Increased Expression of Insulin-like Growth Factor-II Messenger RNA–Binding Protein 1 Is Associated with Tumor Progression in Patients with Lung Cancer

Lung cancer is one of the most common causes of cancer death worldwide, and non–small cell lung cancer (NSCLC) accounts for nearly 80% of those cases (1). Many genetic alterations involved in development and progression of lung cancer have been reported, but the precise molecular mechanisms remain unclear (2). Over the last decade, newly developed cytotoxic agents, including paclitaxel, docetaxel, gemcitabine, and vinorelbine, have emerged to offer multiple therapeutic choices for patients with advanced NSCLC. However, those regimens provide only limited survival benefits compared with cisplatin-based therapies (3, 4). Recently, new agents targeting the epidermal growth factor receptor pathway, erlotinib (Tarceva, OSI Pharmaceuticals, Melville, NY) and gefitinib (Iressa, AstraZeneca, Wilmington, DE), were developed and were shown to be very effective for a subset of NSCLC patients. However, even if all kinds of available treatments are applied, the proportion of patients showing good response is still very limited (5–8). Hence, new therapeutic strategies are eagerly awaited.

Systematic analysis of expression levels of thousands of genes using cDNA microarray is an effective approach to identify molecules involved in carcinogenic pathways that can be candidates for development of novel therapeutics and diagnostics. We have been attempting to isolate potential molecular targets for diagnosis and/or treatment of lung cancer by analyzing genome-wide expression profiles of various types of lung cancer cells on a cDNA microarray containing 27,648 genes, using tumor cell populations purified by laser-capture microdissection (9–11). To verify the biological and clinicopathologic significance of the respective gene products, we have also been performing tumor tissue microarray analysis of clinical lung cancer specimens (12–15). This systematic approach revealed that insulin-like growth factor-II (IGF-II) mRNA-binding protein 1 (IMP-1; alias CRDBP, c-myc coding region determinant binding protein) was frequently overexpressed in primary NSCLCs.

IMP-1 is a member of the zipcode-binding protein family, which are orthologous and paralogous members of the same vertebrate RNA-binding protein family, consisting of two RNA recognition motifs and four K homology domains (16). IMP-1 is expressed in most embryonic tissues. Analysis of total RNA from mouse embryos indicated a peak of IMP-1 expression at embryonic day 12.5 followed by decline toward birth and its disappearance in neonatal mice shortly after birth (17). IMP-1 was overexpressed in several human cancers and has been suggested to play various roles in determining the posttranscriptional fate of its RNA targets and to act as a nucleocytoplasmic shuttling protein exhibiting a distinct pattern of localization in the cytoplasm (16, 18–23). The protein is distributed along with microtubules and is likely to be transported toward the leading edge in motile cells. Its nuclear export and cytoplasmic movement depend on RNA binding, implying that IMP-1 recognizes its targets in the nucleus and thereby defines their cytoplasmic fate. IMP-1 was indicated to play a significant role in polarizing genetic information by defining cytoplasmic RNA localization, a critical mechanism especially in developmental systems for the generation of subcellular asymmetries in protein abundance. H19 RNA colocalizes with IMP-1, and removal of the high-affinity attachment site led to delocalization of the truncated RNA (23), indicating that IMP-1 are involved in cytoplasmic trafficking of mRNA (22).

We here report the identification of IMP-1 as a novel prognostic marker and a potential target for therapeutic agents and also provide evidence for its possible role in human pulmonary carcinogenesis through its binding to various mRNAs that encode proteins related with cell proliferation and invasion.

Lung cancer cell lines and clinical samples. The human lung cancer lines used in this study were as follows: A427, A549, LC319, PC3, PC9, PC14, and NCI-H1373 (adenocarcinomas); NCI-H1666 and NCI-H1781 (bronchioloalveolar carcinomas); NCI-H226 and NCI-H647 (adenosquamous carcinomas); RERF-LC-AI, SK-MES-1, EBC-1, LU61, NCI-H520, NCI-H1703, and NCI-H2170 (lung squamous cell carcinomas); LX1 (lung large cell carcinoma); DMS114, DMS273, SBC-3, and SBC-5 (SCLC). All cells were grown in monolayer in appropriate medium supplemented with 10% FCS and were maintained at 37°C in atmospheres of humidified air with 5% CO₂. Human small airway epithelial cells were grown in optimized medium (SAGM) purchased from Cambrex Bioscience, Inc. (Walkersville, MD). Fourteen primary NSCLCs (7 adenocarcinomas and 7 squamous cell carcinomas) were obtained along with adjacent normal lung tissue samples from patients undergoing surgery at Hokkaido University and its affiliated hospitals (Sapporo, Japan).

A total of 267 formalin-fixed primary NSCLCs (stage I-IIIA) and adjacent normal lung tissue samples used for immunostaining on tissue microarrays had been obtained with informed consent from patients undergoing curative surgical operation at Hokkaido University and its affiliated hospitals (Sapporo, Japan). Histologic classification of tumors was done according to the WHO criteria (24). All tumors were staged based on the pathologic tumor-node-metastasis classification of the International Union Against Cancer (25). Postoperative staging evaluation showed that 101 patients were at stage IA, 88 at stage IB, 8 at stage IIA, 27 at stage IIB, and 43 at stage IIIA. Histopathologic examination of resected tumors revealed that 157 cases were diagnosed as adenocarcinoma, 93 cases as squamous cell carcinomas, 13 as lung large cell carcinomas, and 4 as adenosquamous carcinomas (Table 1). This study as well as the use of all clinical materials described above were approved by individual institutional Ethical Committees.

Semiquantitative reverse transcription-PCR. Total RNA was extracted from cultured cells and clinical tissues using Trizol reagent (Life Technologies, Inc., Gaithersburg, MD) according to the manufacturer's protocol. Extracted RNAs and normal human tissue polyadenylate RNAs were treated with DNase I (Nippon Gene, Tokyo, Japan) and reversely transcribed using oligo (dT) primer and SuperScript II reverse transcriptase (Invitrogen, Carlsbad, CA). Semiquantitative reverse transcription-PCR experiments were carried out with the following synthesized IMP-1–specific primers or with β-actin (ACTB)–specific primers as an internal control: IMP-1, 5′-CAGAAGGGACAGAGTAACCAG-3′ and 5′-GAGATCAGGGTTCCTCACTG-3′; ACTB, 5′-GAGGTGATAGCATTGCTTTCG-3′ and 5′-CAAGTCAGTGTACAGGTAAGC-3′. PCR reactions were optimized for the number of cycles to ensure product intensity within the logarithmic phase of amplification.

Northern blot analysis. Human multiple-tissue blots (BD Biosciences Clontech, Palo Alto, CA) were hybridized with a ³²P-labeled PCR product of IMP-1. The cDNA probes of IMP-1 were prepared by reverse transcription-PCR using the primers described above. Prehybridization, hybridization, and washing were done according to the supplier's recommendations. The blots were autoradiographed at room temperature for 30 h with intensifying BAS screens (Bio-Rad, Hercules, CA).

Preparation of anti-IMP-1 polyclonal antibody. Rabbit antibodies specific for IMP-1 were raised by immunizing rabbits with IMP-1 peptides (IEHSVPKKQRSRKIC and CVKQQHQKGQSNQAQARRK) and purified using standard protocols. On Western blots, we confirmed that the antibody was specific to IMP-1 but do not cross-react with other homologous proteins (IMP-2 and IMP-3) using lysates from NSCLC cell lines transfected with IMP-1, IMP-2, and IMP-3 expressing vector and those from endogenous IMP-1 expressing/non-expressing NSCLC cells.

Western blot analysis. Cells were lysed in lysis buffer [50 mmol/L Tris-HCl (pH 8.0), 150 mmol/L NaCl, 0.5% NP40, 0.5% deoxycholate-Na, 0.1% SDS, plus protease inhibitor; Protease Inhibitor Cocktail Set III; Calbiochem, Darmstadt, Germany]. We used an enhanced chemiluminescence Western blotting analysis system (GE Healthcare Biosciences, Piscataway, NJ), as previously described (13–15). SDS-PAGE was done in 10% polyacrylamide gels. PAGE-separated proteins were electroblotted onto nitrocellulose membranes (GE Healthcare Biosciences) and incubated with a rabbit polyclonal anti-human IMP-1 antibody (1:1,000 dilution). A goat anti-rabbit IgG-horseradish peroxidase antibody (GE Healthcare Biosciences) served as the secondary antibodies for these experiments (1:10,000 dilution).

Tissue microarray construction and immunohistochemistry. Lung cancer tissue microarrays were constructed as published elsewhere, using formalin-fixed NSCLCs (12–15). Tissue areas for sampling were selected based on visual alignment with the corresponding H&E-stained sections on slides. Three, four, or five tissue cores (diameter = 0.6 mm; height = 3-4 mm) taken from donor tumor blocks were placed into recipient paraffin blocks using a tissue microarrayer (Beecher Instruments, Sun Prairie, WI). A core of normal tissue area was punched from each case. Five-micrometer sections of the resulting microarray block were used for immunohistochemical analysis.

To investigate the IMP-1 protein level in tissue microarrays of clinical samples, we stained the sections using ENVISION+ kit/horseradish peroxidase (DakoCytomation, Glostrup, Denmark). A rabbit polyclonal anti-IMP-1 antibody (1:500 dilution) was added and incubated for 1 h at room temperature after blocking endogenous peroxidase and proteins, and the sections were incubated with horseradish peroxidase–labeled anti-rabbit IgG as the secondary antibody. Substrate-chromogen was added, and the specimens were counterstained with hematoxylin. Positivity for IMP-1 was assessed semiquantitatively by three independent investigators without prior knowledge of the clinical follow-up data, each of whom recorded staining intensity as negative (scored as 0) or positive (1+). Cases were accepted as positive only if reviewers independently defined them as such.

Statistical analysis. All analyses were done using statistical analysis software (StatView, version 5.0; SAS Institute, Inc., Cary, NC). We attempted to correlate clinicopathologic variables, such as age, gender, smoking history, pathologic tumor-node-metastasis stage, histologic type, and histopathologic grading, with the expression levels of IMP-1 protein determined by tissue microarray analysis. IMP-1 immunoreactivity was assessed for association with clinicopathologic variables using the following statistical tests, such as the Mann-Whitney U test or χ² test. Multivariate logistic regression analysis was done to examine the clinicopathologic factor(s) that was independently associated with the IMP-1 expression (26, 27). Tumor-specific survival curves were calculated from the date of surgery to the time of death related to NSCLC, or to the last follow-up observation. The Kaplan-Meier method was used to generate survival curves, and survival differences were analyzed with the log-rank test, based on the status of IMP-1 expression. Univariate analysis was done using Cox's proportional hazard regression model.

RNA interference assay. We had previously established a vector-based RNA interference system (psiH1BX3.0) that was designed to synthesize small interfering RNAs (siRNA) in mammalian cells (13–15, 28). Ten micrograms of siRNA-expression vector was transfected using 30 μL LipofectAMINE 2000 (Invitrogen) into NSCLC cell lines A549 and LC319. The transfected cells were cultured for 7 days in the presence of appropriate concentrations of Geneticin (G418), and the number of colonies was counted by colony formation assay, and viability of the cells was evaluated by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay 7 days after the treatment. In 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay, Cell-counting kit-8 solution (DOJINDO, Kumamoto, Japan) was added to each dish at a concentration of one-tenth volume, and the plates were incubated at 37°C for additional 4 h. Absorbance was then measured at 490 nm, and at 630 nm as a reference, with a Microplate Reader 550 (Bio-Rad). The target sequences of the synthetic oligonucleotides for RNA interference were as follows: control 1 [enhanced green fluorescent protein (EGFP) gene, a mutant of Aequorea victoria GFP], 5′-GAAGCAGCACGACTTCTTC-3′; control 2 (5S-16S: chloroplast Euglena gracilis gene coding for 5S and 16S rRNAs), 5′-GCGCGCTTTGTAGGATTCG-3′; siRNA-IMP-1-#1, 5′-GCCATCAGTGTGCACTCCA-3′; siRNA-IMP-1-#2, 5′-GGAGGAGAACTTCTTTGGT-3′; siRNA-IMP-1-#3, 5′-GAATCTATGGCAAACTCAA-3′. To validate our RNA interference system, individual control siRNAs were tested by semiquantitative reverse transcription-PCR to confirm the decrease in expression of the corresponding target genes that had been transiently transfected to COS-7 cells. Down-regulation of IMP-1 expression by functional siRNA, but not by controls, was also confirmed in the cell lines used for this assay.

Matrigel invasion assay. NIH-3T3 cells transfected either with IMP-1 constructs with NH₂-terminal FLAG- or COOH-terminal hemagglutinin-tagged sequences (pCAGGS-n3FH-IMP-1), or with mock plasmids were grown to near confluence in DMEM containing 10% FCS. The cells were harvested by trypsinization, washed in DMEM without addition of serum or proteinase inhibitor, and suspended in DMEM at concentration of 1 × 10⁵ per mL. Before preparing the cell suspension, the dried layer of Matrigel matrix (Becton Dickinson Labware, Mountain View, CA) was rehydrated with DMEM for 2 h at room temperature. DMEM (0.75 mL) containing 10% FCS was added to each lower chamber in 24-well Matrigel invasion chambers, and 0.5 mL (5 × 10⁴ cells) of cell suspension was added to each insert of the upper chamber. The plates of inserts were incubated for 22 h at 37°C. After incubation the chambers were processed, and cells invading through the Matrigel were fixed and stained by Giemsa as recommended by the supplier (Becton Dickinson Labware).

RNA immunoprecipitation and cDNA microarray screening for identification of IMP-1–associated mRNAs. We adopted the RNA immunoprecipitation protocol by Niranjanakumari et al. to analyze RNA(s)-protein interactions involving IMP-1 in vivo (29). To determine the IMP-1 associated mRNA(s), we transfected pCAGGS-n3FH-IMP-1 vector into A549 cells. Using these cell lysates transfected with IMP-1 construct, we further did immunoprecipitation experiments twice: first with monoclonal anti-FLAG M2 and then with monoclonal anti-hemagglutinin antibody. A 2.5-μg aliquot of T7-based amplified mRNA from each immunoprecipitated RNA and from the total RNA were reversely transcribed in the presence of Cy5-dCTP and Cy3-dCTP, respectively, as described previously (9–11), for hybridization to a cDNA microarray representing 27,648 genes (IP-microarray analysis).

Expression of IMP-1 transcripts in lung tumors and normal tissues. To identify target molecules for development of novel therapeutic agents and/or biomarkers for lung cancer, we first screened a cDNA microarray for genes that showed ≥3-fold expression in >50% of NSCLCs analyzed (9, 10). Among 27,648 genes screened, we identified the IMP-1 transcript to show >3-fold expression in 68.8% of NSCLCs compared with normal lung tissue (control). We confirmed its overexpression by semiquantitative reverse transcription-PCR experiments in 6 of 14 additional NSCLC cases (2 of 7 adenocarcinomas and 4 of 7 squamous cell carcinomas; Fig. 1A) as well as in 16 of 23 NSCLC and SCLC cell lines. However, its expression was hardly detectable in small airway epithelial cells derived from normal bronchial epithelium (Fig. 1B).

We subsequently generated rabbit polyclonal antibody against human IMP-1 and confirmed its specificity to IMP-1 by Western blot analysis that showed no cross-reactivity to other homologous proteins (IMP-2 and IMP-3) using lysate from NCI-H520 cells that had been transfected with hemagglutinin-tagged IMP-1, IMP-2, or IMP-3 expression vector (Fig. 1C, left). Using this antibody, we confirmed expression of endogenous IMP-1 protein in six lung cancer cell lines by Western blot analysis (three IMP-1–positive and three IMP-1–negative cell lines; Fig. 1C, right). Northern blot analysis using IMP-1 cDNA as a probe identified strong signals corresponding to 4.5-kb transcript that expressed specifically in placenta and testis (Fig. 1D).

Association of IMP-1 expression with poor prognosis of NSCLC patients. To verify the clinicopathologic significance of IMP-1, we additionally examined the expression of IMP-1 protein by means of tissue microarrays containing lung cancer tissues from 267 patients. Positive tumor cells generally showed a cytoplasmic staining pattern in NSCLC, and no staining was observed in any of their adjacent normal lung tissues (Fig. 2A). We classified patterns of IMP-1 expression as negative (scored as 0) or positive (scored as 1+; Fig. 2A). We found positive staining in 139 (52.1%) of 267 NSCLC cases; 57 of 157 adenocarcinoma tumors (36.3%), 70 of 93 squamous cell carcinoma tumors (75.3%), 10 of 13 lung large cell carcinoma tumors (76.9%), and 2 of 4 adenosquamous carcinoma tumors (50.0%) were judged to be positive (Table 1). We then examined a correlation of IMP-1 expression with various clinicopathologic variables and found its significant correlation with gender (higher in male, P = 0.0001), pT classification (higher in larger tumor, P = 0.0003), histopathologic type (higher in non-adenocarcinoma, P < 0.0001), smoking history (higher in smoker, P = 0.0005), and histopathologic grade (higher in non–well-differentiated tumor, P = 0.0001; Table 1). Multivariate logistic regression analysis for these five significant clinicopathologic variables determined that larger tumor size (pT₂-T₃), non-adenocarcinoma histology, and non–well-differentiated histopathologic grade were independent features associated with IMP-1 overexpression (P = 0.0104, 0.0001, and 0.0019, respectively; Supplementary Table S1).

The Kaplan-Meier analysis indicated a significant association between IMP-1 positivity in NSCLCs and tumor-specific 5-year survival (P = 0.0053 by the log-rank test; Fig. 2B). By univariate analysis using the Cox proportional hazard model, gender (male versus female), pT (T₂-T₄ versus T₁), pN (N₁-N₂ versus N₀), histologic type (non-adenocarcinoma versus adenocarcinoma), smoking history (smoker versus nonsmoker), and IMP-1 expression (positive versus negative) were all significantly related to poor tumor-specific survival among NSCLC patients (P = 0.0286, <0.0001, <0.0001, 0.0003, 0.0220, and 0.0064, respectively; Table 2).

Growth inhibition of NSCLC cells by specific siRNA against IMP-1. To assess whether IMP-1 is essential for growth or survival of lung-cancer cells, we constructed plasmids to express siRNAs against IMP-1 (si-IMP-#1, si-IMP-#2, and si-IMP-#3) and control plasmids (siRNAs for EGFP and 5S-16S) and transfected them into lung cancer cell lines A549 and LC319. The mRNA levels in cells transfected with si-IMP-1-#2 or si-IMP-1-#3 were significantly decreased in comparison with cells transfected with either control siRNAs or si-IMP-1-#1. We observed significant decreases in the number of colonies and in the numbers of viable cells measured by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay, suggesting that up-regulation of IMP-1 is related to growth or survival of cancer cells (representative data of A549 were shown in Fig. 3A and B).

Activation of cellular migration by IMP-1. As the immunohistochemical analysis on tissue microarray had indicated that NSCLC patients with IMP-1–positive tumors showed shorter cancer-specific survival periods than those with IMP-1–negative tumors, we examined a possible role of IMP-1 in cell migration and invasion using Matrigel assays, using NIH-3T3 cells. As shown in Fig. 3C, transfection of IMP-1 cDNA into NIH-3T3 cells significantly enhanced its invasive ability through Matrigel, compared with those transfected with mock vector.

Isolation of mRNAs associated with IMP-1 using RNA immunoprecipitation and cDNA microarray. IMP-1 protein is known to exhibit attachments to at least four RNAs (30). IMP-1 binds specifically to (a) one of the two cis-acting, c-myc mRNA instability elements (31); (b) the 5′-untranslated region of the leader-3 IGF-II mRNA, which represents the major embryonic form of this message (17); (c) the H19 RNA, a gene product exhibiting an oncofetal pattern of expression (23); and (d) the neuron-specific tau mRNA that encodes a microtubule-associated protein localized primarily in the cell body and axon of developing neurons (32). However, expression pattern of these mRNAs in lung cancers we examined were not necessarily concordant with that of IMP-1 (data not shown). Therefore, to elucidate the function of IMP-1 in pulmonary carcinogenesis, we searched for another candidate mRNA(s) that would interact with IMP-1 and might thereby play important roles in growth and/or progression of lung cancer using RNA immunoprecipitation and cDNA microarray (IP-microarray). First, we co-hybridized Cy5-labeled mRNAs that were immunoprecipitated with IMP-1 (IP-mRNA) and Cy3-labeled total RNAs isolated from A549 cells on cDNA microarrays. Then, to identify the up-regulated genes in A549 cells compared with normal lung tissues, we co-hybridized with Cy5-labeled total RNAs isolated from A549 cells and Cy3-labeled polyadenylate RNAs derived from normal lung (Clontech). Among 27,648 genes screened, we identified a total of 22 transcripts that were both enriched in IMP-1-IP-mRNA(s) (>2-fold intensity) and overexpressed (>2-fold intensity) in A549 cell line compared with normal lung (Table 3). The 22 genes represented a variety of functions, including genes involved in signal transduction (SMAD3 and RAN), cell adhesion and cytoskeleton (AMIGO2 and LASP1), ubiquitination (UBE2S and RNF20), and some phosphatases (PTP4A1 and SYNJ2; refs. 33–39). Several of them have been indicated to have important roles in carcinogenesis (e.g., involvement of AMIGO2 LASP1, SYNJ2, and PTP4A1 in cell invasion and migration; refs. 34, 35, 38, 39).

ACTB mRNA is transported to the leading lamellae of chicken embryo fibroblasts and to the growth cones of developing neurons (40, 41). The localization of ACTB mRNA depends on the “zipcode,” a cis-acting element in the 3′ untranslated region of the mRNA (42). The respective trans-acting factor, zipcode-binding protein 1, was identified by affinity purification with the zipcode of ACTB mRNA, and it seems to shuttle this RNA to the leading edge of migrating cells (43). Homologues of zipcode-binding protein 1 have since been identified in a wide range of organisms, including frog, fly, mouse, and human (44–46). Zipcode-binding protein 1–like proteins contain two RNA recognition motifs in the NH₂-terminal region and four ribonucleoprotein K homology domains at the COOH-terminal end. IMP-1, one of the IGF2 mRNA-binding proteins, is considered a member of the zipcode-binding protein 1 family. It exhibits multiple attachments to IGF2 leader-3 mRNA and is overexpressed in several human cancers (18–21). In this study, we confirmed by siRNA and cell invasion experiments that IMP-1 could play a significant role in the tumor cell growth/survival and tumor progression. Furthermore, we undertook RNA immunoprecipitation experiments coupled with cDNA microarrays (IP-microarray) and identified dozens of candidate mRNAs that were likely to be associated with IMP-1 in NSCLC cells (see Table 3). The list included many genes encoding proteins functioning in signal transduction, cell adhesion and cytoskeleton, and those having various types of enzymatic activities. For example, RAN (ras-related nuclear protein) is a small GTP-binding protein belonging to the RAS superfamily that is essential for the translocation of RNA and proteins through the nuclear pore complex (47). Ran system is deregulated in certain cellular contexts: this may represent a favoring condition for the onset and propagation of mitotic errors that can predispose cells to become genetically unstable and facilitate neoplastic growth (48).

Intracellular mRNA transport by RNA-binding proteins has been reported in oocytes and developing embryos of fly and frog and in somatic cells, such as fibroblasts and neurons (49, 50). IMP-1, which is expressed only in cancers and limited normal tissues, such as placenta, testis, and fetal tissues, may be required for the transport of certain mRNAs that play essential roles in embryogenesis and carcinogenesis. Interestingly, induction of exogenous IMP-1 expression into A549 cells did not change the levels of proteins encoded by the IMP-1–associated mRNAs (we used antibodies to EPHA7 and IMP-3 listed in Table 3 for Western blotting; data not shown), indicating that the IMP-1 binding to the various mRNAs is unlikely to affect their protein levels. Therefore, it is rather speculated that proliferating germ cells or cancer cells may actively distribute indispensable mRNAs in cells through the transporting system involved in the IMP-1 protein-mRNA complex. The evidence that IMP-1 associates with various mRNAs encoding proteins involved in cell cycle progression, cell invasion and migration, and various types of enzymatic activities supports that hypothesis. In fact, IMP-1 positivity was correlated with tumor extension factor (pT classification) by clinicopathologic investigation using tissue microarray. Further investigations of IMP-1–associated mRNAs could lead to a better understanding of the development of NSCLCs.

We also showed that IMP-1 is expressed significantly higher in lung cancer cells than normal lung cells. Copy number gains at the IMP-1 loci (17q21) were observed in 18.3% of primary breast cancers (19). To elucidate the mechanism of IMP-1 overexpression, we examined gene amplification of IMP-1 by semiquantitative genomic PCR analysis of a couple of lung cancer materials, but we found no amplification of the IMP-1 loci in these cases (data not shown), implying that IMP-1 overexpression should be not regulated by genetic alterations but by epigenetic mechanisms. We also found that IMP-1 might play an important role in the development/progression of lung cancers. In particular, our results showed that IMP-1 overexpression is associated with lung cancer progression, resulting in a poor prognosis for patients with lung cancer. Concordantly, induction of exogenous expression of IMP-1 enhanced the cellular migration/invasive activity of mammalian cells. IMP-1 overexpression in resected specimens may be a useful index for application of adjuvant therapy to the patients who are likely to have poor prognosis. Furthermore, our data indicated that up-regulation of IMP-1 is related to growth or survival of cancer cells. Although the molecular mechanisms underlying increased IMP-1 expression levels in many cancer cells have not been elucidated, IMP-1 may represent a promising molecular target for human cancer treatment.