Slug increases sensitivity to tubulin-binding agents via the downregulation of βIII and βIVa-tubulin in lung cancer cells

Transcription factor Slug/SNAI2 (snail homolog 2) plays a key role in the induction of the epithelial mesenchymal transition in cancer cells; however, whether the overexpression of Slug mediates the malignant phenotype and alters drug sensitivity in lung cancer cells remains largely unclear. We investigated Slug focusing on its biological function and involvement in drug sensitivity in lung cancer cells. Stable Slug transfectants showed typical morphological changes compared with control cells. Slug overexpression did not change the cellular proliferations; however, migration activity and anchorage-independent growth activity with an antiapoptotic effect were increased. Interestingly, stable Slug overexpression increased drug sensitivity to tubulin-binding agents including vinorelbine, vincristine, and paclitaxel (5.8- to 8.9-fold increase) in several lung cancer cell lines but did not increase sensitivity to agents other than tubulin-binding agents. Real-time RT-PCR (polymerase chain reaction) and western blotting revealed that Slug overexpression downregulated the expression of βIII and βIVa-tubulin, which is considered to be a major factor determining sensitivity to tubulin-binding agents. A luciferase reporter assay confirmed that Slug suppressed the promoter activity of βIVa-tubulin at a transcriptional level. Slug overexpression enhanced tumor growth, whereas Slug overexpression increased drug sensitivity to vinorelbine with the downregulation of βIII and βIV-tubulin in vivo. Immunohistochemistry of Slug with clinical lung cancer samples showed that Slug overexpression tended to be involved in response to tubulin-binding agents. In conclusion, our data indicate that Slug mediates an aggressive phenotype including enhanced migration activity, anoikis suppression, and tumor growth, but increases sensitivity to tubulin-binding agents via the downregulation of βIII and βIVa-tubulin in lung cancer cells.


Introduction
The epithelial-mesenchymal transition (EMT), a developmental process in which epithelial cells reduce intercellular adhesions and acquire fibroblastic properties, has emerged to play important roles in the development of the invasive and metastatic potentials of cancer progression [1][2][3]. During the EMT process, epithelial cells lose the expression of E-cadherin and other components of epithelial cell junctions, adopt a mesenchymal cell phenotype, and acquire motility as well as invasive properties [4]. The snail family of zing finger transcription factors, Snail/SNAI1 (snail homolog 1) and Slug/SNAI2 (snail homolog 2), are highly conserved transcription repressors implicated in embryonic development and tumorigenesis, and they are considered to be key EMT-inducible transcription factors [5].
Many clinicopathological studies have demonstrated positive correlations between the expressions of Slug and poor clinical outcomes in breast, ovary, colorectal cancer, and melanoma [1]. In lung cancer, the elevated expression of Slug messenger ribonucleic acid (mRNA) in cancer tissue was significantly associated with postoperative relapse and a shorter patient survival period [6]. These accumulating data strongly suggest that Slug is a poor prognostic factor in many solid cancers and may be involved in the process of metastasis. Meanwhile, the EMT is thought to be involved in drug sensitivity to several anticancer agents [7]. Regarding sensitivity to gemcitabine, mesenchymal-type cancer cells are reportedly associated with gemcitabine resistance in pancreatic cancer cells [8]. A recent study showed that patterns of sensitivity and resistance to three conventional chemotherapeutic agents (gemcitabine, 5-fluorouracil, and cisplatin) in pancreatic cancer cell lines were closely associated with the EMT phenotype [9]. The mechanism of resistance to gemcitabine has been shown to involve the activation of notch signaling, which is mechanistically linked with the mesenchymal-chemoresistance phenotype of pancreatic cancer cells [10]. Another association with the EMT has been found in epidermal growth factor receptor (EGFR)-targeting drugs for the treatment of lung cancer. A clinical trial has revealed that lung cancer cells with strong E-cadherin expression exhibit a significantly longer time to progression after EGFR-tyrosine kinase inhibitor (EGFR-TKI) treatment [9]. Other studies on EGFR-targeting drugs have demonstrated that mesenchymaltype lung cancer cells exhibit an EMT-dependent acquisition of platelet-derived growth factor, fibroblast growth factor receptor, and transforming growth factor beta receptor signaling pathways [11], and integrin-linked kinase is a novel target for overcoming hepatocellular carcinoma resistance to EGFR inhibition [12]. Thus, baseline cellular characteristics based on the EMT phenotype might be useful not only as prognostic biomarkers for a malignant phenotype, but also for predictive markers of sensitivity to anticancer agents.
Collectively, Slug overexpression likely mediates a malignant phenotype and is related to drug sensitivity in several solid cancers; however, detailed studies on the roles of Slug in lung cancer cells remain largely unclear. In this study, we investigated the biological effects of Slug on the cellular phenotype and drug sensitivity in lung cancer cells.

Materials and Methods
Reagents Vinorelbine (VNR), vincristine (VCR), paclitaxel (PTX), and cisplatin were purchased from Wako Pure Chemical Industries (Osaka, Japan). AG1478, an EGFR-TKI, was purchased from Biomol International (Plymouth Meeting, PA). Each chemical agent was dissolved in dimethylsulfoxide (DMSO) for use in the in vitro experiments, and VNR was dissolved in phosphate-buffered saline (PBS) for use in the in vivo experiment.

Paraffin sections
All primary non-small cell lung cancer specimens were derived from patients who underwent surgery at Kobe University Hospital between April 2001 and December 2005. We examined 10 samples of patients who took chemotherapy including TBAs before or after surgery. Baseline characteristics of the samples are presented in Table S1. The institutional review board approval was obtained for this study.

Plasmid construction, viral production, and stable transfectants
The cDNA fragment encoding human full-length Slug was isolated from human corneal epithelial cells using polymerase chain reaction (PCR) and Prime STAR HS DNA polymerase (TaKaRa, Otsu, Japan) with 5′-GG GAA TTC GGC GCC ATG CCG CGC TCC TTC CTG GTC-3′ and 5′-CC CTC GAG CGT CAC TCA GTG TGC TAC ACA GCA GCC-3′ as sense and antisense primers, respectively. The methods used in this section have been previously described [13]. In brief, the sequences of the PCR-amplified DNAs were confirmed by sequencing after cloning into a pCR-Blunt II-TOPO cloning vector (Invitrogen, Carlsbad, CA). Slug cDNA was cut out and transferred into a pQCLIN retroviral vector (BD Biosciences Clontech, San Diego, CA) together with an enhanced green fluorescent protein (EGFP) following internal ribosome entry site sequence (IRES) to monitor the expression of the inserts indirectly. A pVSV-G vector (Clontech, Palo Alto, CA) for the constitution of the viral envelope and the pQCXIX constructs were cotransfected into the GP2-293 cells using the FuGENE6 transfection reagent (Roche Diagnostics, Basel, Switzerland). The stable transfectants expressing EGFP only or EGFP and Slug in each cell line were designated as A549/EGFP, A549/Slug, Ma1/EGFP, Ma1/Slug, H1299/EGFP, and H1299/Slug. These stable cell lines were established from bulk transfected cells, and not from cloned single cells.

Real-time RT-PCR
The methods used in this section have been previously described [13]. The primers used for real-time RT-PCR were purchased from TaKaRa (Table S2). The experiment was performed in triplicate.

Immunofluorescence staining
The methods used in this section have been previously described [15]. Cells were treated with DAPI (6-diamidino-2-phenylindole) to stain the nucleus and photographed using fluorescent microscopy (IX71; Olympus, Tokyo, Japan). The experiment was performed in triplicate.
In vitro growth inhibition assay and migration assay Growth inhibition was evaluated using an MTT assay, as described previously [16]. The methods used in migration assay have been previously described [13]. The experiment was performed in triplicate.

Anchorage-independent growth assay
Anchorage-independent cellular growth was evaluated using an ultra-low attachment surface 6-well plate (Corning â Inc., Lowell, MA). In brief, cells were seeded onto the plates containing RPMI 1640 with 10% FBS. After incubation for 1 week at 37°C, the cells were stained with trypan blue, then observed immediately using a light microscope. The experiment was performed in triplicate. To show numerical comparisons for viable cells, we also performed an MTT assay.

Luciferase reporter assay
The method used in this section has been previously described [17]. TUBB4 promoter fragments (2.5 kbp) were isolated using PCR with the following primers: sense, 5′-GGG GTA CCG TTG AGA GCC ACT GGT CAA ATT TAG TG-3′ and antisense, 5′-GGA AGC TTG CGG CGA GGG TGG AAG ATG CGG CGG AG-3′ for TUBB4. The TUBB4 promoter fragments were cut between the KpnI and HindIII in TUBB4 promoter restriction sites and were transferred into the luciferase reporter vector pGL4.14 (Promega, Madison, WI). All the sequences were verified using DNA sequencing. The empty and TUBB4 promoter-containing reporter vectors were designated as pGL4.14-mock and pGL4.14-TUBB4, respectively. All the samples were examined in triplicate.

Xenograft studies
Nude mice (BALB/c nu/nu; 6-week-old females; CLEA Japan Inc., Tokyo) were used for the in vivo studies and were cared for in accordance with the recommendations for the handling of laboratory animals for biomedical research compiled by the Committee on Safety and Ethical Handling Regulations for Laboratory Animal Experiments, Kinki University. Mice were subcutaneously inoculated with a total of 4 9 10 6 A549/EGFP (n = 10) or A549/Slug cells (n = 10). Two weeks after inoculation, the mice were randomized according to tumor size into two groups to equalize the mean pretreatment tumor size between the two groups (vehicle control or VNR treatment, n = 5 in each group). The mice were then treated with VNR (5 mg/kg/week, i.p.) or the vehicle control (PBS, i.p.) once a week for 3 weeks (administered on days 1, 8, and 15). On day 60, the mice transplanted with A549/EGFP cells were euthanized and the tumor specimens were collected for immunohistochemistry. The mice transplanted with A549/Slug cells were euthanized on day 25 because of ethical considerations, as the tumors grew much faster in these animals. The tumor volume was calculated as the length 9 width 2 9 0.5 and was assessed every 2-3 days.

Immunohistochemistry analysis
Paraffin sections were immunohistochemical stained with rabbit anti-Slug (1:100, Cell Signaling) using standard methods [18]. Signal detection was achieved using the avidinbiotin peroxidase complex assay (ABC Elite Kit, Vector Laboratories, Burlingame, CA) according to the manufacturer's instructions. Slides were then developed with diaminobenzidine (DAB, Invitrogen, Carlsbad, CA) and counterstained with 10% hematoxylin. Immunohistochemically stained sections were examined microscopically and random fields of tumor tissue were identified, digitally captured, and saved for quantification using ImageJ image analysis software (http://rsb.info.nih.gov/ij/). Separate images consisting of positive or background signals (corresponding to DAB and hematoxylin, respectively) were generated from RBG color space original images using a color deconvolution plugin [19]. Suitable thresholds were adjusted and kept constant for each slide, and binary masks were created. The image analysis (IA) score representing the area of pixels corresponding to DAB (positive cells)/ area of pixels corresponding to hematoxylin (negative cells) was calculated and averaged for each individual case.

Statistical analysis
The statistical analyses were performed using Microsoft Excel (Microsoft) to calculate the SD and to test for statistically significant differences between the samples using a Student t-test. A P value of <0.05 was considered statistically significant.

Slug mediated the EMT in A549 cells
To investigate the potential roles of Slug in lung cancer cells, we retrovirally introduced the Slug gene into A549 cells. The stable transfectants expressing EGFP only or EGFP and Slug and were designated as A549/EGFP and A549/Slug, respectively. Slug overexpression induced EMTlike changes in cell morphology including the elongation of the cell shape and cell scattering, which are characteristic features of cells undergoing EMT (Fig. 1A). Real-time RT-PCR demonstrated that Slug overexpression markedly downregulated the mRNA levels of E-cadherin, whereas the mesenchymal markers fibronectin 1 and vimentin were upregulated, compared with A549/EGFP (Fig. 1B). Western blotting showed that Slug overexpression mediated the downregulation of E-cadherin and the upregulation of N-cadherin, vimentin, and fibronectin (Fig. 1C). Immunofluorescence staining also showed the downregulation of E-cadherin and the upregulation of vimentin (Fig. 1D). These results indicated that Slug induces EMT in A549 cells.

Slug enhanced anchorage-independent growth and cellular migration
We examined the cellular growth, migration activity, anchorage-independent cell growth, and the effect on apoptosis using A549/EGFP and A549/Slug cells. No difference was seen in cellular growth between these cells, indicating that Slug does not enhance cellular growth in vitro ( Fig. 2A). Cellular migration activity of the A549/Slug cells was significantly enhanced, compared with the control cells (Fig. 2B). As anchorage-independent growth is considered a hallmark of the EMT, we evaluated the effect of Slug overexpression. Control cells formed dense spheroid-like clusters under anchorage-independent conditions, whereas A549/Slug cells exhibited notably different loose clusters of cells (Fig. 2C). Trypan blue staining revealed that A549/ EGFP cells included many stained dead cells, but A549/Slug cells did not; these results were confirmed by an MTT assay (Fig. 2C). Western blotting showed that Slug overexpression decreased the protein expression of cleaved caspase and cleaved PARP under anchorage-independent conditions, suggesting that Slug suppressed anoikis (Fig. 2D). These results indicated that Slug increased the migration activity and anchorage-independent growth activity with antiapoptotic effects in lung cancer cells.

Slug sensitized lung cancer cells to tubulinbinding agents
Accumulating data indicate that the baseline mesenchymal phenotype is closely associated with resistance to chemotherapy; however, little is known about how Slug overexpression itself induces changes in drug sensitivity. Therefore, we evaluated the sensitivity of A549/EGFP and A549/Slug cells to VNR, VCR, PTX, cisplatin, and the EGFR-TKI. No significant differences were observed in the IC 50 values of cisplatin and the EGFR-TKI to A549 cells (Fig. 3A). Unexpectedly, Slug overexpression sensitized the cells to VNR, VCR, and PTX by about 10-fold (Fig. 3A).
To confirm whether Slug sensitizes other lung cancer cell lines to tubulin-binding agents (TBAs), we established another line of stable transfectants expressing EGFP only or EGFP and Slug using Ma1 and H1299 lung cancer cell lines (Ma1/EGFP, Ma1/Slug, H1299/EGFP, and H1299/Slug) as well as A549 cells and examined the drug sensitivity. Although the differences were relatively small compared with those in A549 cells, Slug overexpression sensitized both Ma1 and H1299 cells to TBAs (two-to fourfold of IC 50 , Fig. 3B

Slug downregulated the expression levels of bIII and bIV-tubulin
Several studies have shown that the decreased expression of bIII and bIV-tubulin in cancer cells is involved in increased sensitivity to TBAs, and this b-tubulin expression is considered to be the major factor determining sensitivity to TBAs [20][21][22][23][24]. Therefore, we hypothesized that Slug modulates the expression levels of the b-tubulin isotype (symbol, gene name, protein name, and accession number are shown in Table S3), thereby increasing the sensitivity to TBAs. Realtime RT-PCR revealed that Slug overexpression did not change the mRNA expressions of TUBB, TUBB1, and TUBB2B, while Slug significantly downregulated the expressions of TUBB3 and TUBB4 in A549/Slug, compared with in A549/EGFP cells (Fig. 4A). The mRNA expression of TUBB, TUBB1, and TUBB2B was changed in Ma1/Slug cells, but not changed in H1299/Slug cells. TUBB4 expression was downregulated in all three cell lines (Fig. 4A). Western blotting also demonstrated that the protein expressions of b-tubulin, bI-tubulin, and bII-tubulin were not modulated by Slug, but bIII and bIV-tubulin were downregulated in A549/Slug cells (Fig. 4B).

Slug suppresses the promoter activities of TUBB4
To address the question of whether Slug directly suppresses the transcription activities of TUBB4, a luciferase reporter assay was performed. The TUBB4 promoter activity was suppressed by about 30%, when cotransfected with a Slug expression vector, compared with an empty vector (Fig. 4C). The result indicates that Slug directly suppresses the transcriptional activities of TUBB4 at transcriptional levels. Two E2-box sequences (CAGGTG/CAC-CTG) are known as the classical binding site of Slug [25]. The TUBB4 promoter region contains two E2-box sequences at À1221 (E2-box #1) and À618 (E2-box #2) (Fig. 4C). To determine whether Slug directly binds to TUBB4 promoter, we performed the ChIP assay against the two putative E2-box regions. The ChIP assay revealed that Slug was bound for the E2-box1 region and the interaction could not be detected for E2-box2 region (Fig. 4D). The result indicates that Slug binds to the TUBB4 promoter and upregulates TUBB4 transcriptional activity. Regarding the TUBB3, one E-box motif "CAGGTG" was found in the proximal TUBB3 promoter at À643. This suggests that a similar mechanism of transcriptional regulation may exist for TUBB3 transcriptional activation as well as TUBB4, although a detailed analysis was not performed. Further study on TUBB3 promoter is warranted.

Slug enhanced tumor growth but sensitized cells to vinorelbine in vivo
We evaluated the Slug-mediated aggressive phenotype and increased drug sensitivity to VNR in vivo. Slug overexpression markedly enhanced the tumor growth in vivo (A549/EGFP, 378 AE 37 mm 3 and A549/Slug, 1820 AE 491 mm 3 ), supporting the idea that Slug induces an aggressive phenotype (Fig. 5A). On the other hand, VNR treatment weakly inhibited the tumor growth in A549/EGFP tumors, compared with in the vehicle control (19% of tumor reduction), while VNR treatment markedly inhibited tumor growth in A549/Slug tumors (64% of tumor reduction), clearly indicating that Slug overexpression markedly increased the drug sensitivity to VNR in vivo, similar to the results observed in vitro.
Immunohistochemical staining confirmed the forced expression of Slug and the downregulation of E-cadherin in A549/Slug tumors (Fig. 5B). In addition, the marked downregulation of bIII and bIV-tubulin was observed in A549/Slug tumors, compared with in A549/EGFP tumors, similar to the results observed in vitro (Fig. 5B). Collectively, these results indicate that Slug mediates an aggressive phenotype but increases sensitivity to TBAs via the transcriptional downregulation of bIII and bIV-tubulin in lung cancer cells.
Slug expression tended to be involved in response to tubulin-binding agents in clinical lung cancer samples We examined 10 clinical lung cancer samples from patients who undertook chemotherapy including TBAs. Immunohistochemistry of Slug showed that the lowest two patients in Slug staining had response of progressive disease, while other patients whose samples had higher expression of Slug resulted in partial response or stable disease against TBAs (Fig. 6). This result suggested that Slug expression is likely to be involved in drug response to TBAs in clinical samples.

Discussion
The EMT has attracted the attention of researchers because of its involvement in the importance of drug sensitivity to anticancer agents in oncology. Several studies have shown that the EMT may be a determinant of the sensitivity of cancers to several anticancer agents such as EGFR-TKI, 5-flurouracil (5-FU), gemcitabine, and cisplatin [8,10,11,[26][27][28], indicating that cancer cells with acquired resistance to anticancer agents have likely transitioned to a mesenchymal phenotype. However, little is known about the sensitivity of cells to TBAs. Unexpectedly, Slug overexpression did not mediate drug resistance but clearly sensitized lung cancer cells to TBAs in our investigation. Generally, several mechanisms have been suggested to be involved in drug sensitivity or resistance to TBAs, and some tubulin isotypes have been reported to be relevant to drug sensitivity to TBAs [20][21][22][23][24]29]. To date, at least nine a-tubulin and seven b-tubulin isotypes have been described, with a complex pattern of distribution in various tissues. The b-tubulin isotypes differ from each other mainly according to differences in amino acid sequences clustered within the last 15 C-terminal residues, which constitute the isotype-defining domain. Some b-tubulin isotypes are widely expressed in many tissues (bI, IVb, and maybe V), while other isotypes are restricted to specific tissues (bII, III, and IVa, neuronal; bIV, hemopoietic [30]). Among these isotypes of tubulin, the most intensively investigated mechanism of drug sensitivity to TBAs is the expression level of bIII and bIV-tubulin [22][23][24]31]. We found that Slug increased the sensitivity to TBAs by about 10-fold in vitro. In addition, an increased sensitivity to VNR and the downregulation of bIII and bIV-tubulin in tumor specimens were observed in vivo. Our finding suggests that the suppression of bIII and bIVa-tubulin by Slug overexpression may contribute to the increased sensitivity to TBAs. ABC transporters are now considered to be involved in drug   promoter activity was examined using a luciferase reporter assay. Luciferase vectors with either an empty or a TUBB4 promoter (pGL4.14-mock or pGL4.14-TUBB4) that includes two E2-box sequences (E2-box #1 and E2-box #2) were transiently cotransfected with a mock or Slug expression plasmid (pcDNA3.1-mock or pcDNA3.1-Slug) expressing b-galactosidase as an internal control. The results were normalized to b-galactosidase activity and are representative of at least three independent experiments. *P < 0.05. (D) Chromatin immunoprecipitation (ChIP) of Slug on the promoter of TUBB4. A549/EGFP and A549/Slug cells were used for analysis. Two putative regions in the TUBB4 promoter that contain the E2-box sequences (E2#1 and E2#2) were amplified using PCR. Primers to the GAPDH promoter were used as the control. Inputs, 1% of chromatin used for ChIP; H 2 O, no DNA templates. EGFP, enhanced green fluorescent protein; SD, standard deviation; PCR, polymerase chain reaction; mRNA, messenger ribonucleic acid. Lung cancer tissues were stained with the anti-Slug antibody using immunohistochemistry. Diffuse expression of Slug was observed in the specimens of Patients 1, 2, and 3, whereas Slug was not expressed in the lung tissue of Patient 9. Expression levels of Slug were quantified and shown as graph according to the response to tubulin-binding agents. IA Score, the image analysis score; PR, partial response; SD, stable disease; PD, progressive disease. Original images were captured at 9400 magnification. resistance to anticancer agents including TBAs, therefore we examined the mRNA expression levels of ABCB1, ABCG2, ABCA1, ABCA2, and ABCC1 on A549/EGFP and A549/Slug cells using real-time RT-PCR. The mRNA expression of ABCA1 was upregulated whereas expression levels of ABCG2 and ABCC1 were downregulated in A549/Slug cells, while the expression levels of ABCB1 and ABCA2 were very low (data not shown). The regulation of ABC transporters by Slug needs further study.
Anoikis suppression, survival from cell death induced by inappropriate or loss of cell adhesion, is likely to be a prerequisite for cancer cells to successfully metastasize to distant sites [32]. One of the underlying molecular mechanisms of anoikis suppression can be explained by the loss of E-cadherin [33]. Snail and Twist are required for this loss of E-cadherin, but whether Slug is capable of directly inducing anoikis suppression remains largely unclear. We demonstrated that Slug is a key molecule in the induction of anoikis suppression. In line with this finding, the elevated expression of Slug mRNA in lung cancer has been shown to be a poor prognostic factor for survival [6]. Further, clinicopathological analysis of the Slug expression in lung cancer as a prognostic biomarker is warranted.
The recent studies on Slug knockdown in cancer cells identified the cancer-related target genes of slug; for example, Slug knockdown downregulated the expression of miR221, chemokine (C-X-C motif) receptor 4, and matrix metallopeptidase 9, and upregulated the expression of plakoglobin and ubiquitin-conjugating enzyme E2D3 [34][35][36][37]. In terms of the cellular phenotype, Slug enhanced cellular migration and invasion with acquisition of aggressive phenotype, in a manner similar to this study. Thus, accumulating bodies of evidence suggest that Slug regulates the expression levels of a wide variety of cancer-related target genes in addition to promoting EMT to cancer cells.
In conclusion, we demonstrated that Slug sensitizes lung cancer cells to TBAs via the suppression of bIII and bIVa-tubulin. Our findings provide novel insight into the biological function of Slug and suggest that Slug overexpression increases sensitivity to TBAs via the direct transcriptional regulation of bIII and bIVa-tubulin in lung cancer cells.

Supporting Information
Additional Supporting Information may be found in the online version of this article: Table S1. Clinical characteristics of lung cancer patients treated with tubulin binding agents. Table S2. The primers used for real-time RT-PCR. Table S3. Summary of the b-tubulin isotype.