An oncogenic role of eIF3e/INT6 in human breast cancer

Article metrics


Altered expression of the eukaryotic translation initiation factor 3 (eIF3) subunit eIF3e/INT6 has been described in various types of human cancer, but the nature of its involvement in tumorigenesis is not yet clear. Using immunohistochemical analysis of 81 primary breast cancers, we found that high tumor grade correlated significantly with elevated cytoplasmic eIF3e level in epithelial tumor cells. Analysis of protein synthesis after siRNA-mediated knockdown in breast cancer cell lines indicated that eIF3e is not required for bulk translation. Microarray analysis of total and polysomal RNAs nonetheless identified distinct sets of mRNAs regulated either positively or negatively by eIF3e; functional classification of these revealed a marked enrichment of genes involved in cell proliferation, invasion and apoptosis. Validated mRNA targets regulated positively at the translational level by eIF3e included urokinase-type plasminogen activator and apoptotic regulator BCL-XL, whereas synthesis of proteins including the mitotic checkpoint component MAD2L1 was negatively regulated. Finally, eIF3e-depleted breast carcinoma cells showed reduced in vitro invasion and proliferation. Taken together, our study data suggest that eIF3e has a positive role in breast cancer progression. It regulates the translation, and in some cases abundance, of mRNAs involved in key aspects of cancer cell biology.


The overall rate of protein synthesis is an important determinant of cell and tissue function. In addition, alterations in the activities of the translational machinery has key functions in controlling gene-specific expression, through the selective translation of specific subsets of messenger RNAs (Holcik and Sonenberg, 2005; Bramham and Wells, 2007; Paquin and Chartrand, 2008). Deregulation of translation initiation can lead to oncogenic transformation and can support cancer growth, suggesting that targeting translation may be a valuable therapeutic approach (Thumma and Kratzke, 2007; Wendel et al., 2007; Mavrakis and Wendel, 2008). Protein synthesis is controlled by multiple translation initiation factors (eIFs), many of which have been implicated in tumorigenesis (Watkins and Norbury, 2002; Dong and Zhang, 2006). Nonetheless, despite numerous studies of the expression of eIFs in human cancers, a detailed picture of how deregulation of eIF function might modulate the quality of translation in such a way as to contribute to tumor progression is currently lacking.

Translation initiation factor eIF3, comprising 13 protein subunits in human cells, coordinates interactions between mRNA and the 40S ribosomal subunit, governs translation reinitiation and functions as a platform for interactions with other regulatory eIFs (Hinnebusch, 2006). Components of eIF3 have previously been reported to support malignant transformation and tumor growth. Individual overexpression of any of five subunits of human eIF3 (eIF3a, b, c, h or i) promoted malignant transformation of immortal fibroblasts (Ahlemann et al., 2006; Savinainen et al., 2006; Zhang et al., 2007). By contrast, studies of eIF3e/INT6 have suggested various roles for this accessory eIF3 subunit either as an oncoprotein or a tumor suppressor (Marchetti et al., 2001; Rasmussen et al., 2001; Buttitta et al., 2005; Chen et al., 2007).

Here, we describe elevated expression of eIF3e in high-grade breast cancers and identify an eIF3e-regulated subset of messenger RNAs involved in cancer development. Our study data suggest that eIF3e has an oncogenic role in the promotion of breast cancer cell proliferation and invasion.


Elevated eIF3e protein levels in human breast cancer

The murine Int-6 gene, encoding eIF3e, was identified as a site of mouse mammary tumor virus (MMTV) integration in MMTV-induced tumors and a precancerous lesion (Marchetti et al., 1995; Asano et al., 1997). In each case the MMTV provirus was integrated within an Int-6 intron, potentially leading to the production of a C-terminally truncated form of eIF3e. We therefore used western and northern blotting of lysates from cell lines to investigate the possibility that truncated forms of eIF3e might contribute to human breast cancer development. A single eIF3e species was present at both the protein and the RNA levels, and was expressed at a comparable level in all the cell lines tested (Figure 1a), with no detectable expression of truncated forms. To confirm specificity of the eIF3e signals, we transfected cells with two siRNAs against EIF3E; either alone or in combination, their transfection led to substantial eIF3e knockdown as observed by western blotting (Figure 1b) and northern blotting (Figure 1c) analyses. Immunostaining of eIF3e in paraffin-embedded MDA-MB-231 cells was effectively eliminated by EIF3E siRNA transfection (Figure 1d), indicating that our antibody is also suitable for the analysis of eIF3e expression in paraffin sections from primary human breast cancers. Immunohistochemical staining of 81 such cancers showed epithelium-specific, cytoplasmic staining in many tumors (Figure 1e). Interestingly, there was a positive correlation (P=0.001) between expression of eIF3e and tumor grade assessed on the Scarff–Bloom–Richardson scale, which takes into account the degree of tubular differentiation, mitotic index and nuclear pleiomorphism. Grades I–III indicate progression from well differentiated (low grade) to poorly differentiated (high grade); strong cytoplasmic eIF3e staining was observed in 4 of 9 (44%) grade I, in 38 of 45 (84%) grade II and in 26 of 27 (96%) grade III tumors. These data suggest a potential role for eIF3e in human breast cancer progression. Nuclear staining was generally weaker and was present in only 25 of 81 (30%) of the tumors, in line with a primary role for eIF3e in translational regulation; there was no significant correlation between nuclear eIF3e positivity and tumor grade.

Figure 1

Expression of eIF3e in human breast cancer cells. (a) Expression levels of eIF3e protein and EIF3E mRNA were determined by western (upper panels) and northern blotting (lower panels), respectively. Loading controls were provided by staining the western blot with Ponceau S (total protein) and the RNA gel with ethidium bromide (total RNA); the positions of 28S and 18S ribosomal RNAs are indicated. (bd) Specific downregulation of eIF3e expression in MDA-MB-231 cells transfected with one or both of two pairs of EIF3E-specific siRNA oligonucleotides, as indicated (+); the EIF3E 1 duplex was used in c and d. Reversed sequence oligonucleotides were used as controls. (b) Western blotting of extracts from MDA-MB-231 cells performed 72 h after transfection using an eIF3e-specific antibody. Actin-specific antibodies were used as loading control. Expression levels are shown as the ratio of eIF3e/actin signals; ratios in controls were normalized to 1.0. (c) Northern blotting performed 48 h after transfection, using the human EIF3E ORF as a probe. Methylene blue staining of the same filter served as a loading control (lower panel). (d, e) Immunostaining of eIF3e in sections of formalin-fixed, paraffin-embedded, MDA-MB-231 cells 72 h after siRNA transfection (d) and six examples of the 81 primary human breast cancers screened (e). Sections were stained with an eIF3e-specific antibody (brown) and counterstained with hematoxylin (blue). Bar, 50 μm.

eIF3e regulates a specific subset of mRNAs

Previous reports have concluded that, besides binding other components of eIF3, eIF3e can associate separately with subunits of the proteasome (Asano et al., 1997; Yen et al., 2003; Zhou et al., 2005). We therefore further investigated the potential involvement of eIF3e in regulation of global protein levels through translation initiation and/or proteolysis through the ubiquitin–proteasome pathway. Knockdown of eIF3e in MDA-MB-231 or U2OS cells had no significant impact on bulk protein synthesis as measured by 35S-methionine pulse labeling (Figure 2a). Similarly, sucrose density gradient separation revealed no major changes in polysome distribution following eIF3e knockdown in MDA-MB-231 cells (Figure 2b). There was, however, an increase in the relative abundance of free 60S ribosomal subunits in eIF3e-depleted cells. Very similar traces were obtained using lysates from cells transfected with the alternative pair of EIF3E and control siRNA duplexes (data not shown). Protein profiles generated by 2D liquid chromatographic separation of whole cell lysates similarly showed no significant changes on eIF3e knockdown (data not shown) and there was only a minor increase in the level of total protein ubiquitinylation (Figure 2c). Ubiquitination of MCM7 was slightly increased on eIF3e knockdown, whereas total MCM7 level was slightly decreased (Supplementary Figure 1), as reported previously (Buchsbaum et al., 2007). These data indicate that eIF3e does not influence global protein levels in human cells and suggest instead its possible involvement in the regulation of a specific subset of proteins.

Figure 2

eIF3e knockdown has no major impact on bulk protein synthesis in MDA-MB-231 or U2OS cells. (a) Bulk protein synthesis in MDA-MB-231 or U2OS cells, 48 h after transfection with EIF3E-specific siRNA or control oligonucleotides, as indicated, as measured by 35S-methionine incorporation during a 2 h labeling period followed by SDS–PAGE and autoradiography (upper panel) or Coomassie blue staining (lower panel). Lane M, molecular weight markers; lane U, cells incubated in the absence of radiolabeled methionine. (b) Polysome profiles obtained by sucrose density gradient centrifugation of cell lysates prepared 48 h after transfection with siRNA duplex EIF3E 1 or control 1. The A254 peaks corresponding to ribosomal subunits and polysomes are indicated. (c) Total protein ubiquitinylation was assayed by western blotting using an anti-ubiquitin (Ub) antibody and cell lysates prepared 72 h after transfection with the siRNA duplexes indicated. The same membrane was rehybridized with antibodies against eIF3e and tubulin. A lysate from cells treated for 1 h with the proteasome inhibitor MG132 (10 mM) provided a positive control for the accumulation of ubiquitinylated proteins, which were quantified using the ratio between ubiquitin- and tubulin-specific signals. Ratios in control transfected cells were normalized to 1.0.

To investigate this point further, we transfected MDA-MB-231 cells with siRNA against EIF3E or a control duplex for 48 h, at which point total and polysome-associated (that is, actively translated) RNAs were compared separately using microarray hybridization. We identified 300 differentially expressed mRNAs (P<0.05) using polysomal RNA, and 153 using total RNA (Figure 3a; Supplementary Tables 1 and 2). Of these, 24 targets were common to both polysomal and total RNAs. Notably, the abundance and/or polysome association of some mRNAs was decreased on eIF3e knockdown, whereas others were increased. Four groups of genes, namely those positively or negatively regulated by eIF3e in polysomal or total RNA, were subjected to gene ontology (GO) analysis and functional classification using the DAVID web-based tool (Table 1). This analysis strongly suggested that eIF3e positively regulates a group of mRNAs encoding proteins involved in coagulation, taxis and endocytosis, and negatively regulates genes controlling cell division and adhesion. The altered expression of eight of these targets was validated using reverse transcriptase (RT)–PCR analysis of polysomal and total RNAs (Figures 3b and c). The expression of three eIF3e-regulated targets was analyzed at the protein level (Figures 4a and b). Decreased PLAU and BCLXL, and increased MAD2L1 protein levels were found in eIF3e-depleted MDA-MB-231 and U2OS cells. In addition, RT–PCR analysis of RNA immunoprecipitated using an eIF3b-specific antibody showed that eIF3-associated BCLXL mRNA was decreased and MAD2L1 mRNA enriched following eIF3e knockdown in MDA-MB-231 cells (Figures 4c and d). The abundance of eIF3e therefore appears to modulate the activity of the whole eIF3 complex in such a way as to determine the association of mRNAs with eIF3 in a selective manner.

Figure 3

Expression profiling of total and polysome-associated mRNAs from eIF3e-depleted and control transfected MDA-MB-231 cells. (a) Venn diagram summarizing the numbers of mRNAs significantly affected by eIF3e knockdown (for further details see Supplementary Tables 1 and 2). (b) Experimental validation of microarray data by semiquantitative RT–PCR; relative PLAU and ACTB (β-actin control) levels in total and polysomal RNA were measured using 30 or 32 PCR cycles, as indicated. (c) Relative expression of further examples of eIF3e targets involved in taxis (SAA1, CMTM7, CXCL1), adhesion (COL5A1), cell-cycle regulation (MAD2L1, CDC34) and apoptosis (BCL2L1) was determined. PCR products were generated using 25 or 30 cycles for total and polysomal RNAs, respectively.

Table 1 GO analysis of microarray data
Figure 4

Decreased PLAU (uPA) and BCL-XL, and increased MAD2L1 protein levels in eIF3e-depleted cells. MDA-MB-231 and U2OS cells were transfected either with EIF3E-specific siRNA duplexes (EIF3E 1 or EIF3E 2) or with reversed sequence duplexes (control 1 or control 2). (a) At 48 h after transfection, growth medium was collected and PLAU (uPA) levels were determined by ELISA. Results are expressed as means±s.d. of three independent measurements. PLAU levels in control transfected cells were normalized to 1.0. (b) At 72 h after transfection, whole-cell lysates were subjected to western blot analysis using antibodies against MAD2L1, BCL-XL or eIF3e; α-tubulin was used as a loading control. (c, d) Whole-cell lysates from untreated MDA-MB-231 (U) and siRNA-transfected cells were used for immunoprecipitation with an eIF3b-specific antibody. (c) Immunoprecipitates and whole-cell lysates were subjected to western blot analysis using antibodies against eIF3e or eIF3b; as a negative control, one sample from untreated cells was processed as for immunoprecipitation but in the absence of primary antibody (no 1°). (d) Immunoprecipitates were subjected in parallel to RNA extraction followed by RT–PCR using primers specific for BCL2L1 (which encodes BCL-XL), MAD2L1, EIF3E or ACTB (β-actin control).

Inhibition of eIF3e expression leads to reduced cellular invasion and cell proliferation

Our expression profiling data suggested that increased eIF3e levels might promote motility and decrease adhesion, cellular properties relevant to cancer invasion and metastasis. To assess the effect of eIF3e suppression on invasion more directly, we incubated MDA-MB-231 and U2OS cells transfected with control or EIF3E-specific siRNA in Matrigel chambers, which mimic the extracellular matrix and provide the basis for a quantitative assay of invasion (Figure 5a). The invasive capacity of eIF3e-depleted cells 48 h after transfection was reduced to 59% (MDA-MB-231) or 36% (U2OS) that of control transfected cells (P<0.01). At this time point there was no significant difference in the proliferation of the two cell populations as judged by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide-based assay (Figure 5b). Nonetheless, eIF3e-depleted MDA-MB-231 and U2OS cells showed significantly reduced cellular proliferation 72 h after siRNA transfection (P<0.05).

Figure 5

Depletion of eIF3e reduces cancer cell invasion and proliferation. MDA-MB-231 and U2OS cells were transfected either with EIF3E-specific siRNA or with reversed sequence (control) siRNA. (a) Cells invading through Matrigel were quantified 48 h after transfection as a percentage of those seen in the control transfected population (mean±s.d. of three independent experiments). (b) Proliferation of control and EIF3E siRNA-transfected cells was assayed using an MTT-based method at the indicated times after transfection. Data shown are the mean±s.d. values of at least three separate experiments.


The data presented here suggest that eIF3e, expression of which is elevated in advanced human breast cancer (Figure 1e), plays an oncogenic role by regulating the expression of gene products involved in cancer growth and invasion. Despite identification of the Int-6 gene encoding eIF3e as a site of MMTV integration (Marchetti et al., 1995), the biological role of eIF3e in tumorigenesis has proved difficult to determine. The expression of truncated EIF3E mRNA in mammary epithelial cells induced a transformed phenotype (Mayeur and Hershey, 2002) and in a mouse model truncated eIF3e expression led to persistent hyperplasia and tumorigenesis in mammary alveolar epithelium (Mack et al., 2007). These findings could be consistent either with a gain of function for the truncated alleles, or with a dominant negative effect. In line with the latter interpretation, genetic screens detected loss of EIF3E heterozygosity in human breast carcinomas (Miyazaki et al., 1997); analyses of EIF3E mRNA expression in primary human breast and non-small-cell lung carcinomas similarly suggested a potential role of eIF3e as a tumor suppressor (Marchetti et al., 2001). Reduced expression of eIF3e was observed in 37% of breast cancers and 31% of non-small-cell lung carcinomas, and appeared to predict poor prognosis in early-stage non-small-cell lung carcinoma (Buttitta et al., 2005). However, in the latter study 73% of all tumors had EIF3E RNA levels greater than those observed in matched normal lung samples, whereas only 27% had reduced levels of EIF3E RNA.

Our description of a positive correlation between elevated eIF3e protein level and high tumor grade provides evidence of an oncogenic role for eIF3e in breast cancer. In line with this view, other studies of eIF3e expression in breast, colon, lung and ovarian tumors have suggested a role for eIF3e together with TID1 and Patched proteins in cell growth and tumorigenesis (Traicoff et al., 2007). In a zebrafish model we found previously that eIF3e was a tissue-specific positive modulator of the MEK–ERK pathway, a key signaling pathway in the development and progression of human cancers (Grzmil et al., 2007). The apparently discrepant findings of studies of eIF3e function in human cancers might in part reflect the use of different methodologies (for example, measuring mRNA level or loss of heterozygosity rather than protein expression), or could indicate that eIF3e contributes differently to tumorigenesis in different cell types. The applicability to other tumors of our findings in primary breast cancer therefore warrants further study.

Human eIF3e was first identified as the p48 subunit of eIF3 (Asano et al., 1997). In addition, eIF3e has been reported to interact with subunits of the proteasome and the COP9 signalosome, suggesting its possible involvement in the regulation of both protein synthesis and degradation (von Arnim and Chamovitz, 2003). Inactivation of fission yeast eIF3e compromised proteasome function and caused the accumulation of polyubiquitinylated proteins (Yen et al., 2003). In Drosophila eIF3e was shown to be a positive regulator of cullin neddylation, thus regulating degradation through the ubiquitin–proteasome pathway of substrates of cullin-containing ubiquitin ligases (Rencus-Lazar et al., 2008). Other studies reported specific interaction of eIF3e with HIF-2α (Chen et al., 2007) or with MCM7 (Buchsbaum et al., 2007), leading to proteasomal degradation or stabilization of the eIF3e partner protein, respectively. In our experiments, eIF3e knockdown induced a slight accumulation of polyubiquitinylated proteins (Figure 2c), including MCM7 (Supplementary Figure 1), leaving open the possibility that some of the biological consequences observed might reflect an involvement of eIF3e in ubiquitin-dependent proteolysis.

The most striking effect of eIF3e downregulation seen here was, however, at the level of altered polysome association of specific mRNAs (Figures 3 and 4), supporting the view that regulation of translational initiation is a major function of eIF3e. This regulatory function is clearly selective, as eIF3e knockdown had no major effect on global translation (Figure 2a), and inhibited polysome association of some mRNAs while stimulating that of others, but the basis of this selectivity is currently unclear. eIF3 binds specifically to certain viral internal ribosome entry sequences (Buratti et al., 1998; Lopez de Quinto et al., 2001; Siridechadilok et al., 2005) and, with eIF2, can promote mRNA binding to preinitiation complexes (Jivotovskaya et al., 2006), but eIF3e itself does not possess a known RNA-binding domain. One clear possibility is that binding of eIF3e induces a conformational change in the eIF3 complex to alter its affinity for regulatory elements on cellular mRNAs. This change in affinity could be positive or negative, according to the mRNA in question.

Dissociation of eIF3 from preinitiation complexes is associated with 60S ribosomal subunit joining (Benne and Hershey, 1978), though it is unclear whether eIF3 actively inhibits 60S joining. The increased abundance of free 60S subunits seen in eIF3e-depleted cells (Figure 2) suggests that the subunit composition eIF3 can indeed influence 60S joining to some extent. This could reflect a direct interaction between the two complexes, or alternatively an influence of eIF3 on other events required for 60S joining, such as GTP hydrolysis by eIF2 (Merrick, 1979). In agreement with our results, fission yeast strains lacking eIF3e are viable, and polysome profile analysis showed that they have no major defects in translation initiation (Bandyopadhyay et al., 2000; Crane et al., 2000), though a decreased level of de novo protein synthesis was detected by pulse labeling in a more recent study (Udagawa et al., 2008). Fission yeast eIF3e and eIF3m define two distinct eIF3 complexes that may promote the translation of different sets of mRNAs; the eIF3m complex associated with bulk cellular mRNAs, whereas the eIF3e-containing complex associated with a far more restricted set (Zhou et al., 2005). In line with this view, fission yeast eIF3e was found to be required for maintaining the basal level of the Atf1 transcription factor (Udagawa et al., 2008).

Our expression profiling of breast cancer cells following eIF3 knockdown identified 300 mRNAs regulated positively or negatively at the translational level (Figure 3a; Table 1), of which 24 were also altered in abundance in total RNA preparations, possibly reflecting their stabilization by increased translation. Intriguingly, GO analysis revealed that numerous genes involved in cell-cycle-related processes were negatively regulated by eIF3e at the level of translation (Table 1). Although this might initially seem at odds with the positive correlation between eIF3e protein levels and tumor grade and proliferation in vitro (Figure 5), the relationship between cell proliferation index and malignancy is not straightforward; other aspects of tumor cell biology are likely more important determinants of tumor grade. Several of the targets identified, including PLAU, MAD2L1 and BCL2L1, are clearly implicated in such aspects of tumorigenesis. Urokinase-type plasminogen activator (PLAU/uPA) is positively regulated by eIF3e (Figures 3 and 4), and is known to degrade the extracellular matrix, promoting invasion and metastasis of malignant tumors including breast cancer, in which elevated PLAU expression is correlated with poor outcome (Han et al., 2005). The reduced invasion activity of eIF3e-depleted cells (Figure 5) is likely to result, in part at least, from decreased PLAU levels. BCL2L1 mRNA (encoding BCL-XL) was similarly positively regulated by eIF3e at the level of polysome association, and our RNA-IP experiments confirmed its eIF3e-dependent recruitment to eIF3 complexes (Figures 4c and d). Cancer cells overexpressing BCL-XL are generally resistant to a wide range of anticancer drugs (Minn et al., 1995), suggesting that breast cancer cells with high eIF3e expression might be inherently resistant to therapy as a result of increased BCL-XL translation. Interestingly, eIF3e was previously isolated through its ability to induce multidrug resistance when overexpressed in fission yeast (Crane et al., 2000). Although this model organism lacks BCL-2 family proteins, eIF3e was proposed to target selectively mRNAs required for protective cellular stress responses (Zhou et al., 2005). MAD2L1, which was negatively regulated by eIF3e (Figures 3 and 4), is a component of the mitotic spindle attachment checkpoint; its dysfunction, through either reduced or increased expression, can cause aberrant chromosome segregation, and hence contribute to malignancy, in mammalian cells (Michel et al., 2001; Sotillo et al., 2007). Inhibition of eIF3e expression was shown previously to delay mitotic progression in human cells (Morris and Jalinot, 2005); our findings suggest that elevated MAD2L1 expression may contribute to this phenotype.

Taken together, the data presented here suggest a model in which high levels of eIF3e support breast cancer progression by regulating translation of mRNAs involved in cancer growth, invasion and apoptosis. Our RNA immunoprecipitation data (Figures 4c and d) suggest that these effects may be mediated principally through the role of eIF3e as a component of the eIF3 translation initiation factor, rather than through some alternative role of eIF3e. Our immunohistochemistry data indicate that measurement of eIF3e levels could in future provide valuable prognostic information in breast cancer. In the longer term, it may even be possible to design therapeutic agents to reverse eIF3e-mediated changes in the pattern of protein synthesis.

Materials and methods

Chemicals and antibodies

Mouse monoclonal anti-actin antibody (AC-40), horseradish peroxidase secondary antibodies and proteasome inhibitor MG132 were from Sigma-Aldrich (Gillingham, UK). The rabbit polyclonal anti-eIF3e antibody CN24 (Watkins and Norbury, 2004) was used in all experiments unless otherwise indicated. Keith Gull (University of Oxford, UK) generously provided the mouse monoclonal anti-α-tubulin antibody (TAT-1). Goat polyclonal antibodies against eIF3b (N-20), PLAU (C-20), MAD2L1(C-19) and rabbit polyclonal anti-ubiquitin (FL-76) were from Santa Cruz Biotechnology (Santa Cruz, CA, USA), the rabbit monoclonal antibody against BCL-XL was from Cell Signaling (Danvers, MA, USA).

Patients and tissue samples

Formalin-fixed paraffin-embedded tissue blocks and corresponding pathology reports were obtained for sequential patients with breast adenocarcinomas (surgery was performed at the John Radcliffe Hospital, Oxford, UK). Tissue microarrays were assembled as described previously (Bubendorf et al., 2001) with three replicate cores for each tumor. Approval was obtained for the use of all human tissue from the local research ethics committee (reference C02.216).

Cell culture and transfection

Breast cancer MDA-MB-231, MDA-MB-435, T47D, MCF-7 and osteosarcoma U2OS cells were grown in Dulbecco's modified Eagle's medium medium (Invitrogen, Paisley, UK) containing 10% fetal bovine serum. MCF-10A breast epithelial cells were maintained in Dulbecco's modified Eagle's medium/Ham's F12 medium (50:50 ratio) with 10% fetal bovine serum, 5 μg/ml hydrocortisone, 10 ng/ml epidermal growth factor and 10 μg/ml insulin. All cells were cultured at 37 °C in a humidified incubator with 5% CO2 and grown to 50–60% confluence for transfection, which was accomplished using Lipofectamine (Invitrogen) according to the supplier's instructions with either gene-specific siRNA duplexes (EIF3E 1, EIF3E 2) or reversed sequence control RNA oligonucleotides (control 1 and control 2) at a final concentration of 100 nM in Optimem (Gibco). EIF3E 1 sense RNA: 5′-IndexTermCAGGGAUGGUAGGAUGCUCdTdT-3′; EIF3E 2 sense RNA: 5′-IndexTermGAACCACAGUGGUUGCACAUU-3′. Cells and growth medium were collected at 24, 48 and 72 h after transfection for subsequent analysis. For proliferation assays, cells were split into 24-well plates 6 h after transfection. After 2, 24, 48 and 72 h of incubation, EZ4U substrate (50 μl per well; Biozol, Eching, Germany) was added and absorbance at 450 nm was measured with a plate reader (GE Healthcare, Amersham, UK). All assays were performed in triplicate.

Northern blot analysis and RT–PCR reactions

Total RNA was isolated using an RNeasy Mini Kit (Qiagen, Crawley, UK) according to the manufacturer's instructions. For northern blot analysis, total RNA was separated by denaturing agarose gel electrophoresis and transferred to a Hybond-C nylon membrane (GE Healthcare). An EIF3E ORF generated by RT–PCR (1338 bp) was cloned into pGEM-T Easy plasmid (Promega, Southampton, UK), sequenced and used as a probe. The probe was labeled with [α-32P] dCTP using a Rediprime II labeling kit (GE Healthcare) and hybridized to membranes in Rapid-hyb buffer (GE Healthcare) containing 100 μg/ml denatured salmon sperm DNA at 65 °C for 16 h. Membranes were washed at room temperature for 15 min in 2 × SSC, then in 0.5 × SSC, 0.5% (w/v) SDS for 15 min at 65 °C. Hybridization signals were detected with a Fuji FLA-5000 imager (Fujifilm, Milton Keynes, UK) and quantified using Aida software (Raytest Isotopenmessgeräte, Straubenhardt, Germany). RT–PCR and semiquantitative RT–PCR analyses were performed using a one-step RT–PCR kit (Qiagen) with the primers listed in Table 2.

Table 2 Oligonucleotide primers used for RT–PCR analysis

Immunoblotting, RNA immunoprecipitation and ELISA

Whole-cell lysates were prepared using lysis buffer (150 mM NaCl, 10 mM EDTA, 50 mM Tris–HCl (pH 7.6), 1% Triton X-100, 1 μg/ml leupeptin, 1 μg/ml aprotinin, and 1 μg/ml phenylmethylsulfonyl fluoride). Lysates (50 μg per lane) were separated by 10% SDS–polyacrylamide gel electrophoresis (PAGE). Proteins were electrotransferred to Hybond-C nitrocellulose membranes (GE Healthcare) before being incubated with appropriate antibodies and processed for ECL detection (ECL Plus; GE Healthcare) according to the manufacturer's protocol. Signals were analyzed using LabWorks software (UV Products, Cambridge, UK). For immunoprecipitation, cell lysates were prepared at 0 °C in NP-40 lysis buffer (1% NP-40, 150 mM NaCl, 20 mM Tris–HCl (pH 7.5), 2 mM EDTA, 1 μg/ml leupeptin, 1 μg/ml aprotinin and 1 μg/ml phenylmethylsulfonyl fluoride) supplemented with 0.2 U/μl RNaseOUT inhibitor (Invitrogen). After repeated shearing through a 25-gauge needle, lysates were clarified by centrifugation (5 min, 13 000 g) and were precleared with protein G-Sepharose. Clarified lysates were incubated overnight at 4 °C with 1 μg of appropriate antibody and 50 μl of protein G-Sepharose (Cancer Research UK). Immunoprecipitates were washed three times with NP-40 lysis buffer, half of the immunoprecipitates were used for RNA extraction using RNAeasy Mini Kit (Qiagen) and the rest was separated by 10% SDS–PAGE and subjected to immunoblotting as described above. For uPA detection in growth medium human uPA ELISA Kit (Assay Pro, St Charles, MO, USA) was used according to the manufacturer's protocol. All assays were performed in triplicate.


Formalin-fixed, paraffin-embedded tissue or transfected MDA-MB-231 cell pellet sections were treated with DeWax solution (Biogenex, San Ramon, CA, USA) according to user manual followed by incubation with sodium citrate buffer (pH 8) for 2 min and with 0.03% hydrogen peroxide (Dako, Ely, UK) for the next 5 min. Horse serum (2.5%) was used for blocking. Sections were incubated overnight with diluted 1:50 (in phosphate-buffered saline, PBS) eIF3e-specific antibody (CN24) and washed with PBS. The EnVision system (Dako) was used for signal visualization according to the manufacturer's instructions and we counterstained sections with hematoxylin. The levels of cytoplasmic and nuclear eIF3e staining were assessed semiquantitatively to produce an intensity distribution score (IDS) on a 12-point scale as described previously (Winters et al., 2001). Average IDS values were determined by examination of 10 fields. Data were analyzed using STATA (STATA Corporation, College Station, TX, USA) and Pearson's correlation coefficients were used to determine the association between eIF3e IDS (using a cutoff value for IDS of 2 to distinguish between ‘high’ and ‘low’ eIF3e staining) and tumor grade.

Sucrose density gradient centrifugation

Cells were incubated for 3 min at 37 °C with 0.1 mg/ml cycloheximide, washed in PBS, harvested in polysome lysis buffer (1% Triton X-100, 300 mM NaCl, 15 mM MgCl2, 15 mM Tris-HCl (pH 7.4), 0.1 mg/ml cycloheximide and 0.33 U/μl RNAse inhibitor), incubated for 5 min on ice, mixed by vortexing and centrifuged at 2000 g for 5 min. The supernatants were supplemented with heparin to a final concentration of 200 μg/ml, centrifuged at 10 000 g for 5 min to remove cell debris and layered onto 20–50% sucrose gradients for fractionation (90 min at 39 000 r.p.m. in a Beckman SW40.1 rotor and L8 ultracentrifuge at 4 °C). After centrifugation, 30 fractions (0.5 ml) were harvested with continuous monitoring of A254 using a gradient fractionator (Bio-Rad, Hemel Hempstead, UK). The fractions were used for polysomal RNA isolation using TriReagent (Sigma-Aldrich) and the RNA was further purified using RNeasy Mini columns (Qiagen).

Microarray analysis

Before labeling, RNA concentration and integrity were determined using Nanochips on a 2100 Bioanalyzer (Agilent Technologies, Stockport, UK), according to the manufacturer's instructions. Two labeling and hybridization protocols were used in conjunction with two types of arrays. For Human OpArrays (Operon, Huntsville, AL, USA), containing oligonucleotide probes representing approximately 39 600 transcripts, a dendrimer-based system was used according to the manufacturer's instructions, with minor modifications. Briefly, RNA (1 μg of total or polysomal RNA) was labeled using the 3DNA Array 900 kit (Genisphere, Hatfield, PA, USA), using Superscript III reverse transcriptase (Invitrogen). The hybridization and detection steps were performed using a two-step procedure on a SlideBooster (Advalytix, Munich, Germany), with a power setting of 25 and a pulse ratio of 3:7 at 55 °C. The first hybridization was for 16 h using hybridization buffer EB, and the second was for 4 h using SDS buffer. For human 22K genome-wide printed cDNA arrays (v1.0.0) from Cancer Research UK, an Amino Allyl MessageAmp aRNA Kit (Applied Biosystems, Warrington, UK) was used for labeling with Cy3 and Cy5 dyes according to the manufacturer's instructions. Hybridization and washing steps were performed using a Lucidea SlidePro platform (GE Healthcare). The OpArray and 22K Cancer Research UK slides were scanned using a ScanArray ExpressHT system (PerkinElmer, Cambridge, UK) and GenePix 4000B microarray scanner, respectively. Images were obtained using autocalibration with 100% laser power, a variable PMT and a target saturation of 90%. Poor-quality spots were manually flagged, and intensity values were extracted using BlueFuse for microarrays version 2 (BlueGnome, Cambridge, UK). Three OpArrays were used to compare total RNAs from three independent transfections and two OpArrays for polysomal RNA from two independent transfections, in each case using two pairs of siRNAs. In addition, six 22K Cancer Research UK slides were used to analyze total RNA from three independent transfections together with a dye swap for each experiment. Intensity values, extracted using BlueFuse, were analyzed using BASE (Saal et al., 2002). Only median fold ratio values with P<0.05 (Cyber t-test) were used for subsequent analysis. The DAVID web-based tool was used for GO terms identification and functional classification of the differentially expressed genes (Dennis et al., 2003).

Invasion assay

Transfected cells were suspended at 5 × 104 cells per ml in complete medium, and 500 μl of each cell suspension was placed in a BioCoat Matrigel Invasion Chamber (BD Biosciences, Oxford, UK) and incubated for 20 h at 37 °C. Invasive cells adhered to the chamber membranes, which were fixed in methanol, washed with PBS and mounted on microscope slides in Vectashield mounting medium with DAPI (Vector Laboratories, Peterborough, UK). Images were acquired and cells were counted using a Zeiss Axioskop microscope (Carl Zeiss Ltd, Welwyn Garden City, UK) and MetaMorph software (Molecular Devices, Sunnyvale, CA, USA). All experiments were repeated in triplicate, and the noninvasive breast cancer cell line MCF-7 was used as a negative control.


  1. Ahlemann M, Zeidler R, Lang S, Mack B, Munz M, Gires O . (2006). Carcinoma-associated eIF3i overexpression facilitates mTOR-dependent growth transformation. Mol Carcinog 45: 957–967.

  2. Asano K, Merrick WC, Hershey JW . (1997). The translation initiation factor eIF3-p48 subunit is encoded by int-6, a site of frequent integration by the mouse mammary tumor virus genome. J Biol Chem 272: 23477–23480.

  3. Bandyopadhyay A, Matsumoto T, Maitra U . (2000). Fission yeast Int6 is not essential for global translation initiation, but deletion of int6(+) causes hypersensitivity to caffeine and affects spore formation. Mol Biol Cell 11: 4005–4018.

  4. Benne R, Hershey JW . (1978). The mechanism of action of protein synthesis initiation factors from rabbit reticulocytes. J Biol Chem 253: 3078–3087.

  5. Bramham CR, Wells DG . (2007). Dendritic mRNA: transport, translation and function. Nat Rev Neurosci 8: 776–789.

  6. Bubendorf L, Nocito A, Moch H, Sauter G . (2001). Tissue microarray (TMA) technology: miniaturized pathology archives for high-throughput in situ studies. J Pathol 195: 72–79.

  7. Buchsbaum S, Morris C, Bochard V, Jalinot P . (2007). Human INT6 interacts with MCM7 and regulates its stability during S phase of the cell cycle. Oncogene 26: 5132–5144.

  8. Buratti E, Tisminetzky S, Zotti M, Baralle FE . (1998). Functional analysis of the interaction between HCV 5′UTR and putative subunits of eukaryotic translation initiation factor eIF3. Nucleic Acids Res 26: 3179–3187.

  9. Buttitta F, Martella C, Barassi F, Felicioni L, Salvatore S, Rosini S et al. (2005). Int6 expression can predict survival in early-stage non-small cell lung cancer patients. Clin Cancer Res 11: 3198–3204.

  10. Chen L, Uchida K, Endler A, Shibasaki F . (2007). Mammalian tumor suppressor Int6 specifically targets hypoxia inducible factor 2{alpha} for degradation by hypoxia- and pVHL-independent regulation. J Biol Chem 282: 12707–12716.

  11. Crane R, Craig R, Murray R, Dunand-Sauthier I, Humphrey T, Norbury C . (2000). A fission yeast homolog of Int-6, the mammalian oncoprotein and eIF3 subunit, induces drug resistance when overexpressed. Mol Biol Cell 11: 3993–4003.

  12. Dennis Jr G, Sherman BT, Hosack DA, Yang J, Gao W, Lane HC et al. (2003). DAVID: database for annotation, visualization, and integrated discovery. Genome Biol 4: P3.

  13. Dong Z, Zhang JT . (2006). Initiation factor eIF3 and regulation of mRNA translation, cell growth, and cancer. Crit Rev Oncol Hematol 59: 169–180.

  14. Grzmil M, Whiting D, Maule J, Anastasaki C, Amatruda JF, Kelsh RN et al. (2007). The INT6 cancer gene and MEK signaling pathways converge during zebrafish development. PLoS One 2: e959.

  15. Han B, Nakamura M, Mori I, Nakamura Y, Kakudo K . (2005). Urokinase-type plasminogen activator system and breast cancer (review). Oncol Rep 14: 105–112.

  16. Hinnebusch AG . (2006). eIF3: a versatile scaffold for translation initiation complexes. Trends Biochem Sci 31: 553–562.

  17. Holcik M, Sonenberg N . (2005). Translational control in stress and apoptosis. Nat Rev Mol Cell Biol 6: 318–327.

  18. Jivotovskaya AV, Valasek L, Hinnebusch AG, Nielsen KH . (2006). Eukaryotic translation initiation factor 3 (eIF3) and eIF2 can promote mRNA binding to 40S subunits independently of eIF4G in yeast. Mol Cell Biol 26: 1355–1372.

  19. Lopez de Quinto S, Lafuente E, Martinez-Salas E . (2001). IRES interaction with translation initiation factors: functional characterization of novel RNA contacts with eIF3, eIF4B, and eIF4GII. RNA 7: 1213–1226.

  20. Mack DL, Boulanger CA, Callahan R, Smith GH . (2007). Expression of truncated Int6/eIF3e in mammary alveolar epithelium leads to persistent hyperplasia and tumorigenesis. Breast Cancer Res 9: R42.

  21. Marchetti A, Buttitta F, Miyazaki S, Gallahan D, Smith GH, Callahan R . (1995). Int-6, a highly conserved, widely expressed gene, is mutated by mouse mammary tumor virus in mammary preneoplasia. J Virol 69: 1932–1938.

  22. Marchetti A, Buttitta F, Pellegrini S, Bertacca G, Callahan R . (2001). Reduced expression of INT-6/eIF3-p48 in human tumors. Int J Oncol 18: 175–179.

  23. Mavrakis KJ, Wendel HG . (2008). Translational control and cancer therapy. Cell Cycle 7: 2791–2794.

  24. Mayeur GL, Hershey JW . (2002). Malignant transformation by the eukaryotic translation initiation factor 3 subunit p48 (eIF3e). FEBS Lett 514: 49–54.

  25. Merrick WC . (1979). Evidence that a single GTP is used in the formation of 80 S initiation complexes. J Biol Chem 254: 3708–3711.

  26. Michel LS, Liberal V, Chatterjee A, Kirchwegger R, Pasche B, Gerald W et al. (2001). MAD2 haplo-insufficiency causes premature anaphase and chromosome instability in mammalian cells. Nature 409: 355–359.

  27. Minn AJ, Rudin CM, Boise LH, Thompson CB . (1995). Expression of bcl-xL can confer a multidrug resistance phenotype. Blood 86: 1903–1910.

  28. Miyazaki S, Imatani A, Ballard L, Marchetti A, Buttitta F, Albertsen H et al. (1997). The chromosome location of the human homolog of the mouse mammary tumor-associated gene INT6 and its status in human breast carcinomas. Genomics 46: 155–158.

  29. Morris C, Jalinot P . (2005). Silencing of human Int-6 impairs mitosis progression and inhibits cyclin B-Cdk1 activation. Oncogene 24: 1203–1211.

  30. Paquin N, Chartrand P . (2008). Local regulation of mRNA translation: new insights from the bud. Trends Cell Biol 18: 105–111.

  31. Rasmussen SB, Kordon E, Callahan R, Smith GH . (2001). Evidence for the transforming activity of a truncated Int6 gene, in vitro. Oncogene 20: 5291–5301.

  32. Rencus-Lazar S, Amir Y, Wu J, Chien CT, Chamovitz DA, Segal D . (2008). The proto-oncogene Int6 is essential for neddylation of Cul1 and Cul3 in Drosophila. PLoS One 3: e2239.

  33. Saal LH, Troein C, Vallon-Christersson J, Gruvberger S, Borg A, Peterson C . (2002). BioArray Software Environment (BASE): a platform for comprehensive management and analysis of microarray data. Genome Biol 3: software0003.1–software0003.6.

  34. Savinainen KJ, Helenius MA, Lehtonen HJ, Visakorpi T . (2006). Overexpression of EIF3S3 promotes cancer cell growth. Prostate 66: 1144–1150.

  35. Siridechadilok B, Fraser CS, Hall RJ, Doudna JA, Nogales E . (2005). Structural roles for human translation factor eIF3 in initiation of protein synthesis. Science 310: 1513–1515.

  36. Sotillo R, Hernando E, Diaz-Rodriguez E, Teruya-Feldstein J, Cordon-Cardo C, Lowe SW et al. (2007). Mad2 overexpression promotes aneuploidy and tumorigenesis in mice. Cancer Cell 11: 9–23.

  37. Thumma SC, Kratzke RA . (2007). Translational control: a target for cancer therapy. Cancer Lett 258: 1–8.

  38. Traicoff JL, Chung JY, Braunschweig T, Mazo I, Shu Y, Ramesh A et al. (2007). Expression of eIF3-p48/INT6, TID1 and Patched in cancer, a profiling of multiple tumor types and correlation of expression. J Biomed Sci 14: 395–405.

  39. Udagawa T, Nemoto N, Wilkinson CR, Narashimhan J, Jiang L, Watt S et al. (2008). Int6/eIF3e promotes general translation and Atf1 abundance to modulate Sty1 MAPK-dependent stress response in fission yeast. J Biol Chem 283: 22063–22075.

  40. von Arnim AG, Chamovitz DA . (2003). Protein homeostasis: a degrading role for Int6/eIF3e. Curr Biol 13: R323–R325.

  41. Watkins SJ, Norbury CJ . (2002). Translation initiation and its deregulation during tumorigenesis. Br J Cancer 86: 1023–1027.

  42. Watkins SJ, Norbury CJ . (2004). Cell cycle-related variation in subcellular localization of eIF3e/INT6 in human fibroblasts. Cell Prolif 37: 149–160.

  43. Wendel HG, Silva RL, Malina A, Mills JR, Zhu H, Ueda T et al. (2007). Dissecting eIF4E action in tumorigenesis. Genes Dev 21: 3232–3237.

  44. Winters ZE, Hunt NC, Bradburn MJ, Royds JA, Turley H, Harris AL et al. (2001). Subcellular localisation of cyclin B, Cdc2 and p21(WAF1/CIP1) in breast cancer association with prognosis. Eur J Cancer 37: 2405–2412.

  45. Yen HC, Gordon C, Chang EC . (2003). Schizosaccharomyces pombe Int6 and Ras homologs regulate cell division and mitotic fidelity via the pro. Cell 112: 207–217.

  46. Zhang L, Pan X, Hershey JW . (2007). Individual overexpression of five subunits of human translation initiation factor eIF3 promotes malignant transformation of immortal fibroblast cells. J Biol Chem 282: 5790–5800.

  47. Zhou C, Arslan F, Wee S, Krishnan S, Ivanov AR, Oliva A et al. (2005). PCI proteins eIF3e and eIF3m define distinct translation initiation factor 3 complexes. BMC Biol 3: 14.

Download references


We thank Ben Thomas and Sasha Akoulitchev for help and advice with 2D liquid chromatography, Cheng Han for help with the statistical analysis, Dan Scott and other members of the laboratory for their comments on the article. This work was supported by Cancer Research UK, the Association for International Cancer Research and the Wellcome Trust (through grant 075491/Z/04 to JR).

Author information

Correspondence to C J Norbury.

Ethics declarations

Competing interests

The authors declare no conflict of interest.

Additional information

Supplementary Information accompanies the paper on the Oncogene website

Supplementary information

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Grzmil, M., Rzymski, T., Milani, M. et al. An oncogenic role of eIF3e/INT6 in human breast cancer. Oncogene 29, 4080–4089 (2010) doi:10.1038/onc.2010.152

Download citation


  • MAD2L1
  • BCL-XL
  • PLAU
  • urokinase-type plasminogen activator
  • breast cancer

Further reading