Article Text

Original article
Overexpression of EIF5A2 promotes colorectal carcinoma cell aggressiveness by upregulating MTA1 through C-myc to induce epithelial–mesenchymaltransition
  1. Wei Zhu1,2,
  2. Mu-Yan Cai1,3,
  3. Zhu-Ting Tong1,2,
  4. Sui-Sui Dong4,
  5. Shi-Juan Mai1,2,
  6. Yi-Ji Liao1,2,
  7. Xiu-Wu Bian5,
  8. Marie C Lin1,6,
  9. Hsiang-Fu Kung1,6,
  10. Yi-Xin Zeng1,2,
  11. Xin-Yuan Guan1,4,
  12. Dan Xie1,2
  1. 1State Key Laboratory of Oncology in South China, Cancer Center, Sun Yat-Sen University, Guangzhou, China
  2. 2Departments of Experimental Research, Cancer Center, Sun Yat-Sen University, Guangzhou, China
  3. 3Department of Pathology, Cancer Center, Sun Yat-Sen University, Guangzhou, China
  4. 4Department of Clinical Oncology, the University of Hong Kong, Hong Kong, China
  5. 5Institute of Pathology and Southwest Cancer Center, Southwest Hospital, Third Military Medical University, Chongqing, China
  6. 6State Key Laboratory of Oncology in South China, the Chinese University of Hong Kong, Hong Kong, China
  1. Correspondence to Dr Dan Xie, State Key Laboratory of Oncology in South China, Cancer Center, Sun Yat-Sen University, No 651, Dongfeng Road East, 510060 Guangzhou, China; xied{at}mail.sysu.edu.cn

Abstract

Background and Aims The authors have previously isolated a putative oncogene, eukaryotic initiation factor 5A2 (EIF5A2) from 3q26. In this study, EIF5A2 was characterised for its role in colorectal carcinoma (CRC) aggressiveness and underlying molecular mechanisms.

Methods The expression dynamics of EIF5A2 were examined by immunohistochemistry in a cohort of carcinomatous and non-neoplastic colorectal tissues and cells. A series of in-vivo and in-vitro assays was performed to elucidate the function of EIF5A2 in CRC and its underlying mechanisms.

Results The overexpression of EIF5A2 was examined by immunohistochemistry in 102/229 (44.5%) CRC patients, and it was significantly correlated with tumour metastasis and determined to be an independent predictor of shortened survival (p<0.05). Ectopic overexpression of EIF5A2 in CRC cells enhanced cell motility and invasion in vitro and tumour metastasis in vivo, and induced epithelial–mesenchymal transition (EMT). The depletion of EIF5A2 expression prevented CRC cell invasiveness and inhibited EMT. Importantly, the metastasis-associated protein 1 (MTA1) gene was identified as a potential downstream target of EIF5A2 in CRC cells, and knockdown of MTA1 eliminated the augmentation of carcinoma cell migration, invasion and EMT by ectopic EIF5A2. The overexpression of EIF5A2 in CRC cells substantially enhanced the enrichment of c-myc on the promoter of MTA1, and MTA1 upregulation by EIF5A2 was partly dependent on c-myc.

Conclusion The data suggest that EIF5A2 plays an important oncogenic role in CRC aggressiveness by the upregulation of MTA1 to induce EMT, and EIF5A2 could be employed as a novel prognostic marker and/or effective therapeutic target for CRC.

  • Colorectal carcinoma
  • EIF5A2
  • epithelial–mesenchymal transition
  • metastasis
  • MTA1

Statistics from Altmetric.com

Significance of this study

What is already known about this subject?

  • Previously, we have isolated a novel oncogene, EIF5A2 from 3q26 in ovarian carcinoma. Amplification of EIF5A2 has been detected in human CRC and overexpression of EIF5A2 has been documented to be associated with CRC metastasis.

What are the new findings?

  • Overexpression of EIF5A2 in CRC is important in the acquisition of an aggressive and/or poor prognostic phenotype.

  • Ectopic overexpression of EIF5A2 in CRC cells can enhance both cell motility and invasion in vitro and tumour metastasis in vivo, and induces EMT.

  • Overexpression of MTA1 in CRC patients is positively correlated with EIF5A2 expression, and knockdown of MTA1 in CRC cells does eliminates the augmentation of cell migration, invasion and EMT by ectopic EIF5A2. The upregulation of MTA1 by EIF5A2 in CRC cells is partly dependent on c-myc.

How might it impact on clinical practice in the foreseeable future?

  • Our data highlight an important oncogenic role of EIF5A2 in the promotion of the aggressive nature of human CRC, and suggest that EIF5A2 could be employed as a novel prognostic marker and/or as an effective therapeutic target for CRC.

Colorectal carcinoma (CRC) is one of the leading causes of cancer death in developed countries.1 In China, CRC is the fifth leading cause of cancer-related death and the incidence continues to increase.2 Clinically, approximately 60% patients with newly diagnosed CRC develop metastases.3 The main cause of death in CRC patients is tumour metastasis, but the underlying molecular mechanisms responsible for the development of metastasis are still not fully understood. It has been suggested that CRC cell dissemination is the result of a complex progression regulated by universal metastatic mediators encompassing extracellular matrix remodelling, invasion through tissue boundaries and seeding of target organs.4 5 This vulnerability of metastatic cascade, including a coordinated series of rate-limiting steps,5 predicts that the disruption of one or more of these underlying mechanisms may greatly improve the prognosis of patients with CRC.

Amplification of 3q is one of the most frequent chromosomal aberrations in CRC6 7 and many other human cancers including ovarian,8 lung,9 oesophageal,10 gastric,11 cervical12 and prostate13 carcinomas. These imply that human chromosome 3q contains oncogenes related to the tumorigenesis and/or progression of human cancers. Previously, we have isolated a novel oncogene, eukaryotic initiation factor 5A2 (EIF5A2) from 3q26, one of the most frequent chromosomal alterations in ovarian carcinoma.14 15 Further function studies demonstrated an oncogenic role of EIF5A2 in tumorigenesis.15 We have also recently found that the overexpression of EIF5A2 is associated with tumour metastasis in colorectal16 and hepatocellular carcinomas (HCC)17 and poor prognosis in ovarian18 and bladder cancers.19 To date, however, the impact of EIF5A2 expression on CRC patient survival and its potential oncogenic role and molecular mechanisms in CRC has not been elucidated. In the present study, to investigate whether abnormalities of EIF5A2 are involved in the pathogenesis of CRC, the protein expression dynamics of EIF5A2 were first examined in a series of carcinomatous and non-neoplastic human colorectal tissues and cells. The clinical/prognostic significance of EIF5A2 expression in our CRC cohort was assessed. In addition, the tumorigenicity of EIF5A2 in CRC and the underlying molecular mechanisms involving the oncogenic role of EIF5A2 were investigated.

Herein, we report that overexpression of EIF5A2 in CRC is important in the acquisition of an aggressive and/or poor prognostic phenotype. Knockdown of EIF5A2 in CRC cells inhibits cell migration and invasion in vitro, whereas overexpression of EIF5A2 is sufficient to promote CRC cell invasive and/or metastatic capacity both in vitro and in vivo. More importantly, we demonstrate, for the first time, that EIF5A2 induces CRC cell epithelial–mesenchymal transition (EMT) by the upregulation of metastasis-associated protein 1 (MTA1), ultimately leading to the enhanced invasiveness of cancer cells. Upregulation of MTA1 by EIF5A2 in CRC cells is partly dependent on c-myc. Our results provide functional and mechanistic links between the putative oncogene EIF5A2 and EMT in the aggressive nature of CRC.

Materials and methods

CRC cell lines and cell cultures

Six CRC cell lines (ie, LOVO, SW480, DLD1, HT29, HCT116 and SW1116) were selected and cultured in this study. See supplementary materials and methods (available online only) for details.

Patients and TMA

In this study, the paraffin-embedded pathological specimens from 229 patients with CRC were obtained from the archives of the Department of Pathology, Cancer Center, Sun Yat-Sen University and Guangdong Provincial People's Hospital, Guangzhou, China, between January 2000 and November 2006. The cases selected were based on a distinctive pathological diagnosis of CRC, undergoing primary and curative resection for CRC, availability of resection tissue, follow-up data, and had not received preoperative anticancer treatment.These CRC cases included 142 (62.0%) men and 87 (38.0%) women, with mean age of 57.3 years. Average follow-up time was 55.42 months (median 60.0 months; range 0.5–98 months). Patients whose cause of death remained unknown were excluded from our study. Clinicopathological characteristics for these patients were detailed in table 1. The tissue microarray (TMA) was constructed according to a method described previously.20 In this CRC TMA, three cores of sample were selected from each tumour or normal tissue. In addition, 10 pairs of fresh CRC and adjacent normal colorectal mucosa specimens were collected in 2010. Tumour grades were defined in accordance with the criteria of the WHO (2000). The pathological TNM status of all CRC was assessed according to the criteria of the sixth edition of the TNM classification of the International Union Against Cancer (2002). The Institute Research Medical Ethics Committee of Sun Yat-Sen University granted approval for this study.

Table 1

Correlation between expression of EIF5A2 and clinicopathological features in 229 cases of CRC

Western blotting

See supplementary materials and methods (available online only) for details.

Immunohistochemistry and selection of cut-point score

See supplementary materials and methods (available online only) for details.

RNA interference

Short interfering RNA specifically against EIF5A217 and MTA121 genes and corresponding scrambled siRNA (Ambion, Austin, Texas, USA) were transfected into CRC cells in six-well plates using Lipofectamine 2000 reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions. The gene silencing effect was measured by western blotting 48 h post-transfection.

Plasmid constructs and transfection

The construction of a mammalian plasmid expressing EIF5A2 (pcDNA-EIF5A2) was described previously.15 The MTA1 complementary DNA (Fulengen, Guangzhou, China) was cloned into pcDNA3.1 plasmid. Cells were transfected with pcDNA-EIF5A2/pcDNA-MTA1 or the control plasmid pcDNA3.1(+) using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. For the establishment of the EIF5A2-SW1116 cell line stably expressing EIF5A2, 48 h after transfection, the cells were split at a ratio of 1:10. Next, cells were maintained in Leibovitz's L-15 medium containing 200 µg/ml of G418 (Calbiochem, San Diego, CA, USA). After 6 weeks of selection, resistant colonies stably transfected with pcDNA-EIF5A2 (SW1116 pcDNA-EIF5A2) or pcDNA3.1(+) [SW1116 pcDNA3.1(+)] were pooled.

Wound-healing and cell invasion assays

See supplementary materials and methods (available online only) for details.

Immunofluorescence staining

See supplementary methods (available online only) for details.

Experimental in-vivo metastasis model

Eight 4-week-old male severe combined immunodeficient (SCID-Beige) mice in each experimental group were injected with SW1116-EIF5A2 and SW1116-Vec cells separately. Briefly, 2×105 cells were injected intravenously through the tail vein into each SCID mouse in a laminar flow cabinet. Six weeks after cell injection, mice were killed and examined. Animals were housed under standard conditions and cared for according to the institutional guidelines for animal care. All the procedures are in accordance with the guidelines of the laboratory animal ethics committee of Sun Yat-Sen University.

Rho-GTPase activation assay

See supplementary methods (available online only) for details.

Real-time PCR gene array

RNA was extracted from SW1116-EIF5A2 and SW1116-Vec using Trizol (Invitrogen) and we cleaned them using the RNeasy MinElute cleanup kit (Qiagen, Valencia, California, USA). Subsequently, total RNA was reverse transcribed using SuperScript III reverse transcriptase (Invitrogen) and cDNA was amplified by PCR using 2× Super Array PCR master mix (SuperArray Bioscience, Frederick, Maryland, USA). Real-time PCR was then performed on each sample using the Human Tumor MetastasisRT2 Profiler PCR array (Super Array Bioscience, Frederick, MD, USA) in an Opticon DNA Engine ABI PRISM7900 system (Applied Biosystems, Foster City, California, USA), according to the manufacturer's instructions. Data were normalised for GAPDH levels by the ΔΔCt method.

Chromatin immunoprecipitation assays

See supplementary methods (available online only) for details.

Dual luciferase reporter assay

See supplementary methods (available online only) for details.

Statistical analysis

Statistical analysis was performed using the SPSS statistical software package (standard version 13.0). Receiver operating characteristic (ROC) curve analysis was applied to determine the cut-off scores for the overexpression of EIF5A2 and MTA1 and low-level expression of CTNNA1.22 The correlation between EIF5A2 expression and clinicopathological features of CRC patients was analysed by the χ2 test or Fisher's exact test. For univariate survival analysis, survival curves were obtained using the Kaplan–Meier method. The Cox proportional hazards regression model was performed for multivariate survival analysis. The independent Student's t test was performed to analyse the statistical significance between two preselected groups. p Values of less than 0.05 were considered statistically significant.

Results

Western blotting analysis of EIF5A2 expression in CRC cells and colorectal tissues

Of the six CRC cell lines analysed by western blotting, endogenous EIF5A2 expression was examined in two lines (ie, LOVO and SW480), whereas the other four lines (ie, DLD1, HT29, HCT116 and SW1116) showed undetectable or very low levels of endogenous EIF5A2 (figure 1A, left). In primary CRC tissues, six out of 10 (60%) had upregulated EIF5A2 expression, when compared with adjacent normal colorectal mucosal tissues (figure 1A, right).

Figure 1

Expression of eukaryotic initiation factor 5A2 (EIF5A2) in colorectal carcinoma (CRC) cell lines and colorectal tissues and its prognostic significance in CRC patients. (A) Left: the levels of EIF5A2 protein in six CRC cell lines examined by western blotting. Right: EIF5A2 expression was upregulated in three of the four CRC tissues compared with paired normal colorectal mucosal tissues, N, normal tissue; T, carcinoma tissue. (B) Receiver operating characteristic curve analysis was employed to determine the cut-off score for the overexpression of EIF5A2. The sensitivity and specificity for each clinical outcome were plotted. (C) Representative immunohistochemistry images show the overexpression of EIF5A2 in a CRC tissue (left) and normal expression of EIF5A2 in adjacent normal colorectal mucosal tissue (right). (D) Kaplan–Meier survival analysis according to EIF5A2 expression in 229 patients with CRC (log-rank test). Probability of survival of patients: normal expression of EIF5A2, n=127; overexpression of EIF5A2, n=102 (p<0.0001).

Immunohistochemical staining of EIF5A2 expression in colorectal tissues and its correlation with CRC patients' clinicopathological features and survival

To investigate the expression status of EIF5A2 in CRC, we conducted immunohistochemical staining for EIF5A2 on a CRC TMA containing 229 pairs of CRC specimens and corresponding normal colorectal mucosal tissues. According to the ROC curve (figure 1B), overexpression of EIF5A2 was examined in 102 of 229 (44.5%) primary CRC compared with only 12 of 229 (5.2%) normal mucosal tissues (p<0.0001, Fisher's exact test). Immunohistochemical staining of EIF5A2 in representative samples of CRC and normal colorectal mucosal tissues is shown in figure 1C.

Correlation analysis demonstrated that the overexpression of EIF5A2 was positively correlated with CRC lymph node and/or distant metastasis and a more advanced clinical stage (p<0.05, table 1). Kaplan–Meier analysis showed that the mean survival time for patients with CRC having an overexpression of EIF5A2 was 64.1 months compared with 83.3 months for patients with CRC having a normal expression of EIF5A2 (p<0.0001, log-rank test, figure 1D, table 2). Further multivariate Cox regression analysis determined that EIF5A2 expression is an independent prognostic factor for the poor survival of CRC patients (RR 2.241, CI 1.227 to 4.095, p=0.009, table 2).

Table 2

Univariate and multivariate analysis of different prognostic parameters in 229 patients with CRC

The expression levels of EIF5A2 influenced the invasive capacity of CRC cell lines in vitro

As the overexpression of EIF5A2 examined by immunohistochemistry was positively associated with CRC metastasis and/or ascending clinical stage, to investigate the impact of EIF5A2 on CRC cell invasiveness, two CRC cell lines, LOVO and SW480, were treated with two specific siRNA against EIF5A2 (siRNA-1 and siRNA-2). Both siRNA could efficiently knock down endogenous EIF5A2 in CRC cells (figure 2A). The results showed that knockdown of EIF5A2 caused an apparent suppression of cell migration in both LOVO and SW480 cell lines using a wound-healing assay (p<0.01, figure 2B). Matrigel invasion assays also demonstrated that ablation of endogenous EIF5A2 markedly reduced the invasive ability of both LOVO and SW480 cell lines (p<0.05, figure 2C).

Figure 2

Silence of eukaryotic initiation factor 5A2 (EIF5A2) by RNA interference inhibits colorectal carcinoma cell migration and invasion. (A) Western blotting reveals that EIF5A2 was efficiently knocked down by the treatment of EIF5A2-siRNA-1 or EIF5A2-siRNA-2. (B) Wound-healing assays show that EIF5A2-silenced LOVO and SW480 cells had lower motility compared with that in control cells. Data are the means±SE of three independent experiments. *p<0.05, **p<0.01 by Student's t test. (C) Cell invasion was evaluated using a matrigel invasion chamber. Silence of EIF5A2 decreased LOVO and SW480 cell invasive capacity. The numbers of invaded cells in siEIF5A2 and control siSCR groups are shown in the right panel. Error bars indicate ±SE. *p<0.05, **p<0.01 by Student's t test.

To determine whether ectopic overexpression of EIF5A2 could enhance the invasive capacity of CRC cells, we constructed a SW1116-EIF5A2 cell line, which overexpressed EIF5A2 (figure 3A right). Wound-healing assay demonstrated that ectopic EIF5A2 enhanced SW1116 cell migration at the edge of exposed regions (p<0.05, figure 3A). The matrigel invasion assay showed that SW1116-EIF5A2 cells had significantly increased invasive capacity, compared with our control SW1116-vector cells (p<0.05, figure 3B). Collectively, these results provide evidence that elevated expression levels of EIF5A2 are important for the aggressive phenotype of CRC cells.

Figure 3

Ectopic overexpression of eukaryotic initiation factor 5A2 (EIF5A2) enhances colorectal carcinoma SW1116 cell migration and invasion in vitro. (A) Right: ectopic expression of EIF5A2 was substantially increased in SW1116-EIF5A2 cells compared with that in SW1116-vector cells by western blotting (upper panel). Left: representative results of wound-healing assays demonstrate that SW1116-EIF5A2 cells had higher motility than that in SW1116-vector cells. The numbers of migrative cells in SW1116-vector and SW1116-EIF5A2 groups are shown in the right, bottom panel. Error bars indicate ±SE. *p<0.05, **p<0.01 by Student's t test. (B) Ectopic overexpression of EIF5A2 enhanced SW1116 cell invasion in a transwell assay. Columns: mean of triplicate experiments; *p<0.05; independent Student's t test. (C) Overexpression of EIF5A2 promoted SW1116 cell invasion and metastasis in severe combined immunodeficient (SCID-Beige) mice in vivo. Left, representative metastatic nodules on the surface of the liver of SCID mice (indicated by arrows); right, number of metastatic nodules formed in the liver of SCID mice 6 weeks after tail vein injection of SW1116-vector and SW1116-EIF5A2 cells (eight mice per group; *p=0.016; independent Student's t test). (D) Examples of haematoxylin and eosin staining in two liver nodule samples originating from SW1116-vector and SW1116-EIF5A2 cell-injected mice. Left: tumour originating from SW1116-vector cells had a clear boundary between the tumour and non-tumour tissue. Right: representative microsatellite tumour formation was observed in the tumour originating from SW1116-EIF5A2 cells.

Overexpression of EIF5A2 enhanced metastatic potential of CRC cell line in vivo

To assess whether EIF5A2 overexpression is causative in an experimental metastasis model, we injected SW1116-EIF5A2 cells into the tail vein of SCID mice, while SW1116-vector cells were used as a control (eight mice per group). Six weeks after cell injection, mice were killed and metastatic tumour nodules formed in the lung and in the liver were examined. We did not detect tumour nodule formation in the lungs of all mice examined. However, metastatic tumour nodules were frequently found in the livers of mice (figure 3C left, indicated by arrows), and the overexpression of EIF5A2 increased SW1116 cell liver metastasis threefold (p=0.016, figure 3C right). In addition, nodule samples originating from SW1116-EIF5A2-injected mice showed more cancer cells invading the surrounding liver tissue (figure 3D right).

The expression levels of EIF5A2 influenced CRC cell line EMT

As our recent study provided evidence that ectopic expression of EIF5A2 in HCC cells resulted in an enhancement of EMT,17 we asked whether EIF5A2 induces EMT in CRC cells. In this study, after the silence of EIF5A2 in LOVO and SW480 cells, the levels of the two epithelial markers E-cadherin and β-catenin were upregulated, while the levels of the two mesenchymal markers fibronectin and vimentin were downregulated (figure 4A). On the other hand, we observed that after ectopic overexpression of EIF5A2 in SW1116 cells, the expression of E-cadherin and β-catenin decreased, whereas the expression of fibronectin and vimentin increased, as evidenced by both western blotting (figure 4B) and immunofluorescence staining assays (figure 4C). These findings indicate that the expression levels of EIF5A2 influence CRC cells EMT in vitro.

Figure 4

The expression levels of eukaryotic initiation factor 5A2 (EIF5A2) influence colorectal carcinoma cell line epithelial–mesenchymal transition both in vitro and in vivo. (A) Western blotting reveals that knock down of EIF5A2 by siEIF5A2 resulted in an increased expression of epithelial makers (E-cadherin and β-catenin) and a decreased expression of mesenchymal markers (fibronectin and vimentin) in both LOVO and SW480 cells in vitro, compared with that in control siSCR-treated cells. (B) Western blotting assay shows decreased levels of the epithelial markers (E-cadherin and β-catenin) and increased levels of the mesenchymal markers (fibronectin and vimentin) in SW1116-EIF5A2 cells compared with that in SW1116-vector cells. (C) Immunofluorescence staining shows a downregulated expression of E-cadherin and β-catenin and an upregulated expression of vimentin and fibronectin in SW1116-EIF5A2 cells in vitro. (D) Immunohistochemistry staining shows a decreased expression of E-cadherin and β-catenin and an increased expression of vimentin and fibronectin in tumour tissues of the livers of mice originating from SW1116-EIF5A2 cells, compared with that originated from SW1116-vector cells.

In our SCID mouse metastasis model, further immunohistochemical staining demonstrated that the metastatic tumour nodules originating from SW1116-EIF5A2 cells had increased the expression of vimentin and fibronectin and decreased the expression of E-cadherin and β-catenin, compared with that from SW1116-vector cells (figure 4D). These data reveal that the overexpression of EIF5A2 enhanced SW1116 cell EMT in vivo.

The expression levels of EIF5A2 did not affect Rho/Rac Gtpases activation in CRC cells

Previously we found that the overexpression of EIF5A2 in HCC cells could modulate Rho/Rac GTPase activation, and thus promote cancer cell invasiveness.17 However, no apparent differences in the levels of active RhoA, Rac1 and Cdc42 were observed between SW1116-EIF5A2 cells and control SW1116-vector cells (see supplementary figure S1, available online only). In addition, similar levels of active RhoA, Rac1 and Cdc42 were also examined between the EIF5A2 knocked down and control LOVO and SW480 cells (data not shown). These observations virtually excluded the possibility that EIF5A2 modulates Rho/Rac GTPase activation in CRC cells.

EIF5A2 upregulated MTA1 expression in CRC cells

To obtain further insight into the mechanisms of EIF5A2 in CRC cell invasion and/or metastasis, messenger RNA expression profiles of SW1116-EIF5A2 cells were compared with those of control SW1116-vector cells using a human tumour metastasis RT2 profiler PCR array containing 84 cell metastasis-related genes. The results showed that a total of four upregulated (ie, MMP2, IGF1, METAP2 and MTA1) and two downregulated (NR4A3 and CTNNA1) genes, which showed more than a twofold change in mRNA levels, were identified in SW1116-EIF5A2 cells, compared with those in control SW1116-vector cells (figure 5A, table 3). Subsequently, these downstream targets were selected and further validated by a western blotting assay. Consistent with those of mRNA expression in real-time PCR array, increased MTA1 and decreased CTNNA1 in protein levels were examined by western blotting in SW1116 cells after EIF5A2 overexpression (figure 5B).

Figure 5

Eukaryotic initiation factor 5A2 (EIF5A2) upregulated metastasis-associated protein 1 (MTA1) expression in colorectal carcinoma (CRC) cells and tissues. (A) The six genes, MMP2, IGF1, METAP2, MTA1, NR4A3 and CTNNA1, showed more than a twofold mRNA differential expression in SW1116-EIF5A2 cells compared with that in SW1116-vector cells by using a human tumour metastasis RT2 profiler PCR array. (B) Ectopic overexpression of EIF5A2 substantially upregulated MTA1 expression and downregulated CTNNA1 expression in SW1116-EIF5A2 cells detected by western blotting. (C) Overexpression of EIF5A2 and MTA1 was examined by immunohistochemistry in a CRC (case 63).

Table 3

List of genes differentially expressed in CRC SW1116 cells after EIF5A2 overexpression using a human tumour metastasis real-time PCR array

In addition, a significant positive correlation between the overexpression of EIF5A2 and MTA1 was evaluated in our large cohort of CRC tissues (figure 5C; p=0.003, table 4). There were no significant differences in CTNNA1 expression between the EIF5A2 overexpression and normal expression groups (p=0.305, table 4).

Table 4

Association between expression of EIF5A2 and MTA1 and CTNNA1 in CRC

MTA1 is responsible for EIF5A2-induced CRC cell EMT and invasiveness

To investigate whether MTA1 is required for EIF5A2-induced CRC cell EMT and invasiveness, RNA interference was used to silence MTA1 expression in SW1116-EIF5A2 cells. After siMTA1 treatment, EIF5A2-induced EMT was inhibited, as evidenced by the increased expression of epithelial markers (E-cadherin and β-catenin) and the decreased expression of mesenchymal markers (fibronectin and vimentin) in SW1116-EIF5A2 cells (figure 6A,B). In addition, wound-healing and transwell assays showed that the migrative and invasive capacities of SW1116-EIF5A2 cells were all dramatically inhibited after the silencing of MTA1 (figure 6C,D). In contrast, when MTA1 was introduced into EIF5A2-silenced LOVO and SW480 cells, the inhibited CRC cell EMT and migrative/invasive capacities were substantially re-enhanced (see supplementary figure S3, available online only). These data, taken together, provide evidence that MTA1 is responsible for the EIF5A2-induced invasiveness and/or EMT in CRC cells.

Figure 6

Eukaryotic initiation factor 5A2 (EIF5A2)-mediated colorectal carcinoma (CRC) SW1116 cell epithelial–mesenchymal transition, migration and invasion are partly inhibited after silence of metastasis-associated protein 1 (MTA1) or c-myc by specific siRNA. (A) Western blotting shows that after silence of MTA1 or c-myc in SW1116-EIF5A2 cells, the levels of the epithelial markers (E-cadherin and β-catenin) increased, while the levels of the mesenchymal markers (fibronectin and vimentin) decreased. (B) Immunofluorescence staining shows an upregulated expression of E-cadherin and β-catenin and a downregulated expression of vimentin and fibronectin in SW1116-EIF5A2 cells, after knock down of MTA1 or c-myc by specific siRNA. (C) Wound-healing assays show that the enhanced migrative ability in SW1116-EIF5A2 cells was inhibited by silence of MTA1 or c-myc. (D) The invasive ability of SW1116-EIF5A2 cells was dramatically inhibited after siMTA1 or sic-myc treatment in a transwell assay. Data are the means±SE of three independent experiments. *p<0.05, **p<0.01 by Student's t test.

Upregulation of MTA1 by EIF5A2 in CRC cells is partly dependent on c-myc

As a previous study revealed that c-myc could directly regulated MTA1 expression by binding to the promoter of MTA1,23 we wondered whether EIF5A2 upregulated MTA1 expression is mediated by c-myc in CRC cells. As anticipated, the ChIP results demonstrated that the enrichment of c-myc, as well as histone acetyltransferase GCN5, TIP60, acetylated H3 (AC-H3) and acetylated H4 (AC-H4) on the promoter of MTA1, was substantially enhanced in the SW1116-EIF5A2 cell line, concurrent with increased levels of histone H3 and H4 acetylation, compared with that in control SW1116-vector cells (figure 7A). However, we did not observe the altered levels of c-myc, histone acetyltransferase GCN5 and TIP60 after the ectopic overexpression of EIF5A2 (figure 7B). In addition, the enrichment of GCN5, TIP60, AC-H3 and AC-H4 on the MTA1 promoter was virtually reduced when endogenous c-myc was knocked down by siRNA (figure 7C). Furthermore, the dual luciferase reporter assay showed that knocking down c-myc by specific siRNA partly inhibited the transcriptional activity of EIF5A2 on the MTA1 promoter (upregulated ratio 4.15 vs 2.17, figure 7D). Further studies demonstrated that EIF5A2-mediated the upregulation of MTA1 expression, EMT and the migrative/invasive capacity of SW1116-EIF5A2 cells were all prevented, when c-myc was knocked down (figure 6).

Figure 7

Upregulation of metastasis-associated protein 1 (MTA1) by eukaryotic initiation factor 5A2 (EIF5A2) in colorectal carcinoma cells is partly dependent on c-myc. (A) The ChIP assay was performed to analyse the enrichment of c-myc, GCN5, TIP60 and acetylated histone H3 and H4 on the promoter of MTA1 between SW1116-vector and SW1116-EIF5A2 cells. (B) The expression levels of c-myc, GCN5 and TIP60 in SW1116-vector and SW1116-EIF5A2 cells was examined by western blotting. (C) SW1116-EIF5A2 cells were transfected with negative control or si c-myc. Forty-eight hours later, the ChIP experiment was carried out to evaluate the enrichment of c-myc, GCN5, TIP60 and acetylated histone H3 and H4 on the MTA1 promoter. The knocking down efficiency of c-myc was assessed by western blotting (insert). (D) SW1116 cells were first transfected with sic-myc or control siRNA. After 36 h, the cells were co-transfected with the pGL3-MTA1 promoter luciferase, pRL-TK Renilla luciferase construct and control plasmid pcDNA3.1(+) or pcDNA-EIF5A2 plasmid. Thirty-six hours later, the luciferase activity of the MTA1 promoter was measured and normalised relative to luciferase activity (RLA). The bar shows the average ±SD of three independent experiments.

Discussion

Previously, we isolated a novel candidate oncogene EIF5A2 from 3q26 using chromosome microdissection and hybrid selection.14 Our recent study showed that EIF5A2 was frequently amplified and overexpressed in CRC and the overexpression of EIF5A2 may be important in the acquisition of a metastatic phenotype of CRC.16 However, the impact of EIF5A2 expression on CRC patient survival and the role and molecular mechanism by which EIF5A2 regulates CRC cell aggressiveness remain unclear.

In the present study, the protein expression of EIF5A2 was first examined in a series of carcinomatous and non-neoplastic human colorectal tissues. Our results, using western blotting analysis, clearly showed that the majority of the CRC tissues examined had high levels of EIF5A2 expression. Subsequently, the expression dynamics of EIF5A2 were examined by immunohistochemistry, using a CRC TMA with complete follow-up data. The results validated our previous findings that EIF5A2 was frequently overexpressed in CRC tissues16 and it was significantly correlated with tumour metastasis and advanced clinical stage, suggesting that the upregulated expression of EIF5A2 in CRC may facilitate the invasive/metastatic phenotype. Importantly, we further found that the overexpression of EIF5A2 in CRC was a strong and an independent predictor of short cancer-specific survival. The examination of EIF5A2 expression by immunohistochemistry could thus be used as an additional tool in identifying those CRC patients at increased risk of tumour invasion and/or metastasis. These findings underscore a potentially important role of EIF5A2 as an underlying biological mechanism in the development and/or progression of CRC.

The gene EIF5A2 is located at chromosome 3q26.2 and was recently recognised as a novel member of the EIF5A gene family.14 24 EIF5A2 shares 82% identical amino acid sequence with its family member EIF5A1, including the minimum domain needed for EIF5A1 maturation by hypusine modification at the lysine-50 residue.25–27 To date, EIF5A1 and EIF5A2 are the only known proteins that require posttranslational modification by DHPS. DHPS has previously been implicated as one of the metastasis signature genes.28 It is therefore a logical hypothesis that one or both of the proteins may be involved in the pathogenesis of cancer metastasis. To confirm this hypothesis, a series of in-vitro and in-vivo assays, such as wound-healing, Matrigel invasion and experimental metastasis assays, were employed to investigate the role of EIF5A2 in regulating CRC cell motility, invasion and metastasis. The results showed that siRNA mediated EIF5A2 knockdown in the SW480 and LOVO cells, inhibited the ability of either cell migration or invasion and reduced EMT. In contrast, the ectopic overexpression of EIF5A2 in SW1116 cells substantially promoted cell motility and invasiveness and activated EMT. In a tail vein injection mouse model of cancer metastasis, overexpression of EIF5A2 led to a significant increase in the number of lesions of liver metastasis, and also induced SW1116 cells EMT in metastatic lesions in vivo. These data support our emerging view that EIF5A2 is an important factor in CRC cell aggressiveness, which probably undergoes EMT to achieve higher motility and invasiveness.

To date, however, the molecular mechanisms by which EIF5A2 regulates cancer cell migration/invasiveness remain unclear. In a recent study of ours using HCC cells, we suggested that EIF5A2 might stimulate cell cytoskeleton rearrangement through the activation of Rho/Rac GTPase, thus inducing cancer cell invasiveness.17 In this study, however, we did not examine the altered levels of active RhoA, Rac1 and Cdc42 before and after EIF5A2 overexpression or knockdown. It does appear, therefore, that in our CRC cells, EIF5A2 promotes invasiveness by the regulation of targets and/or pathways other than the activation of Rho/Rac GTPase, suggesting that the mechanism(s) by which EIF5A2 regulates cancer progression may be tumour-type specific.

To investigate the downstream molecular events involving EIF5A2 and CRC invasiveness and/or metastasis, we compared mRNA expression profiles between SW1116-EIF5A2 cells and SW1116-vector cells using a human tumour metastasis real-time PCR array, containing 84 well-known cell invasion/metastasis-related genes. Of the 84 genes, six genes were differentially expressed in mRNA levels by twofold or more (ie, upregulated: IGF-1, METAP2, MMP2 and MTA1; downregulated: NR4A3, CTNNA1). Subsequently, upregulated MTA1 was validated in protein levels by western blotting in SW1116-EIF5A2 cells. On the other hand, EIF5A5 knockdown by RNA interference decreased the protein expression of MTA1 in SW480 and LOVO cells (data not shown). Furthermore, we did observe a significant positive correlation between the overexpression of EIF5A2 and MTA1 in our large cohort of CRC tissues. These results, collectively, suggest that in CRC cells, EIF5A2 might regulate cell migration/invasion by the regulation of MTA1.

The MTA1 gene was initially isolated based on its induction in metastatic mammary adenocarcinoma cells.29 During recent years, MTA1 has been demonstrated to be a critical regulator of the invasive and/or metastatic process in a series of different human cancers.30 Previous studies suggested that MTA1 regulates metastatic potential as part of the multiprotein Mi-2/nucleosome remodelling and deacetylating complex by controlling EMT.31 32 In human CRC, greater expression of MTA1 mRNA was frequently observed in tumour tissues compared with the normal counterpart tissue,32 33 and it was significantly correlated with the depth of CRC invasion and lymph node metastasis.33 In our present study, we did observe that the overexpression of MTA1 protein was not only closely associated with CRC metastasis and poor survival (see supplementary table S1 and figure S2, available online only), but was also positively correlated with EIF5A2 expression. To determine if MTA1 is a downstream target involved in EIF5A2-induced CRC cell aggressiveness, we silenced MTA1 by siRNA in SW1116-EIF5A2 cells. We found that the EIF5A2-mediated migrative/invasive capacities and enhanced EMT were dramatically prevented when MTA1 was knocked down. On the other hand, when ectopic MTA1 was overexpressed in EIF5A2-silenced LOVO and SW480 cells, the migrative/invasive and EMT phenotypes of CRC cells were substantially rescued. These data suggest that MTA1 might be a critical downstream target of EIF5A2 and is responsible for the EIF5A2-induced EMT and invasiveness in CRC cells.

To date, however, the mechanisms by which EIF-5A2 regulates MTA1 expression have not been elucidated. Previous study revealed that c-myc can directly bind to the promoter of MTA1 and increases the levels of histone H3 and H4 acetylation to upregulate MTA1 expression.23 Other groups reported that c-myc may recruit two histone acetyltransferases, GCN5 and TIP60, to acetylate H3 and H4, and thus activates the transcription of certain target genes.34–36 In the present study, we further found that in SW1116 CRC cells, the enrichment of c-myc on the promoter of MTA1 was substantially enhanced after the overexpression of EIF5A2, concurrent with increased levels of acetylated H3 and H4. Although we did not observe the altered levels of c-myc, as well as GCN5 and TIP60 after EIF5A2 overexpression, we did find that the recruitment of c-myc, GCN5 and TIP60 to the promoter of MTA1 were all increased. Furthermore, when c-myc was knocked down by siRNA in SW1116-EIF5A2 cells, we observed that the EIF5A2-mediated upregulation of MTA1, enhanced cell EMT and migration/invasion were all prevented. Our data, together with the findings of others,23 34–36 suggest that EIF5A2 might enhance the binding of c-myc to the promoter of MTA1; such could recruit the two histone acetyltransferases GCN5 and TIP60 to acetylate H3 and H4, and ultimately upregulating MTA1 expression to promote the aggressive phenotypes of CRC cells. It is noteworthy here that our luciferase reporter assays showed that silencing of c-myc by siRNA in EIF5A2-SW1116 cells only partly inhibited the transcriptional activity of EIF5A2 on the MTA1 promoter, and EIF5A2-mediated MTA1 upregulation was inhibited partly by knockdown of c-myc (figure 6A). These data imply that besides c-myc, other unknown mechanisms might be involved in the regulation of MTA1 by EIF5A2 in CRC cells. Clearly, further work is needed to clarify the mechanisms of EIF-5A2 regulating MTA1 in detail.

In summary, our report describes the expression pattern of EIF5A2 in human CRC and the overexpression of EIF5A2 may be important in the acquisition of an aggressive/poor prognostic phenotype of the tumour. Furthermore, functional and/or mechanistic studies of EIF5A2, as provided in this report, suggest a critical role of EIF5A2 in the control of cell invasion/metastasis and EMT by regulating MTA1 through c-myc, an activity that might be responsible, at least partly, for the development and/or progression of human CRC.

References

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:

Footnotes

  • See Commentary, p 473

  • WZ, MYC and ZTT contributed equally to this work.

  • Funding This work is supported by grants from the Nature Science Foundation of China (no 30972884), the 973 Project of China (2010CB529400 and 2010CB912802) and the Program for Excellent Young Talents in Sun Yat-Sen University Cancer Center (no 303045134001).

  • Competing interests None.

  • Ethics approval This study was approved by the Institute Research Medical Ethics Committee of Sun Yat-Sen University, Guangzhou, China.

  • Provenance and Peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles