HOXD3 Enhances Motility and Invasiveness Through the TGF-B-Dependent and -Independent Pathways in A549 Cells

HOXD3 enhances motility and invasiveness through the TGF-b-dependent and -independent pathways in A549 cells

Yasumasa J Miyazaki1,2, Jun-ichi Hamada*,1, Mitsuhiro Tada1, Keiji Furuuchi1, Yoko Takahashi1, Satoshi Kondo2, Hiroyuki Katoh2 and Tetsuya Moriuchi1

1Division of Cancer-Related Genes, Institute for Genetic Medicine, Hokkaido University, Kita-15, Nishi-7, Kita-ku, Sapporo 060-0815, Japan; 2Surgical Oncology, Cancer Medicine, Division of Cancer Medicine, Hokkaido University Graduate School of Medicine, Kita-15, Nishi-7, Kita-ku, Sapporo 060-8638, Japan

Homeobox genes regulate sets of genes that determine regulates the transcription of genes relevant to the cellular fates in embryonic morphogenesis and maintenance formation of speci®c segmental architecture (McGinnis of adult tissue architecture by regulating cellular motility and Krumlauf, 1992). Homeobox-containing genes are and cell-cell interactions. Our previous studies showed that subdivided into more than 20 classes according to their a speci®c member, HOXD3, when overexpressed, upregu- primary sequences. Class I homeobox-containing genes lates integrin b3 expression in human erythroleukemia were the ®rst to be discovered and have been most HEL cells and lung cancer A549 cells, and enhances their extensively studied. In mammals, 39 class I homeobox motility and invasiveness. We performed a microarray genes are clustered in a similar arrangement of 13 study of over 7075 genes to determine the mechanisms paralog groups on four dierent chromosomal/genomic underlying the HOXD3-enhanced motility and invasiveness regions, HOXA, B, C, and D (Graham et al., 1989; in A549 cells. RT ± PCR-based tracking gene analyses Apiou et al., 1996; Mark et al., 1997). They are highlighted a set of TGF-b-upregulated genes, which expressed in a spatiotemporal manner during embryo- included matrix metalloproteinase-2, syndecan-1, CD44, nic morphogenesis, each regulating a group of genes and TGF-b-induced 68 kDa protein. Exogenous TGF-b involved in modeling a speci®c segmental architecture. also caused this pattern of upregulation in A549 cells and Class I homeobox-containing genes have also been enhanced their migratory and invasive activity, con®rming demonstrated in normal adult tissues with character- the involvement of TGF-b signaling. However, HOXD3 istic patterns, suggesting their possible role in the reduced the expression of TGF-b-independent genes coding maintenance of tissue-speci®c architecture (Cillo et al., for desmosomal components such as desmoglein, desmo- 1992; Cillo, 1994 ± 1995). plakin and plakoglobin which are known to suppress tumor The deregulated expressions of HOX genes have invasion and metastasis. These results suggest that HOXD3 been observed in certain cancers. In acute myeloid enhances the invasive and metastatic potential of cancer leukemia, a chromosomal translocation results in the cells through the TGF-b-dependent and -independent fusion of the nuclear pore complex protein NUP98 and pathways. HOXA9 protein, which seems to promote leukemogen- Oncogene (2002) 21, 798 ± 808. DOI: 10.1038/sj/onc/ esis through inhibition of HOXA9-mediated dierentia- 1205126 tion (Borrow et al., 1996; Nakamura et al., 1996a). Proviral activation of Hoxa9 and Hoxa7 by retro- Keywords: homeobox gene; HOXD3; microarray; lung viruses has been shown to be involved in leukemia cancer cells; TGF-b; invasion development in a mouse myeloid leukemia model (Nakamura et al., 1996b). In solid tumors, HOX genes exhibit altered expression patterns in human kidney, Introduction colon and lung cancers, compared to those in normal organs (Cillo et al., 1992, 1999; De Vita et al., 1993; Homeobox-containing genes are the master regulators Tiberio et al., 1994). Altered HOX gene expression is of cell dierentiation and morphogenesis in animals also noted in metastatic lesions of lung and colon (Gehring and Hiromi, 1986). They contain a common cancers, compared to those in their primary lesions sequence element of 183 bp, the homeobox, which (Cillo, 1994 ± 1995; De Vita et al., 1993). encodes a highly conserved 61-amino-acid homeodo- We previously showed that overexpression of the main. The homeodomain is responsible for recognition HOXD3 gene enhanced integrin b3 expression in both and binding of sequence-speci®c DNA motifs, and cis- human erythroleukemia HEL cells and lung carcinoma A549 cells, and that these cells acquired strong ability to adhere to and migrate toward the integrin b3 ligands. However, this ®nding was not observed in the *Correspondence: J-i Hamada; E-mail: [email protected] Received 30 April 2001; revised 2 October 2001; accepted 29 control cells unexpressing HOXD3 gene (Taniguchi et October 2001 al., 1995; Hamada et al., 2001). The HOXD3-over- HOXD3-response genes in human cancer cells YJ Miyazaki et al 799 expressing A549 cells acquired ability to produce large of 6185 genes (87.4%) did not dier between HOX+2 amounts of extracellular matrix-degrading enzymes and Neo1 cells. The remaining 433 genes (6.1%) were including urokinase-type plasminogen activator (uPA) not expressed in either samples. and matrix metalloproteinase-2 (MMP-2), resulting in To verify the microarray results, we performed semi- the enhancement of in vitro cell invasion of Matrigel quantitative RT ± PCR on RNA extracted from the (Hamada et al., 2001). Boudreau et al. (1997) noted the two HOXD3-overexpressing clones (HOX+1 and HOXD3-mediated conversion of endothelium from HOX+2) and two control clones transfected with the resting to an activated angiogenic state. They showed empty vector (Neo1 and Neo2). PCR products that stimulation of endothelial cells with basic separated by agarose gel electrophoresis were stained ®broblast growth factor (bFGF) caused increased with ethidium bromide, and observed under UV light expression of HOXD3, and enhanced expression of (Figure 1). The intensity of each band relative to both the integrin avb3 and uPA in endothelial cells. control was analysed by densitometry. As shown in These lines of evidence suggest that HOXD3 plays a Table 1, the HOXD3-responsive genes indicated by the pivotal role in the regulation of genes related to microarray were classi®ed into ®ve groups: (1) invasion and metastasis. However, these studies remain extracellular matrix (ECM) components; (2) cell incomplete in our attempt of understanding down- adhesion molecules; (3) molecules associated with stream genes of the HOXD3 gene. To better under- ECM-degradation; (4) cytoskeletal system-associated stand the mechanisms of HOXD3-mediated molecules; and (5) growth factors, cytokines and their enhancement of invasive and metastatic potential, we related molecules. Most genes pro®led dierentially by monitored eects of HOXD3-overwxpression on global the microarray analysis were likewise characterized by gene expression by using a human cDNA microarray RT ± PCR. However, in seven gene expressions (Devel- of 7,075 genes. This analysis highlighted the involve- opmental endothelial locus 1, tissue transglutaminase 2, ment of many eectors, especially molecules associated galectin 3, annexin VIII, interferon g-inducible protein with cell-cell and cell-extracellular matrix interactions 16, bone morphogenetic protein 5 and neurotensin), and the activation of a TGF-b-regulated pathway in HOXD3-overexpressing A549 cells.

Results

Identification of genes responsive to HOXD3 transduction by cDNA microarray analysis We analysed the downstream eects of HOXD3 in A549 cells, using 7075 human cDNA microarrays. A HOXD3-transfected A549 clone (HOX+2) and a control vector-transfected clone (Neo1) were investigated as representative cell populations. Of the 7075 genes analysed, 6438 (91.0%) satis®ed the examination criterion (see Materials and methods). In HOX+2 cells, the signal intensities (¯uorescence units) ranged from 127 (Human PDGF-associated protein mRNA, com- plete cds, U41745) to 27 785 (solute carrier family 24 member 1, AF062921). In Neo1 cells, the ¯uorescence signal ranged from 86 (H sapiens mRNA for RP3 gene, AI885178) to 10 638 (solute carrier family 24 member 1, AF062921). The ratios of the relative signal of HOX+2/Neo1 varied from 5.6 (thrombospondin 1, X14787) to 1/3.4 (cystein-rich protein 1, AI433969). Our previous study had revealed that HOXD3-overexpression enhanced the expression of the integrin b3inHEL cells and A549 cells (Taniguchi et al., 1995; Hamada et al., 2001). As expression of the integrin b3 was Figure 1 Expression patterns of representative genes responding upregulated 1.6-fold in the microarray analysis, we to HOXD3-overexpression or TGF-b stimulation. RNA was isolated from HOXD3-overexpressing cells (HOX+1 and regarded the increases in the ratio by more than 1.6-fold HOX+2) and the control transfectant cells (Neo1 and Neo2) as upregulation and the decreases to less than 1/1.6 as which had been treated with TGF-b (0, 0.4, 2 or 10 ng/ml) for downregulation by HOXD3-overexpression. We identi- 24 h. (a) The expression of TGF-b-induced 68 kDa protein (big- ®ed 74 genes (70 cDNAs and four ESTs) (1.0%) as h3) was upregulated by HOXD3 and TGF-b;(b) vitronectin was upregulated (more than 1.6-fold) and 383 genes (167 down-regulated by HOXD3 and TGF-b;(c) HOXD3 was not aected by TGF-b, and integrin b3 was upregulated by HOXD3 cDNAs and 216 ESTs) (5.4%) as downregulated (less but not by TGF-b;(d) plakoglobin was down-regulated by than 1/1.6) by HOXD3-overexpression. The expression HOXD3 whereas it was not aected by TGF-b

Oncogene HOXD3-response genes in human cancer cells YJ Miyazaki et al 800 Table 1 Genes that were upregulated or downregulated in A549 cells overexpressing HOXD3 gene Expression ratio Response (HOX+2/Neo1)a RT ± PCRb to TGF-bc

Extracellular matrix components Thrombospondin-1 5.6 : UP TGF-b-induced 68 kDa protein 5.4 : UP Developmental endothelial locus-1 4.3 ::: ± Collagen X, a1 1.7 : ± Syndecan 1 1.6 : UP Vitronectin 1/1.7 ; DOWN Latent TGF-b binding protein-1 1/1.7 ;; ± Cell adhesion molecules CD24 4.1 : ± CD44 2.1 :: ± Caveolin-2 1.7 : ± Integrin b3 1.6 :: ± Desmoglein-1 1/1.6 ; ± Plakoglobin 1/1.6 ;; ± Desmoplakin 1/2.3 ; ± Extracullular matrix-related enzymes and inhibitors Plasminogen activator inhibitor I 2.9 :: UP Tissue transglutaminase 2 2.8 ::: ± Lysyl hydroxylase 2 2.8 : ± Matrix metalloproteinase 2 2.7 :: UP Proteinase inhibitor 12 1/1.9 ? ± Cytoskeleton-related molecules Keratin 1 4.3 : ± Galectin 3 2.6 ::: ± Quiescin Q6 2.4 : ± Transgelin 2.4 :: UP Keratin 7 2.3 N.D. N.D. Keratin 5 2.3 N.D. N.D. Keratin 6B 2.2 N.D. N.D. Transmembrane 4 superfamily member 2 2.2 ? UP Annexin VIII 2.1 ::: ± Keratin 2A 2.0 N.D. N.D. Desmoyokin 1.9 : N.D. Tropomyosin 2 1.9 ? ± Vimentin 1.9 : ± a-actinin 4 1.8 N.D. N.D. Rab 13 small G protein 1.7 : ± Actin-binding LIM protein 1/2.3 ; ± Cytokines, humoral factors and their related molecules Insulin-like growth factor binding protein 3 2.4 : UP Interferon g-inducible protein 16 2.0 ::: ± Complement component 3 1.8 :: ± Insulin-like growth factor binding protein 4 1.7 ? ± Bone morphogenetic protein 5 1/2.0 ;;; ± Clusterin 1/3.0 ; ± Neurotensin 1/3.3 ;;; ±

aListed are genes with 51.6 or 41/1.6 of the relative expression levels (HOX+2/Neo1). bThe microarray results were con®rmed by semiquantitative duplex RT ± PCR. PCR products were electrophoresed in an agarose gel electrophoresis and intensity of the bands observed under a UV illuminator was analysed by Scion Image. Eects of HOXD3-overexpression on the expression levels of the genes listed as assessed by the RT ± PCR are indicated as follows; ?, no apparent eects; :/;, 2- to 5-fold induction/reduction; ::/;;, 5- to 10-fold; :::/;;;, more than 10- fold; and N.D., not determined. cEects of exogenous TGF-b (0.4, 2, and 10 ng/ml) on the expression levels of the genes listed as assessed by the RT ± PCR are indicated as follows: Up, upregulated; Down, downregulated; ± , no apparent eects; and N.D., not determined.

the ratios of the relative signal of HOX+2/Neo1 by tissue transglutaminase (2.8-fold), matrix metallopro- RT ± PCR were dierent from those by the microarray teinase 2 (2.7-fold) and CD44 (2.1-fold). These analysis. Nonetheless, none of them showed an inverse expressions were previously documented to be upregu- pro®le. lated by TGF-b (Negoescu et al., 1995; Skonier et al., 1992; Keski-Oja et al., 1988; George et al., 1990; Overall et al., 1991; Nakashio et al., 1997), therefore Production of active TGF-b by HOXD3-overexpressing we investigated whether the TGF-b signaling pathway A549 cells was activated in these cells. First, we measured TGF-b The genes of which expression were altered by HOXD3 in the medium conditioned with HOXD3-overexpres- included thrombospondin-1 (5.6-fold), TGF-b-induced sing cells (HOX+1 and HOX+2) and control 68 kDa protein (big-h3) (5.4-fold), PAI-1 (2.9-fold), transfectant cells (Neo1 and Neo2), using a growth

Oncogene HOXD3-response genes in human cancer cells YJ Miyazaki et al 801 inhibition assay with Mv1Lu cells (Figure 2a). The cells treated with any of the inhibitors and those from bioassay revealed that the medium conditioned with the control cells. HOXD3-overexpressing cells contained approximately 3 ng/ml of the active form of TGF-b (the concentra- Responses of endogenous genes to active TGF-b in control tion of total-latent and active form-TGF-b was 6 ± transfectant (Neo1 and Neo2) cells 9 ng/ml). Although the media conditioned with the parent, Neo1 and Neo2 cells, respectively, contained an To identify the genes which were upregulated or equivalent total TGF-b to that of the HOXD3- downregulated by TGF-b signaling in HOXD3-over- overexpressing cell media, they contained less amounts expressing clones (HOX+1 and HOX+2), we treated of active TGF-b. The growth of Mv1Lu cells, which the control transfectant (Neo1 and Neo2) cells with was inhibited by the media conditioned with HOX+1 active TGF-b at concentrations of 0, 0.4, 2 or 10 ng/ml and HOX+2 cells, was completely recovered by an for 24 h. Isolated total RNA was used for semi- addition of neutralizing anti-TGF-b antibody (Figure quantitative RT ± PCR on 35 genes that were up- or 2e). We also determined the amounts of TGF-b in the down-regulated by HOXD3 in the microarray analysis. conditioned media by an enzyme-linked immunosor- As represented by big-h3 in Figure 1a, thrombospon- bent assay. Although there were no dierences in din-1, syndecan 1, PAI-1, MMP-2, and transgelin were production of the latent form of TGF-b1, the media upregulated by TGF-b-treatment in a dose-dependent conditioned with HOX+1 and HOX+2 cells con- manner. These genes were also upregulated in HOXD3- tained higher levels of active TGF-b1 protein than overexpressing clones (Table 1 and Figure 1a). The those with the parent, Neo1 and Neo2 cells (Figure expression of vitronectin, which was downregulated in 2b). The immunoblot analysis using an anti-TGF-b the HOXD3-overexpressing clones, was reduced by antibody demonstrated a 25 kDa band which repre- TGF-b (Figure 1b). TGF-b treatment did not aect the sented the active form of TGF-b in the media expressions of HOXD3, integrin b3, developmental conditioned with HOXD3-overexpressing cells, but endothelial locus-1 (Del-1), tissue transglutaminase, not in the media with parent cells or neo-transfected CD24, CD44 and quiescin Q6; except HOXD3, they cells (Figure 2c). To examine whether the HOXD3- were upregulated by HOXD3-overexpression (Table 1 overexpressing cells cleaved latent form of TGF-b to and Figure 1c). The expression of desmosomal produce active TGF-b, we checked the latency protein components such as desmoglein 1, desmoplakin and by immunoblotting with the use of an anti-TGF-b1- plakoglobin was repressed by HOXD3-overexpression latency-associated peptide (b1-LAP) antibody. As but not by TGF-b-treatment (Table 1 and Figure 1d). shown in Figure 2d, three bands (Mr 4200 kDa, &110 kDa and &85 kDa) were detected in the TGF-b enhancement of in vitro cell motility and conditioned media from every cell clone. The media invasiveness conditioned with HOX+1 and HOX+2 cells contained a more amount of &85 kDa protein which To examine whether TGF-b in¯uences metastasis- represented a b1-LAP homodimer (Miyazono et al., related properties of A549 cells, we assessed the eects 1991) than those with Neo1 and Neo2 cells. On the of TGF-b on the in vitro invasion and motility of contrary, the media conditioned with Neo1 and Neo2 parent, HOXD3-overexpressing (HOX+1 and cells contained a greater amount of &110 kDa protein HOX+2) and control neo-transfected (Neo 1 and which represented a TGF-b1/b1-LAP complex (Miya- Neo2) cells. HOX+1 and HOX+2 cells showed zono et al., 1991) than those with HOX+1 and aggressive invasion of Matrigel (a reconstituted base- HOX+2 cells. ment membrane) and type I collagen gel compared to MMP-2 and TSP-1 which were upregulated by the parent, Neo1 or Neo2 cells (Figure 3a). When the HOXD3-overexpression (Table 1) are known as parent, Neo1 or Neo2 cells were treated with TGF-b, activators of TGF-b (Schultz-Cherry et al., 1995; Yu the invasion of type I collagen gels by these cells was and Stamenkovic, 2000). Our previous study demon- enhanced, whereas their invasion of Matrigel was strated that HOX+1 and HOX+2 cells expressed a unchanged (Figure 3b). To con®rm that endogenous higher level of urokinase-type plasminogen activator active TGF-b was responsible for the increased (uPA) than parental, Neo1 and Neo2 cells (Hamada et invasion of type I collagen gel, we tested the eects al., 2001). uPA converts plasminogen into plasmin of neutralizing antibodies against TGF-b and TGF-b which enzymatically activates TGF-b (Khalil, 1999). receptor type II and a human recombinant soluble Therefore, we collected conditioned media from TGF-b receptor type II (sTGFRII) on the invasion of HOX+1 and HOX+2 cells which had been cultured type I collagen gel by HOX+1 and HOX+2 cells. As in the presence of TIMP-2 (a natural inhibitor of shown in Figure 3c, these two antibodies and sTGF- MMP-2), a GGWSHW peptide (a peptide to block the RII reduced the invasion. binding of TSP-1 to CD36) or aprotinin and leupeptin HOXD3-overexpressing cells showed a highly hap- (as inhibitors of uPA and plasmin). The amount of totactic response to vitronectin, ®bronectin and type I active TGF-b in the conditioned media was measured collagen compared to parent, Neo1 or Neo2 cells by the growth inhibition assay and ELISA. As shown (Figure 4a ± c). TGF-b stimulated the haptotaxis of the in Figure 1e,f, there was no dierence in the amount of parent and vector-transfectants to type I collagen but active TGF-b between the conditioned media from the not to vitronectin or ®bronectin, whereas it did not

Oncogene HOXD3-response genes in human cancer cells YJ Miyazaki et al 802

Figure 2 Production of TGF-b by A549 cells transfected with HOXD3.(a) TGF-b activity in the medium conditioned with parental A549 cells and the transfectant cells. TGF-b activity was assessed by growth inhibition assay using mink lung epithelial Mv1Lu cells, and expressed as the concentration (ng/ml) of recombinant human TGF-b1 which showed equivalent growth inhibition on the standard curve. A half of the conditioned medium was acid-treated for activation of latent form of TGF-b (total TGF-b, open columns). Growth inhibition by non-treated conditioned medium indicates active TGF-b (solid columns). The columns represent mean+standard deviation for quadruplicate samples. *P50.001 compared to the parental or neomycin-resistant gene transfectant cells by one-way ANOVA followed by Fisher's probable least-squares dierence analysis as a post hoc test. (b) The amounts of TGF-b protein in these samples were assayed with the use of a TGF-b1-ELISA kit. The amount of TGF-b1 protein in acid-treated conditioned media (total TGF-b) is shown by open columns, and that in acid-non-treated conditioned media (active TGF-b) by solid columns. The columns represent mean+standard deviation for quadruplicate samples. *P50.001 compared to the parental or neomycin-resistant gene transfectant cells by one-way ANOVA followed by Fisher's probable least-squares dierence analysis as a post hoc test. (c) Immunoblot analysis of active form of TGF-b in medium conditioned with each cell line. Samples pulled down with heparin-Sepharose beads from the medium conditioned with each cell line were resolved with SDS ± PAGE, and electrotransferred to Immobilon P membrane. The membrane was probed with a mouse monoclonal antibody to TGF-b (active form), and visualized with Enhanced Chemiluminescent Detection System. As a positive control, 4 ng of recombinant human TGF- b (rhTGF-b) was used. (d) Immunoblot analysis of TGF-b1-latency-associated peptide (b1-LAP) in medium conditioned with each cell clone. The membrane was probed with mouse monoclonal antibody to b1-LAP. (e,f) The amounts of active TGF-b in media conditioned with HOXD3-overexpressing cells treated with proteinase inhibitors or a blocking peptide to bind TSP-1 to CD36. The conditioned media were collected from the cells cultured in the presence of 50 nM TIMP-1 ( ), 50 nM TIMP-2 ( ), 20 mM GGWSHW ( ), 20 mM GGYSHW as a control peptide ( ), 10 mg/ml aprotinin ( ), 10 mg/ml leupeptin ( ) or nothing ( ). (e) TGF-b activity without acid-treatment was measured by the growth inhibition assay using Mv1Lu cells. Mv1Lu cells were incubated with the medium conditioned with HOX+1 or HOX+2 cells in the presence of either a neutralizing antibody to TGF-b ( )or normal rabbit IgG ( ) as control. *P50.001 compared to the medium conditioned with the cells treated with normal rabbit IgG or the non-treated cells by one-way ANOVA followed by Fisher's probable least-squares dierence analysis as a post hoc test. (f) The amount of TGF-b1 protein in acid-non-treated conditioned media was measured by TGF-b1-ELISA

Oncogene HOXD3-response genes in human cancer cells YJ Miyazaki et al 803 aect the haptotaxis of HOXD3-overexpressing cells to any of the extracellular matrix components (Figure 4a ± c). To con®rm that endogenous active TGF-b was responsible for the increased haptotaxis to type I collagen, we tested the eects of neutralizing antibodies against TGF-b and TGF-b receptor type II and sTGFRII on haptotaxis of HOX+1 and HOX+2 cells to type I collagen. As shown in Figure 4d, these two antibodies and sTGFRII reduced the haptotactic migration.

Discussion

We revealed by the cDNA microarray analysis that gene expression patterns in HOXD3-overexpressing cells related to alterations in (1) the cell-cell interactions, (2) the cell-extracellular matrix (ECM) interactions, (3) the cytoskeletal system, and that the HOXD3-overexpression activated a speci®c signaling pathway (TGF-b-regulated pathway). Interestingly, HOXD3-overexpression downregulated the gene expression of desmosomal components including desmoglein, desmoplakin and plakoglobin. Immunohistochemical studies have shown that a loss of staining for desmosomal components correlates with invasive and metastatic potential in both transitional cell carcinoma (Conn et al., 1990) and squamous cell carcinoma (Harada et al., 1992; Hiraki et al., 1996; Shinohara et al., 1998). Transfection of the cDNA encoding desmosomal components into highly invasive cells inhibits their in vitro invasion of collagen gels (Tselepis et al., 1998). These reports indicate that desmosomes have a role in suppression of tumor spreading, and therefore a reduction in the expression of desmosomal components in HOXD3-overexpressing cells may aid their dissociation or migration from the primary tumor mass. Interaction of cells with the extracellular matrix is also important in tumor invasion and metastasis as well as in embryonic morphogenesis. Our microarray analysis and subsequent semi-quantitative RT ± PCR Figure 3 In vitro invasiveness of parental A549 cells and the analysis revealed that the HOXD3-overexpression transfectant cells when treated with TGF-b. In vitro invasion upregulated the gene expression of (1) adhesion assay was performed by using Transwell chambers with Matrigel- coated membranes (a) or with membranes embedded in type I molecules (integrin b3, CD44), (2) ECM components collagen gel (b). The cells which had been treated with TGF-b for (TSP-1, bI-GH3, Del-1) and (3) the molecules 24 h at the indicated concentrations ( , no TGF-b; , 0.4 ng/ associated with ECM-degradation (PAI-1, MMP-2). ml; , 2 ng/ml; , 10 ng/ml) were placed in the upper Vitronectin, one of the ECM components, was down- compartment of the chamber. The same concentrations of TGF- b were added to the upper compartment during the invasion regulated by HOXD3-overexpression. We have already assay. After 24 h-incubation, invaded cells were counted under a shown that the integrin b3 subunit induced by microscope. The columns represent mean+standard deviation for HOXD3-overexpression forms the integrin avb3 hetero- triplicate assays. #P50.001 compared to the parental or neomycin-resistant gene transfectant cells, *P50.001 compared to the cells untreated with TGF-b,**P50.001 compared to the cells treated with 0.4 ng/ml of TGF-b by one-way ANOVA followed by Fisher's probable least-squares dierence analysis as receptor type II ( ) for 30 min at 48C. The cells (26104) treated a post hoc test. (c) Invasion of type I collagen gel by HOXD3- with each antibody or soluble TGF-b receptor type II were placed overexpressing cells (HOX+1 and HOX+2) was inhibited by in the upper compartment of the chamber. After 24 h-incubation, treatment with antibody to TGF-b or TGF-b receptor II, or invaded cells were counted under a microscope. The columns soluble TGF-b receptor type II. The cells were incubated with represent mean+standard deviation for triplicate assays. DME/F12 containing 0.1% BSA ( ) with anti-TGF-b antibody *P50.001 compared to the non-treated or normal IgG treated ( ), anti-TGF-b receptor type II ( ), normal rabbit IgG ( ), cells by one-way ANOVA followed by Fisher's probable least- normal goat IgG ( ), and recombinant human soluble TGF-b squares dierence analysis as a post hoc test

Oncogene HOXD3-response genes in human cancer cells YJ Miyazaki et al 804

Figure 4 Haptotactic activities of parental A549 cells and the transfectant cells to ®bronectin (a), vitronectin (b) and type I collagen (c,d) when treated with TGF-b. Haptotaxis assay was performed by using Transwell chambers. The lower surface of the membranes of transwell chambers was coated with 10 mg of ®bronectin, vitronectin or type I collagen. The cells (26104) which had been treated with TGF-b for 24 h at the indicated concentrations (a, b and c: , no TGF-b; , 0.4 ng/ml; , 2 ng/ml; , 10 ng/ ml) were placed in the upper compartment of the chamber. The same concentrations of TGF-b were added to the upper compartment during the haptotaxis assay. After 6 h-incubation, migrated cells were counted under a microscope. The columns represent mean+standard deviation in randomly selected 20 ®elds per well at 6200 magni®cation. #P50.001 compared to the parental, Neo1 or Neo2 cells, }P50.001 compared to the parental or Neo1 cells, *P50.001 compared to the cells untreated with TGF-b, }P50.001 compared to the cells treated with 0.4 ng/ml of TGF-b.(d) Inhibition of haptotaxis of HOXD3-overexpressing cells (HOX+1 and HOX+2) to type I collagen by treatment with antibody to TGF-b or TGF-b receptor type II, or soluble TGF-b receptor type II. The cells were incubated with DME/F12 containing 0.1% BSA ( ) with anti-TGF-b antibody ( ), anti-TGF-b receptor type II ( ), normal rabbit IgG ( ), normal goat IgG ( ), and recombinant human soluble TGF-b receptor type II ( ) for 30 min at 48C. The cells (26104) treated with each antibody or soluble TGF-b receptor type II were placed in the upper compartment of the chamber. After 6 h-incubation, migrated cells were counted by using a microscope. *P50.001 compared to the non-treated or normal IgG treated cells by one-way ANOVA followed by Fisher's probable least-squares dierence analysis as a post hoc test

dimer and enhances migration in the presence of their sion. Of the genes dierentially expressed between the ligands, vitronectin and ®brinogen. Furthermore, HOXD3-overexpressing cells and the control cells, phagokinetic track assay in the absence of avb3 ligands some were documented as being TGF-b-regulated shows that HOXD3-overexpressing cells have a motile genes. These included TSP-1, big-h3, PAI-1 and activity higher than those of the parental or control MMP-2 (Negoescu et al., 1995; Skonier et al., 1992; cells (Hamada et al., 2001). The present study indicated Keski-Oja et al., 1988; Overall et al., 1991). We a possible missing link. Namely, Del-1, one of the con®rmed by RT ± PCR analysis that the expression ECM components, upregulated by HOXD3-overex- patterns of these genes in the control cells stimulated pression, has recently been identi®ed as a novel ligand with exogenous TGF-b were similar to those in the for the integrin avb3 (Hidai et al., 1998). And the Del-1 HOXD3-overexpressing cells. The bioassay and ELISA protein has been shown to promote avb3-dependent for TGF-b and the immunoblot analysis analysis using endothelial cell attachment and migration, which antibodies to TGFb1 and b1-LAP demonstrated that suggests that Del-1 functions as a motility-stimulating this activation of the TGF-b pathway was due to an factor for endothelial cells during angiogenesis (Penta accelerated conversion of latent TGF-b to an active et al., 1999). We therefore infer that the cell motility of form in the HOXD3-overexpressing cells. It has been HOXD3-overexpressing A549 cells is stimulated by proposed that TGF-b activation occurs in vivo through Del-1 through the integrin avb3 in an autocrine the pathways of plasminogen/plasmin, TSP-1/CD36, manner. integrin avb3/MMP-2 (or MMP-9) or integrin avb6 We found the participation of the TGF-b-regulated (Khalil, 1999; Lyons et al., 1988; Schultz-Cherry et al., pathway in the conversion of the cells to a more motile 1995; Yehualaeshet et al., 1999; Munger et al., 1999; and invasive phenotype during HOXD3-overexpres- Yu and Stamenkovic, 2000). To examine whether the

Oncogene HOXD3-response genes in human cancer cells YJ Miyazaki et al 805 pathways of MMP-2, TSP-1 or plasminogen/plasmin ®bronectin or their invasion of Matrigel, which is were involved in the activation of TGF-b, we collected composed of laminin-1, collagen type IV and heparan conditioned media in the presence of tissue inhibitor of sulfate proteoglycan (Kleinman et al., 1982). In metalloproteinase-2, a natural inhibitor of MMP-2, a addition, none of the expressions of integrin b3, Del- peptide (GGWSHW) which disturbs the binding of 1, or desmosomal components (desmoglein, desmopla- TSP-1 with CD36, or aprotinin and leupeptin which kin and plakoglobin) was aected by TGF-b-stimula- are inhibitors of plasmin. However, none of the tion. Taken together, the TGF-b signaling activated by inhibitors prevented the activation of TGF-b achieved HOXD3-overexpression seems to mainly facilitate by HOX+1 or HOX+2 cells. Further, the involve- ligand (COL-I)-dependent cell migration and partly ment of integrin avb6 is unlikely as expression of avb6 to participate in gain of more invasive and metastatic was not observed in any of the parent cell line and the potential of cells. Type I collagen is the major transfected clones (J Hamada, unpublished data). component in tumor stroma. TGF-b-mediated hapto- Thus, we have no evidence to demonstrate that tactic and invasive responses may facilitate invasion of MMP-2, TSP-1, plasminogen/plasmin or avb6 was tumor cells from the in vivo primary tumor tissues. involved in HOXD3-mediated TGF-b activation. We In conclusion, the phenotypic and transcriptional need further experiments to elucidate the mechanism alteration of A549 cells by HOXD3-overexpression responsible for the HOXD3-mediated TGF-b activa- may be caused by signals through multiple pathways tion. However, it is of interest that HOXD3 is including the TGF-b-regulated pathway. Our study implicated in the regulation of the conversion from provides the ®rst evidence that HOXD3 homeobox latent to active TGF-b, since TGF-b is known to be an gene regulates many downstream eectors, and alters important regulator in embryonic development (Moses the invasive and metastatic potential of cancer cells. and Serra, 1996; Massague and Chen, 2000; Kimelman and Grin, 2000). Finally we have shown that the activation of TGF-b signaling by HOXD3-overexpression alters the cellular Materials and methods properties of migration and invasion. Exogenous TGF- b promoted haptotaxis to type I collagen (COL-I) in Antibodies and reagents the parent and control transfectant cells, consistent A rabbit pan-speci®c antibody to TGF-b, a mouse mono- with a previous report that TGF-b stimulated the clonal antibody to human TGF-b1-latency-associated peptide invasion of COL-I gels by A549 cells in a dose- (b1-LAP) and recombinant human TGF-b1 were purchased dependent manner (Mooradian et al., 1992). Hapto- from R&D (Minneapolis, MN, USA). A goat antibody to tactic activity of HOX+1 and HOX+2 cells to type I human TGF-b receptor type II, a rabbit antibody to TGF-b collagen was suppressed by an addition of blocking and recombinant human TGF-b soluble receptor type II were antibodies or soluble TGF-b receptor type II; however, from Genzyme (Minneapolis, MN, USA). A mouse mono- their activity was still higher than that of parent, Neo1 clonal antibody to thrombospondin was from Oncogene Research Products (Cambridge, MA, USA). Procine type I and Neo2 cells. This phenomenon suggests that collagen was from Nitta Gelatin (Osaka, Japan). Normal haptotaxis of HOXD3-overexpressing cells is stimu- rabbit IgG was from Chemicon (Temecula, CA, USA). lated through not only a TGF-b-mediated pathway but Normal goat IgG was from Cappel (Durham, NC, USA). also other signaling pathway(s). Like haptotaxis to Bovine vitronectin was from Yagai (Yamagata, Japan). type I collagen, invasion of type I collagen gel by Bovine serum albumin was from Boehringer Mannheim HOX+1 and HOX+2 cells was also partially (Mannheim, Germany). Matrigel was from Becton Dickinson suppressed by the blocking antibodies and sTGFRII. (Bedford, MA, USA). Transwell chamber was from Costar 1 Further, exogeneous active TGF-b stimulated the (Cambridge, MA, USA). Geneticin (G418 sulfate), Molony invasion of parent, Neo1 and Neo2 cells, but did not murine leukemia virus reverse transcriptase, random primer reach the same level as those of HOX+1 and HOX+2 and Trizol, human tissue inhibitor of metalloproteinase (TIMP)-1 and TIMP-2 were from Life Technologies (Rock- cells. It seems that involvement of TGF-b signaling in ville, MD, USA). Taq polymerase was from Nippon Gene the invasion of type I collagen gel is relatively low (Tokyo, Japan). Glutaraldehyde, aprotinin and leupeptin compared to that in haptotaxis to the collagen. Unlike were from Wako (Tokyo, Japan). Crystal violet was from haptotaxis, the cells invading type I collagen gel need Kanto Chemical (Tokyo, Japan). Heparin-Sepharose, CL-6B to exert more biological ability, for example ability to was from Amersham Pharmacia Biotech (Uppsala, Sweden). digest the collagen. In not only migration to type I Oligopeptide, GGWSHW and GGYSHW were synthesized collagen but also degradation of type I collagen, the by Sigma Genosys Japan (Ishikari, Hokkaido, Japan). HOXD3-overexpressing cells may use TGF-b-independent pathway(s) as well as TGF-b-dependent one. We Cells and cell culture previously demonstrated that the HOXD3-overexpres- Human lung cancer A549 cells and mink lung epithelial Mv1Lu sing A549 cells showed high haptotactic activity to cells were obtained from the Japanese Cancer Research vitronectin and ®brinogen and a strong invasiveness to Resources Bank (JCRB, Tokyo, Japan). These cells were Matrigel compared to the parent and control transfec- grown on tissue culture dishes in a 1 : 1 (v/v) mixture of tant cells (Hamada et al., 2001). In the present study, Dulbecco's modi®ed minimum essential medium and Ham's however, exogenous TGF-b did not aect the F12 medium (DME/F12) supplemented with 5% fetal bovine haptotaxis of the control cells to vitronectin and serum (FBS). Cloned A549 cell lines expressing HOXD3 gene

Oncogene HOXD3-response genes in human cancer cells YJ Miyazaki et al 806 (HOX+1 and HOX+2) or neomycin-resistant gene (Neo1 and Bioassay of TGF-b activity Neo2) were established by the use of lipofection with a mammalian expression vector, pMAMneo/HOX4A(HOXD3) In vitro bioassay to measure TGF-b activity was based on the or empty vector pMAMneo (Hamada et al., 2001). These cells method described by Lucas et al. (1991) with some were grown on tissue culture dishes in DME/F12 supplemented modi®cation. Brie¯y, mink lung epithelial Mv1Lu cells were with 5% FBS, Geneticin (400 mg/ml). The cell lines were plated on ¯at-bottomed 96-well tissue culture plates in 50 ml 3 cultured at 378C in a humidi®ed 5% CO2 atmosphere. of DME/F12 supplemented with 5% FBS at 3.2610 cells/ well. Prior to addition of samples (conditioned media from each cell line and DME/F12) to the wells, a half of each cDNA microarray analysis sample was acid-treated for activation of latent form of TGF- Target cDNA was generated from 1 mg polyadenylated b by incubating each 200 ml sample with 1.5 ml of 6N HCl for mRNA, which was reverse-transcribed and labeled either 30 min at room temperature and neutralized with 3 mlof with Cy5 (HOX+2) or Cy3 (Neo1) dUTP. The average neutralizing solution (1 : 1=6 M NaOH : 1 M HEPES). Fifty intensity of the Cy5 ¯uorescence divided by the average ml of the conditioned media, acid-treated-CM and DME/F12 intensity of the Cy3 ¯uorescence equaled 0.97 (balance were added to each quadruplicate well. Recombinant human coecient), indicating similar labeling eciency for each set TGF-b1 (rhTGF-b1) diluted with DME/F12 at serial of target cDNAs. Target cDNA was hybridized on concentrations (0 ± 20 ng/ml) was added to the wells (50 ml/ IncyteGEM, microarrays containing 7075 probes with well) to establish a standard curve. After 3 days of sequences complementary to 4107 human genes and 2968 incubation, the assay was terminated and cell numbers were human expressed sequence tags (ESTs) (UniGem V, Genome quanti®ed by using a colorimetric crystal violet-staining Systems, St. Louis, MO, USA). Following the hybridization procedure (Kueng et al., 1989). To determine cell number, and washing, the relative expression levels of both cDNA the absorbance of each well was read at 595 nm on a populations were measured and compared by obtaining the Microplate reader (BIO-RAD, Model 550, Hercules, CA, Cy5/Cy3 ¯uorescence ratio for each target cDNA to satisfy USA). Data of absorbance at 595 nm were converted to cell the examination criterion. The inclusion criterion was based numbers according to the pre-tested linear correlation on an image recognition algorithm for each cDNA in the between the cell number and the absorbance. Percentage of analysis and included a ¯uorescence signal from the cDNA growth inhibition was calculated as %=1006{17(no. of exceeding a signal to background ratio of 2.5 and the cDNA cells in sample well with CM or rhTGF-b)/(no. of cells in covering its grid location on the microarray for 440%. In control well)}. For each experiment, a standard curve for per this study, genes were considered dierentially expressed if an cent growth inhibitions to rhTGF-b1 concentrations was increase was more than 1.6-fold or a decrease was to less established. The concentration of TGF-b in CM was deduced than 1/1.6. from the rhTGF-b1 concentration (ng/ml) of the equivalent growth inhibiting activity on the standard curve. Semiquantitative duplex RT ± PCR analysis Enzyme-linked immunosorbent assay (ELISA) for TGF-b For RT ± PCR analysis, total RNA was extracted from monolayer cultures of HOX+1, HOX+2, Neo1, Neo2 cells The amounts of TGF-b1 protein in conditioned media were and TGF-b (0.4, 2 or 10 ng/ml)-treated-Neo1 and -Neo2 cells assayed with the use of a TGF-b1-ELISA kit (AN' ALYZA with Trizol, according to the manufacturer's instruction. Two human TGF-b1, Genzyme Techne, Minneapolis, MN, USA). mg of total RNA sample was subjected to cDNA synthesis for To activate latent TGF-b1 to be immunoreactive, the 2 h at 378Cin50ml of reaction mixture containing 4 U/ml of conditioned media were treated with 1 N HCl and then Molony murine leukemia virus reverse transcriptase, 7.5 mM neutralized with 1.2 N NaOH/0.5 M HEPES, according to dithiothreitol, 0.5 mM MgCl2, 0.5 mM dNTP and 2 mM the manufacturer's instruction. random primer. PCR ampli®cation of cDNA was performed in 50 ml of reaction mixture containing 1 ml of cDNA sample, Immunoblot analysis 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 2 mM MgCl2, 0.125 U/ml of Taq polymerase and dierent primer sets For the detection of active form of TGF-b, the media (10 nM each). Co-ampli®cation of the speci®c gene and conditioned with each cell line (1.2 ml) were mixed with 50 ml human b-actin gene, as an internal control, was achieved by of heparin-Sepharose CL-6B beads and rotated at 48C for using two primer sets in a single reaction mixture. Each 18 h (McCarey et al., 1992). The heparin-Sepharose beads primer was designed to encompass an exon junction for were washed ®ve times with DME/F12; bound TGF-b was prevention of templating possibly contaminated genomic eluted with SDS ± PAGE sample buer and boiled for 4 min. DNA. PCR products were electrophoresed in a 2.5% agarose Each sample was resolved with 12.5% SDS ± PAGE and gel and intensity of the bands observed under a UV electrotransferred to a polyvinylidene di¯uoride membrane illuminator was analysed by Scion Image (Scion, Frederick, (Immobilon P, Millipore, Bedford, MA, USA). For the MD, USA). detection of latent form of TGF-b, the media conditioned with each cell line were concentrated to 2 mg/ml of protein using a Centricon-10 (Amicon, Millipore, Bedford, MA, A549 cell-conditioned medium USA). Ten ml each of the concentrated conditioned media Tumor cells (26105/dish) were plated on 60-mm tissue was subjected to SDS ± PAGE in a 7.5% gel under non- culture dish in DME/F12 supplemented with 5% FBS. After reducing conditions and electrotransferred to the membrane. 24 h, the cultures were washed twice with DME/F12 medium The membranes were blocked in TBS-T (20 mM Tris-HCl, without FBS and then incubated with DME/F12 medium pH 7.5, 137 mM NaCl, 0.1% Tween 20) with 5% skim milk without FBS for 24 h. The medium was collected and overnight at 48C, and then incubated with primary antibodies centrifuged at 800 g for 10 min, and the supernatants were for 1 h at room temperature. The membranes were then recentrifuged at 20 000 g for 10 min. All processes were done incubated with horseradish peroxidase-conjugated antibodies in sterile condition. for 1 h at room temperature, and developed with reagents of

Oncogene HOXD3-response genes in human cancer cells YJ Miyazaki et al 807 the Enhanced Chemiluminescent Detection System (Amer- sterile deionized-distilled water. The Matrigel-coated Trans- sham Pharmacia, Little Chalfont, UK). well chambers were air-dried in a hood overnight. The Matrigel-coated membranes were then washed twice with 100 ml of DME/F12 and incubated with DME/F12 for 1 h at Flow cytometry room temperature. Before the assay, the medium in the upper The cultured cells were harvested by trypsinization and compartment of the Transwell chamber was replaced with washed twice with PBS containing 0.065% sodium azide. The 100 ml of A549 cell suspension (26105 cells/ml) in DME/F12 cells were incubated with primary antibodies for 1 h at 48C, supplemented with 1% FBS. Into the lower compartment of and then with secondary antibodies conjugated with FITC. the chamber, 600 ml of medium conditioned with human lung After incubation, the cells were washed three times with PBS ®broblast MRC-5 cells was placed. After 24 h of incubation, containing 0.065% sodium azide. The stained cells were membranes was ®xed with 10% neutral-buered formalin resuspended in 1 ml of PBS containing 0.065% sodium azide and stained in Giemsa solution. The cells attached to the and analysed with the use of FACSCalibur (Becton upper side of the membrane were wiped out, and those Dickinson, San Jose, CA, USA). attached to the lower side of the membrane were counted under a microscope. Invasiveness was evaluated by the number of cells invaded per membrane (mean+s.d., n=3). Haptotaxis assay Haptotaxis assay was performed by using Transwell Invasion of type I collagen gel by A549 cells chambers. The lower surface of the membranes with 8-mm pores of Transwell chambers was coated with 10 mg each of Type I collagen gel (Cellmatrix type I-P, Nitta Gelatin, Yao, FN, VN, or FB, and dried up in a hood overnight. Before the Japan), 10-fold concentrated DME/F12 and NaHCO3/ assay, the membranes coated with the matrix components HEPES were mixed in volume at 9 : 1 : 1, respectively. The had been washed twice with DME/F12. Six hundreds mlof mixed collagen solution (600 ml) was placed in each well of DME/F12 containing 0.1% BSA was placed into the lower 24-well plates, and a Transwell was placed onto each well. compartment of the Transwell chambers, and then 100 mlof The Transwells were pressed down to assure their tight the cell suspension (26105 cells/ml in DME/F12 containing contact to the collagen solution. The 24-well plates with 0.1% BSA) was placed into the upper compartment. After collagen sol and Transwells were incubated at 378C for 6 h-incubation, each membrane was ®xed with 10% neutral- 30 min for gel-formation of the collagen solution. After buered formalin and stained in Giemsa solution. After the gelation, 90 ml of cell suspension (2.26105/ml in DME/F12 cells attached to the upper side of the membrane were containing 0.1% BSA) was placed in the upper compartment removed by wiping with a cotton swab, those attached to the of each Transwell. Immediately after the cell placement, 10 ml lower side of the membrane were counted under a of recombinant human TGF-b diluted at a series of microscope. Haptotactic activity was evaluated by the concentrations (0, 4, 20, 100 ng/ml) with DME/F12 contain- number of cells per ®eld at 6200 magni®cation (mean+s.d., ing 0.1% BSA was added into the upper compartment. After n=20). In some experiments, to examine the involvement of 24 h of incubation at 378CinaCO2 incubator, the Transwell TGF-b in the haptotaxis to type I collagen, the cells were was removed and the collagen gel was washed twice with incubated in DME/F12 containing 0.1% BSA and recombi- 2 ml of DME/F12. The cells in the collagen gel were nant human soluble TGF-b receptor type II (10 mg/ml), anti- observed under an inverted phase-contrast microscope TGF-b antibody (200 mg/ml), anti-TGF-b receptor type II (Eclipse TE300, Nikon, Tokyo, Japan). Invasiveness was antibody (100 mg/ml), normal rabbit IgG (200 mg/ml) or evaluated by the number of cells in the collagen gel in normal goat IgG (100 mg/ml) for 30 min at 48C. The cell triplicate assay. suspension (100 ml) was placed into the upper compartment of Transwell as described above.

In vitro invasion of reconstituted basement membrane, Matrigel, Acknowledgments by A549 cells The authors wish to thank Ms. Masako Yanome for help In vitro invasion was assayed by the method reported by in preparing the manuscript. This work was supported by a Albini et al. (1987) with some modi®cation (Nagayasu et al., Grant-in-AidforScienti®cResearch(B)andaGrant-in- 1998). Brie¯y, membranes with 8-mm pores of Transwell Aid for Scienti®c Research on Priority Areas (C) from the chambers were coated with 100 ml of 20-times diluted Ministry of Education, Culture, Sports, Science and Matrigel (Becton Dickinson, Bedford, MA, USA) in cold Technology of Japan.

References

Albini A, Iwamoto Y, Kleinman HK, Martin GR, Aaronson Cillo C, Barba P, Freschi G, Bucciarelli G, Magli MC and SA, Kozlowski JM and McEwan RN. (1987). Cancer Res., Boncinelli E. (1992). Int. J. Cancer, 51, 892 ± 897. 47, 3239 ± 3245. Cillo C. (1994 ± 1995). Invasion Metastasis, 14, 38 ± 49. Apiou F, Flagiello D, Cillo C, Malfoy B, Poupon MF and Cillo C, Faiella A, Cantile M and Boncinelli E. (1999). Exp. Dutrillaux B. (1996). Cytogenet. Cell Genet., 73, 114 ± 115. Cell. Res., 248, 1±9. Borrow J, Shearman AM, Stanton Jr VP, Becher R, Collins Conn IG, Vilela MJ, Garrod DR, Crocker J and Wallace T, Williams AJ, Dube I, Katz F, Kwong YL, Morris C, DM. (1990). Br.J.Urol.,65, 176 ± 180. Ohyashiki K, Toyama K, Rowley J and Housman DE. De Vita G, Barba P, Odartchenko N, Givel JC, Freschi G, (1996). Nat. Genet., 12, 159 ± 167. Bucciarelli G, Magli MC, Boncinelli E and Cillo C. (1993). Boudreau N, Andrews C, Srebrow A, Ravanpay A and Eur. J. Cancer, 29A, 887 ± 893. Cheresh DA. (1997). J. Cell Biol., 139, 257 ± 264.

Oncogene HOXD3-response genes in human cancer cells YJ Miyazaki et al 808 Gehring WJ and Hiromi Y. (1986). Annu. Rev. Genet., 20, Moses HL and Serra R. (1996). Curr.Opin.Genet.Dev.,6, 147 ± 173. 581 ± 586. George MD, Vollberg TM, Floyd EE, Stein JP and Jetten Munger JS, Huang X, Kawakatsu H, Griths MJ, Dalton AM. (1990). J. Biol. Chem., 265, 11098 ± 11104. SL, Wu J, Pittet JF, Kaminski N, Garat C, Matthay MA, Graham A, Papalopulu N and Krumlauf R. (1989). Cell, 57, Rifkin DB and Sheppard D. (1999). Cell, 96, 319 ± 328. 367 ± 378. Nagayasu H, Hamada J, Kawano T, Konaka S, Nakata D, Hamada J, Omatsu T, Okada F, Furuuchi K, Okubo Y, Shibata T, Arisue M, Hosokawa M, Takeichi N and TakahashiY,TadaM,MiyazakiYJ,TaniguchiY,Shirato Moriuchi T. (1998). Br. J. Cancer, 77, 1371 ± 1377. H, Miyasaka K and Moriuchi T. (2001). Int. J. Cancer, 93, Nakamura T, Largaespada DA, Lee MP, Johnson LA, 516 ± 525. Ohyashiki K, Toyama K, Chen SJ, Willman CL, Chen IM, Harada T, Shinohara M, Nakamura S, Shimada M and Oka Feinberg AP, Jenkins NA, Copeland NG and Shaugh- M. (1992). Int. J. Oral. Maxillofac. Surg., 21, 346 ± 349. nessy Jr JD. (1996a). Nat. Genet., 12, 154 ± 158. HidaiC,ZupancicT,PentaK,MikhailA,KawanaM, Nakamura T, Largaespada DA, Shaughnessy Jr JD, Jenkins Quertermous EE, Aoka Y, Fukagawa M, Matsui Y, NA and Copeland NG. (1996b). Nat. Genet., 12, 149 ± 153. Platika D, Auerbach R, Hogan BL, Snodgrass R and Nakashio T, Narita T, Akiyama S, Kasai Y, Kondo K, Ito Quertermous T. (1998). Genes Dev., 12, 21 ± 33. K, Takagi H and Kannagi R. (1997). Int. J. Cancer, 70, Hiraki A, Shinohara M, Ikebe T, Nakamura S, Kurahara S 612 ± 618. and Garrod DR. (1996). Br. J. Cancer, 73, 1491 ± 1497. Negoescu A, Lafeuillade B, Pellerin S, Chambaz EM and Keski-Oja J, Raghow R, Sawdey M, Loskuto DJ, Feige JJ. (1995). Exp. Cell Res., 217, 404 ± 409. Postlethwaite AE, Kang AH and Moses HL. (1988). J. Overall CM, Wrana JL and Sodek J. (1991). J. Biol. Chem., Biol. Chem., 263, 3111 ± 3115. 266, 14064 ± 14071. Khalil N. (1999). Microbes Infect., 1, 1255 ± 1263. Penta K, Varner JA, Liaw L, Hidai C, Schatzman R and Kimelman D and Grin KJ. (2000). Curr. Opin. Genet. Dev., Quertermous T. (1999). J. Biol. Chem., 274, 11101 ± 11109. 10, 350 ± 356. Schultz-Cherry S, Chen H, Mosher DF, Misenheimer TM, Kleinman HK, McGarvey ML, Liotta LA, Robey PG, Krutzsch HC, Roberts DD and Murphy-Ullrich JE. Tryggvason K and Martin GR. (1982). Biochemistry, 21, (1995). J. Biol. Chem., 270, 7304 ± 7310. 6188 ± 6193. Shinohara M, Hiraki A, Ikebe T, Nakamura S, Kurahara S, Kueng W, Silber E and Eppenberger U. (1989). Anal. Shirasuna K and Garrod DR. (1998). J. Pathol., 184, Biochem., 182, 16 ± 19. 369 ± 381. Lucas C, Fendly BM, Mukku VR, Wong WL and Palladino Skonier J, Neubauer M, Madisen L, Bennett K, Plowman MA. (1991). Meth. Enzymol., 198, 303 ± 316. GD and Purchio AF. (1992). DNA Cell Biol., 11, 511 ± 522. Lyons RM, Keski-Oja J and Moses HL. (1988). J. Cell Biol., Taniguchi Y, Komatsu N and Moriuchi T. (1995). Blood, 85, 106, 1659 ± 1665. 2786 ± 2794. Mark M, Rijli FM and Chambon P. (1997). Pediatr. Res., 42, Tiberio C, Barba P, Magli MC, Arvelo F, Le Chevalier T, 421 ± 429. Poupon MF and Cillo C. (1994). Int. J. Cancer, 58, 608 ± Massague J and Chen Y-G. (2000). Genes Dev., 14, 627 ± 644. 615. McCarey TA, Falcone DJ and Du B. (1992). J. Cell. Tselepis C, Chidgey M, North A and Garrod D. (1998). Proc. Physiol., 152, 430 ± 440. Natl. Acad. Sci. USA, 95, 8064 ± 8069. McGinnis W and Krumlauf R. (1992). Cell, 68, 283 ± 302. Yehualaeshet T, O'Connor R, Green-Johnson J, Mai S, Miyazono K, Olofsson A, Colosetti P and Heldin CH. Silverstein R, Murphy-Ullrich JE and Khalil N. (1999). (1991). EMBO J., 10, 1091 ± 1101. Am. J. Pathol., 155, 841 ± 851. Mooradian DL, McCarthy JB, Komanduri KV and Furcht Yu Q and Stamenkovic I. (2000). Genes Dev., 14, 163 ± 176. LT. (1992). J. Natl. Cancer Inst., 84, 523 ± 527.

Oncogene