A High-Throughput Study in Melanoma Identifies Epithelial- Mesenchymal Transition As a Major Determinant of Metastasis

Research Article

A High-Throughput Study in Melanoma Identifies Epithelial- Mesenchymal Transition as a Major Determinant of Metastasis

Soledad R. Alonso,1 Lorraine Tracey,1 Pablo Ortiz,4 Beatriz Pe´rez-Go´mez,5 Jose´ Palacios,1 Marina Pollań,5 Juan Linares,6 Salvio Serrano,7 Ana I. Saéz-Castillo,6 Lydia Sańchez,2 Raquel Pajares,2 Abel Sańchez-Aguilera,1 Maria J. Artiga,1 Miguel A. Piris,1 and Jose´ L. Rodrı´guez-Peralto3

1Molecular Pathology Programme and 2Histology and Immunohistochemistry Unit, Centro Nacional de Investigaciones Oncolo´gicas; Departments of 3Pathology and 4Dermatology, Hospital Universitario 12 de Octubre; 5Centro Nacional de Epidemiologı´a, Instituto de Salud Carlos III, Madrid, Spain; and Departments of 6Pathology and 7Dermatology, Hospital Universitario San Cecilio, Granada, Spain

Abstract with a less favorable prognosis as potential candidates for adjuvant Metastatic disease is the primary cause of death in cutaneous or novel therapies. malignant melanoma (CMM) patients. To understand the Currently, the prognosis of primary CMM is mainly based mechanisms of CMM metastasis and identify potential on histopathologic criteria. The most important of these is the predictive markers, we analyzed gene-expression profiles of Breslow index, although it is merely a measure of tumor depth. 34 vertical growth phase melanoma cases using cDNA micro- New molecular markers that correlate with melanoma genesis and/or progression are continuously being identified but, to date, arrays. All patients had a minimum follow-up of 36 months. Twenty-one cases developed nodal metastatic disease and 13 most of them have been obtained in experimental models and did not. Comparison of gene expression profiling of metastatic have not yet been confirmed in series of human samples. and nonmetastatic melanoma cases identified 243genes with The development of high-throughput screening techniques in a >2-fold differential expression ratio and a false discovery genomics and proteomics has enabled the analysis of the rate of <0.2 (206 up-regulated and 37 down-regulated). This expression of multiple genes and proteins in large series set of genes included molecules involved in cell cycle and of tumor samples (1) and may contribute to the resolution of apoptosis regulation, epithelial-mesenchymal transition some specific issues of clinical relevance, such as identifying the (EMT), signal transduction, nucleic acid binding and tran- steps for melanoma progression and metastasis. Until now, most scription, protein synthesis and degradation, metabolism, and studies characterizing CMM have been done using cell lines (2, 3), mouse models (4), or metastatic tumor samples (2, 5). However, a specific group of melanoma- and neural-related proteins. Validation of these expression data in an independent series data obtained from primary CMM samples are difficult to obtain, of melanomas using tissue microarrays confirmed that the due in part to the lack of retrospective collections of frozen expression of a set of proteins included in the EMT group primary melanomas samples or adequate follow-up (5–7). (N-cadherin, osteopontin, and SPARC/osteonectin) were sig- Molecular changes associated with the acquisition of metastatic nificantly associated with metastasis development. Our results capacity in vertical growth phase melanoma are still to be fully suggest that EMT-related genes contribute to the promotion described. We might expect a repertoire of differentially expressed of the metastatic phenotype in primary CMM by supporting genes defining the metastatic phenotype for primary CMM cases. specific adhesive, invasive, and migratory properties. These The present study aimed to investigate the metastasis signature data give a better understanding of the biology of this in primary CMM. To this end, we compared the gene expression aggressive tumor and may provide new prognostic and patient profile in a series of vertical growth phase primary CMMs that had stratification markers in addition to potential therapeutic developed metastatic disease with vertical growth phase CMM targets. [Cancer Res 2007;67(7):3450–60] without metastatic disease by the end of follow-up. Our results confirm the multifactorial genesis of melanoma metastasis and identify the epithelial-mesenchymal transition (EMT) and the Introduction relation with the extracellular matrix as key steps in melanoma progression. In primary cutaneous malignant melanoma (CMM) patients, it is essential to determine the molecular changes associated with metastasis and to apply this knowledge to the fields of outcome prediction and targeted treatment. This information will lead to a Materials and Methods better understanding of the biology of this tumor, and will probably Patients and tissue samples. This study featured a series of 34 primary provide prognostic information for defining subgroups of patients CMMs provided by the Hospital San Cecilio (Granada, Spain) and the Hospital 12 de Octubre (Madrid, Spain). Patients were recruited between 1990 and 2002 to ensure a minimum follow-upof 36 months. Samples were collected and frozen according to standard protocols, and their histology Note: Supplementary data for this article are available at Cancer Research Online was reviewed by two pathologists (S.R.A. and J.L.R-P.) to confirm that they (http://cancerres.aacrjournals.org/). contained at least 50% tumor cells. All selected cases corresponded to Requests for reprints: Miguel A. Piris, Programa de Patologı´a Molecular, Centro consecutive melanoma tumors with potential metastatic capacity (i.e., Nacional de Investigaciones Oncolo´gicas, C/Melchor Ferna´ndez Almagro 3, Madrid vertical growth phase cases with a Breslow index >1 mm). The cutoff was 28029, Spain. Phone: 34-91-224-69-00; Fax: 34-91-224-69-23; E-mail: [email protected]. I2007 American Association for Cancer Research. selected taking into account the criteria of the American Joint Committee doi:10.1158/0008-5472.CAN-06-3481 on Cancer by which thin melanomas (V1 mm) have an excellent prognosis

Cancer Res 2007; 67: (7). April 1, 2007 3450 www.aacrjournals.org

Table 1. Antibodies used in the study indicating clone, source, dilution, visualization method, scoring, threshold, and positive controls

Protein Clone Source Dilution Visualization Scoring Threshold Positive control system

Glypican 3 Polyclonal Santa Cruz 1:25 LSAB/DAB Pos/neg >10% positive cells Hepatocarcinoma Biotechnology N-cadherin 3B9 Zymed 1:10 LSAB/DAB Pos/neg z5% positive cells, Ovarian carcinoma, membranous expression cardiac muscle SPARC/ 15G12 Novocastra 1:25 LSAB/DAB Pos/neg >10% positive cells, cytoplasm Endothelial cells in osteonectin malignant tumors Osteopontin Polyclonal Abcam 1:1,500 LSAB/DAB Pos/neg >10% positive cells, cytoplasm Stromal cells in normal skin PKCa H7 Santa Cruz 1:25 LSAB/DAB High/low z50% positive cells, cytoplasm Small lymphocytes, nevus Biotechnology

Abbreviations: Pos/neg, positive/negative; LSAB, peroxidase-labeled streptavidin biotin; DAB, diaminobenzidine.

(>90% survival at 5 years) compared with melanomas with a penetration online.8 Sample hybridizations were done as described elsewhere (12). depth of >1 mm (8). Histologic review of the cases was done on paraffin- After washing, the two fluorescent signals on the slides were scanned with embedded tissue, whereas frozen sections were examined for assessing the a standard two-color microarray scanner (Scanarray 5000XL, GSI proportion of the tumoral cells and the character of the neoplastic Lumonics, Kanata, Ontario, Canada). Images were analyzed with the infiltrate. GenePix 4.1 software (Axon Instruments, Inc., Union City, CA). The clone For tissue microarray (TMA) analysis, we used a retrospective cohort sequences of all the genes included in the OncoChipand the repro- of patients representing 127 primary vertical growth phase melanoma ducibility of the expression data of multiple genes have been previously cases (formalin-fixed and paraffin-embedded tissue) collected from 1980 verified (12). to 2000. These cases were obtained from the Hospital 12 de Octubre Data extraction and analysis. Data from each hybridization were (Madrid) and were included in six separate TMAs (1.5-mm core maintained in a database for analysis. Fluorescence intensity measurements diameter), with two representative duplicate cores for each case (9), were subjected to automatic background subtraction. The Cy3/Cy5 ratios and constructed with a manual tissue arrayer (Beecher Instruments, Sun were normalized to the value of the median ratio of all spots in the array. Prairie, WI) using a standard method (10). Patient medical records were The sum of the median background for each channel was calculated, reviewed to gather information on age, gender, localization, tumor and spots with total intensities less than the calculated sum of median thickness, distant invasion (lymph node or skin), and follow-up of at backgrounds were discarded. Additionally, spots with background- least 36 months. Patients were not treated before the development of subtracted signal intensities <500 fluorescence units (sum of the two metastasis. The work was conducted in accordance with the Declaration channels) and bad spots were excluded from the analysis. All ratio values of Helsinki Principles and under the supervision of the Hospital 12 de were log-transformed (base 2), and duplicated spots in the array were Octubre Ethics Committee. averaged. Inconsistent duplicates were discarded and all consistent RNA isolation and amplification. Total RNA was extracted from duplicate spots and genes were averaged. In addition, genes for which frozen samples of primary CMM using TRIzol reagent (Invitrogen, fewer than 70% of the potential data were available were excluded from Carlsbad, CA) followed by purification with the RNeasy Mini kit (Qiagen further analysis. The median expression of each gene was calculated for Inc., Valencia, CA) and digestion with RNase free DNase I according to each patient group (metastatic and nonmetastatic primary CMM). Differ- the manufacturer’s instructions. RNA quality and integrity were verified ences in expression between the two groups were analyzed by the Wilcoxon using the Bioanalyzer system (Agilent Technologies, Palo Alto, CA). test. To account for the effect of multiple hypothesis testing on the iden- Double-stranded cDNA was synthesized from 5 Ag of total RNA using tification of significance, adjusted P values were also computed using the the Superscript Choice System for cDNA synthesis (Invitrogen) with an method proposed by Hochberg and Benjamini (13) for controlling the false- oligo-dT primer containing a T7 RNA polymerase promoter. In vitro discovery rate (FDR). A single value for each gene was obtained. For our transcription was carried out using the T7 Megascript in vitro trans- purposes, a gene was deemed to be up-regulated or down-regulated if there cription kit (Ambion, Austin, TX) as previously described (11, 12). The was an at least 2-fold difference in expression and if the FDR value was <0.2. quality of the amplified RNA produced was checked by electrophoresis The regulated genes were functionally classified manually on the basis of and its concentration was measured. exhaustive searches in PubMed, the Genecards database, and Gene Microarray procedures: preparation/synthesis of fluorescent cDNA Ontology. and hybridization. For each melanoma sample, 5 Ag of amplified RNA TMA immunohistochemistry. We undertook immunohistochemical (aRNA) were directly labeled with cyanine 3-conjugated dUTP (Cy3), analysis to identify at the protein level a selection of the statistically and whereas 5 Ag of aRNA from the Universal Human Reference RNA biologically significant genes obtained in the cDNA microarray study. This (Stratagene, La Jolla, CA) were labeled with cyanine 5–conjugated dUTP was done in a TMA series of 127 vertical growth phase melanoma cases (Cy5). The Centro Nacional de Investigaciones Oncolo´gicas (CNIO) with known follow-up. The proteins selected for validation featured the OncoChip platform (v.1.4) was used in all cases to perform the cDNA groupthat provedto be biologically relevant in melanoma metastasis microarray procedure. Basically, the CNIO OncoChip is a cDNA microarray development, including glypican 3 (polyclonal, Santa Cruz Biotechnology, especially designed for the analysis of genes involved in cancer. It includes Santa Cruz, CA), N-cadherin (3B9, Zymed, San Francisco, CA), osteonectin/ 2,489 cancer-relevant genes in addition to genes involved in drug response, tissue-specific genes, and control genes. The platform used has a total of 6,386 genes represented by 7,237 human clones purchased from Research Genetics (Huntsville, AL). A list of these genes can be found 8 http://bioinfo.cnio.es/data/oncochip www.aacrjournals.org 3451 Cancer Res 2007; 67: (7). April 1, 2007

SPARC (15G12, Novocastra, Newcastle, United Kingdom), osteopontin results and selection of the thresholds, internal controls for the antibody (polyclonal, Abcam, Novus Biologicals, Littleton, CO), and protein kinase reactivity, and tissue controls for the series were done according previously Ca (PKCa; H7, Santa Cruz Biotechnology, Santa Cruz, CA). A heat-induced, published methods (9). epitope-retrieval step was done in a solution of sodium citrate buffer Scoring systems. Immunostaining results were evaluated by two (pH 6.5) in the case of osteonectin/SPARC, osteopontin, and PKCa, and with different pathologists (S.R.A. and J.L.R-P.) and scored using clear cutoff EDTA buffer in the case of glypican 3 and N-cadherin. The slides were then criteria to facilitate the reproducibility of the method. Discrepancies were heated for 2 min in a conventional pressure cooker and rinsed in cool resolved by simultaneous reevaluation. Briefly, the result was recorded as running water for 5 min. They were then quickly washed in TBS (pH 7.4), positive or negative and high versus low expression, taking into account and incubated with the selected primary antibodies glypican 3 (1:25), the expression in tumoral cells and the specific cutoff for each marker N-cadherin (1:10), osteonectin/SPARC (1:25), osteopontin (1:1500), and (see Table 1 for description of thresholds). As a general criterion, the cutoffs PKCa (1:25). Immunodetection was done with biotinylated secondary were selected to facilitate reproducibility and, when possible, to translate antibodies, followed by peroxidase-labeled streptavidin biotin (Dakocyto- biological events. Scoring for the sample replicates was highly reproducible mation, Glostrup, Denmark) visualization in the case of glypican, in this series (96%). osteopontin, and PKCa and Envision (Dakocytomation) for N-cadherin Statistical analysis. To validate protein expression in the series analyzed and SPARC. Diaminobenzidine chromogen was used as peroxidase by TMA, the relationship between marker expression in the patient samples substrate. All immunostaining was done in a TechMate 500 automatic and disease-free survival (DFS) curves were derived by the Kaplan- immunostaining device (DAKO, Glostrup, Denmark). Incubations omitting Meier method. Statistical significance of associations between individual the specific antibody were used as a control of the technique. Scoring of the variables and DFS was determined using the log-rank test. Cox univariate

Figure 1. Representative diagram of the study profile. Metastatic disease was considered as end point for DFS. VGP, vertical growth phase; CMM, cutaneous malignant melanoma; EMT, epithelial mesenchymal transition; FDR, false discovery rate; DFS, disease-free survival; TMA, tissue microarray. K-M, Kaplan-Meier.

Cancer Res 2007; 67: (7). April 1, 2007 3452 www.aacrjournals.org

Downloaded from cancerres.aacrjournals.org on October 1, 2021. © 2007 American Association for Cancer Research. Metastasis Signature in Melanoma proportional hazard analysis was also done independently for each variable. From the 5,253 clones, corresponding to 2,945 known genes Significance was concluded for values of P < 0.05. All statistical tests were and 1,984 expressed-sequence tags that were suitable for analysis two-sided. A multivariate model including the Breslow index as a after filtering steps, 243 genes were differentially expressed continuous variable was also developed by backward elimination. All (>2-fold, FDR<0.2) in primary vertical growth phase melanomas statistical analyses were carried out using the Stata statistical program with metastases compared with primary nonmetastatic vertical (StataCorp2001, release 9.0). growth phase melanomas. Of these, 206 genes were up-regulated and 37 genes were down-regulated (see Supplementary Table S1 for details). These genes were categorized according to their Results main biological function and/or mechanism of action. As seen in Clinical features. The study included two sets of patients. Table 2, six biological process classes (cell cycle and apoptosis, The training set was composed of 34 patients (14 males and 20 EMT, immune modulation, metabolism, nucleic acid binding and females) with primary vertical growth phase CMM and >1 mm transcription, protein synthesis and degradation, and signal (Breslow index). After a median follow-upof 67.3 months (range transduction) were identified as relevant in our metastatic gene 16–166 months), 21 developed metastasis (median 9.4 months, signature. A groupof genes with unknown function was also range 0.73–139). identified (see Supplementary Data for details). Figure 2 shows the The validation set of patients, analyzed by TMAs, included 127 expression profile of the significant genes in the training series patients (57 males and 70 females) with primary vertical growth classified by these gene-functional categories. phase CMM and a median age at diagnosis of 60 years (range 21–91 Interestingly, one of the most important functional clusters years). The median follow-upwas 116.8 months (range 1–276.5 recognized here as markers of metastatic melanoma included months), during which 73 patients developed metastatic disease genes involved in EMT, a category that includes genes associated and 54 did not (median 52.7 months, range 1–222). Figure 1 shows with cell adhesion, cell motility, migration, and extracellular matrix an outline of the study. interaction and organization (Table 3). This cDNA microarray Global changes in gene expression between metastatic and screening identified several genes involved in the EMT phenom- nonmetastatic primary vertical growth phase CMM cases. To enon (N-cadherin, SPARC, and WNT receptor frizzled) and the identify genes that might be potential markers for melanoma physical link between the actin cytoskeleton, the extracellular metastases, cDNA microarray studies were done to compare matrix (mainly mediated by receptors of the integrin family, such primary invasive CMM cases with and without metastatic disease. as integrin aV), and supporting vessels (endothelin receptor B).

Table 2. Functional categories of differentially expressed genes

Functional categories Gene name

Cell cycle and apoptosis MAPRE1, SKP1A, FAIM, PMP22, SET, CDKN1A, CDC5L, WEE1, CASP5, DAD1, APEX1, VDAC1, BCCIP, CKS1B, CYC1, HBXIP, TP53I3, KIF11, BCL2L10, TOB2 EMT LUM, SDCBP, MFAP1, CTSB, PRKCA, RAB1A, H2-ALPHA, RRAGA, TUBB3, CLIC4, SPARC, ANLN, RAN, ENC1, EMP1, DSG2, SEPP1, ITGAV, TUBA2, HMMR, MASA, TUBA3, CX3CR1, CDH2, EDNRB, CSPG2, CD63, SMARCA1, KIT, KRT8, ANXA5, FGG, PFN1, SPA17, FZD1, APLP2, SERPINA3, SEMA3B, MMP2, TUBA1, EDN2, PCDH9, CDH10, WNT2, PAPPA Immune modulation CXCL12, MAGEA6, IFITM2, CCBP2 Metabolism PRDX5,VKORC1, ACADM, ALDH1A2, GLO1, HIBADH, PMM1, NDUFAB1, NDUFB11, ENO1, MDH1, ACSL3, ATP6V0B, MGC4172, OSBP, PFKM, SLC39A1, ATP6V0D1, YWHAQ, LRPAP1, ADK, SLC30A6, GGH, GSTA4, DCK, ATP5L2, LIPA, COX8A, CA9, AKR7A2, SCD, AK3L1, TXNRD1, SLC25A17, NPC2, NDUFB3, PAICS, COX6B1, ADSL, CYP7B1, CYP19A1, IDS Nucleic acid binding and transcription factors NCOA4, HIST1H4C, RBM4, MAP4K3, NOLA2, BZW1, ARID5B, CBFB, POLR3K, TRIP3, RDBP, ASH2L, CDYL, XRCC5, HIST1H4B, KLF10, TCEB1, HIST1H3D, GTF2H2, FOXO1A, REST, OLIG2 Protein synthesis and degradation HSPA8, MRPL15, CCT5, EIF5, UBE2D2, MRPL37, MRPS10, RPL10A, VBP1, RPS22, UBL5, CCT6A, GC20, EIF3S3, RPL35A, SECISBP2 Signal transduction CAV1, NME2, SRI, PPP1R3C, S100A10, WBP5, GABARAP, IGFBP1, PPP1R8, C11orf15, PRKAR1A, PGR1_GPR153, SEC5L1, CGI-141, MYB, SPRY2, PPP2CB, STK24, GNG12, GRB2, MAL, SHB, DUSP12, TGFBR2, PC4, IDE, CGA, IL1RL1, PPP2R5C, TIMM10, PDE5A, RABAC1, GIMAP6, TXK, EPS15, NTF5, AKAP13

NOTE: Categories are based on Gene Cards, National Center for Biotechnology Information, HUGO gene nomenclature committee, and PubMed. EMT is a category that includes genes associated with cell adhesion, cell motility, migration, and extracellular matrix organization.

www.aacrjournals.org 3453 Cancer Res 2007; 67: (7). April 1, 2007

Figure 2. Results of global expression profiling between primary CMM with metastasis at follow-up (red squares) and primary CMM without metastasis (yellow squares). Expression data for each gene in each profiled sample are presented in the form of a heat map (green to red scale) of log2-transformed ratios. Names on the right represent the main functional category for each gene. Up-regulated genes are at the top and down-regulated genes are at the bottom.

Cancer Res 2007; 67: (7). April 1, 2007 3454 www.aacrjournals.org

Table 3. EMT-related genes

Gene name Gene description n-fold change Log2 change

LUM Lumican 12.278 3.618 SDCBP Syndecan binding protein (syntenin, melanoma differentiation-associated 6.753 2.756 protein 9-MDA; pro-TGF-a cytoplasmic domain-interacting protein 18; scaffold protein Pbp1) MFAP1 Microfibrillar-associated protein 1 4.147 2.052 CTSB Cathepsin B 3.487 1.802 PRKCA Protein kinase Ca 3.207 1.681 RAB1A RAB1A member RAS oncogene family 2.986 1.578 H2-ALPHA a-Tubulin isotype H2-a 2.934 1.553 RRAGA Ras-related GTP binding A 2.917 1.545 TUBB3 Tubulin h3 (other designation: tubulin h4) 2.874 1.523 CLIC4 Chloride intracellular channel 4 2.870 1.521 SPARC Secreted protein, acidic, cysteine-rich (osteonectin) 2.840 1.506 ANLN Anillin, actin binding protein 2.764 1.467 RAN RAN, member RAS oncogene family 2.745 1.457 ENC1 Ectodermal-neural cortex (with BTB-like domain). Other designations: 2.660 1.412 nuclear restricted protein, BTB domain-like (brain); tumor protein p53 inducible protein 10 EMP1 Epithelial membrane protein 1 2.645 1.403 DSG2 Desmoglein 2 2.629 1.395 SEPP1 Selenoprotein P, plasma, 1 2.628 1.394 ITGAV Integrin aV (vitronectin receptor, a polypeptide, antigen CD51) antigen 2.563 1.358 identified by monoclonal antibody L230 TUBA2 Tubulin a2 2.457 1.297 HMMR Hyaluronan-mediated motility receptor (RHAMM) intracellular hyaluronic 2.439 1.286 acid binding protein MASA L1 neuronal cell adhesion molecule, MASA (mental retardation, aphasia, 2.425 1.278 shuffling gait and adducted thumbs) TUBA3 Tubulin a3 2.400 1.263 CX3CR1 Chemokine (C-X3-C motif) receptor 1 2.380 1.251 CDH2 Cadherin 2, type 1, N-cadherin [neuronal N-cadherin 1; cadherin 2, N-cadherin 2.375 1.248 (neuronal); cadherin 2, type 1; calcium-dependent adhesion protein, neural cadherin] EDNRB Endothelin receptor type B 2.315 1.211 CSPG2 Versican (chondroitin sulfate proteoglycan 2) 2.261 1.177 CD63 Melanoma 1 antigen 2.205 1.141 SMARCA1 SWI/SNF-related matrix-associated, actin-dependent regulator of chromatin, 2.199 1.137 subfamily a, member 1 FGG Fibrinogen g chain 2.149 1.104 PFN1 Profilin 1 2.142 1.099 SPA17 Sperm autoantigenic protein 17 2.129 1.090 FZD1 Frizzled homologue 1 (Drosophila) 2.121 1.085 APLP2 Amyloid h (A4) precursor-like protein 2 2.098 1.069 SERPINA3 Serine (or cysteine) proteinase inhibitor, clade A (a-1 antiproteinase, antitrypsin), member 3 2.076 1.054 MMP2 Matrix metallopeptidase 2 (gelatinase A, 72 kDa gelatinase, 72 kDa type IV collagenase) 2.046 1.033 TUBA1 Tubulin a 1 (testis-specific) 2.018 1.013 EDN2 Endothelin 2 À2.279 À1.189 PCDH9 Protocadherin 9 À2.345 À1.230 CDH10 Cadherin 10 type 2 (T2-cadherin) À2.397 À1.262 WNT2 Wingless-type MMTV integration site family member 2 À3.274 À1.711 PAPPA Pregnancy-associated plasma protein A, pappalysin 1 À3.323 À1.733

Interestingly, the increase in the expression of N-cadherin is identified in the nonmetastatic group, in spite of not exhibiting associated with the loss of the type II cadherin 10 (CDH10). differences in the Clark level or presence of ulceration, when A substantial number of the genes in the metastatic signature compared with the rest of nonmetastatic cases. These two cases could be included in a generic category named melanoma- and had in common the expression of the transforming growth factor neural-related genes, which was created consistently with the h receptor II (TGFBR2), a signal transduction protein whose neural origin of the melanocytes (Table 4). mutations have been associated with tumor progression in some This series included two primary cases with metastatic disease solid tumors (14). Moreover, case mel-08 had altered cell cycle and samples whose expression profiles somewhat resembled those apoptosis regulator genes with a demonstrated role in melanoma www.aacrjournals.org 3455 Cancer Res 2007; 67: (7). April 1, 2007

Downloaded from cancerres.aacrjournals.org on October 1, 2021. © 2007 American Association for Cancer Research. Cancer Research progression, such as overexpression of p21 (CDKN1A; ref. 9) or Immunostaining for these proteins confirmed that they were down-regulation of BCL2L10. The mel-09 case showed up- expressed by tumoral melanoma cells. The adhesion molecule N- regulation of ENC1 (ectodermal neural cortex 1), a gene associated cadherin was found with a membrane and cytoplasmic pattern. related with the Wingless (WNT) pathway and whose expression Osteonectin (SPARC) was recognized in tumoral melanocytes and has been linked with colorectal carcinogenesis (15), and meningi- in endothelial vessels close to the tumor, as described in other oma progression (16). neoplasms (21, 22). Osteopontin is an acidic matrix protein (bone Validation of gene expression results by immunohistochem- syaloprotein) commonly localized within normal elastic fibers of istry on TMAs. To corroborate the gene expression data and to the skin and secreted by many transformed cells. Immunolocaliza- identify in situ the cells expressing the various markers, tion of this protein in our melanoma cases was found in tumoral immunohistochemical studies of an independent series of primary cells and also in the stroma (see Fig. 3B). As shown in the table in vertical growth phase CMMs were done using TMAs. The relevance Fig. 3A, the expression of these three proteins was significantly and intensity in the changes of expression in genes linked with associated with an increased incidence of melanoma metastases EMT in the training series led us to concentrate on validating some [P = 0.013 for N-cadherin; P = 0.006 for osteonectin (SPARC); significantly up-regulated proteins (>2-fold change and FDR<0.2) in P = 0.05 for osteopontin], thereby confirming its potential this category (N-cadherin, osteonectin/SPARC, and PKCa). Promp- prognostic value in CMM. ted by a literature search and the nearly significant results reported PKCa, a protein involved in melanoma progression (23), was also here, we also studied glypican 3, which has recently been described expressed in the cytoplasm of tumoral cells, as previously described as a novel tumor marker for melanoma (17), and osteopontin,a (24), but the univariate study failed to show any significant protein belonging to this EMT group that has a crucial role in association with tumor progression in this series. However, the melanoma (18, 19) and seems to be diagnostically useful in the multivariate analysis including N-cadherin, osteopontin, osteonec- detection of mesothelioma in serum obtained from peripheral tin, PKCa, and Breslow index showed that only PKCa was blood (20). independent of the Breslow index (Table 5). Figure 3 shows the results of the univariate analyses of selected proteins and representative immunohistochemistry for N-cadherin, osteonectin/SPARC, and osteopontin, which were the statistically Discussion significant markers. These results are closely consistent with the The findings reported here confirm microarray data previously cDNA microarray data. obtained in cell lines (e.g., lumican; ref. 3) or in other experimental

Table 4. Melanoma and neural-related genes

Gene name Gene description n-fold change Log2 change

SDCBP Syndecan binding protein (syntenin, melanoma differentiation-associated 2.7555 6.753 protein 9-MDA; pro-TGF-a cytoplasmic domain-interacting protein 18; scaffold protein Pbp1) CHN1 Chimerin 1 (GTPase-activating protein, U, 2); n-chimerin 1.523 2.874 PMP22 Peripheral Myelin Protein 22 (growth arrest-specific 3) 1.505 2.838 S100A10 S100 calcium binding protein A10 [Annexin II ligand, calpactin I, light 1.491 2.811 polypeptide (p11)] ENC1 Ectodermal-neural cortex (with BTB-like domain). Other designations: 1.4115 2.66 nuclear restricted protein, BTB domain-like (brain); tumor protein p53 inducible protein 10 EMP1 Epithelial membrane protein 1 1.403 2.645 SEPP1 Selenoprotein P, plasma 1 1.394 2.628 MASA L1 neuronal cell adhesion molecule, MASA (mental retardation, aphasia, 1.278 2.425 shuffling gait and adducted thumbs) PTPRZ1 Protein tyrosine phosphatase, receptor-type, Z polypeptide 1 1.255 2.387 MITF Microphthalmia-associated transcription factor 1.2505 2.379 MAGEA6 Melanoma antigen family A, 6 1.2075 2.309 AKT1S1 AKT1 substrate 1 (proline-rich) 1.1705 2.251 CD63 Melanoma 1 antigen 1.141 2.205 MAL Mal, T-cell differentiation protein 1.138 2.201 KIT V-kit Hardy-Zuckerman 4 feline sarcoma viral oncogene homologue, 1.123 2.178 C-Kit, CD117, PBT, SCFR APLP2 Amyloid h (A4) precursor-like protein 2 1.069 2.098 PDGFRA Platelet-derived growth factor receptor, a polypeptide 1.055 2.078 ALS4 Amyotrophic lateral sclerosis 4 1.053 2.075 PCDH9 Protocadherin 9 À1.2295 À2.345 CDH10 Cadherin 10 type 2 (T2-cadherin) À1.2615 À2.397 NTF5 Neurotrophin 5 (neurotrophin 4/5) À1.791 À3.461 OLIG2 Oligodendrocyte lineage transcription factor 2 À1.9555 À3.879

Cancer Res 2007; 67: (7). April 1, 2007 3456 www.aacrjournals.org

Figure 3. A, Kaplan-Meier analysis of DFS among the validation set of 127 vertical growth phase melanoma cases included in the TMA. The genes evaluated were N-cadherin, osteonectin (SPARC), osteopontin, glypican 3, and PKCa. Kaplan-Meier plots show the curves for significant proteins. 95% CI, confidence interval. B, representative immunohistochemical staining of significantly up-regulated EMT candidate genes identified from the microarray data. Brown staining denotes a positive signal. Magnification, Â4 (inset Â20). assays (e.g., osteopontin; ref. 19), ENDRB (25), PKCa (26), and seem this functional cluster. EMT is the process by which an epithelial to imply that the metastatic melanoma signature in human cell suffers transitory changes in cell structure and becomes a more samples is recognized by the increased expression of clusters of motile mesenchymal cell with migratory and invasive properties genes involved in the control of EMT, cell cycle and apoptosis, (27). As a result, cell-cell junctions are altered; cells lose polarity, immune response, metabolism, transcription regulation, protein express mesenchymal markers, and the actin cytoskeleton becomes synthesis and degradation, and signal transduction. reorganized. Consequently, tumoral cells lose contact with In accordance with the results obtained, we have focused our neighboring cells, become motile, and interact with extracellular attention on the validation stepin the EMT because a large and matrix, invading surrounding territories and acquiring capacity for highly significant groupof up-regulatedgenes were members of metastasis. The relevance of the EMT signature to the origin of www.aacrjournals.org 3457 Cancer Res 2007; 67: (7). April 1, 2007

Downloaded from cancerres.aacrjournals.org on October 1, 2021. © 2007 American Association for Cancer Research. Cancer Research metastasis is not restricted to melanoma, and has been identified transfected with N-cadherin are morphologically transformed from as a determinant of local invasion and metastasis in other tumor an epithelial-like shape to a fibroblast-like shape; by contrast, types, such as breast (28), gastrointestinal, or prostate carcinomas adenoviral reexpression of E-cadherin in melanoma cells down- (29), among others. regulates endogenous N-cadherin and reduces their malignant The present results suggest that migration and invasion in potential (37). The present study not only confirms the role of melanoma are the result of a specific interaction between tumoral N-cadherin in melanoma progression but also shows for the first and stromal cells (30, 31) associated with the expression of a set time how gain of N-cadherin and loss of class II type cadherin can of molecules involved in EMT (31). Thus, the study reveals facilitate metastatic dissemination in human primary CMM N-cadherin, lumican, glypican, osteonectin (SPARC), osteopontin, samples. metalloproteinases, and integrins, among others, as being involved Osteonectin (SPARC or BM40) is a secreted extracellular matrix in the melanoma metastatic process. Although it is difficult to glycoprotein involved in tumor cell migration, invasion, and verify EMT experimentally in vivo due to the reversible nature of angiogenesis (38). Data presented here are consistent with previous the process, expression profiling identifies a molecular signature reports of SPARC overexpression as a marker of aggressiveness or linked with this phenomenon. Moreover, the immunohistochemical local invasion in CMM (39, 40). studies of N-cadherin, osteopontin, glypican 3, osteonectin Osteopontin is a secreted phosphoprotein with a crucial role (SPARC), and PKCa validate the data presented and confirm that in tumor progression and metastasis in many kinds of tumors, at least these components of the EMT signature are expressed by including melanoma (19). It promotes antiapoptotic signaling the melanoma cells themselves. and angiogenesis and induces matrix metalloproteinases (MMP; One of most remarkable findings of this study is the presence ref. 18). Osteopontin expression may be acquired in the early stages of what has been called the switch of the cadherin class: loss of of melanoma invasion, with moderate to high levels in primary epithelial cadherins (E-cadherin) with gain of neural cadherins invasive and metastatic melanoma samples and null or low levels (N-cadherin). Cadherins are calcium-dependent, cell-adhesion of expression in benign nevi, dysplastic nevi, or in situ melanoma molecules that are critical for the development and maintenance (19). Here, cytoplasmic osteopontin expression in invasive mela- of epithelial architecture. As described in carcinoma models, nomas was associated with metastatic disease, mimicking previous N-cadherin is a crucial molecule in the EMT event, acting as an experimental models in which osteopontin allowed murine oncogene in many tumors by promoting tumor invasiveness and melanocytes to adhere, spread, and survive in three-dimensional progression (32). In our study, increased N-cadherin was matched collagen gels (41). by loss of cadherin 10 expression, a type II cadherin with near- The serine/threonine kinase PKCa is a protein with a well- equivalent binding strength to that of E-cadherin (33) as described recognized role in regulating cell growth and progression (42). previously in cell line models (34, 35) and in desmoplastic Melanocytes and melanoma cells express several PKC isoforms, melanoma (36). Although melanoma cells are not epithelial in with proposed roles in melanoma genesis, invasion, and metastasis nature, the EMT for this tumor is well recognized and the relevance through the WNT signaling pathway or by regulating integrin of the cadherin switch has been previously supported by molecule expression of integrin. The specific role of PKCa in CMM experimental approaches, demonstrating that melanoma cell lines has been a controversial issue, with some results suggesting that

Table 5. Kaplan-Meier, log-rank test, and Cox regression univariate and multivariate analyses of DFS

Univariate and Breslow-adjusted analyses

Marker No. cases Cases with metastatic disease Hazard ratio (95% CI) P Adjusted by Breslow

Hazard ratio (95% CI) P

N-cadherin 108 60 1.95 (1.15–3.31) 0.013 1.49 (0.86–2.58) 0.155 Osteonectin 112 66 1.99 (1.21–3.25) 0.006 1.48 (0.89–2.46) 0.126 Osteopontin 102 61 1.88 (1.00–3.55) 0.05 1.72 (0.91–3.23) 0.095 Glypican 3 107 64 1.51 (0.80–2.83) 0.199 1.19 (0.62–2.26) 0.603 PKCa 119 69 1.21 (0.75–1.95) 0.429 2.03 (1.19–3.44) 0.009 Multivariate analysis

Marker Hazard ratio SE zP> z 95% CI

PKCa 2.313 0.7188 2.89 0.004 1.3092–4.0870 Breslow 1.415 0.0655 7.50 0.000 1.2925–1.5497

NOTE: Data were measured from the time of primary vertical growth phase CMM to clinical or histopathologic metastatic disease. The P values are shown for the univariate analysis of the validation TMA set before and after introducing the Breslow index (bivariate Cox model). Significant markers (P < 0.05) are shown in bold. Abbreviation: 95% CI, 95% confidence interval.

Cancer Res 2007; 67: (7). April 1, 2007 3458 www.aacrjournals.org

PKCa and PKCy isoforms suppress cell growth, whereas others each functional cluster of genes, as identified here, requires further imply its involvement in cell motility (42, 43) or in metastasis study, both in human samples and experimental models. promotion (44). Expression profiling data in this series identifies Interestingly, the multivariate study showed that three of the PKCa as one of the genes up-regulated in melanoma with protein markers analyzed in TMA were associated with local metastasis, and the multivariate analysis showed that it is Breslow invasion and metastasis, thereby linking the two phenomena. independent, implying that there is a relationship between tumor Overall, our results lead us to propose that metastasis in melanoma growth and proliferation. This awaits further investigation. is determined by the interaction of sets of molecules involved in Lumican, the most highly up-regulated gene in melanoma cases the regulation of EMT immune response, cell metabolism, nucleic with metastasis, is a proteoglycan of the extracellular matrix acid binding and transcription, protein synthesis and degradation, involved in collagen-fiber organization, epithelial-cell migration, and signal transduction. Moreover, validation of protein expression and tissue repair. Lumican is an established EMT marker (45) in TMA confirms how some of the up-regulated genes are whose protein expression level is associated with tumorigenesis predictors of metastasis development. Taking all our results into and progression in a variety of tumors (46, 47). Data presented here consideration, it seems that the control of the EMT plays a pivotal are also consistent with those obtained from the comparison of role in the metastatic process in melanoma. This confirms and normal melanocytes with melanoma cell strains from advanced builds on findings linking EMT genes with melanoma progression lesions (3). obtained in experimental models and human samples studies, such Cell-matrix interactions are closely related to EMT category. The as cadherins, MMPs, integrins, lumican, osteopontin, EDNRB, Snail, integrins are a family of cell surface adhesion molecules that and others (6, 49). coordinate cell-cell and cell-matrix interactions. This study identified up-regulation of integrin aV(ITGAV) as a marker of melanoma metastases, confirming a previous report (48). Closely Acknowledgments related are the MMP family of proteins, which are involved in the Received 9/19/2006; revised 1/17/2007; accepted 1/31/2007. Grant support: FIS grant PI 040641, MMA 2005-085, and European Commission breakdown of the extracellular matrix in normal physiologic (Translational and Functional Onco-Genomics) grant LSHC-CT-2004-503438. S.R. processes and in cancer invasion and metastasis. Here, MMP2 Alonso was supported in part by FIS grant CM03/00034 and by the CNIO (Madrid, (which degrades type IV collagen) and ADAM9 (disintegrin and Spain). L. Tracey received support though grants from the CNIO and the Higher Education Authority of Ireland through the Department of Haematology/Institute of metalloproteinase domain 9) are associated with melanoma Molecular Medicine, St. James Hospital, Dublin, Ireland. progression, confirming previous findings regarding invasion (49). The costs of publication of this article were defrayed in part by the payment of page Two of the metastatic cases analyzed here (mel-08 and mel-09) charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. displayed a nonmetastatic phenotype, with the exception of the We thank the Tumour Bank of the Hospital San Cecilio; the Spanish National increased expression of TGFBR2, p21 (CDKN1A), or ENC1. It is Tumour Bank Network, CNIO; Laura Cereceda (CNIO) and Alicia Maroto (CNIO, known that melanoma and endothelial cells express type 2 TGF-h Hospital 12 de Octubre) for their valuable help in providing specific tissue samples for this study; Pilar Sandoval (CNIO, Hospital Gregorio Maranõń) for her time with receptor (50), which can enhance adhesion of melanoma cells to gridding; Dr. David Hardisson and Raquel Marcos (CNIO, Hospital Universitario La endothelium and favor invasion through activation of the TGF-h Paz) for helping with the SPARC immunostaining; Ramoń Diaz for helping with the Wilcoxon test; Amancio Carnero for their valuable guidance with gene classification; pathway. In the same way, p21 overexpression has been proposed and Phil Mason for his helpwith the English edition of the final version of the as a marker of melanoma progression (9). The specific relevance of manuscript.

References 9. Alonso SR, Ortiz P, Pollan M, et al. Progression in 17. Nakatsura T, Kageshita T, Ito S, et al. Identification of cutaneous malignant melanoma is associated with glypican-3 as a novel tumor marker for melanoma. Clin 1. Zhang L, Zhou W, Velculescu VE, et al. Gene distinct expression profiles: a tissue microarray-based Cancer Res 2004;10:6612–21. expression profiles in normal and cancer cells. Science study. Am J Pathol 2004;164:193–203. 18. Samanna V, Wei H, Ego-Osuala D, et al. a-V- 1997;276:1268–72. 10. Rimm DL, CampRL, Charette LA, et al. Tissue dependent outside-in signaling is required for the 2. Bittner M, Meltzer P, Chen Y, et al. Molecular microarray: a new technology for amplification of tissue regulation of CD44 surface expression, MMP-2 secre- classification of cutaneous malignant melanoma by resources. Cancer J 2001;7:24–31. tion, and cell migration by osteopontin in human gene expression profiling. Nature 2000;406:536–40. 11. Eberwine J. Amplification of mRNA populations melanoma cells. ExpCell Res 2006;312:2214–30. 3. Hoek K, Rimm DL, Williams KR, et al. Expression using aRNA generated from immobilized oligo(dT)-T7 19. Zhou Y, Dai DL, Martinka M, et al. Osteopontin profiling reveals novel pathways in the transformation primed cDNA. Biotechniques 1996;20:584–91. expression correlates with melanoma invasion. J Invest of melanocytes to melanomas. Cancer Res 2004;64: 12. Tracey L, Villuendas R, Ortiz P, et al. Identification of Dermatol 2005;124:1044–52. 5270–82. genes involved in resistance to interferon-a in cutane- 20. Pass HI, Lott D, Lonardo F, et al. Asbestos exposure, 4. McGill GG, Horstmann M, Widlund HR, et al. Bcl2 ous T-cell lymphoma. Am J Pathol 2002;161:1825–37. pleural mesothelioma, and serum osteopontin levels. regulation by the melanocyte master regulator Mitf 13. Benjamini Y, Yekutieli D. Quantitative trait Loci N Engl J Med 2005;353:1564–73. modulates lineage survival and melanoma cell viability. analysis using the false discovery rate. Genetics 2005; 21. Chen X, Higgins J, Cheung ST, et al. Novel endothelial Cell 2002;109:707–18. 171:783–90. cell markers in hepatocellular carcinoma. Mod Pathol 5. Haqq C, Nosrati M, Sudilovsky D, et al. The gene 14. Fernandez-Peralta AM, Nejda N, Oliart S, et al. 2004;17:1198–210. expression signatures of melanoma progression. Proc Significance of mutations in TGFBR2 and BAX in 22. Kato Y, Lewalle JM, Baba Y, et al. Induction of SPARC Natl Acad Sci U S A 2005;102:6092–7. neoplastic progression and patient outcome in sporadic by VEGF in human vascular endothelial cells. Biochem 6. Winnepenninckx V, Lazar V, Michiels S, et al. Gene colorectal tumors with high-frequency microsatellite Biophys Res Commun 2001;287:422–6. expression profiling of primary cutaneous melanoma instability. Cancer Genet Cytogenet 2005;157:18–24. 23. Gruber JR, Desai S, Blusztajn JK, et al. Retinoic acid and clinical outcome. J Natl Cancer Inst 2006;98: 15. Fujita M, Furukawa Y, Tsunoda T, et al. Up-regulation specifically increases nuclear PKC a and stimulates 472–82. of the ectodermal-neural cortex 1 (ENC1) gene, a AP-1 transcriptional activity in B16 mouse melanoma 7. Talantov D, Mazumder A, Yu JX, et al. Novel genes downstream target of the h-catenin/T-cell factor cells. ExpCell Res 1995;221:377–84. associated with malignant melanoma but not benign complex, in colorectal carcinomas. Cancer Res 2001;61: 24. Timar J, Liu B, Bazaz R, et al. Association of protein melanocytic lesions. Clin Cancer Res 2005;11:7234–42. 7722–6. kinase-C-a with cytoplasmic vesicles in melanoma cells. 8. Retsas S, Henry K, Mohammed MQ, et al. Prognostic 16. Wrobel G, Roerig P, Kokocinski F, et al. Microarray- J Histochem Cytochem 1996;44:177–82. factors of cutaneous melanoma and a new staging based gene expression profiling of benign, atypical and 25. Lahav R, Endothelin receptor B is required for the system proposed by the American Joint Committee on anaplastic meningiomas identifies novel genes associ- expansion of melanocyte precursors and malignant Cancer (AJCC): validation in a cohort of 1284 patients. ated with meningioma progression. Int J Cancer 2005; melanoma. Int J Dev Biol 2005;49:173–80. Eur J Cancer 2002;38:511–6. 114:249–56. 26. Lahn MM, Sundell KL. The role of protein kinase www.aacrjournals.org 3459 Cancer Res 2007; 67: (7). April 1, 2007

C-a (PKC-a) in melanoma. Melanoma Res 2004;14: sion of invasion-related adhesion receptors. Am J Pathol 42. Oka M, Kikkawa U. Protein kinase C in melanoma. 85–9. 2000;156:1515–25. Cancer Metastasis Rev 2005;24:287–300. 27. Grunert S, Jechlinger M, Beug H. Diverse cellular and 35. Hsu MY, Wheelock MJ, Johnson KR, et al. Shifts in 43. Szalay J, Bruno P, Bhati R, et al. Associations of PKC molecular mechanisms contribute to epithelial plasticity cadherin profiles between human normal melanocytes isoforms with the cytoskeleton of B16F10 melanoma and metastasis. Nat Rev Mol Cell Biol 2003;4:657–65. and melanomas. J Investig Dermatol SympProc 1996;1: cells. J Histochem Cytochem 2001;49:49–66. 28. Blanco MJ, Moreno-Bueno G, Sarrio D, et al. 188–94. 44. Dennis JU, Dean NM, Bennett CF, et al. Human Correlation of Snail expression with histological grade 36. Gallego Attis M, Burchette JL, Angelica Selim M, et al. melanoma metastasis is inhibited following ex vivo and lymph node status in breast carcinomas. Oncogene Differential expression of N-cadherin distinguishes treatment with an antisense oligonucleotide to protein 2002;21:3241–6. aggressive spindle cell melanomas from desmoplastic kinase C-a. Cancer Lett 1998;128:65–70. 29. Tomita K, van Bokhoven A, van Leenders GJ, et al. melanomas. Am J Dermatopathol 2006;28:232. 45. Saika S, Kono-Saika S, Ohnishi Y, et al. Smad3 Cadherin switching in human prostate cancer progres- 37. Krengel S, Groteluschen F, Bartsch S, et al. Cadherin signaling is required for epithelial-mesenchymal transi- sion. Cancer Res 2000;60:3650–4. expression pattern in melanocytic tumors more likely tion of lens epithelium after injury. Am J Pathol 2004; 30. Hendrix MJ, Seftor EA, Kirschmann DA, et al. depends on the melanocyte environment than on tumor 164:651–63. Remodeling of the microenvironment by aggressive cell progression. J Cutan Pathol 2004;31:1–7. 46. Ping Lu Y, Ishiwata T, Asano G. Lumican expression melanoma tumor cells. Ann N Y Acad Sci 2003;995: 38. Jendraschak E, Sage EH. Regulation of angiogenesis in a cells of islets in pancreas and pancreatic cancer 151–61. by SPARC and angiostatin: implications for tumor cell cells. J Pathol 2002;196:324–30. 31. Johnson JP. Cell adhesion molecules in the develop- biology. Semin Cancer Biol 1996;7:139–46. 47. Naito Z, Ishiwata T, Kurban G, et al. Expression and ment and progression of malignant melanoma. Cancer 39. Rumpler G, Becker B, Hafner C, et al. Identification of accumulation of lumican protein in uterine cervical Metastasis Rev 1999;18:345–57. differentially expressed genes in models of melanoma cancer cells at the periphery of cancer nests. Int J Oncol 32. Hazan RB, Phillips GR, Qiao RF, et al. Exogenous progression by cDNA array analysis: SPARC, MIF and a 2002;20:943–8. expression of N-cadherin in breast cancer cells induces novel cathepsin protease characterize aggressive phe- 48. Kageshita T, Hamby CV, Hirai S, et al. a(v)h3 expres- cell migration, invasion, and metastasis. J Cell Biol 2000; notypes. Exp Dermatol 2003;12:761–71. sion on blood vessels and melanoma cells in primary 148:779–90. 40. Ledda F, Bravo AI, Adris S, et al. The expression lesions: differential association with tumor progression 33. Shimoyama Y, Tsujimoto G, Kitajima M, et al. of the secreted protein acidic and rich in cysteine and clinical prognosis. Cancer Immunol Immunother Identification of three human type-II classic cadherins (SPARC) is associated with the neoplastic progres- 2000;49:314–8. and frequent heterophilic interactions between different sion of human melanoma. J Invest Dermatol 1997; 49. Redondo P, Lloret P, Idoate M, et al. Expression and subclasses of type-II classic cadherins. Biochem J 2000; 108:210–4. serum levels of MMP-2 and MMP-9 during human 349:159–67. 41. Geissinger E, Weisser C, Fischer P, et al. Autocrine melanoma progression. Clin Exp Dermatol 2005;30:541–5. 34. Hsu MY, Meier FE, Nesbit M, et al. E-cadherin stimulation by osteopontin contributes to antiapoptotic 50. Teti A, De Giorgi A, Spinella MT, et al. Transforming expression in melanoma cells restores keratinocyte- signalling of melanocytes in dermal collagen. Cancer growth factor-h enhances adhesion of melanoma cells to mediated growth control and down-regulates expres- Res 2002;62:4820–8. the endothelium in vitro. Int J Cancer 1997;72:1013–20.

Cancer Res 2007; 67: (7). April 1, 2007 3460 www.aacrjournals.org

Downloaded from cancerres.aacrjournals.org on October 1, 2021. © 2007 American Association for Cancer Research. A High-Throughput Study in Melanoma Identifies Epithelial-Mesenchymal Transition as a Major Determinant of Metastasis

Soledad R. Alonso, Lorraine Tracey, Pablo Ortiz, et al.

Cancer Res 2007;67:3450-3460.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/67/7/3450

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2009/03/27/67.7.3450.DC1

Cited articles This article cites 50 articles, 10 of which you can access for free at: http://cancerres.aacrjournals.org/content/67/7/3450.full#ref-list-1

Citing articles This article has been cited by 35 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/67/7/3450.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/67/7/3450. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.