Oncogene (2008) 27, 5578–5589 & 2008 Macmillan Publishers Limited All rights reserved 0950-9232/08 $32.00 www.nature.com/onc ORIGINAL ARTICLE Identification of SOX4 target using phylogenetic footprinting-based prediction from expression microarrays suggests that overexpression of SOX4 potentiates metastasis in hepatocellular carcinoma

Y-L Liao1,6, Y-M Sun2,6, G-Y Chau3, Y-P Chau4, T-C Lai5, J-L Wang5, J-T Horng2, M Hsiao5 and A-P Tsou1

1Department of Biotechnology and Laboratory Science in Medicine, National Yang-Ming University, Taipei, Taiwan, ROC; 2Institute of Computer Science and Information Engineering, National Central University, Chung-Li, Taiwan, ROC; 3Department of Surgery, Taipei Veterans General Hospital, Taipei, Taiwan, ROC; 4Institute of Anatomy and Cell Biology, National Yang-Ming University, Taipei, Taiwan, ROC and 5Genomics Research Center, Academia Sinica, Nankang, Taipei, Taiwan, ROC

A comprehensive microarray analysis of hepatocellular Introduction carcinoma (HCC) revealed distinct synexpression patterns during intrahepatic metastasis. Recent evidence has Hepatocellular carcinoma (HCC) is one of the most demonstrated that synexpression group member genes common malignant neoplasms in humans and is are likely to be regulated by master control (s). Here prevalent among Asian populations. Molecular analyses we investigate the functions and gene regulation of the have shown that the HCC pathogenesis is a multi- SOX4 in intrahepatic metastatic factorial and multistep process reflecting alterations HCC. SOX4 is important in tumor metastasis as RNAi derived from epigenetic instability, chromosomal in- knockdown reduces tumor cell migration, invasion, in vivo stability and deregulated transcription (Chen et al., tumorigenesis and metastasis. A multifaceted approach 2002; El-Serag and Rudolph, 2007). HCC death has integrating gene profiling, binding site computation and been ranked the third most common cancer death empirical verification by chromatin immunoprecipitation worldwide as a result of poor prognosis due to the and gene ablation refined the consensus SOX4 binding disease’s heterogeneous nature and the high rate of motif and identified 32 binding loci in 31 genes with high intrahepatic metastasis that occurs in the dense vascu- confidence. RNAi knockdown of two SOX4 target genes, lature of the liver (Tang et al., 2004; Kim et al., 2005). neuropilin 1 and semaphorin 3C, drastically reduced Recently, critical molecular signature genes were shown cell migration activity in HCC cell lines suggesting that to be involved in organ-specific metastasis (Kang et al., SOX4 exerts some of its action via regulation of these 2003; Minn et al., 2005), supporting the notion that two downstream targets. The discovery of 31 previously tumor cell–host interaction is the primary key to tumor unidentified targets expands our knowledge of how SOX4 metastasis as was originally postulated in the seed-and- modulates HCC progression and implies a range of novel soil hypothesis (Paget, 1889). Intrahepatic metastasis is SOX4 functions. This integrated approach sets a para- unique in that the liver microenvironment is the milieu digm whereby a subset of member genes from a for HCC cells both before intravasation and after synexpression group can be regulated by one master extravasation during the metastasis cascade. Thus, control gene and this is exemplified by SOX4 and liver-specific signature genes for local invasion are to advanced HCC. be expected. Initial efforts to search for genes that Oncogene (2008) 27, 5578–5589; doi:10.1038/onc.2008.168; functionally contribute to intraheptic metastasis have published online 26 May 2008 identified a few candidates (Ye et al., 2003) but a comprehensive portrayal of intraheptic metastasis pro- Keywords: SOX4; HCC metastasis; synexpression; gression is still lacking. target prediction We compared the expression profiles of six HCCs undergoing intraheptic metastasis and four primary HCCs. Among the differentially expressed genes identi- Corresponding authors: Dr A-P Tsou, Department of Biotechnology fied, we further characterized a transcription factor and Laboratory Science in Medicine, National Yang-Ming University, SOX4, a member of a highly conserved transcription Taipei 112, Taiwan, ROC or Dr J-T Horng, Institute of Computer Science and Information Engineering, National Central University, factor SOX (Sry-box) family known to have a Chung-Li 320, Taiwan, ROC or Dr M Hsiao, Genomics Research characteristic DNA-binding HMG domain (Bowles Center, Academia Sinica, Nankang, Taipei, 115, Taiwan, ROC et al., 2000). SOX4 has been shown to be a transcrip- E-mails: [email protected] or [email protected] or tional activator involved in the development of the [email protected] 6These authors contributed equally to this work. cardiac outflow tract, of pro-B cell expansion (Schilham Received 11 December 2007; revised 31 March 2008; accepted 17 April et al., 1996) and of the central nervous system (Cheung 2008; published online 26 May 2008 et al., 2000). SOX4 is absent in adult normal livers Overexpression of SOX4 potentiates metastasis in HCC Y-L Liao et al 5579 (Hunt and Clarke, 1999), hence it is not involved in To substantiate the significance of SOX4 overexpres- normal liver functions. Recently, overexpression of sion, further validation using a larger set of HCC SOX4 has been found to be associated with several samples was performed. The expression of SOX4 in T3 human cancer types (Aaboe et al., 2006; Liu et al., 2006; tumors (n ¼ 24) was significantly higher than adjacent Pramoonjago et al., 2006). Although SOX4 was the first normal tissues (n ¼ 61, adjacent-average) and T1 tumors ‘classical’ transcription factor identified with separable (n ¼ 15). No statistically significant difference was found DNA-binding and transactivation domains among the between T3 tumors and T2 tumors (n ¼ 22) suggesting SOX family members, information on its target genes is that elevated expression of SOX4 correlates with tumor scanty. Only two SOX4-binding motifs, 50-AACAAAG-30 progression (Figure 1c). and 50-AACAATA-30, have been empirically verified in the promoters of p56lck (McCracken et al., 1997) and human CD2 (Wotton et al., 1995), respectively. SOX Suppression of SOX4 expression reduces depend on requisite partners for target specificity migration/invasion of HCC cells in vitro and combinatorial control with partner factors is the main A high level of SOX mRNA was detected in the fetal theme for gene regulation by these proteins (Kamachi liver and in five HCC cell lines, with the exception of et al., 2000; Wilson and Koopman, 2002). Whether HepG2 (Figure 2a). This pattern of expression is induced expression of SOX4 leads to a different and reminiscent of known oncofetal antigens detected in pathological transcription program in HCC compared HCC (Grizzi et al., 2007). RNA interference with VSV- with the normal hepatocytes from which they originate is pseudolentiviral shRNA was used to reduce SOX4 not known. In this study, we determined a functional role expression in Mahlavu, SK-Hep1 and HuH7, which for SOX4 in liver tumor progression. Furthermore, we are cell lines known to manifest invasion activity investigated transcription regulation by SOX4 using (Shouval et al., 1988; Ye et al., 2003). shSOX4 reduced phylogenetic footprinting to obtain target prediction. SOX4 expression at both mRNA (Figure 2d, insets) and Finally, we carried out target verification by real-time level (Figure 2b); this happened in both the quantitative PCR, chromatin immunoprecipitation assay parental and HA-SOX4-transfected Mahlavu cells. To and promoter reporter activity assay. overcome the poor quality of commercial anti-SOX4 antibody available, which is able to detect endogenous SOX4 protein (’ in Figure 2b), the western analysis Results used HA-SOX4-transfected Mahlavu cells and anti-HA antibody to detect HA-SOX4 fusion protein (b in Identification of genes associated with intrahepatic Figure 2b). The SOX4 protein has a molecular weight of metastatic HCCs by transcription profiling 47 kDa, but migrated at about 70kDa when over- To explore the genes contributing to intrahepatic expressed in the transfected cells probably due to the metastasis, we conducted a global transcriptome analy- effect of the highly acidic serine-rich domain on protein sis using Affymatrix chips (HG_U133 Plus 2.0) with mobility (Hur et al., 2004). Downregulation of SOX4 RNA samples from four primary HCC (T1), six did not significantly affect cell growth (Figure 2c) but intrahepatic metastatic HCC (T3) and two normal liver rather induced distinct morphological changes in the samples (HNL). We focused on a set of 600 genes that HCC cells (Figure 2d). Five to seven days after shSOX4 were highly expressed in T3 tissues (Supplementary lentiviral infection, HuH7 cells exhibited a less-trans- Figure 1 and Table 1). This gene group was marked as formed pavement-like cell arrangement (Lee et al., 2003) T3>T1. Fourteen genes with sixfold higher expression whereas giant fused cells were noticed in the SK-Hep1 in the T3 tissues were identified and these were enriched and Mahlavu cultures. Upon close examination with with adhesion molecules (NRCAM, CD24, CTHRC1 phallodine staining, F-actin positive membrane protru- and ROBO1) and transcription factors (FOXQ1, sions were significantly reduced in addition to a loss of CDCA7 L, MYBL2 and SOX4); thus, these genes are F-actin organization in the shSOX4-treated cells candidate genes for direct involvement in tumor (Figure 2e). The level of vimentin, a marker for the metastasis (Figures 1a and b). Overexpression of epithelial-mesenchymal transition (EMT), was drasti- SOX4 has been shown to promote apoptosis in bladder cally reduced in shSOX4-treated Mahlavu cells carcinoma (Aaboe et al., 2006); in contrast, RNAi (Figure 3a). This suggested that elevated expression of knockdown of SOX4 induces apoptosis in adenoid SOX4 impacts on the EMT of HCC cells. Therefore, the cystic carcinoma and prostate cancer (Liu et al., 2006; effect of shSOX4 on in vitro migration/invasion was Pramoonjago et al., 2006). This suggests that the SOX4 examined and a 30–65% reduction in both migration molecular activity during caracinogenesis is influenced and invasion activity was found in SOX4 knockdown by additional factors that are enriched in a tissue- cells (Figure 3b). Mahlavu showed the greatest response specific manner. The requirement for SOX4 in a wide to the SOX4 RNA interference. range of developmental processes (van De et al., 1993; Ya et al., 1998; Cheung et al., 2000) further argues for a Prediction of potential SOX4 target genes synergy between SOX4 and tissue-selective partner using computational approach proteins. Among the identified highly upregulated genes A transcription factor exerts its actions via regulation of in intrahepatic liver tumors, SOX4 was an attractive downstream target genes. To fully understand the roles candidate for further investigation. of SOX4 in HCC progression, a comprehensive knowl-

Oncogene Overexpression of SOX4 potentiates metastasis in HCC Y-L Liao et al 5580

Figure 1 Microarray analysis identified SOX4 as upregulated in intrahepatic metastatic liver tumors. (a) Candidate genes associated with intrahepatic metastases. Fourteen candidate genes with sixfold higher expression in T3 tissues were identified in the microarray analysis and their expression levels were verified by reverse transcription (RT)–PCR (b). (c) Further confirmation of increased SOX4 mRNA expression in an additional 61 pairs of hepatocellular carcinoma (HCC) samples (T1, n ¼ 15; T2, n ¼ 22; and T3, n ¼ 24). Results were normalized against the expression level of GAPDH mRNA in each sample. The box plot shows the data distribution across the group classification and presents the 75th and 25th percentile (upper and lower quartile) with the median value in between. N: an average of expression level of all 61 normal adjacent tissues (adjacent-average). There is a statistically significant difference between the level of SOX4 mRNA when T3 tumors are compared to their adjacent tissue and T1 tumors.

edge of its targets is needed. Our assumption was that immunoprecipitation (ChIP) assays were performed to the genes regulated by SOX4 are likely to be present in discover if these motifs are functional binding sites for the SOX4 synexpression group, namely, the T3>T1 SOX4. gene group. The binding motifs of the evolutionarily conserved transcription factors tend to reside in human- mouse conserved noncoding blocks (CNBs) within the Validation of the predicted SOX4 target genes using ChIP promoter regions of the putative target genes (Hardison and siRNA assays et al., 1997; Wasserman et al., 2000). To unravel the ChIP is a powerful technique for analysing transcription SOX4 targets, we developed a bioinformatics screening factor binding sites in living cells. We performed ChIP strategy based on phylogenetic footprinting to identify assays with Mahlavu cells transfected with HA-SOX4 SOX4-binding motifs in the T3>T1 gene group (Fickett using anti-HA antibody to detect sites binding the HA- and Wasserman, 2000). A degenerate consensus motif, SOX4 fusion protein. In parallel, control ChIPs were WWCAAWG (A/T A/T CAA A/T G; Wilson and done using commercially available anti-immunoglobulin Koopman, 2002), which accommodates the known G (IgG) antibody. As shown in Figure 4a, SOX4 protein SOX4 binding motifs, AACAA A/T G was used. formed complexes with the predicted motifs of 31 genes A group of 41 genes containing at least one binding but failed to bind to the predicted motifs of 10genes motif in their CNBs was retrieved from the T3>T1 gene (Table 1, genes underlined). These results indicated a group (Supplementary Table 2a). In CNBs, 47 binding 76% positive prediction rate. The specificity of the ChIP motifs were identified (Table 1). The motifs demon- assay was further confirmed using HCC cells transfected strated a broad distribution of spatial locations. Sixteen with an unrelated HA-FOXQ1 gene construct (data not motifs were located >1 kb upstream from the transcrip- shown). Occurrences of the motif patterns in ChIP- tion start site (TSS), six sites were in intron 1 and verified target genes were tallied. Although the sample three sites were in the 50-UTR regions. Chromatin size is small, nevertheless the frequency of AACAAAG,

Oncogene Overexpression of SOX4 potentiates metastasis in HCC Y-L Liao et al 5581

Figure 2 Suppression of SOX4 expression by RNA interference resulted in morphological changes and actin organization of hepatocellular carcinoma (HCC) cells. Suppression of SOX4 in different HCC cell lines was achieved with shSOX4 lentiviral infection. A control shLuc lentivirus was also used. (a) Expression level of SOX4 in fetal liver, adult liver and six HCC cell lines detected by reverse transcription (RT)–PCR. (b) The effect of SOX4 knockdown was verified in the parental (Mahlavu) and in HA-SOX4- transfected Mahlavu (Mahlavu/HA-SOX4) cells by western blotting using anti-HA and anti-SOX4 antibodies. Cytosolic (C) or nuclear (N) proteins (30 mg) were loaded. HA: cells transfected with pcDNA3.1-HA vector. Arrowhead (b): HA-SOX4 fusion protein; arrow (’): endogenous SOX4 protein. Total actin was used as the internal control. An antibody against actin was used as the control in this study; MAB1501 is a pan-actin antibody that can react with all six isoforms of vertebrate actin. (c) SOX4 knockdown did not affect cell proliferation of Mahlavu cells. Cell growth was measured using the MTT assay. (d) The morphologies of HuH7, Mahlavu and SK-Hep 1 cells after downregulation of SOX4 by shRNA. Knockdown of SOX4 in HCC cells was confirmed by RT–PCR (insets). The presence of altered morphologies of the HCC cells was revealed by phase contrast microscopy. Images were taken 5–7 days after infection. Original magnification: Â 200. (e) Loss of actin organization in shSOX4-treated cells. Control shLuc-treated and shSOX4-treated HuH7 and Mahlavu cells were fixed and stained with rhodamine-phalloidine or 46-diamidino-2-phenyl indole (DAPI) to detect F-actin and the nucleus, respectively. Right column: merged image of F-actin and nucleus. Bar, 20 mm for shLuc-treated Mahlavu; 10 mm for other treatments.

TTCAAAG and ATCAAAG is slightly higher than all genes, MAP4, NAV3, NPNT and PAM, were not the other patterns combined (Supplementary Table 2b). affected (Figure 4d). Unexpectedly, a reduction of Based on the percentage of each nucleotide at each SOX4 expression would seem to have a discordant position within the core motif (WWCAAWG) and effect on 13 target genes between HuH7 and Mahlavu the three flanking nucleotides, the consensus binding cells (Table 3). motif for SOX4 is A/T T/A CAA A/t G (Table 2a). Motif specificity was validated by comparing the SOX4 motif with the consensus SOX9 motif (Mertin et al., NRP1 and SEMA3C, two SOX4 target genes, regulate 1999). In contrast to SOX9, no predominant 50 or 30 in vitro cell migration flanking nucleotides were observed for the SOX4 Our results revealed that SOX4 was a very potent binding motifs in the 31 target genes (Table 2b). This transcription factor, which switched on a considerable result demonstrates that the binding of the SOX4 downstream transcriptional cascade involving multiple protein is specific. cellular pathways (Table 1). A detailed knowledge of the To further characterize these 31 genes as SOX4 direct affected genes will require a comprehensive investiga- targets, we analysed the expression levels of all 31 genes tion. Genes implicated in mobility and/or metastasis are in Mahlavu and HuH7 cells. The expression level of 20 often associated with axon guidance, cell differentiation, genes was decreased (Figure 4b), whereas that of 7 genes Wnt signaling and microtubule dynamics. In neuronal was slightly increased in the SOX4 knockdown Mahlavu cells, guidance molecules modulate growth cone motility cells (Figure 4c). However, the expression levels of four through cytoskeletal changes (Guan and Rao, 2003).

Oncogene Overexpression of SOX4 potentiates metastasis in HCC Y-L Liao et al 5582

Figure 3 Suppression of SOX4 resulted in differential expression of vimentin and reduction in migration/invasion activity of Mahlavu cells. (a) Mahlavu cells express high level of vimentin but E-cadherin was undetectable. Less expression of vimentin was detected after SOX4 knockdown (western blot and immunofluorescence). Antibodies to E-cadherin and vimentin were used. Right column of the lower panel: merged image of vimentin and nucleus. Bar, 20 mm for shLuc-treated cells; 10 mm for shSOX4-treated cells. (b) SOX4 knockdown resulted in a 30–65% reduction in migration (upper panel) and invasion (lower panel) activity in three hepatocellular carcinoma (HCC) cell lines. Cells were infected with either control shLuc or shSOX4 for 16 h, which was followed by 2-day puromycin (2 mg/ml) selection before the cells were plated for the migration or invasion assay. Cell migration and invasiveness were assessed as described in Materials and methods. Thirty-five fields were counted for every filter. Data are an average of triplicates for each condition. **Po0.01. White bars, Mahlavu cells; black bars, SK-Hep1 cells; gray bars, HuH7 cells.

Both semaphorin 3C (SEMA3C) and neuron navigator this observation is not clear. The cloned promoter 3 (NAV3) are involved in axonal guidance (Maes et al., fragments of NRP1 and SEMA3C contain overlapping 2002; Gonthier et al., 2007) where neuropilin-1 (NRP1) binding sites for SOX4, FOXA1, HNF4 and POU2F1. is a for the semaphorins (He and Tessier- The CAA to TGG mutation may have created poor Lavigne, 1997). accessibility for other transcription factors because a Identification of SOX4 target genes in the category of general reduction of reporter activity was detectable axon guidance is of great interest. To clarify the SOX4 under SOX4 knockdown. This effect needs further transcriptional regulatory mechanism, we conducted investigation. Nevertheless, this result indicates that promoter reporter assays with the NRP1 and SEMA3C SOX4 can directly induce NRP1 and SEMA3C tran- promoter regions and also generated a series of mutant scription activation through the SOX4 binding motifs constructs with a core binding-site mutation found within their promoters. (WWCAAWG to WWTGGWG). In the presence of To understand whether NRP1 and SEMA3C expres- exogeneous SOX4, the promoter activity of NRP1 was sion confers biological activity on SOX4, we examined stimulated by 2.5-fold and this was reduced significantly the effect of silencing NRP1 and SEMA3C on cell by a mutated SOX4 binding site (Figure 5a). The migration. Knockdown of NRP1 and SEMA3C SEMA3C promoter has three motifs, one (À845 to (Figure 5c) drastically reduced cell-migration activity À839) in the conserved block and two additional motifs in HCC cells (Figure 5d). This result lends support to (À917 to À911 and À1104 to À1098) in nonconserved the idea that SOX4 affects cell migration partly via regions. Mutations of the core binding motifs resulted in regulation of its downstream target genes; thus NRP1 small reductions in luciferase activity but a 50% and SEMA3C are effectors whereas SOX4 is an reduction was detected when all three binding sites were upstream instigator in the chain of regulation. mutated (M3; Figure 5b). This indicates that all three SOX4 binding motifs in the SEMA3C promoter directly SOX4 knockdown suppresses in situ tumor invasion and influence SEMA3C transcription activation. The basal metastasis levels of both NRP1 and SEMA3C promoter activity To determine whether SOX4 knockdown affects meta- were partially reduced in SOX4 knockdown cells static potentials, we initially infected the Mahlavu and (Figures 5a and b). Surprisingly, shSOX4 also reduced HuH7 cells with shLuciferase (shLuc) or shSOX4 the luciferase activities of the promoter reporters lentiviruses. The infected cells were injected into the bearing the mutated binding sites. A definitive cause to flanks of nude mice or, alternatively, orthotopically into

Oncogene Overexpression of SOX4 potentiates metastasis in HCC Y-L Liao et al 5583 Table 1 Forty-one genes in the T3>T1 group had SOX4-binding site(s) in the promoter region Function Gene symbol T3/T1 Predicted SOX4 Positions of binding Locations of motifs relative ratioa binding motifb motif from ATGc (TSS)d to location of ATGe

Metabolism AKR1B1012.17 TTCAAAG À2017 (À320) U (exon 1) PAM 4.22 ATCAAAG, À2231, À2216 (À373) U U (exon 1) ATCAAAG GYG 2.94 TTCAAAG À2282 (À89) U (exon 1) RCC2 2.92 ATCAAAG À2782 (À1210) U (exon 2) DHRS13 2.69 TTCAATG À2837 (À393) U (exon 2) ALDH18A1 2.35 TTCAATG, À2149, À1362 (À3248) Intron 1, Intron 1 (exon 2) AACAAAG Kinase HUNK 3.56 AACAATG À1517 (À360) U (exon 1) PFKFB4 2.20TTCAAAG À1370( À17) U (exon 1) CKMT1 2.18 ATCAAAG À954 (À163) U (exon 1) Transcription factor FOXQ1 9.22 ATCAAAG À1897 (À265) U (exon 1) TEAD2 3.65 AACAAAG À2369 (À2382) 50UTR (exon 2) CREB3L1 2.31 TTCAATG À2556 (À103) U (exon 1) MYEF2 2.01 TTCAATG À2565 (À56) U (exon 1) Axon guidance NAV3 3.39 AACAATG À918 (À45) U (exon 1) SEMA3C 3.03 ATCAAAG À3415 (À2570) U (exon 2) NRP1 2.03 AACAATG À1251 (À160) U (exon 1) Regulation of POU2AF1 4.27 TTCAAAG À34 (À515) 50UTR (exon 1)* transcription HIC2 2.69 TTCAAAG À26314 (À25428) U (exon 2) VGLL4 2.35 TACAAAG À891 (À17529) Intron 1 (exon 2) G-protein signaling GNG4 2.32 AACAAAG À68337 (À66916) U (exon 2) pathway GPRC5B 2.02 AACAAAG À1596 (À11984) Intron 1 (exon 2) Cell adhesion CSPG2 3.43 ATCAATG, À13252, À12909 (À11755) U U (exon 2) TTCAAAG LAMB1 3.40AACAAAG À217 (À474) U (exon 2) Transport SLC22A15 3.10TACAAAG À1986 (À130) U (exon 1) SLC2A1 2.84 AACAATG À2800 (À179) U (exon 1) Cell differentiation SERPINE2 5.30TACAATG À2504 (À37376) Intron 1 (exon 2) Wnt signaling pathway DKK1 4.84 TTCAAAG À945 (À137) U (exon 1) DNA repair NEIL3 2.81 ATCAAAG À1421 (À80) U (exon 1) Actin cytoskeleton FHOD3 2.75 TTCAAAG À1420( À119) U (exon 1) organization snRNA processing INTS8 2.37 TACAAAG À2473 (À5679) 50UTR (exon 5) Proteolysis CTSC 2.33 TACAATG À1514 (À98) U (exon 1) Phosphatase PTPN14 2.24 ATCAAAG À2173 (À86496) Intron 1 (exon 2) Calcium ion binding NPNT 2.20ATCAAAG À1844 (À202) U (exon 1) Regulation of microtubule depolymerization MAP4 2.08 TACAAAG À91232 (À90414) U (exon 2) Cell cycle CHFR 2.05 ATCAAAG À426 (À284) U (exon 2) RNA binding RBM102.04TTCAAAG, À4457, À2449 (À2232) U U (exon 2) TTCAAAG Regulation of translation SMG5 2.03 TTCAATG, À2789, À1341(À94) U U (exon 1) AACAAAG Protein folding HSPBAP1 2.01 AACAAAG À197 (À123) U (exon 1) Unclear CCDC97 2.89 AACAAAG À665 (À122) U (exon 1) UBAP2L 2.47 AACAAAG, À6595, À1063 (À3786) U Intron 1 (exon 2) AACAAAG UBAP2 2.19 AACAAAG À33410( À31801) U (exon 2)

The promoter region was defined as a 3-kb upstream sequence from both the transcription start site (TSS) and the translation start site (ATG). TSS was positioned according to the translation start site of the gene. TSSs of several genes are located at a distance of >3000 bp from the ATG site. The sequences and the positions of the binding motifs are listed. The genes are ranked by their relative levels between the T3 and T1 patient groups (T3/T1 ratio). Functional classification of the genes was derived from the database and a literature search. Underlined: Genes that failed to show SOX4 binding by ChIP analysis. aT3/T1 ratio: Relative gene expression levels between T3 and T1 patient groups revealed by the microarray experiments. bPredicted SOX4 binding motifs identified in the conserved non-coding blocks. cPositions of the motifs from the translation start site (ATG), in bp. dPosition of TSS from the translation start site in bp. eLocation of motifs (U: upstream from TSS; intron or 50UTR) and location of translation start site in the gene (exon no.). the livers of nude mice. Knockdown of SOX4 signifi- lentivirus (data not shown). Local invasion with visible cantly reduced Mahlavu tumor growth (Figure 6a). multiple tumor foci in the liver parenchyma was Invasion into deeper smooth muscles layers (Figure 6b, observed after intrahepatic injection of shLuc-infected blue arrowheads) was only found with the shLuc tumors Mahlavu cells (blue arrowheads, Figure 6c, mice 1 and but not with the shSOX4 tumors. Similar findings were 2; Figure 6d, mouse 2) but not in mice injected with also observed in HuH7 cells infected with shSOX4 shSOX4-Mahlavu cells (Figure 6c, mice 3 and 4:

Oncogene Overexpression of SOX4 potentiates metastasis in HCC Y-L Liao et al 5584

Figure 4 Identification of SOX4 targets in hepatocellular carcinoma (HCC) cells. (a) Chromatin immunoprecipitation (ChIP) assays with anti-HA antibody showed binding of HA-SOX4 to the promoters of 31 genes in Mahlavu cells transfected with HA-SOX4. All of these genes have SOX4 binding motifs in the human-mouse Conserved noncoding blocks (CNBs). LCK is a confirmed SOX4 target gene. Input: input DNA; HA: immunoprecipitation with anti-HA antibody from cells expressing HA-SOX4; IgG: immunoprecipita- tion with mouse immunoglobulin G (IgG) antibody was used as a negative control. The promoter of CSPG2 and TEAD2 failed to bind to SOX4. Expression of SOX4 target genes was detected using real-time quantitative PCR by the SYBR Green I protocol. All values were normalized against GAPDH mRNA. After shSOX4 knockdown of endogenous SOX4 (b) expression of 20SOX4 target genes in Mahlavu cells was downregulated, (c) expression of 7 target genes in Mahlavu cells was upregulated and (d) no significant change in expression was found for 4 target genes. C: shLuc was used as the negative control; sh: shSOX4. In all cases, shSOX4 knockdown reduced the level of endogenous SOX4 mRNA by X60%. White bars, shLuc treatment; shaded bars, shSOX4 treatment. The standard deviation is indicated. *Po0.05; **Po0.01.

Table 2 Analysis of SOX4 binding motifs in confirmed target genes Only a remnant in the needle track (Figure 6e, lower-left a panel, black arrow) and the residual tumor mass (Figure 6e; lower-middle panel, yellow arrowheads) A 36.1 27.8 52.8 44.4 0 100 100 69.4 0 38.9 C 11.1 22.2 0 0 100 0 0 0 0 8.3 were observed in mouse 3 and mouse 4, respectively. G 27.8 16.7 0 0 0 0 0 0 100 22.2 Our findings suggest that SOX4 plays an important T 25.0 33.3 47.2 55.6 0 0 0 30.6 0 30.6 role in regulating progression of liver tumors. The Consensus binding motif for SOX4 potential of SOX4 as a novel pathological staging

a t A/T T/A C A A A/t G a marker and a therapeutic target for liver cancer merits further investigation. b SOX4 a t A/T T/A C A A A/t G a SOX9* A G A A C A A T G G Discussion

(a) The percentage of each nucleotide at each position within the core In an effort to identify genes differentially expressed in motif (boxed) and three flanking nucleotides was calculated. The core binding motif is boxed. A nucleotide appearing at a position with a intrahepatic metastasis, we found that SOX4 is over- percentage of >40% is written in an upper-case letter otherwise they expressed in the liver tumors with local invasion. The are in lower-case letters. The consensus binding motif for SOX4 is A/T pattern of expression was validated in a larger set of T/A CAA A/t G. (b) Comparison of the consensus binding motifs of HCC samples and the clinical relevance of SOX4 SOX4 and SOX9 (Mertin et al., 1999). expression explored using mechanistic studies. Empirical evidence suggested that SOX4 plays an important Figure 6d, mouse 3, Figure 6e, mice 3 and 4). The edges function in liver tumor metastasis as RNAi knockdown of the Mahlavu tumor foci were found to be actively reduced HCC cell migration, invasion and intrahepatic invading the normal parenchyma and also the major metastasis in an orthotopic liver cancer model. This blood vessel cavities of the liver (Figure 6e, mouse 1). SOX4 function also seems to operate during breast

Oncogene Overexpression of SOX4 potentiates metastasis in HCC Y-L Liao et al 5585 Table 3 Discordant gene expression after SOX4 RNA knockdown in cancer metastasis (Tavazoie et al., 2008) but not in other Mahlavu and HuH7 cells human cancer types (Aaboe et al., 2006; Liu et al., 2006; Symbol Gene expression after Pramoonjago et al., 2006). This suggests that SOX4 has SOX4 knockdown diverse activities across various human cancers and that these are likely to be cell context-dependent, involving Mahlavu HuH7 differential target activation. SOX4 knockdown neither AKR1B10Down Up induced apoptosis (data not shown) nor decreased cell CREB3L1 Down Up growth in Mahlavu cells (Figure 2d). It is tempting to GPRC5B Down Up speculate that shSOX4 reduced in vivo tumor growth ALDH18A1 Up Down and metastasis by a failure of angiogenesis as reported HSPBAP1 Up Down RBM10Up Down for siNRP1 (Hong et al., 2007). RCC2 Up Down To delineate the impact of SOX4 overexpression in UBAP2L Up Down the transcription regulation of HCC, a target gene CCDC97 Up No change search platform incorporating doctrines frequently HIC2 Up No change MAP4 No Change Down implemented in eukaryotic transcription regulation NAV3 No Change Down was developed. These were (1) co-expressed genes are PAM No Change Down likely to be coregulated (Niehrs and Pollet, 1999; Karaulanov et al., 2004) and (2) noncoding regulatory Down: Reduced gene expression after SOX4 knockdown. Up: sequences tend to be evolutionarily conserved (Hardison increased gene expression after SOX4 knockdown. No change: Gene expression level not affected by SOX4 knockdown. Gene expression et al., 1997; Wasserman et al., 2000). The assumption was analyzed by real-time quantitative PCR. that overexpression of SOX4 modulates the transcrip-

Figure 5 SOX4 target genes, NRP1 and SEMA3C regulate cell migration. The promoters of NRP1 and SEMA3C are transactivated by SOX4. HEK293 T cells were transiently co-transfected with HA-SOX4 or pcDNA3.1-HA vector and (a) wild-type pGL3-NRP1 promoter (’) or pGL3-NRP1 promoter-mutant ( ), (b) wild-type pGL3-SEMA3C promoter (’) or pGL3-SEMA3C promoter- mutants. The SEMA3C promoter constructs with the mutated core binding sequences were designated as pGL3-SEMA3C promoter- M1 ( ), -M2 ( )orM3( ). Luciferase activity was measured 48 h after transfection. The transfection efficiency was normalized against pRL-TK activity. Normalized luciferase activity from triplicate samples is presented relative to that of cells transfected with the pGL3-basic construct (&). The experiment was repeated twice with same results. *Po0.05; **Po 0.01. (c) Suppression of either NRP1 or SEMA3C expression effectively reduced cell migration activity in hepatocellular carcinoma (HCC) cell lines. The effect of shRNA was verified by reverse transcription (RT)–PCR. HuH7 cells were not treated with shSEMA3C due to the low expression level of SEMA3C in these cells. (d) Knockdown of NRP1 or SEMA3C resulted in a nearly 50% reduction of migration activity in HCC cell lines. Data are an average of triplicates for each condition. **Po0.01. White bars, Mahlavu cells; black bars, SK-Hep1 cells; gray bars, HuH7 cells.

Oncogene Overexpression of SOX4 potentiates metastasis in HCC Y-L Liao et al 5586

Figure 6 Effects of SOX4 knockdown in subcutaneous tumors and orthotopic liver tumors. (a) In vivo subcutaneous tumor growth curves of shLuc and shSOX4 infected Mahlavu cells (n ¼ 7). * Po0.05; ** Po0.01. (b) H&E stained images of representative subcutaneous tumors that were extracted at 6 weeks after cell injection. Note shLuc control infected Mahlavu cells showed muscular infiltration. It should also be noted that tumor cells after shSOX4 knockdown formed multiple layers of cells with a flat morphology (yellow arrowheads) but did not invade into the smooth muscle. (c) Gross morphology of livers and abdominal cavity organs from representative mice after intrahepatic injection of shLuc (mice 1 and 2) or shSOX4 (mice 3 and 4) infected Mahlavu cells. (d) Gross morphology of representative livers from mice 2 and 3 in (c) after formaldehyde fixing for one day. (e) H&E stained images of representative livers that were extracted at 7 weeks after cell injection. Metastatic tumor foci (upper-left and middle-left panels) and invasive edges (upper-right and middle-middle panels, blue arrowheads) in mice 1 and 2 are identified. In the upper-middle panel, it is important to note the presence of tumor cells invading the hepatic vessels and forming small metastatic foci. Middle-right panel shows the morphology of the disseminating tumor in the abdominal cavity. Lower-right panel shows no invasion of tumor cells into liver parenchyma. Bar represents 100 mm. White arrows represent the injection sites of tumor cells. Black arrow shows the needle track. Blue arrowheads mark the tumor foci and invasive edges of tumor cells in the shLuc group. Yellow arrowheads represent the smooth and flat edges of the tumor cells in shSOX4 knockdown group.

tional activity of member genes within its synexpression for only p56lck and CD2 are known (Wotton et al., 1995; group was approached using both a bioinformatics and McCracken et al., 1997; Liu et al., 2006). This study is a biochemical strategy. Experimental evidence from the first to report 31 experimentally confirmed direct gene profiling, binding-site computation prediction, target genes for SOX4 in one publication. SOX4 is a ChIP verification and gene ablation led to the recogni- pleiotropic transcription factor regulating a wide variety tion of previously unidentified SOX4 target genes. of biological processes. Two SOX4 target genes, Stringent adherence to cross-species conserved sequence SEMA3C and NRP1, are associated with tumorigenesis paradigm improved substantially the motif prediction. or tumor progression (Bielenberg et al., 2006; Herman In the T3>T1 group, SOX4 binding motifs were and Meadows, 2007) but only their role in HCC recognized in 343 genes, of which 41 genes had motifs progression is explored in this study. High levels of in the human–mouse CNBs. ChIP verification further SEMA3C and NRP1 correlated with HCC cell migra- confirmed 31 genes as true targets. Although the tion, supporting a favorable role in promoting intrahe- importance of several SOX family proteins is well patic HCC metastasis; the implications of these results characterized in vertebrate developmental processes needs to be further tested. (Bowles et al., 2000), fully defined SOX target genes A dichotomy of SOX4 transcriptional activity, where- are few. There are four known SOX4 targets, p56lck, by it acts as both an activator and a suppressor, was CD2, TLE1 and PUMA but the SOX4 binding motifs observed in two HCC cell lines with dissimilar

Oncogene Overexpression of SOX4 potentiates metastasis in HCC Y-L Liao et al 5587 differentiation statuses (Table 3). Several questions CA, USA) and designated HA-SOX4. Luciferase reporter remained to be addressed. These include how SOX4 constructs containing the promoter regions of NRP1 (nucleo- modulates the transcriptional activity of genes with tides À1toÀ1216 with the SOX4 binding site AACAATG at divergent biological activities and how is the mode of position À1091 to À1097 bp) and SEMA3C (nucleotides À1to SOX4-mediated transcription activity, either activation À1173) were subcloned in pGL3-basic vector (Promega, Madison, WI, USA) and designated wild-type pGL3-NRP1 or suppression, determined. Combinatorial control promoter and pGL3-SEMA3C promoter, respectively. SOX4 integrating the concerted actions of multiple proteins core binding site mutations (CAA) within the promoter to achieve transcription repression or activation is the regions of NRP1 and SMEA3C were generated using a key to the higher eukaryote complex regulatory net- QuickChange Site-Directed Mutagenesis Kit (Stratagene, La works. The molecular roles in the combinatorial Jolla, CA, USA). The SEMA3C promoter has three SOX4 assembly of transcription regulators are exemplified by binding sites at À845 to À839, À917 to À911 and À1104 to the POU and SOX protein partnership and others À1098. The CAA of the core binding site was mutated at the (Kamachi et al., 2000; Wilson and Koopman, 2002; À845 site (M1), at both the À845 and À917 sites (M2) and at Remenyi et al., 2004). The mechanistic basis of SOX all three sites (M3). The RT–PCR primers used in mutagenesis functional specificity mainly rests on context-specific are listed in Supplementary Table 4. binding partner proteins (Wilson and Koopman, 2002). Syntenin has been shown to directly associate with Sox4 RNA interference with shRNA and this association is necessary for IL-5-mediated HCC cells were plated and infected with lentiviruses expressing activation of Sox4 transcriptional activity in B cells shSOX4, shNRP1, shSEMA3C or shLuc in the presence of (Geijsen et al., 2001). Whether this particular coopera- 8 mg/ml protamine sulfate for 24 h, which was followed by tion occurs with other SOX4-mediated gene regulation puromycin (2 mg/ml; 48 h) selection. RT–PCR and/or western events has not been tested. We explored the binding blotting were performed to validate the knockdown efficiency. motifs for potential partner proteins that bind in the The shRNA constructs are described in the Supplementary vicinity of the SOX4 binding site (o50bp). Several Materials and methods. POU2F1 binding motifs emerge as common cis elements near the SOX4 motif in the majority of SOX4 target Antibodies, immunoblotting, immunostaining and staining genes identified in this study (data not shown). Whether for F-actin SOX4 and POU2F1 synergistically regulate SOX4 Immunoblotting was performed as described previously (Lee targets is under investigation. A comprehensive analysis et al., 2003). Nuclear and cytosolic lysates were prepared using of the regulatory modules surrounding SOX4 binding a CelLytic Nuclear Extraction System Kit (Sigma). Protein sites ought to shed further light on the molecular basis lysate (30 mg) was electrophoresed on 10% SDS polyacryla- of SOX4 regulation. mide gels and transferred onto polyvinylidene difluoride membranes (Millipore, Bedford, MA, USA). The membranes were incubated with primary antibodies overnight at 4 1C and then with horseradish peroxidase-conjugated secondary anti- Materials and methods body (PerkinElmer Life Sciences, Boston, MA, USA). The primary antibody against HA was purchased from Covance Cell lines and human liver tissues (Berkeley, CA, USA), SOX4 from Abnova (Taipei, Taiwan), The human HCC cell lines, Mahlavu, HuH7, SK-Hep1 and vimentin from Lab Vision (Fremont, CA, USA), E-cadherin human HEK293T cells were cultured as described previously from BD Biosciences (Franklin Lakes, NJ, USA) and actin (Chau et al., 2007). Paired samples of tumor/nontumorous from Chemicon (Temecula, CA, USA). Signals were detected liver tissues (71) were obtained from patients who had by an enhanced chemiluminescence kit (PerkinElmer, Wal- undergone primary HCC curative hepatic resection at Taipei tham, MA, USA). The relative level of protein expression was Veterans General Hospital, Taiwan. The study was approved normalized against actin. Immunostaining was performed as by the Committee for the Conduct of Human Research and described previously (Chau et al., 2007). Incubation with patient informed consent was obtained. On analysis, 19, 22 rhodamine-phalloidin (1 U/ml; Invitrogen) was done at room and 30of the samples were classified at stages 1 (T1), 2 (T2) temperature for 15 min followed by counterstaining with 46- and 3 (T3) HCC, respectively (AJCC, 6th edn, TNM diamidino-2-phenyl indole. The fluorescent images were classification). Immediately after surgical resection all tissues captured using an Olympus FV1000 confocal microscope were snap frozen in liquid nitrogen and stored at –80 1C ( Â 60oil immersion lens). until use.

Analysis of gene expression Cell migration assay and invasion assay Gene expression of SOX4 in cell lines and after shSOX4 Cells were tested for migration and invasion abilities in vitro in knockdown was examined using RT–PCR and standard gel a Minicell (Millipore, Billerica, MA, USA). The lower side or electrophoresis. Expression of SOX4 target genes was detected the upper side of the polycarbonate membranes (containing with real-time quantitative PCR (RT-qPCR) by the SYBR 8-mm pores) of the Minicell coated with 50 mg/ml of type I Green I protocol (Bio-Rad, Hercules, CA, USA). All values collagen or 80 mg per well of Matrigel was used for migration were normalized against GAPDH mRNA. The primer assay or invasion assay, respectively. Cells were added to the sequences are listed in Supplementary Table 3. upper chamber of a Minicell. After incubation for 16 h at 37 1C, the cells at the lower side were prepared for Gimsa stain. Plasmid constructs The level of migration or invasion was determined using a The full-length human SOX4 (NM_003107) gene was sub- microscope at  200 magnification. All experiments were cloned into pcDNA3.1(B) HA vector (Invitrogen, Carlsbad, repeated three times.

Oncogene Overexpression of SOX4 potentiates metastasis in HCC Y-L Liao et al 5588 Prediction of SOX4 binding sites using cross-species comparison measured using the Dual-Luciferase Reporter Assay System A bioinformatics method that estimates the binding site Kit (Promega). conservation tendency by phylogenetic analysis of the hu- man–mouse orthologous promoters to predict the conserved In vivo turmorigenesis assay transcription factor binding sites was developed. In total, Mahlavu and HuH7 cells (107) infected with shLuc or shSOX4 19073 orthologous gene pairs were analysed. The promoters of lentiviruses were implanted subcutaneously into the flanks of the human and mouse orthologous gene pairs were retrieved the nude mice. Primary tumor growth rate was analysed by from Ensembl genome database V32 (Homo sapiens core measuring tumor length (L) and width (W), and tumor volume 32.35e and Mus musculus core 32.34). The promoter region was calculated according to V ¼ 0.4 Â ab2 as previously was defined as a 3-kb upstream sequence from both the TSS described (Attia and Weiss, 1966). Orthotopical intrahepatic and from the translation start site. The upstream sequences of injection was conducted as previously described (Lu et al., paired genes that shared 80% conservation based on local 2007). Using 20 ml of serum-free medium containing 50% alignment using BLAST (version 2.2.12) were identified as Matrigel (BD Biosciences, San Jose, CA, USA), 106 cells were CNBs. The putative promoter sequences of CNBs were slowly injected into the liver of anesthetized mice with a 27- scanned for the SOX4 binding motif, WWCAAWG. The gauge needle. Mice were killed 6–7 weeks later for histopatho- genes scored were classified into three categories, namely logical evaluation. ‘Motifs in CNB’, ‘Motifs not in CNB’ and ‘No Motif’. The final category was that where no SOX4-binding motif was Statistical analysis found. All data are expressed as mean±s.d. and compared among groups using the Student’s t-test. Categorical variables were Chromatin immunoprecipitation reactions compared using the Wilcoxon rank test. A P-value o0.05 was The ChIP assays were performed as described previously considered statistically significant. (Nelson et al., 2006). Cells were fixed with 1% formaldehyde, harvested and lysed in SDS buffer containing 50m M Tris-HCl Acknowledgements (pH 8.1), 0.5% SDS, 100 mM NaCl, 5 mM EDTA and protease inhibitors. The pellet was sonicated using 1 s pulses separated We acknowledge the technical supports provided by the by 5 s, for 4 min at output level 6 using a Sonicator3000 Microarray & Gene Expression Analysis Core Facility of the (Misonix, Farmingdale, NY, USA). The sheared chromatin VGH National Yang-Ming University Genome Research was precleared with 30 ml protein A beads (Amersham Center (VYMGC). The Gene Expression Analysis Core Biosciences, Piscataway, NJ, USA) followed by incubation Facility is supported by National Research Program for with 5 mg anti-HA mAb or mouse IgG (PerkinElmer). Genomic Medicine (NRPGM), National Science Council. Untreated sonicated chromatin was saved as input. Purified RNAi reagents were obtained from the National RNAi Core DNA was subjected to PCR reactions using the primers Facility located at the Institute of Molecular Biology/Genomic described in Supplementary Table 5. Research Center, Academia Sinica (supported by the National Research Program for Genomic Medicine Grants of NSC Promoter reporter assay (NSC 94-3112-B-001-003 and NSC 94-3112-B-001-018-Y). We HEK293 T cells (5 Â 104 per well) were seeded in 24-well plate thank Dr Hua-Chien Chen for providing the PCR primers for and co-transfected with 0.25 mg of HA-control vector or HA- the RT-qPCR assays. This work was supported in part by SOX4 construct, 0.25 mg of pGL3-NRP1 (wild-type or mutant) grant (NSC 95-2752-B-010-002-PAE) from National Science or pGL3-SEMA3C (wild-type or mutant) and 0.05 mg of pRL- Council (Program for Promoting Academic Excellence of TK (Promega) using jetPEI reagent (Polyplus-Transfection, Universities Phase II) and a grant from the Ministry of Illkirch, France). After 48 h, the luciferase activity was Education, Aim for the Top University Plan to APT and YPC.

References

Aaboe M, Birkenkamp-Demtroder K, Wiuf C, Sorensen FB, Thykjaer Cheung M, bu-Elmagd M, Clevers H, Scotting PJ. (2000). Roles of T, Sauter G et al. (2006). SOX4 expression in bladder carcinoma: Sox4 in central nervous system development. Brain Res Mol Brain clinical aspects and in vitro functional characterization. Cancer Res Res 79: 180–191. 66: 3434–3442. El-Serag HB, Rudolph KL. (2007). Hepatocellular carcinoma: Attia MA, Weiss DW. (1966). Immunology of spontaneous mammary epidemiology and molecular carcinogenesis. Gastroenterology 132: carcinoma in mice. V. Acquired tumor resistance and enhancement 2557–2576. in strain A mice infected with mammary tumor virus. Cancer Res 26: Fickett JW, Wasserman WW. (2000). Discovery and modeling of 1787–1800. transcriptional regulatory regions. Curr Opin Biotechnol 11: 19–24. Bielenberg DR, Pettaway CA, Takashima S, Klagsbrun M. (2006). Geijsen N, Uings IJ, Pals C, Armstrong J, McKinnon M, Raaijmakers Neuropilins in neoplasms: expression, regulation, and function. Exp JA et al. (2001). Cytokine-specific transcriptional regulation through Cell Res 312: 584–593. an IL-5Ralpha interacting protein. Science 293: 1136–1138. Bowles J, Schepers G, Koopman P. (2000). Phylogeny of the SOX Gonthier B, Nasarre C, Roth L, Perraut M, Thomasset N, Roussel G family of developmental transcription factors based on sequence and et al. (2007). Functional interaction between matrix metalloprotei- structural indicators. Dev Biol 227: 239–255. nase-3 and semaphorin-3C during cortical axonal growth and Chau GY, Lee AF, Tsay SH, Ke YR, Kao HL, Wong FH et al. guidance. Cereb Cortex 17: 1712–1721. (2007). Clinicopathological significance of survivin expression Grizzi F, Franceschini B, Hamrick C, Frezza EE, Cobos E, Chiriva- in patients with hepatocellular carcinoma. Histopathology 51: Internati M. (2007). Usefulness of cancer-testis antigens as 204–218. biomarkers for the diagnosis and treatment of hepatocellular Chen X, Cheung ST, So S, Fan ST, Barry C, Higgins J et al. (2002). carcinoma. J Transl Med 5:3. Gene expression patterns in human liver cancers. Mol Biol Cell 13: Guan KL, Rao Y. (2003). Signalling mechanisms mediating neuronal 1929–1939. responses to guidance cues. Nat Rev Neurosci 4: 941–956.

Oncogene Overexpression of SOX4 potentiates metastasis in HCC Y-L Liao et al 5589 Hardison RC, Oeltjen J, Miller W. (1997). Long human-mouse Minn AJ, Gupta GP, Siegel PM, Bos PD, Shu W, Giri DD et al. sequence alignments reveal novel regulatory elements: a reason to (2005). Genes that mediate breast cancer metastasis to lung. Nature sequence the mouse genome. Genome Res 7: 959–966. 436: 518–524. He Z, Tessier-Lavigne M. (1997). Neuropilin is a receptor for the Nelson JD, Denisenko O, Sova P, Bomsztyk K. (2006). Fast chromatin axonal chemorepellent Semaphorin III. Cell 90: 739–751. immunoprecipitation assay. Nucleic Acids Res 34: e2. Herman JG, Meadows GG. (2007). Increased class 3 semaphorin Niehrs C, Pollet N. (1999). Synexpression groups in eukaryotes. expression modulates the invasive and adhesive properties of Nature 402: 483–487. prostate cancer cells. Int J Oncol 30: 1231–1238. Paget S. (1889). The distribution of secondary growth in cancer of the Hong TM, Chen YL, Wu YY, Yuan A, Chao YC, Chung YC et al. breast. Lancet 1: 571–573. (2007). Targeting neuropilin 1 as an antitumor strategy in lung Pramoonjago P, Baras AS, Moskaluk CA. (2006). Knockdown of cancer. Clin Cancer Res 13: 4759–4768. Sox4 expression by RNAi induces apoptosis in ACC3 cells. Hunt SM, Clarke CL. (1999). Expression and hormonal regulation of the Oncogene 25: 5626–5639. Sox4 gene in mouse female reproductive tissues. Biol Reprod 61: 476–481. Remenyi A, Scholer HR, Wilmanns M. (2004). Combinatorial control Hur EH, Hur W, Choi JY, Kim IK, Kim HY, Yoon SK et al. (2004). of gene expression. Nat Struct Mol Biol 11: 812–815. Functional identification of the pro-apoptotic effector domain in Schilham MW, Oosterwegel MA, Moerer P, Ya J, de Boer PA, van de human Sox4. Biochem Biophys Res Commun 325: 59–67. WM et al. (1996). Defects in cardiac outflow tract formation and Kamachi Y, Uchikawa M, Kondoh H. (2000). Pairing SOX off: with pro-B-lymphocyte expansion in mice lacking Sox-4. Nature 380: partners in the regulation of embryonic development. Trends Genet 711–714. 16: 182–187. Shouval D, Schuger L, Levij IS, Reid LM, Neeman Z, Shafritz DA. Kang Y, Siegel PM, Shu W, Drobnjak M, Kakonen SM, Cordon- (1988). Comparative morphology and tumourigenicity of human Cardo C et al. (2003). A multigenic program mediating breast cancer hepatocellular carcinoma cell lines in athymic rats and mice. metastasis to bone. Cancer Cell 3: 537–549. Virchows Arch A Pathol Anat Histopathol 412: 595–606. Karaulanov E, Knochel W, Niehrs C. (2004). Transcriptional Tang ZY, Ye SL, Liu YK, Qin LX, Sun HC, Ye QH et al. (2004). regulation of BMP4 synexpression in transgenic Xenopus. EMBO A decade’s studies on metastasis of hepatocellular carcinoma. J J 23: 844–856. Cancer Res Clin Oncol 130: 187–196. Kim Y, Sills RC, Houle CD. (2005). Overview of the molecular biology Tavazoie SF, Alarco´ n C, Oskarsson T, Padua D, Wang Q, Bos PD of hepatocellular neoplasms and hepatoblastomas of the mouse et al. (2008). Endogenous human microRNAs that suppress breast liver. Toxicol Pathol 33: 175–180. cancer metastasis. Nature 451: 147–152. Lee YC, Tang YC, Chen YH, Wong CM, Tsou AP. (2003). Selenite- van de WM, Oosterwegel M, van NK, Clevers H. (1993). Sox-4, an induced survival of HuH7 hepatoma cells involves activation of Sry-like HMG box protein, is a transcriptional activator in focal adhesion kinase-phosphatidylinositol 3-kinase-Akt pathway lymphocytes. EMBO J 12: 3847–3854. and Rac1. J Biol Chem 278: 39615–39624. Wasserman WW, Palumbo M, Thompson W, Fickett JW, Lawrence Liu P, Ramachandran S, Ali SM, Scharer CD, Laycock N, Dalton WB CE. (2000). Human-mouse genome comparisons to locate regula- et al. (2006). Sex-determining region Y box 4 is a transforming tory sites. Nat Genet 26: 225–228. oncogene in human prostate cancer cells. Cancer Res 66: 4011–4019. Wilson M, Koopman P. (2002). Matching SOX: partner proteins and Lu YS, Kashida Y, Kulp SK, Wang YC, Wang D, Hung JH et al. co-factors of the SOX family of transcriptional regulators. Curr (2007). Efficacy of a novel histone deacetylase inhibitor in murine Opin Genet Dev 12: 441–446. models of hepatocellular carcinoma. Hepatology 46: 1119–1130. Wotton D, Lake RA, Farr CJ, Owen MJ. (1995). The high mobility Maes T, Barcelo A, Buesa C. (2002). Neuron navigator: a human gene group transcription factor, SOX4, transactivates the human CD2 family with homology to unc-53, a cell guidance gene from enhancer. J Biol Chem 270: 7515–7522. Caenorhabditis elegans. Genomics 80: 21–30. Ya J, Schilham MW, de Boer PA, Moorman AF, Clevers H, Lamers McCracken S, Kim CS, Xu Y, Minden M, Miyamoto NG. (1997). An WH. (1998). Sox4-deficiency syndrome in mice is an animal model alternative pathway for expression of p56lck from type I promoter for common trunk. Circ Res 83: 986–994. transcripts in colon carcinoma. Oncogene 15: 2929–2937. Ye QH, Qin LX, Forgues M, He P, Kim JW, Peng AC et al. (2003). Mertin S, McDowall SG, Harley VR. (1999). The DNA-binding Predicting hepatitis B virus-positive metastatic hepatocellular specificity of SOX9 and other SOX proteins. Nucleic Acids Res 27: carcinomas using gene expression profiling and supervised machine 1359–1364. learning. Nat Med 9: 416–423.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

Oncogene