,0.01), suggesting that E 2 enhanced the transcription of vav1 gene.
We further explored the effects of time and concentration of E 2 on Vav1 expression. MCF7 and T47D cells were exposed to E 2 (10 27 mol/L) for different time points, or to the indicated concentration of E 2 , respectively. As shown in Figure 2B, the E 2 -induced Vav1 expression increased to about 2.2 fold at 48 h (P,0.01) and plateaued at 72 h in both cell lines. The expression of Vav1 reached to nearly maximum at 10 27 mol/L of E 2 , as only limited increases of Vav1 from 10 27 mol/L to 10 26 mol/L were observed, namely from 2.29 to 2.59 fold for MCF7 and 2.15 to 2.21 fold for T47D, respectively (Fig. 2C). The above data indicated that the induction of Vav1 expression is dependent on the time and dose of the ER ligand treatment.
Given that ICI 182,780 and Tamoxifen have been applied in endocrine therapy for ER-positive breast cancer due to their inhibitory effects on ER activation [35][36][37], we used these drugs to address the role of ER in the estrogen regulation of Vav1. As shown in Figure 2D, E 2 alone induced Vav1 protein expression by 2.39-fold and 2.08-fold in MCF7 and T47D cells (Fig. 2D, Lane 2 and 5, P,0.01), respectively. However, this expression was Figure 2. ER-mediated Vav1 expression. (A) MCF7 and T47D cells were treated with E 2 (10 27 mol/L) or DMSO for 24 h. The relative level of vav1 mRNA was determined by qRT-PCR and was presented by the ratio of vav1 mRNA of E 2 -treated samples to that of DMSO-treated control samples and presented as y-axis. The data represented the mean value6S.D. of three independent experiments. (B and C) MCF7 and T47D cells were exposed to E 2 (10 -7 mol/L) for 0 to 72 h (B), or increasing concentration of E 2 for 48 h (C). (D and E) MCF7 and T47D cells were pre-treated with ICI 182,780 (4610 27 mol/L) (D) or increasing concentration of Tamoxifen for 30 min (E) before adding E 2 (10 27 mol/L) for 48 h. The DMSO treatment was used as a solvent control. The Vav1 expression in above treated samples was analyzed by Western Blot with anti-Vav1 antibody, with tubulin as protein loading control. The bar chart below each example blot represents the normalized protein level of Vav1 to Tubulin of three independent experiments. The DMSO treatment was set as 1 to indicate the basal level of Vav1 expression. ''**'' indicates P,0.01 versus DMSO treatment and ''a**'' indicates P,0.01 versus E 2 treatment by unpaired student T test. doi:10.1371/journal.pone.0099052.g002 suppressed by ICI 182,780 to the basal level (Fig. 2D, Lane 3 and 6, P,0.01). Similarly, Tamoxifen counteracted E 2 and inhibited Vav1 expression to the level below the baseline at high concentrations (Fig. 2E, lanes 3-5 and 8-10, P,0.01). These data suggested that Vav1 expression was not only dependent on the time and dose of estrogen, but also required the activation of ERs in breast cancer cell lines.
ER increases the promoter activity of vav1 gene As E 2 -ER efficiently enhanced Vav1 protein as well as mRNA expression, we predicted that ERs would function as a transcriptional activator for vav1 gene promoter. The minimal regulatory sequences of vav1 proximal promoter region, which covered nucleotide (nt) 2287 to +301 relative to Transcription Start Site (TSS), was constructed in plasmid pGL3 as described [30], and the resulting plasmid was named pVav1-Luc. In agreement with Vav1 protein expression in Figure 2, E 2 treatment induced a maximal activation of vav1 promoter, nearly 3 fold above the control (DMSO treatment) (Fig. 3A, P,0.01). The presence of Tamoxifen decreased the vav1 promoter activity to the basal level, and the promoter activity exhibited a negative correlation with the increasing concentration of the drug (Fig. 3A). Similarly, ICI 182,780 suppressed the reporter gene by 1.33 fold (Fig. 3B, P, 0.01 versus E 2 treatment). The above results suggested that ERs were involved in the activation of vav1 promoter activity, and thus the transcriptional activation of vav1 gene.
The transcription of vav1 gene is mediated by the a form of ER By far, two ER forms, ERa and ERb, were known to play significant roles in diverse tissues, and both were detected in MCF7 and T47D cell lines (Fig. 1). We attempted to identify the one which modulated vav1 transcription by using chemical agonist specific for ERa (PPT), or ERb (DPN). The cells transfected with pVav1-Luc were treated with DMSO, E 2 , PPT, or DPN, respectively. The luciferase activity was measured and presented as fold of induction to that treated by DMSO. As shown in Figure 3C, E 2 and PPT induced the promoter activation to the same extent (,3-fold, P,0.01 versus DMSO treatment and N.S. between E 2 and PPT groups), whereas, the ERb agonist DPN did not show any significant induction above the control (DMSO). Thus, only ERa, but not ERb, was responsible for E 2 -induced vav1 transcription.

ERa is involved in the complex associated with vav1 promoter
The involvement of ERa in estrogen-induced vav1 transcription led us to examine the vav1 proximal promoter for conserved ERE sequence in silico by rVista2.0 (http://rvista.dcode.org/) and TRANSFAC (http://www.cbrc.jp/htbin/nph-tfsearch). However, the search result revealed no perfect ERE at the vav1 promoter region, rather, there were two half-ERE sites (hERE) located at the positions +165 to +169 bp and +273 to +277 bp to TSS, respectively (Fig. 4A). As previously reported, ERE-like sequence, such as two half ERE sites, can bind with estrogen activated ER even though they were separated by hundreds of base pairs [9] [38]. Thus we set to verify if ERa bound to the hERE sites at vav1 promoter by ChIP analysis. The primers corresponding to the region spanning the two hERE sites (+59 to +340) were designed accordingly. As shown in Figure 4B upper panel, the sample prior to immunoprecipitation (Input) exhibited a positive hERE region, whereas was detected negative in the post-immunoprecipitated sample (ERa), indicating that ERa did not interact with the hERE sites. Unexpectedly, the region 2232 to +71 was found in association with ERa (Fig. 4B, lower panel, third lane from the left), though there was no consensus binding site for ER.
Furthermore the recruitment of ERa was increased by ,1.7 fold upon E 2 treatment (Fig. 4B, lower panel, sixth lane from the left, P,0.01), and reduced by Tamoxifen treatment (Fig. 4C, P,0.01 and ChIP analysis was performed with anti-ERa antibody or control IgG. Two sets of primers specific for +59 to +340 region containing hERE sites (upper panel) or the 2232 to +71 region of the vav1 promoter (lower panel) were used in PCR. The PCR products were detected by agarose gel electrophoresis. The input represented the DNA in crude cell extract before the immunoprecipitation. (C) T47D Cells were treated with the reagents as indicated in the left side, and ChIP assay were carried on using primers specific for 2232 to +71 of vav1 promoter. The PCR products were resolved by agarose gel electrophoresis. The bar chart beside each example blot represents the normalized DNA level of 2232 to +71 to Input of three independent experiments. ''**'' indicates P,0.01 versus DMSO treatment and ''a**'' indicates P,0.01 versus E 2 treatment by unpaired student T test. doi:10.1371/journal.pone.0099052.g004 ERa associates with 238 to 25 region at vav1 promoter via other transcription factors The above results indicated that ERa was in complex with the 59 region of vav1 gene promoter. Several transcription factors were predicted to bind at the 59 minimal regulatory region of the human vav1 gene, including ETF, Sp1, E2F, NF-e, c-Myb, TCFa, PU.1, and ELF-1 [30]. We therefore attempted to locate the regions that respond to estrogen. The wild type vav1 promoter (WT) and the truncated mutants (D1, D2, D3) that lack the predicted transcription factor binding sites were depicted in Figure 5A, and the reporter plasmids were constructed [30]. As shown in Figure 5B, the wild type promoter activity was elevated to 3 fold by E 2 . The deletion mutant D1 that lacks region -(143,152) exhibited similar extent of induction (2.6 fold), implying that the region 2(143,152) was dispensable in E 2induced vav1 expression. In contrast, the E 2 induction of truncated promoters D2 and D3 was severely suppressed to less than 1.5 fold (P,0.01), indicating that these two regions, 2(25,38) and 2 (5,22), were required for E 2 -induced vav1 transcription. As these regions were reported to possess putative binding sites for transcription factors E2F/NF-e/c-Myb at - (25,38) and TCFa/ PU.1/ELF-1 at 2 (5,22), respectively, ERa may associate with certain transcription factors within these regions. As reported previously, c-Myb affects vav1 transcription in lung cancer cells [30] and is also involved in the E 2 -ER regulated gene expression in breast cancer cells [39]. Meanwhile, another breast cancer related transcription factor, ELF-1, is identified to interact with the promoter of vav1 (Genome browser, http://genome.ucsc.edu/) [40]. Firstly, to confirm the binding of these two transcription factors to vav1 promoter, the ChIP analysis was performed. As shown in Figure 5C, both c-Myb and ELF-1 presented positively in complex with the vav1 promoter (Fig. 5C, upper two panels), and the presence of E 2 or Tamoxifen had no effects on the complex (Fig. 5C, bottom panel).
Next we investigated whether ERa associated with c-Myb and/ or ELF-1 to form the transcriptional complex, and if the complex formation was E 2 -dependent. By co-immunoprecipitation analyses shown in Figure 5D, detectable amount of ERa was pulled down with anti-c-Myb (upper panels) or anti-ELF-1(lower panels) antibodies, respectively, in comparison with control IgG (Fig. 5D, lane 2 versus lane 1). And the amount of co-immunoprecipitated ERa increased with the presence of E 2 (Lane 3 in both upper and lower panels), indicating the E 2 -activated ERa associated with the transcription factors c-Myb and ELF-1. In the presence of Tamoxifen, the co-immunoprecipitated ERa was barely detectable (Fig. 5D, lane 4), further suggesting that it was the activated form of ERa that bound to these transcription factors.
Given that the existing interaction of c-Myb/ELF-1 with vav1 promoter was constitutive (Fig. 5C), and ERa association with -232 to +71 region was E 2 inducible (Fig. 4C), we proposed that the E 2 dependent activation of vav1 promoter was achieved by the association of resident transcription factors such as c-Myb and ELF-1, with E 2 -activated ERa. As the model shown in Figure 5E, E 2 induced activation of ERa, which, instead of binding directly to vav1 promoter, interacted with the existing transcription factors to control the vavl transcription. Of course, our results did not exclude the involvement of other regulators in the complex. Nevertheless, the recruitment of ERa and the formation of the transcriptional complex enhanced the transcription of vav1 gene.
The expression of Vav1 promotes cell cycle progression in breast cancer cells E 2 is identified as a causative factor of breast cancers and wellcharacterized to induce cell growth in ER-positive breast tumors [34]. As Vav1 is also involved in cell proliferation in lung cancer and pancreatic cancer cells [22,23], we speculated that Vav1 participated in cell cycle progression of breast cancer cells under the control of E 2 . Stable cell lines, T47D-ShVav1 expressing short hairpin RNA for Vav1, and T47D-Ctrl expressing a scrambled sequence, were established by lentivirus-based transduction. The expression of Vav1 in these cell lines was verified as shown in Figure 6A, and the level of Vav1 expressed in shVav1 cells was reduced (Fig. 6A, top panel, the left two lanes). The E 2 treatment elevated the Vav1 expression proportionally (Fig. 6A, top panel, the right two lanes). The expression of Cyclin D1 was also determined as a commonly recognized factor for cell cycle progression (Fig. 6A, middle panel). In the absence of E 2 , shVav1 reduced Cyclin D1 by 50% (P,0.01) in comparison with the control (left lane). E 2 induced a significant 2.31-fold increase of Cyclin D1 (Fig. 6A, third lane from the left, P,0.01) in coordination with Vav1, and that was reduced by the shRNA knockdown of Vav1 (P,0.01 versus E 2 treatment of Control).
Furthermore, we examined the effects of Vav1 on the cell proliferation by flow cytometry analysis. The shRNA-transduced breast cancer cells were synchronized to G0/G1 phase and the G0/G1 arrest was released by E 2 or DMSO treatment as described in Methods. The decreased percentage of cells in G0/ G1 phase represented the cells progressing to cell cycle. As in Figure 6B, in the absence of E 2 , 5.20% of T47D-Ctrl cells and 1.41% of T47D-ShVav1 cells were released from the G0/G1 checkpoint (Fig. 6B, left bars), respectively. With E 2 treatment, 13.36% of the T47D-Ctrl cells reentered cell cycle, whereas only 3.91% of T47D-ShVav1 cells were progressed (Fig. 6B, right bars). Thus, knockdown of Vav1 resulted in a 3.79% reduction in DMSO group and a 9.45% reduction in E 2 group. The differences of the percentage of proliferating cells with shVav1 or not reckoned for the relative amount of Cyclin D1 of those cells (Fig. 6A), and revealed that knockdown of Vav1 decreased E 2induced cell proliferation. In addition, overexpression of Vav1 in T47D cells also exhibited an enhanced Cyclin D1 by 2.5 fold (Fig.  S1A), and a higher cell growth rate was also shown by WST-1 Myb (left panels) or ELF-1 (right panels), with preimmune IgG as control. The precipitated DNAs were analyzed by PCR with the primers corresponding to position 2232 to +71 of vav1 promoter. The bar chart below the example blot represents the normalized DNA level of 2232 to +71 to Input of three independent experiments. ''N.S.'' indicates P.0.05 versus DMSO treatment by unpaired student T test. (D) The association of ERa with c-Myb and ELF-1. T47D cells were treated with E 2 (10 27 mol/L) or DMSO for 4 h or pretreated with Tamoxifen (10 26 mol/L) for 30 min in prior to E 2 treatment and lysed. Antibodies against c-Myb (upper panels) and ELF-1 (lower panels) or control IgG (Lane 1) were used to immunoprecipitate protein complex, which were then resolved by Western Blot with indicated antibodies. (E) The proposed model for ERa modulating the vav1 promoter activity. E 2 -activated ERa interacts with vav1 promoter via interacting with transcription factors such as c-Myb, ELF-1, or perhaps other unknown coregulators (labeled ''?'') to promote the vav1 transcription. doi:10.1371/journal.pone.0099052.g005 analysis (Fig. S1B). The implications from the above experiments are 2-fold, 1) Vav1 expression level significantly influences the cell cycle progression and cell proliferation; and 2) The amount of Vav1 protein contributes the E 2 -upregulated cell proliferation.

Discussion
Vav1 has been recognized as a hematopoietic-specific protein and plays important roles in T cell activation. The nonhematopoietic expression of Vav1 has been reported recently in association with several human malignancies, including pancreatic ductal adenocarcinomas [23], lung cancer [22], neuroblastoma [21], melanoma [41], and breast cancer [24][25][26][27]. Our present study revealed that Vav1 protein expression was observed in some breast cancer cell lines by Western Blot, especially in ER+ cell lines, though with discrepancy in MCF-7 cell line [27]. It was mutually believed that such discrepancy may result from the resources and passage numbers of the cell line (personal communication with Dr. Katzav). Nevertheless, Vav1 was aberrantly expressed in breast cancer tissue and cell lines.
Several studies are attempted to explore the mechanisms involved in non-hematopoietic expression of Vav1. Epigenetic indication of vav1 expression is proposed. For example, the demethylation of the vav1 gene promoter is detected in primary pancreatic adenocarcinomas [23]; methylation of CpG in 59regulatory sequences of the vav1 promoter is addressed in lung cancer cells [30]; and degradation of Vav1 through Cbl ubiquitination is proposed in breast cancer cells [27]. Given a positive correlation between Vav1 and ER expression in breast cancer tissue [27], we were motivated to explore the Vav1 expression along the E 2 -ER axis. Our data unveiled the transcriptional control of the vav1 gene under the estrogen-ER pathway, and the required isoform, ERa, was the dominant form in breast tissue [42].
A plethora of genes are reported to be modulated by estrogen-ER pathway via estrogen response elements. We analyzed the vav1 promoter sequence in silico, and identified two half ERE sites. However, it was not supported by the ChIP analysis, as these sequences did not bind and recruit ERa (Fig. 4B). Rather, promoter sequences stretching from 2232 to +71 bp to TSS associated with ERa indirectly. Further, two regions in the vav1 promoter were found indispensable for E 2 -induced reporter activity, which contained the binding sites for c-Myb and ELF-1, respectively. Accumulating evidence does support this scenario that ER indirectly activates gene transcription via binding to other DNA-bound transcription factors [10][11][12]43]. For example, ERa complexes with c-Myc to mediate the expression of Noxa in breast cancer cells [12]. The association of ER with Sp1 is also reported to modulate c-fos expression [11]. In our data, ERa, by interacting with c-Myb and ELF-1, conferred estrogen responsiveness to vav1 gene (Fig. 5E). Indeed, c-Myb and ELF-1 possess the leucine-rich motif: LXXLL, which can be recognized by ERa [43]. Of course, other coregulators involved in complex with ERa remain to be identified.
Two activation functions (AFs) mediate the transcriptional activation of ER, the N-terminal AF-1, and the AF-2 in ligand binding domain (LBD) [44]. The biological ligand, E 2 , binds to LBD and induces its conformational change to trigger the activity of AF-2 which can be recognized by coactivators [43]. Previous studies suggest that Tamoxifen, the non-steroidal type I ER antagonist, induces a different conformation from that induced by E 2 , and thus blocks the binding of coactivators and inhibits AF-2 activity [45,46]. By contrast, ICI 182,780, a steroidal type II ER antagonist, binds competitively to the E 2 binding site, blocks ER Figure 6. Effect of Vav1 in cell cycle progression. (A) T47D cells were infected with lentivirus particles that expressed Vav1-specific shRNA or shRNA with a scramble sequence (served as a control). The homogenous cells were first synchronized to G0/G1 phase and then treated with DMSO as control or E 2 (10 27 mol/L) for 36 h before harvest. The expression of Vav1 and Cyclin D1 were analyzed by Western Blot with indicated antibodies respectively. The density of the bands was quantitated by Quantity One software (Bio-Rad, version 4.4.0, CA, USA). The bar chart represents the normalized protein level of Cyclin D1 to tubulin of three independent experiments. ''**'' indicates P,0.01 versus DMSO treatment of Control, and ''a**'' indicates P,0.01 versus E 2 treatment of Control by unpaired student T test. (B) Another aliquots of above cells were stained by PI and analyzed by flow cytometer for DNA contents. The decrease in the percentage of cells in G0/G1-phase before and after the treatment by DMSO or E 2 was calculated and plotted as y-axis and labeled on the bar graph. The differences between two T47D-Ctrl and T47D-ShVav1 cells were also marked. The data represented the mean 6S.D. of three independent experiments. ''**'' indicates P,0.01 versus corresponding control group by unpaired student T test. doi:10.1371/journal.pone.0099052.g006 activation, and leads to a rapid degradation of ER [35]. As both Tamoxifen and ICI 182,780 eliminated E 2 -induced vav1 expression (Fig. 3), and the Co-IP analysis revealed that the association of ERa with the two cofactors was disrupted by Tamoxifen (Fig. 5D), we speculated that the correct conformation of ERa is required for its complex with the DNA-bound c-Myb and ELF-1 to access vav1 promoter, thus control vav1 transcription.
Elevated expression of Vav1 has been demonstrated to affect cell proliferation in lung cancer and pancreatic cancer cells [22,23]. As a GDP/GTP exchange factor, Vav1 is also involved in CXCL12-promoted invasion of melanoma cells [41]. In this study, we observed that the aberrant expression of Vav1 correlated well with the production of Cyclin D1, a critical mediator of estrogenstimulated cell cycle progression [47,48], thus contributing to the proliferation of breast cancer cells (Fig. 6, Fig. S1). We could not rule out other mechanisms that Vav1 played in corresponding to the growth and development cancer, as Vav1 might also protect cells from apoptosis by enhancing anti-apoptotic Bcl2 transcription that we reported in leukemia cells [49].
In summary, our data revealed that E 2 promoted the expression of vav1 in a dose-and time-dependent manner in ER-positive breast cancer cells. E 2 -ER indirectly enhanced vav1 transcription, perhaps via the interactions with other transcription factors such as c-Myb and ELF-1. The E 2 -induced Vav1 level in breast cancer cells was in favor of promoting Cyclin D1 expression and accelerating the cell proliferation, and Vav1 might partially contribute to the pathogenesis and prognosis of breast cancer. This study emphasized the involvement of Vav1 ectopic expression in ER positive breast cancer cells, which reinforced the hypothesis that Vav1 could exert its oncogenic role in human breast cancer. Figure S1 Effect of Vav1 overexpression on cell proliferation. T47D cells were transduced with lentivirus particles encoding Vav1 (pCDH-Vav1) or the control vector backbone (pCDH). The expression of Vav1 and Cyclin D1 were analyzed by Western Blot and cell proliferation was determined by WST-1 proliferation assay in these cells.