Twist Promotes Tumor Cell Growth through YB-1 Expression

Cisplatin is widely used in the treatment of many solid tumors. Cisplatin can induce oxidative stress, endoplasmic reticulum stress, and DNA-damaging stress, leading to the activation of multiple signal transduction pathways. These are involved in mediating cisplatin-induced apoptosis and exert antitumor activities. Thus, the mechanisms by which cancer cells can develop cisplatin resistance can be complex. Cisplatin resistance arises through multiple mechanisms, such as decreased accumulation of cisplatin, inactivation of cisplatin by glutathione conjugation, and increased DNA repair activity. However, little is known about the relationships and interactions among these mechanisms. To resolve and elucidate these complex mechanisms as a whole, we have started to identify the critical transcription-related factors and their interacting partners (1).

We have reported that several transcription factors are up-regulated in cisplatin-resistant cells, including YB-1 (2), ATF4 (3), Clock (4), and ZNF143 (5). These have been reported to be transcription factors directly involved in cisplatin resistance (1). YB-1 is an important molecule for brain development and null-mutant mice show embryonic lethality (6, 7). YB-1 has been proposed to possess various biological activities in both nucleus and cytoplasm (8). YB-1 expression closely correlates with cell proliferation and is up-regulated in both proliferating adult tissues and cancer cells. It is well known that YB-1 overexpression is closely associated with an unfavorable clinical outcome (8, 9). We have previously shown that several E-boxes are located in the YB-1 promoter (10). YB-1 expression is induced by cisplatin and is regulated by the E-box–binding transcription factor c-Myc (11). YB-1 is overexpressed in cisplatin-resistant cells but c-Myc expression is down-regulated in these cells, suggesting that other E-box–binding transcription factors might participate in YB-1 expression.

Among E-box–binding transcription factors, Twist belongs to the family of basic helix-loop-helix transcription factors. Twist inhibits the p53-mediated response to cellular stress (12). Twist shows vague sequence recognition and binds to the nucleotide sequence 5′-CANNTG-3′. Several binding sites are present in the core promoter of the YB-1 gene. Here, we show that YB-1 is a target gene of Twist and that Twist is overexpressed in cisplatin-resistant cells. We also show that knockdown of either YB-1 or Twist expression can induce cell cycle arrest, and YB-1 is a major downstream target of Twist.

These results contribute to the elucidation of the roles of both YB-1 and Twist in tumor progression and suggest that these proteins may represent promising molecular targets for cancer chemotherapy.

Cell culture. Human epidermoid cancer KB cells and human prostate cancer PC3 cells were cultured in Eagle's MEM. Human breast cancer MCF7 cells were cultured in DMEM. These media were purchased from Nissui Seiyaku and contained 10% fetal bovine serum. The cisplatin-resistant KB/CP4 and P/CDP6 cells were derived from KB and PC3 cells as previously described (3) and found to be ∼23- to 63-fold more resistant to cisplatin than their parental cells (3). Human dermal fibroblasts were purchased from Cell Applications, Inc., and cultured in fibroblast basal medium (Cell Applications). Cell lines were maintained in a 5% CO₂ atmosphere at 37°C.

Antibodies. Antibodies against signal transducer and activator of transcription 3 (STAT3), Clock, and Twist were purchased from Santa Cruz Biotechnology. Anti-Flag (M2) antibody, anti-Flag (M2) affinity gel, and anti–β-actin (AC-15) antibody were purchased from Sigma. Anti-YB-1 antibody was prepared as previously described (2).

Plasmid construction. To obtain the full-length cDNA for Twist, PCR was carried out on a SuperScript cDNA library (Invitrogen) using the following primer pair: 5′-ATGATGCAGGACGTGTCCAGCT-3′ and 5′-CTAGTGGGACGCGGACATGGACC-3′. The PCR product was cloned into the pGEM-T easy vector (Promega). For construction of pcDNA3-Flag-Twist, the NH₂-terminal Flag-tagged Twist cDNA was ligated into the pcDNA3 vector (Invitrogen). The construction of pcDNA3-Flag-YB-1 has previously been described (13).

The core promoter and partial first exon (−661 to +78) of the wild-type YB-1 gene were amplified by PCR using placenta DNA and the primer pair 5′-AGATCTCTATCACGTGGCTGTTGC-3′ and 5′-AAGCTTATCAGTCCTCCATTCTCATTGG-3′. PCR was also done to introduce the E-box mutations into the YB-1 promoter using the following primer pairs: MT, 5′-AGATCTCTATCAGCTGGCTGTTGC-3′ and 5′-AAGCTTATCAGTCCTCCATTCTCATTGG-3′; MM, 5′-AGATCTCTATACCGGTGCTGTTGC-3′ and 5′-AAGCTTATCAGTCCTCCATTCTCATTGG-3′; M1, 5′-AGATCTCTATACCGTGGCTGTTGC-3′ and 5′-AAGCTTATCAGTCCTCCATTCTCATTGG-3′; and M2, 5′-AGATCTCTATACACGGTCTGTTGC-3′ and 5′-AAGCTTATCAGTCCTCCATTCTCATTGG-3′. Underlined nucleotides indicate a wild-type E-box and mutated E-boxes. These PCR products were cloned and ligated into the BglII-HindIII sites of the pGL3 basic vector (Promega).

Western blot analysis. Western blot analysis was done as previously described (4, 5) with antibodies for STAT3 (1:1,000), Clock (1:1,000), Twist (1:500), Flag (M2; 1:5,000), and β-actin (1:10,000).

Knockdown analysis using small interfering RNAs. The following double-stranded RNA 25-bp oligonucleotides were commercially generated (Invitrogen): Clock small interfering RNA (siRNA), 5′-UAAAGUCUGUUGUUGUAUCAUGUGC-3′ (sense) and 5′-GCACAUGAUACAACAACAGACUUUA-3′ (antisense); Twist siRNA #1, 5′-UUGAGGGUCUGAAUCUUGCUCAGCU-3′ (sense) and 5′-AACUGAGCAAGAUUCAGACCCUCAA-3′ (antisense); Twist siRNA #2, 5′-UUGAGGGUCUGAAUCUUGCUCAGCU-3′ (sense) and 5′-AGCUGAGCAAGAUUCAGACCCUCAA-3′ (antisense); YB-1 siRNA #1, 5′-AAAGCAAGCACUUUAGGUCUUCAGC-3′ (sense) and 5′-GCUGAAGACCUAAAGUGCUUGCUUU-3′ (antisense); and YB-1 siRNA #2, 5′-UUUGCUGGUAAUUGCGUGGAGGACC-3′ (sense) and 5′-GGUCCUCCACGCAAUUACCAGCAAA-3′ (antisense). PC3 or MCF7 cells (1 × 10⁶) were transfected with siRNA as previously described (4, 5). Aliquots of 2.5 × 10⁴, 2.5 × 10⁵, 3 × 10², or 2.5 × 10³ cells were used in cell proliferation assays, flow cytometry, colony formation assays, or WST-8 assay, respectively, as described in the following sections. The remaining cells were seeded in 100-mm dishes with 10 mL of culture medium and harvested after 72 h of culture for Western blot analysis as described earlier.

Luciferase assay. Transient transfection and a luciferase assay were done. Cells were cotransfected with the indicated amounts of YB-1 reporter plasmid and expression plasmids using Superfect reagent (Qiagen); 48 h posttransfection, a luciferase assay was done as previously described (4, 5). The results shown are normalized to protein concentration measured using the Bradford method and are representative of at least three independent experiments.

Chromatin immunoprecipitation assay. The chromatin immunoprecipitation assay was done as previously described (4, 5). Briefly, PC3 cells were transiently transfected with Flag or Flag-Twist plasmid and cultured for 48 h as described above. Soluble chromatin from 1 × 10⁶ cells was incubated with 2 μg of anti-Flag (M2) affinity gel or antimouse immunoglobulin G (IgG). Purified DNA was dissolved in 20 μL of distilled water and 2 μL of DNA were used for PCR analysis with the following primer pairs: YB-1 #1 (−2,112 to −1,730), 5′-ATTAAGCAGGCAAAGAGGAAGG-3′ (forward) and 5′-GTGGTTTACTGGACTCTATGAC-3′ (reverse); YB-1 #2 (−1,708 to −1,329), 5′-TAAGGGTAATAGTAGTCACTGG-3′ (forward) and 5′-TACTGTAGGCCATCATCACC-3′ (reverse); YB-1 #3 (−892 to −422), 5′-AGCCCTCCACCTTCTCCCTGC-3′ (forward) and 5′-CTATGGCAGCCCGGGTTCAGC-3′ (reverse); YB-1 #4 (−664 to +77), 5′-AGATCTCTATCACGTGGCTGTTGC-3′ (forward) and 5′-AAGCTTATCAGTCCTCCTCCATTCTCATTGG-3′ (reverse); YB-1 #5 (+409 to +936), 5′-GCCCGGCACTACGGGCTGCG-3′ (forward) and 5′-GTGTGCGCAGGCCGCGGACG-3′ (reverse); YB-1 #6 (+1,336 to +1,923), 5′-AAGGCGTTTACTACCTCTGG-3′ (forward) and 5′-CTAATAAGCTACAGCCAGGG-3′ (reverse); and Rad51 promoter region (−471 to +221), 5′-AGATCTGCGATGGTGAGAACTCGCGGACC-3′ (forward) and 5′-AAGCTTCACCCCGCGGGCGTGGCACG-3′ (reverse). PCR products were separated by electrophoresis on 2% agarose gels and stained with ethidium bromide.

Cytotoxicity analysis. The water-soluble tetrazolium salt (WST-8) assay was done as previously described (4, 5). PC3 or MCF7 cells (2.5 × 10³) transfected with the indicated amounts of siRNA were seeded in 96-well plates. The following day, the indicated concentrations of the drugs were applied. After 72 h, the surviving cells were stained with TetraColor ONE (Seikagaku Corporation) for 90 min at 37°C. The absorbance was then measured at 450 nm.

Cell proliferation assay. PC3, MCF7, or human dermal fibroblast cells (2.5 × 10⁴) were seeded in 12-well plates and transfected with siRNA as described above. Twelve hours after transfection was set to 0 h. The cells were harvested with trypsin and counted daily with a Coulter-type cell size analyzer (CDA-500, Sysmex).

Flow cytometry. PC3 or MCF7 cells (2.5 × 10⁵) were seeded in six-well plates, transfected with siRNA, and cultured for 72 h. The cells were harvested, washed twice with ice-cold PBS with 0.1% bovine serum albumin (BSA), and resuspended in 70% ethanol. After washing twice with ice-cold PBS, cells were resuspended in PBS with 0.1% BSA, incubated with RNase (Sigma), and stained with propidium iodide (Sigma). Cells were analyzed using an EpicsXL-MCL flow cytometer (Beckman Coulter).

Colony formation assay. For colony formation assay, 3 × 10² PC3 cells transfected with the indicated amounts of siRNA were seeded in 35-mm dishes with 2-mL culture medium. Twelve hours later, transfection of indicated expression plasmid was done. Seven days posttransfection, the number of colonies was counted following 2% Giemsa staining.

Statistical analysis. The t test was used for statistical analysis and significance was set at the 5% level.

Twist is up-regulated in cisplatin-resistant cells and regulates YB-1 gene expression. We have previously shown that YB-1 is overexpressed in cisplatin-resistant cells. Because several E-boxes are found in the promoter region of the YB-1 gene, we first examined the level of expression of E-box binding proteins. A number of transcription factors have been shown to bind to E-boxes. The expression of the circadian transcription factor Clock, but not that of c-Myc or the upstream stimulatory factor 1, was up-regulated in cisplatin-resistant cells (4). Initially, we examined whether Clock is responsible for the transcriptional regulation of the YB-1 gene. However, YB-1 expression was unaffected by the down-regulation of Clock (Fig. 1A). Then, the expression of another E-box binding protein, Twist, was examined. We found that Twist expression was significantly up-regulated in two cisplatin-resistant cells that were independently isolated (Fig. 1B). This prompted us to test the hypothesis that Twist might affect YB-1 gene expression. We verified the existence of a relationship between YB-1 expression and Twist expression using two different siRNAs targeting different regions of Twist. Inactivation of Twist was shown to suppress YB-1 expression (Fig. 1C).

To determine whether the YB-1 expression was due to E-box–dependent transcriptional activation, we analyzed the activity of the YB-1 promoter region as well as several constructs containing E-box mutations (Fig. 2A). PC3 cells ectopically expressing Twist showed an ∼4-fold increase in luciferase activity driven by YB-1-WT-Luc containing a wild-type E-box (CACGTG; Fig. 2B). In addition, PC3 cells cotransfected with Twist expression plasmids and the YB-1-MT-Luc construct containing a mutated E-box (CAGCTG) also displayed a 4-fold higher level of luciferase activity. However, three E-box mutants (MM, M1, and M2) did not respond to Twist expression. These data indicate that E-box sequences (i.e., 5′-CANNTG-3′) contribute to the activation of transcription by Twist, which is also supported by previous report (14). To determine if the transcription activation of Twist on the YB-1 promoter was a result of direct recruitment of Twist to the promoter, we carried out chromatin immunoprecipitation assays covering different segments of the promoter and 5′ end of the YB-1 gene. We tried this assay using a commercial antibody against Twist, but it did not work well. Subsequently, transient transfection of human prostate cancer PC3 cells with a Flag-tagged Twist expression plasmid was used. Forty-eight hours after transfection, a chromatin immunoprecipitation assay was done with an anti-Flag antibody. When control primers from irrelevant Rad51 gene promoter were used, no binding of Twist was observed in Rad51 gene promoter (Fig. 2C). As shown in Fig. 2C, we found that Twist was present in three segments; −1,708/−1,329, −892/−422, and −664/+77 bp. These results indicate that E-boxes in the proximal promoter region are the target of Twist binding in vivo.

Knockdown of Twist expression sensitizes cells to cisplatin. It has been shown that YB-1 expression is involved in cellular sensitivity to DNA-damaging agents (2, 7). To determine whether Twist also contributes to the cytotoxic effect of anticancer agents, cells were initially exposed to a range of concentration of anticancer agents in the presence or absence of siRNA against Twist. Knockdown of Twist expression sensitizes cells to cisplatin but not to etoposide and 5-fluorouracil (5-FU; Fig. 3).

Down-regulation of both Twist and YB-1 induces G₁ arrest and decreases cell proliferation. To get further insight into the biological relevance of both Twist and YB-1, we analyzed their effect on the proliferation of PC3 and MCF7 cells. When Twist expression was knocked down by transfection of cells with a Twist-specific siRNA, growth retardation was induced in both PC3 and MCF7 cells (Fig. 4A). Similar results were observed when YB-1 expression was down-regulated (Fig. 4B). To understand the implications of this growth retardation, cells were treated with siRNA and assayed for DNA content at 72 h posttransfection by laser scanning flow cytometry (Fig. 4C and D). The silencing of either Twist or YB-1 expression led to a significant and reproducible increase in the proportion of PC3 and MCF7 cells in the G₁ fraction. Furthermore, we observed a substantial increase in the proportion of cells in the sub-G₁ fraction. These data are consistent with previous reports (12, 15). Thus, both Twist and YB-1 are essential mediators of growth regulation in cancer cells. Further, we used primary human fibroblasts to examine the molecular basis of both YB-1 and Twist in cellular proliferation. Murine embryonic fibroblasts from YB-1–null mouse embryo showed greatly reduced proliferation (6). Western blot analysis showed that Twist and YB-1 protein levels were significantly decreased on transfection of siRNAs targeting Twist and YB-1, respectively (Fig. 5A). Knockdown of Twist again decreased the YB-1 expression in primary fibroblasts. We found that knockdown of Twist and YB-1 resulted in the inhibition of cellular proliferation (Fig. 5B and C).

YB-1 is a major downstream target of Twist. To assess whether YB-1 expression modulates the growth of cancer cell that was transfected with siRNA for Twist, we then carried out colony formation assays. About 20% to 30% reduction of colony forming activity was observed in the cells treated with either YB-1 or Twist siRNA (Fig. 6A and B). Cotransfection of YB-1 expression plasmid increased the colony number to ∼90% of that in the control. Thus, YB-1 reexpression almost completely rescued Twist siRNA–induced growth arrest. However, cotransfection of Twist expression plasmid could not rescue YB-1 siRNA–induced growth arrest. Western blot analysis confirmed that siRNA and expression plasmids functioned properly in these experiments (Fig. 6C). These results suggest that YB-1 is a major downstream target of Twist and necessary for cell growth.

The Y-box–binding protein family contains a cold shock domain, which is a highly conserved nucleic acid binding domain (8, 9). Y-box proteins are multifunctional and are involved in both transcription and translation (8, 9, 16). YB-1 promotes cell proliferation through the transactivation of target genes such as proliferating cell nuclear antigen, epidermal growth factor receptor, DNA topoisomerase II, thymidine kinase, and DNA polymerase α (1). We have previously reported that YB-1 functions as a positive transcription factor to activate human multidrug resistance 1 in response to various environmental stimuli (8, 9). YB-1 is overexpressed in cisplatin-resistant cells (2, 8, 9) and is involved in brain development (6). The transcriptional regulation of Y-box proteins is poorly understood. We have also characterized the structure and function of the promoter of two Y-box proteins, YB-1 (10) and Contrin/dbpC. Interestingly, E-boxes are found in the promoter regions of both genes. Among E-box–binding proteins, Clock has been found to be overexpressed in cisplatin-resistant cells (4). However, YB-1 expression is not affected by the knockdown of Clock expression with siRNA (Fig. 1A). We found that another E-box binding protein, Twist, is overexpressed in cisplatin-resistant cells (Fig. 1B) and positively regulates YB-1 expression (Fig. 1C). Both reporter and chromatin immunoprecipitation assays clearly showed that Twist positively regulated YB-1 gene expression (Fig. 2). Using whole-genome oligonucleotide microarrays, Twist was also found to be overexpressed in cisplatin-resistant ovarian cancer cells (17). Functional analysis of Twist provides considerable insight into the epigenetics of cisplatin-resistance. Twist depletion in cells using siRNA confers sensitivity to cisplatin but not to etoposide and 5-FU (Fig. 3). These results correspond to the previous reports that YB-1 depletion in cells confers sensitivity to DNA-damaging agents (2, 7).

Twist has been suggested to be a potential oncogene interfering with p53-related pathways (11). We have shown that siRNA-mediated YB-1 knockdown results in reduced cell growth (6). Inactivation of Twist expression was sufficient to provoke massive apoptosis in neuroblastoma cell lines (15). We have also found that siRNA-mediated Twist knockdown results in complete inhibition of the proliferation not only of cancer cell lines but also of primary cultured cells (Figs. 4 and 5). Fluorescence-activated cell sorting (FACS) analysis also showed that silencing of Twist expression by siRNA induces G₁ arrest and apoptosis (Fig. 4C and D). Further, cotransfection of YB-1 expression plasmid can rescue the Twist siRNA–induced growth arrest (Fig. 6A and B), suggesting that YB-1 is a major downstream target gene of Twist. Twist was discovered as a Drosophila gene whose mutation causes the TWISTED phenotype in embryos (18). In humans, Twist germ-line mutations and reduced Twist expression are responsible for Saethre-Chotzen syndrome (19). Interestingly, we have reported that YB-1 is required for cranial neural tube morphogenesis in mice (6). Like YB-1–null mice, Twist-null embryos also exhibit failure of neural tube closure (20), suggesting that YB-1 might be a major target gene of Twist. We have also confirmed that there are several E-boxes in the promoter region of the mouse YB-1 gene (data not shown), suggesting that Twist also regulates YB-1 expression in mice. In clinical studies, the level of YB-1 expression has been shown to correlate with tumor growth and prognosis (8, 9). Although there is no sufficient evidence to indicate the role of Twist in human cancers, it will be important to examine whether Twist expression correlates with YB-1 expression. The serine/threonine kinase AKT/protein kinase B has been shown to play a key role in cell proliferation. AKT2 is a Twist target gene and a functional mediator of Twist as well as YB-1 (21). AKT1 has recently been shown to promote YB-1 phosphorylation and nuclear translocation (22). Based on these reports, Twist can regulate the expression of both YB-1 and AKT2 and might be involved in promoting cell proliferation, migration, and invasion. Twist regulates epithelial-mesenchymal transition and is involved in tumor metastasis. It has been shown that inhibition of Twist expression in highly metastatic cells suppresses their metastatic ability (23). This indicates that Twist expression might correlate with malignant progression and prognosis.

In summary, we have identified that YB-1 is a major target gene of Twist and that YB-1 expression is required for tumor cell growth. This functional link between YB-1 and Twist suggests that Twist and its downstream effector YB-1 will provide promising molecular targets for the treatment of various human cancers.

Grant support: The Ministry of Education, Culture, Sports, Science, and Technology of Japan (MEXT); Kakenhi grants 13218132 and 17590257; Kobayashi Institute for Innovative Cancer Chemotherapy; and a Grant-in-Aid for Cancer Research from the Fukuoka Cancer Society, Japan.

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 Satoko Takazaki and Yukiko Yoshiura for their technical assistance.