Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Identification of a metastasis signature and the DLX4 homeobox protein as a regulator of metastasis by combined transcriptome approach


Although widespread metastasis is the major cause of human lung cancer-related deaths, its underlying mechanism remains largely unclear. Our genome-wide comparison of the expression profiles of a highly metastatic lung cancer cell line, NCI-H460-LNM35 (LNM35), and its parental clone, NCI-H460-N15 (N15), resulted in the identification of a cancer metastasis signature composed of 45 genes. Through gene ontology analysis, our study also provided insights into how this 45-gene metastasis signature may contribute to the acquisition of metastatic potential. By applying the signature to datasets of human cancer cases, we could demonstrate significant associations with a subset of cases with poor prognosis not only for the two datasets of cancers of the lung but also for cancers of the breast. Furthermore, we were able to show that enforced expression of the DLX4 homeobox gene, which was identified as a gene with significant downregulation in LNM35 as well as with significant association with favorable prognosis for lung cancer patients, markedly inhibited in vitro motility and invasion as well as in vivo metastasis via both hematogenous and lymphogenous routes. Taken together, these findings indicate that our combined transcriptome analysis is an efficient approach in the search for genes possessing both clinical usefulness in terms of prognostic prediction in human cancer cases and clear functional relevance for studying cancer biology in relation to metastasis.


Cancer metastasis is the major cause of cancer-related deaths. It has long been understood that a minute fraction of primary cancer cells acquire multiple genetic and/or epigenetic changes in a multi-step manner, which eventually confer the ability to metastasize to distant organs (Fidler, 2003). This concept of an evolution of metastatic clones within the primary tumor site was proposed mostly on the basis of the findings in experimental models, but a provocative study by Ramaswamy et al recently provided intriguing evidence for the possible acquisition of metastatic potential early in carcinogenesis, which may consequently confer metastatic capabilities to the bulk of primary tumors (Ramaswamy et al., 2003; Pantel and Brakenhoff, 2004).

Considerable advances have been attained in the understanding of the molecular carcinogenesis of lung cancer (Osada and Takahashi, 2002), but it still remains the leading cause of cancer death in economically developed countries including Japan (Jemal et al., 2005). The long-term survival rate continues to be unsatisfactory, and no more than 50% of the cases who successfully undergo potentially curative resection survive for more than 5 years after surgery, the remaining cases eventually suffering widespread metastases or local recurrence. Unfortunately, very little is known about how lung cancer cells give rise to distant metastasis. Identification of molecules with a crucial role in distant spread is therefore of the utmost importance to reduce the intolerable number of deaths owing to this devastating disease.

To this end, we previously established a highly metastatic subline, NCI-H460-LNM35 (hereafter referred to as LNM35), which consistently and spontaneously metastasizes through both hematogenous and lymphogenous routes when injected either subcutaneously or orthotopically (Kozaki et al., 2000). In the present study, we aimed at identifying a metastasis signature with both functional and clinical relevance, first selecting genes with abundant and differential expression between LNM35 and N15 cells, and then questioning whether they could be used to discriminate lung cancer patients in terms of postoperative prognoses. Finally, we determined whether the DLX4 homeobox gene, one of the metastasis signature genes with uncharacterized features in relation to metastasis, could functionally play a role in metastasis-related abilities both in vitro and in vivo.


Identification of genes associated with the acquisition of metastatic potential in LNM35

In order to identify genes differentially expressed between highly metastatic LNM35 and low metastatic N15 cell lines, we compared the expression profiles of 11 168 probes. A total of 4849 probes remained after the first filtering, which omitted probes with expression values below 0.1. Two-hundred and forty-nine probes further passed the 2.0-SD variability threshold filtering, and 60 probes corresponding to 45 unique genes also met the ‘robustness criteria’. Twenty-seven genes were overexpressed, whereas 18 were significantly reduced in association with the acquisition of a metastatic capability during in vivo selection of LNM35 (Table 1). The 27 examples overexpressed in LNM35 included various genes previously suggested to play a functional role in metastasis such as chemokine (C-X-C motif) ligand 1 (CXCL1) (Bobrovnikova-Marjon et al., 2004; Gallagher et al., 2005), pituitary tumor-transforming 1 (PTTG1) (Heaney et al., 2000; Bernal et al., 2002), immediate early response 3 (IER3) (Goswami et al., 2004; Nicholson et al., 2004), defender against cell death 1 (DAD1) (Ayala et al., 2004; Goswami et al., 2004; Kim et al., 2006) and methionyl aminopeptidase 2 (METAP2) (Dasgupta et al., 2005; Morowitz et al., 2005). Similarly, genes with higher expression in N15 also included some previously implicated in metastasis such as cadherin 1 (E-cadherin) (Cavallaro and Christofori, 2004) and lectin galactoside-binding soluble 3 (galectin-3) (Liu and Rabinovich, 2005).

Table 1 Genes differently expressed in LNM35 vs. N15

Characterization of the 45-gene metastatic signature using gene ontology terms

In order to gain functional insight into how the metastasis signature genes might contribute in the acquisition of metastatic potential in LNM35, we therefore employed our Gene Ontology (GO) term identifier of functions and other characteristics that are utilized in association with certain phenotypes of interest (Takeuchi et al., 2006). With this identifier, we found that GO terms related to four biological processes, three molecular functions and one cellular component were significantly more frequently observed than expected (P<0.005), when the frequency of a GO term appearing in the 45-gene metastatic signature was compared with that in the entire set of 8644 unique genes on the microarrays used in this study (Table 2). The acquisition of metastatic potential in LNM35 was consequently suggested to be associated with changes in microtubule-related biological processes (P<0.0001 for microtubule polymerization and P<0.0001 for microtubule-based movement) and GTP-related molecular functions (P<0.0001 for GTPase activity and P= 0.0015 for GTP binding). In addition, it is worth noting that morphogenesis (P=0.0029) was selected as a statistically significant term.

Table 2 Gene Ontology (GO) terms significantly related to the metastasis signature

Clinical significance of the 45-gene metastatic signature for postoperative prognosis for human lung and breast cancers

As the next step towards investigating the relevance and generality of the metastasis signature, we asked whether the 45-gene metastasis signature identified in our metastasis model could be applied to human lung cancer specimens to predict clinical behavior. We analyzed a dataset of 50 surgically treated cases of non-small cell lung cancers (NSCLCs), acquired in our previous study, using the same microarray platform (Aichi dataset) (Tomida et al., 2004). Unsupervised hierarchical clustering, using the 45 metastasis-signature gene set, clearly differentiated the 50 NSCLC cases into two distinct clusters, cluster 1 and cluster 2 (Figure 1a). It was intriguing that patients in cluster 1 showed significantly more unfavorable prognosis than those in cluster 2 (Figure 1b; P=0.0234, log-rank test). It is worth noting that 18 of 22 genes (82%) clustered in the upper branch of the side tree were derived from LNM35-overexpressed genes (indicated in yellow), whereas 14 of 23 genes (61%) in the lower branch of the side tree were derived from N15-overexpressed/LNM35-downregulated genes (indicated in green) (P=0.0058, Fisher's exact test). In addition, multivariate analysis revealed that metastasis signature defined cluster to be a significant prognostic factor independent of disease stage in terms of five-year postoperative survival in this dataset (Table 3). These findings suggested that the 45-gene expression signature, which was captured as a metastasis signature through analysis of our metastasis model, may also play a role in human lung cancer metastasis. It should be noted that the 45-gene metastasis signature did not overlap with our previous 25-gene prognosis prediction classifier, which was constructed specifically for analyzing the Aichi dataset, possibly because of the use of distinct filtering criteria (Tomida et al., 2004). However, clusters 1 and 2 in Figure 1a predominantly include cases with predictions of fatal and favorable prognoses, which seems to show consistency with the notion that most lung cancer-related deaths are attributable to metastasis.

Figure 1

Kaplan–Meier survival curves based on favorable/unfavorable categorization of hierarchical clustering analysis using the 45-gene metastasis signature. (a) Analysis of the Aichi dataset comprising 50 NSCLC cases, showing clear distinctions between the two resulting clusters. Columns represent individual patients and rows represent genes constituting the 45-gene metastasis signature. Yellow and green boxes on the left indicate upregulated and downregulated expression in LNM35, respectively, in comparison with N15. Boxes below indicate the results of prognosis prediction (red for fatal and blue for favorable), using our previous 25-gene prognosis classifier for NSCLC constructed with the Aichi dataset. (b) Kaplan-Meier survival curves of cluster 1 and cluster 2 for Aichi 50 NSCLC cases.

Table 3 Multivariate Cox regression analysis of potential prognostic factors for survival 5 years after surgery

The general applicability of the 45-gene metastasis signature was further examined with the aid of an additional set of 62 Dana-Faber/MIT (Boston) stage I/II lung adenocarcinoma cases, which includes those with distinctive postoperative prognoses, i.e., either ‘alive without recurrence for at least 4 years’ or ‘dead with recurrence’ (Ramaswamy et al., 2003). Again, we observed the presence of two major branches with distinctly different postoperative survival (P=0.0554, Figure 2a). The applicability of the signature to other solid tumors was similarly examined by using a Netherlands Cancer Institute (Amsterdam) dataset of breast cancer as well as Dana-Faber/MIT (Boston) datasets of medulloblastoma and diffuse large B-cell lymphoma (Ramaswamy et al., 2003), all of which contained information of either overall survival or time to metastasis. The Amsterdam dataset of 78 cases with stage I primary breast adenocarcinomas with at least 5-year clinical follow-up was analyzed by means of unsupervised hierarchical clustering analysis. Our metastasis signature again clearly showed two major groups with a significant difference in postoperative prognosis (P=0.0173, Figure 2b). We also observed a similar difference between two major clusters in the hierarchical clustering analysis of the Boston dataset of 60 medulloblastomas with at least 5-year clinical follow-up after multimodality treatment (P=0.0507, Figure 2c). Interestingly, however, the analysis of the Boston dataset of 58 diffuse large B-cell lymphomas did not identify any statistically significant differences (P=0.876, Figure 2d).

Figure 2

Analysis of Boston dataset of 62 lung adenocarcinomas (a) and Amsterdam dataset of 78 breast cancers (b) as well as Boston datasets of 60 medulloblastomas (c) and 58 diffuse large B-cell lymphomas (d), showing significantly different prognosis in solid tumors but not in hematologic malignancy, according to the clusters based on the expression profiles of the 45-gene metastasis signature.

Identification of DLX4 as a metastasis signature gene with the functional involvement in metastasis of LNM35

The purpose of the combined transcriptome analysis is to identify genes that are both clinically useful for prognosis for human cancer cases and functionally relevant for studying cancer biology in relation to metastasis. We therefore performed in silico screening by applying permutation t-test analysis to the Aichi dataset used for hierarchical clustering analysis, and found that the distal-less homeobox 4 (DLX4) was the most significantly favorable prognosis-associated gene (P=0.0234) (Supplementary Table). As the functional relationship between DLX4 and metastasis has not yet been explored, we further investigated whether augmentation of the expression level of DLX4 could affect the invasive and metastatic nature of LNM35.

Our microarray analysis revealed that low-metastatic N15 cells expressed the favorable-prognosis-associated gene, DLX4, at a level 2.5-fold higher than highly metastatic LNM35 cells, which was also confirmed at the protein level by Western blot analysis showing readily detectable expression of the DLX4 protein in N15 cells in contrast to negligible expression in LNM35 cells (Figure 3a). A DLX4 expression construct was then stably transfected into LNM35, yielding DLX4-transfected LNM35 cell clones stably expressing DLX4 at a level comparable to N15 cells. Their morphologies and growth rates were not noticeably altered (Figure3b and data not shown for morphologies), but DLX4 transfectants had significantly reduced capabilities in terms of both motility and invasion in vitro (Figures 3c and d). Motility of two DLX4 transfectants was significantly reduced down to less than 30% of the LNM35 level, and invasiveness of both transfectants was similarly reduced. Furthermore, the introduction of DLX4 into the highly metastatic LNM35 cells clearly abrogated metastasis in vivo, resulting in the marked reduction in the abilities to form metastases via both hematogenous and lymphogenous routes (Figures 3e and f).

Figure 3

Introduction of DLX4 into the highly metastatic LNM35 cells, showing inhibition of the invasive and metastatic phenotype. (a) DLX4 protein expression in LNM35, DLX4-transfected LNM35 clones 17 and 18 as well as in the low-metastatic N15, a parental clone of LNM35. (b) Comparison of in vitro growth characteristics. (c and d) Comparison of in vitro motility and invasion. Marked reduction is evident in DLX4-transfected clones 17 and 18, but not in vector control (VC1), to levels comparable with those for N15. Significant reduction of in vivo metastasis via both hematogenous (e) and lymphogenous (f) routes with DLX4-transfected LNM35 clones 17 and 18.


It is well accepted that the acquisition of metastatic phenotype, the most deadly manifestation, is a highly complex process involving multiple genes. Our study clearly showed our 45-gene metastasis signature, selected on the basis of differential expression between highly metastatic- and low-metastatic LNM35 sublines, to be significantly associated with postoperative prognosis in two independent sets of lung cancers, the Aichi and Boston datasets. Furthermore, the 45-gene metastasis signature was also significantly associated with prognosis for breast cancers and medulloblastomas, indicating that it is generic in nature. Interestingly, however, this signature did not appear to be related to outcome for diffuse large B-cell lymphomas, which is consistent with the idea that metastasis of hematopoietic tumors may employ a mechanism or mechanisms specific for navigating the hematologic and lymphoid cells (Ramaswamy et al., 2003). It is of note that this metastasis signature overlaps partly with that identified by Ramaswamy et al. (2003) through comparisons of expression profiles between primary and metastatic sites of human cancers. For example, PTTG1 (also known as securin) and small nuclear ribonucleoprotein (SNRPF) were upregulated in common, and downregulation of nuclear hormone receptor (NR4A1) was included as metastatic phenotype-associated genes in both studies.

A few previous expression profiling studies have used paired metastatic and non-metastatic cell lines for identifying a gene set involved in the process of metastasis. As expected, they identified partially overlapping gene sets, but clearly distinct sets of genes were also selected depending on the models used by each research group. One at Memorial Sloan-Kettering Cancer Center found a gene set associated with either bone or lung metastasis in breast cancer patients using a highly metastatic clone of the MDA-MB-231 human breast cancer cell line, and certain genes such as CXCL1 were shared with our gene set (Minn et al., 2005). Similarly, our metastasis signature gene set contains related to the ubiquitin-proteasome pathway, including UBE1 and PSMD2, whereas others like USP22 and RNF2 were included in the 11-gene signature identified by Glinsky et al. (2005) through analysis of highly metastatic prostate cancer cell line. These findings suggest that certain functions are shared in common between organs for acquiring metastatic phenotype, but at the same time there are also many other constituents that are not shared.

Our GO term analysis also sheds light on the constituents of the 45-gene metastasis signature as a whole. It is interesting that the microtubule-related biological process was found to be significantly associated with the acquisition of metastatic potential, considering that the involvement of microtubules in cell extensions that participate in active cell migration is well established, and that some of the agents that bind to tubulin and disrupt microtubule dynamics are in clinical use for cancer treatment (Jordan and Wilson, 2004). Although the GO term selected as being significant is associated with GTPase-related activities, GTPases are known to be involved in cell morphogenesis through the induction of specific types of actin cytoskeleton and the alignment and stabilization of microtubules. In addition, the selection of morphogenesis as a statistically significant term is intriguing, as tumors are often viewed as corrupt forms of normal developmental processes, and genes that play an important role in embryonic development are frequently found to be altered in cancer (Kang and Massague, 2004). Epithelial-mesenchymal transition (EMT) is a vital step in morphogenesis during embryonic development, and it is also thought to be involved in the conversion of early stage tumors into invasive malignancies (Kang and Massague, 2004). In this connection, it is worth noting that loss of E-cadherin, a hallmark of EMT in cancer, appears to have occurred in LNM35 as part of the process of acquisition of metastatic potential during in vivo selection.

We further demonstrated that the DLX4 homeobox gene, which was reduced in LNM35 and was most significantly associated with favorable prognosis in lung cancer patients, could reduce motility and invasion in vitro as well as their metastasis in vivo through both hematogenous and lymphogenous routes. Our findings, therefore, indicate that DLX4 plays a functional role affecting invasive and metastatic capabilities, and is not a mere surrogate. Homeobox genes of the Distal-less (Dlx) family are expressed in vertebrate embryos at the contact of epithelial cell layers and adjacent mesenchymal cells, playing a role in epithelial-mesenchymal cell interactions (Quinn et al., 1998). In this context it is of particular interest that DLX4 has also been suggested to regulate trophoblast invasion (Quinn et al., 1998). We should also mention that a CpG island in the 5′ upstream region of DLX4 was recently reported to be methylated in breast cancer cell lines (Miyamoto et al., 2005).

In conclusion, expression-profiling analyses of a metastasis model, in combination with human cancer specimens, appear to be a highly relevant way to investigate genes of both functional and clinical importance. Functional clarification of additional components of the 45-gene metastasis signature, including the two as yet uncharacterized genes, should provide a clue to better understanding of processes underlying tumor progression, and may ultimately lead to the development of novel treatment or preventative approaches.

Materials and methods

Cell lines and animals

Establishments of LNM35, a highly metastatic subline of the NCI-H460 human large cell lung cancer cell line, and its parent, N15, have been reported previously (Kozaki et al., 2000). They were maintained in Rosewell's Park Memorial Institute media (RPMI) 1640 medium, supplemented with 10% fetal bovine serum, as described earlier (Kozaki et al., 2000). Five-week-old female severe combined immunodeficient (SCID) mice were purchased from CLEA Japan, Inc. (Tokyo, Japan) and maintained under specific-pathogen-free conditions.

Microarray data acquisition and analyses

Two sets of membrane cDNA microarrays (GeneFilter Human Microarrays Release I and Release II; Invitrogen, Carlsbad, CA, USA), containing a total of 11 168 spots corresponding to 8644 independent genes, were analyzed in duplicate using five microgram of total RNA as detailed in our previous report, with microarray data acquisition as detailed previously (Tomida et al., 2004). The raw data were re-scaled to account for the differences in individual hybridization intensities as follows. First, array spots that showed expression values below 0.1 were omitted from the analysis before normalization. Then, expression levels for each gene were normalized within each of the independent hybridizations by using loess nonlinear normalization in the statistical package R, available at (Gentleman et al., 1994). The average expression values of LNM35 were compared to those of N15, and array spots that met the following ‘robustness criteria’ were selected; expression values of at least 1.0 in the two independent hybridizations in LNM35 and/or N15 as well as differential expression at the level of more than 2.0-SD of that considering all the spots. Averaged expression values of probes corresponding to the same unique gene were used for further analysis.

Hierarchical clustering and survival analysis

Expression profiles of the 45-gene metastasis signature in 50 NSCLC cases were extracted from our previous expression profiling dataset (Tomida et al., 2004), which is available at GEO as GSE4705 and GSE4716. The microarray datasets of Boston lung cancers, Amsterdam breast cancers and Boston medulloblastomas and diffuse large B-cell lymphomas, described in a paper by Ramaswamy (Ramaswamy et al., 2003), are publicly available at as datasets B, C, E and F, respectively. The GPL91 platform file, available at GEO (, was used to annotate each Affymetrix probe set ID of dataset B with the UniGene ‘Gene Symbol’. We used the UniGene Gene Symbols for our cross-platform mapping of genes, and 40 genes in dataset B were mapped from among the 45-gene set. These genes were then used for hierarchical clustering of the Boston lung data set. The gene symbols contained in the ‘Description’ column of dataset C were similarly used for mapping, with the result that 41 genes in dataset C were mapped from among the 45-gene set, and then used for the Amsterdam breast data set. For datasets E and F, the GPL80 platform file was used for cross-platform mapping of the genes, resulting in the mapping of 28 genes from among the 45-gene set. Averaged expression values of probes corresponding to the same unique gene were further analyzed. The CLUSTER (Eisen et al., 1998) program was used for average linkage hierarchical clustering of both genes and cases by means of median centering and normalization. The results were displayed with the aid of TREEVIEW (Eisen et al., 1998), and the resultant two clearly distinct clusters containing predominantly LNM-35-upregulated and downregulated genes were termed ‘Cluster 1’ and ‘Cluster 2’, respectively. The Kaplan–Meier method was used to estimate survival as a function of time, and survival differences were analyzed with the log-rank test. Cox proportional hazards modeling was performed to identify which independent factors might jointly have a significant effect on survival. All the statistical analyses were performed with Stata software (version 7; Stata Corp, College Station, TX, USA).

Bioinformatic analyses

GO (Ashburner et al.,. 2000) analysis was employed to highlight functionally distinct biological features of a gene set associated with the acquisition of invasive and metastatic capabilities in LNM35, as described previously (Takeuchi et al., 2006). Briefly, database files used for this GO analysis were downloaded from the UniGene ftp site. Eventually, 5346 unique genes on the microarray were linked to about 30 000 GO terms by parsing the database files including Hs.seq.all, and LL_tmpl with the program written by Perl. These terms were subjected to Fisher's exact test to identify which GO terms were over- or under-represented in a gene set of interest.

Permutation t-test analysis was used for selecting genes that were significantly different in the favorable and unfavorable prognosis patient subgroups. We then randomly permuted the class labels among the samples and re-computed the t-statistics. This randomization was repeated 1000 times, and the P-value for the observed class label was calculated from the distribution of 1000 sets of t-statistics.

Construction and transfection of the DLX4 expression vector and Western blot analyses

The IMAGE clone (clone ID 3907376) containing a full-length DLX4 cDNA was purchased from Invitrogen. The DLX4 expression vector was constructed by inserting the SmaI-digested DNA fragment from IMAGE clone into the EcoRV site of pcDNA3 (pcDNA3-DLX4) and then sequenced. LNM35-DLX4 stable transfectants (clones 17 and 19) were generated by transfection of 2 μg of pcDNA3-DLX4, using the FuGENE 6 reagent (Roche Applied Science, Indianapolis, IN, USA), according to the manufacturer's instructions. Transfected cells were selected with 1 mg/ml G418, and Western blot analyses were performed with an anti-DLX4 polyclonal antibody (Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA) and a horseradish peroxidase-conjugated secondary antibody (Cell Signaling Technology, Beverly, MA, USA).

In vitro motility and invasion assay and in vivo spontaneous metastasis assay

To quantify in vitro motility and invasion, transwell-chamber culture systems were employed. The upper surface of 6.4 mm-diameter filters with 8 mm pores (BD Biosciences, Bedford, MA, USA) was coated with 0.1 ml of 0.1 mg/ml Matrigel (Collaborative Research Inc., Bedford, MA, USA) for the invasion assay, and the upper chambers were filled with 0.4 ml of serum-free RPMI 1640 medium and placed on culture plates with 24 wells containing 1 ml of the medium. Cells (1.0 × 104 cells for the motility assay, 1.0 × 105 cells for the invasion assay) in 0.1 ml of serum-free RPMI 1640 medium were then added to the upper chambers and cultured for 24 h. The filters were fixed with 70% ethanol and stained with Giemsa, and the cells on the lower surface of the filters were counted in triplicate. Three independent experiments gave similar results, and the representative results are shown for each analysis.

Spontaneous metastasis assay was performed as follows. 1.0 × 107 cells in 0.1 ml of serum-free RPMI 1640 medium were injected into subcutaneous tissue of the right abdominal wall of 6-week-old female SCID mice. Three mice each were transplanted with LNM35, N15 and VC1, respectively, whereas five mice were subcutaneously injected with the LNM35-DLX transfectant. One mouse injected with LNM35-DLX4-19 cells died before the end of the 40-day observation period because of peritonitis carcinomatosa. Forty days after injection, the mice were sacrificed by cervical dislocation under deep anesthesia and their lungs, lymph nodes and subcutaneous tumors were resected, weighed and fixed with 4% formaldehyde. The lung-metastatic nodules were examined and counted under a dissecting microscope. All the results presented in this study were obtained by averaging these data.

Accession codes




  1. Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM et al. (2000). Gene ontology: tool for the unification of biology The Gene Ontology Consortium. Nat Genet 25: 25–29.

    CAS  Article  Google Scholar 

  2. Ayala G.E, Dai H, Ittmann M, Li R, Powell M, Frolov A et al. (2004). Growth and survival mechanisms associated with perineural invasion in prostate cancer. Cancer Res 64: 6082–6090.

    CAS  Article  Google Scholar 

  3. Bernal JA, Luna R, Espina A, Lazaro I, Ramos-Morales F, Romero F et al. (2002). Human securin interacts with p53 and modulates p53-mediated transcriptional activity and apoptosis. Nat Genet 32: 306–311.

    CAS  Article  Google Scholar 

  4. Bobrovnikova-Marjon EV, Marjon PL, Barbash O, Vander Jagt DL, Abcouwer SF . (2004). Expression of angiogenic factors vascular endothelial growth factor and interleukin-8/CXCL8 is highly responsive to ambient glutamine availability: role of nuclear factor-kappaB and activating protein-1. Cancer Res 64: 4858–4869.

    CAS  Article  Google Scholar 

  5. Cavallaro U, Christofori G . (2004). Cell adhesion and signalling by cadherins and Ig-CAMs in cancer. Nat Rev Cancer 4: 118–132.

    CAS  Article  Google Scholar 

  6. Dasgupta B, Yi Y, Hegedus B, Weber JD, Gutmann DH . (2005). Cerebrospinal fluid proteomic analysis reveals dysregulation of methionine aminopeptidase-2 expression in human and mouse neurofibromatosis 1-associated glioma. Cancer Res 65: 9843–9850.

    CAS  Article  Google Scholar 

  7. Eisen MB, Spellman PT, Brown PO, Botstein D . (1998). Cluster analysis and display of genome-wide expression patterns. Proc Natl Acad Sci USA 95: 14863–14868.

    CAS  Article  Google Scholar 

  8. Fidler IJ . (2003). The pathogenesis of cancer metastasis: the ‘seed and soil’ hypothesis revisited. Nat Rev Cancer 3: 453–458.

    CAS  Article  Google Scholar 

  9. Gallagher PG, Bao Y, Prorock A, Zigrino P, Nischt R, Politi V et al. (2005). Gene expression profiling reveals cross-talk between melanoma and fibroblasts: implications for host-tumor interactions in metastasis. Cancer Res 65: 4134–4146.

    CAS  Article  Google Scholar 

  10. Gentleman RC, Lawless JF, Lindsey JC, Yan P . (1994). Multi-state Markov models for analysing incomplete disease history data with illustrations for HIV disease. Stat Med 13: 805–821.

    CAS  Article  Google Scholar 

  11. Glinsky GV, Berezovska O, Glinskii AB . (2005). Microarray analysis identifies a death-from-cancer signature predicting therapy failure in patients with multiple types of cancer. J Clin Invest 115: 1503–1521.

    CAS  Article  Google Scholar 

  12. Goswami S, Wang W, Wyckoff JB, Condeelis JS . (2004). Breast cancer cells isolated by chemotaxis from primary tumors show increased survival and resistance to chemotherapy. Cancer Res 64: 7664–7667.

    CAS  Article  Google Scholar 

  13. Heaney AP, Singson R, McCabe CJ, Nelson V, Nakashima M, Melmed S . (2000). Expression of pituitary-tumour transforming gene in colorectal tumours. Lancet 355: 716–719.

    CAS  Article  Google Scholar 

  14. Jemal A, Murray T, Ward E, Samuels A, Tiwari RC, Ghafoor A et al. (2005). Cancer statistics, 2005. CA Cancer J Clin 55: 10–30.

    Article  Google Scholar 

  15. Jordan MA, Wilson L . (2004). Microtubules as a target for anticancer drugs. Nat Rev Cancer 4: 253–265.

    CAS  Article  Google Scholar 

  16. Kang Y, Massague J . (2004). Epithelial-mesenchymal transitions: twist in development and metastasis. Cell 118: 277–279.

    CAS  Article  Google Scholar 

  17. Kim YH, Girard L, Giacomini CP, Wang P, Hernandez-Boussard T, Tibshirani R et al. (2006). Combined microarray analysis of small cell lung cancer reveals altered apoptotic balance and distinct expression signatures of MYC family gene amplification. Oncogene 25: 130–138.

    CAS  Article  Google Scholar 

  18. Kozaki K, Miyaishi O, Tsukamoto T, Tatematsu Y, Hida T, Takahashi T . (2000). Establishment and characterization of a human lung cancer cell line NCI-H460-LNM35 with consistent lymphogenous metastasis via both subcutaneous and orthotopic propagation. Cancer Res 60: 2535–2540.

    CAS  PubMed  Google Scholar 

  19. Liu FT, Rabinovich GA . (2005). Galectins as modulators of tumour progression. Nat Rev Cancer 5: 29–41.

    CAS  Article  Google Scholar 

  20. Minn AJ, Gupta GP, Siegel PM, Bos PD, Shu W, Giri DD et al. (2005). Genes that mediate breast cancer metastasis to lung. Nature 436: 518–524.

    CAS  Article  Google Scholar 

  21. Miyamoto K, Fukutomi T, Akashi-Tanaka S, Hasegawa T, Asahara T, Sugimura T et al. (2005). Identification of 20 genes aberrantly methylated in human breast cancers. Int J Cancer 116: 407–414.

    CAS  Article  Google Scholar 

  22. Morowitz MJ, Barr R, Wang Q, King R, Rhodin N, Pawel B et al. (2005). Methionine aminopeptidase 2 inhibition is an effective treatment strategy for neuroblastoma in preclinical models. Clin Cancer Res 11: 2680–2685.

    CAS  Article  Google Scholar 

  23. Nicholson BE, Frierson HF, Conaway MR, Seraj JM, Harding MA, Hampton GM et al. (2004). Profiling the evolution of human metastatic bladder cancer. Cancer Res 64: 7813–7821.

    CAS  Article  Google Scholar 

  24. Osada H, Takahashi T . (2002). Genetic alterations of multiple tumor suppressors and oncogenes in the carcinogenesis and progression of lung cancer. Oncogene 21: 7421–7434.

    CAS  Article  Google Scholar 

  25. Pantel K, Brakenhoff RH . (2004). Dissecting the metastatic cascade. Nat Rev Cancer 4: 448–456.

    CAS  Article  Google Scholar 

  26. Quinn LM, Latham SE, Kalionis B . (1998). A distal-less class homeobox gene, DLX4, is a candidate for regulating epithelial-mesenchymal cell interactions in the human placenta. Placenta 19: 87–93.

    CAS  Article  Google Scholar 

  27. Ramaswamy S, Ross KN, Lander ES, Golub TR . (2003). A molecular signature of metastasis in primary solid tumors. Nat Genet 33: 49–54.

    CAS  Article  Google Scholar 

  28. Takeuchi T, Tomida S, Yatabe Y, Kosaka T, Osada H, Yanagisawa K et al. (2006). Expression profile-defined classification of lung adenocarcinoma shows close relationship with underlying major genetic changes and clinicopathologic behaviors. J Clin Oncol 24: 1679–1688.

    CAS  Article  Google Scholar 

  29. Tomida S, Koshikawa K, Yatabe Y, Harano T, Ogura N, Mitsudomi T et al. (2004). Gene expression-based, individualized outcome prediction for surgically treated lung cancer patients. Oncogene 23: 5360–5370.

    CAS  Article  Google Scholar 

Download references


We thank Keishi Katoh for his helpful discussion on bioinformatic analysis. This work was supported by a Grant-in-Aid for Scientific Research on Priority Areas from the Ministry of Education, Culture, Sports, Science and Technology of Japan and a Grain-in-Aid for Scientific Research (B) from the Japan Society for the Promotion of Science.

Author information



Corresponding author

Correspondence to T Takahashi.

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

Tomida, S., Yanagisawa, K., Koshikawa, K. et al. Identification of a metastasis signature and the DLX4 homeobox protein as a regulator of metastasis by combined transcriptome approach. Oncogene 26, 4600–4608 (2007).

Download citation


  • lung cancer
  • metastasis
  • expression profile
  • DLX4
  • gene ontology

Further reading


Quick links