Global Gene Expression Profiling Unveils S100A8/A9 As Candidate Markers in H-Ras-Mediated Human Breast Epithelial Cell Invasion

Global Gene Expression Profiling Unveils S100A8/A9 as Candidate Markers in H-Ras-Mediated Human Breast Epithelial Cell Invasion

Aree Moon,1 Hae-Young Yong,1 Jae-In Song,1 Daniela Cukovic,2 Sridevi Salagrama,2 David Kaplan,3 David Putt,3 Hyesook Kim,2,3 Alan Dombkowski,2 and Hyeong-Reh Choi Kim4

1College of Pharmacy, Duksung Women’s University, Seoul, Korea; 2Institute of Environmental Health Sciences, 3Detroit R&D, Inc., and 4Department of Pathology, Karmanos Cancer Institute, Detroit, Michigan

Abstract potential of breast epithelial cells. Our gene profile The goal of the present study is to unveil the gene data provide useful information which may lead to the expression profile specific to the biological processes of identification of additional potential targets for the human breast epithelial cell invasion and migration prognosis and/or therapy of metastatic breast cancer. using an MCF10A model genetically engineered to (Mol Cancer Res 2008;6(10):1544–53) constitutively activate the H-ras or N-ras signaling pathway. We previously showed that H-Ras, but not N-Ras, induces MCF10A cell invasion/migration, Introduction whereas both H-Ras and N-Ras induce cell proliferation The major cause of death from breast cancer is the metastatic and phenotypic transformation. Thus, these cell lines spread of the disease from the primary tumor to distant sites in provide an experimental system to separate the gene the body, yet predictive means of metastatic potential or expression profile associated with cell invasion apart strategy for therapeutic intervention are poorly established at from cell proliferation/transformation. Analysis of whole present. Recent studies of the gene expression profiles between human genome microarray revealed that 412 genes were human primary tumors and metastases provided molecular differentially expressed among MCF10A, N-Ras MCF10A, insight into potential biomarkers that are indicative of and H-Ras MCF10A cells and hierarchical clustering invasiveness of the disease (1-5). However, interpretation of separated 412 genes into four clusters. We then tested comparative analyses of patient tissues are complicated due to whether S100A8 and S100A9, two of the genes which are difficulties controlling degrees of stromal elements in the tumor most highly up-regulated in an H-Ras–specific manner, samples, differences in the genetic background among patients, play a causative role for H-Ras–mediated MCF10A cell tumor heterogeneity within an individual, timing of biopsy in invasion and migration. Importantly, small interfering relation to the chemotherapy, and/or radiation therapy, etc. RNA–mediated knockdown of S100A8/A9 expression Mutations of the ras oncogene are among the most frequent significantly reduced H-Ras–induced invasion/ genetic alterations in human tumors. The single point mutation migration. Conversely, the induction of S100A8/A9 at amino acid residue 12 (Gly to Asp) of H-Ras is more often expression conferred the invasive/migratory phenotype found in mammary carcinoma, whereas the same mutation of to parental MCF10A cells. Furthermore, we provided N-Ras is detected in teratocarcinoma and leukemia (6). To evidence of signaling cross-talk between S100A8/A9 and investigate the molecular mechanisms by which breast the mitogen-activated protein kinase signaling pathways epithelial cells undergo phenotypic transformation and acquire essential for H-Ras–mediated cell invasion and invasive potential, we have previously generated the MCF10A migration. Taken together, this study revealed model system in which H-Ras or N-Ras is constitutively S100A8/A9 genes as candidate markers for metastatic activated by mutating Gly to Asp in the amino acid codon 12 in H-Ras and N-Ras, respectively. We showed that although both H-Ras and N-Ras induce phenotypic transformation of MCF10A cells assessed by anchorage-independent cell growth Received 4/19/08; revised 6/12/08; accepted 6/30/08. and foci formation, only H-Ras activation induces invasive and Grant support: Microarray and Bioinformatics Core of the WSU EHS Center, migratory phenotypes in these cells (7). H-Ras–induced NIEHS Center grant P30 ES06639 and by NCI R41 CA105578 (H-R.C. Kim and H. Kim). This work was also supported by the MOE, the MOCIE, and the invasiveness was associated with its activation of p38 MOLAB of Korea through the fostering project of the Lab of Excellency, and mitogen-activated protein kinases (p38 MAPK) and extracel- partly by the KOSEF grant funded by the Korea government (MOST, no. R11- lular signal-regulated kinases (ERK), resulting in the induction 2007-107-01002-0) and the KOSEF NRL Program grant funded by the Korea government (MEST, no. ROA-2008-000-20070-0; A. Moon). of matrix metalloproteinase (MMP)-2 and MMP-9, whereas The costs of publication of this article were defrayed in part by the payment of N-Ras activated ERK but not p38 MAPK leading to MMP-9 page charges. This article must therefore be hereby marked advertisement in induction with little effect on MMP-2 expression (8, 9). In order accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Requests for reprints: Aree Moon, College of Pharmacy, Duksung Women’s to unveil the molecular signature specific to the biological University, 419 Ssangmun-Dong, Tobong-Ku, Seoul 132-714, Korea. Phone: 82- processes of invasive breast cancer, the present study analyzed 2901-8394; Fax: 82-2901-8386. E-mail: [email protected] Copyright D 2008 American Association for Cancer Research. the gene expression profile of our control, N-Ras–transfected doi:10.1158/1541-7786.MCR-08-0189 and H-Ras–transfected MCF10A cells, an experimental model

1544 Mol Cancer Res 2008;6(10). October 2008 Downloaded from mcr.aacrjournals.org on September 23, 2021. © 2008 American Association for Cancer Research. Candidate Markers for Breast Cell Invasion 1545

system representing the phenotypic conversion of transforma- in Fig. 1C. Cluster 1 consists of 84 genes that are markedly tion (control versus N-Ras– or H-Ras–transfected cells) and down-regulated in H-Ras MCF10A cells (H) compared with conversion from noninvasive/migratory to invasive/migratory MCF10A (10A) and N-Ras MCF10A (N) cells. A list of 20 breast epithelial cells (control or N-Ras versus H-Ras). Analysis genes with the lowest values of log2H/10A is shown in Table 1. of whole human genome microarray revealed that the Cluster 2 consists of 77 genes that are progressively up- expression levels of 412 genes were differentially modulated regulated in N compared with H and 10A. Twenty genes with among MCF10A, N-Ras MCF10A, and H-Ras MCF10A cells. the highest values of log2N/10A are listed in Table 1. Seventy- Among 173 genes that were uniquely up-regulated in invasive eight genes were grouped into cluster 3, which are progres- H-Ras MCF10A cells compared with noninvasive control or sively down-regulated in N compared with H and 10A. A list N-Ras MCF10 cell line, two calcium-binding proteins from of 20 genes with the lowest values of log2N/10A is shown in the S100 family, S100A8 (myeloid-related protein-8, MRP8, Table 1. Cluster 4 consists of 173 genes that are markedly or calgranulin A) and S100A9 (MRP14 or calgranulin B; up-regulated in H compared with 10A and N. Twenty genes refs. 10-12), were prominently regulated. The microarray with the highest values of log2H/10A are listed in Table 1. results of a few potential biomarkers, including S100A8 and S100A9, were validated by real-time reverse transcription-PCR Validation of Microarray Analysis by Real-time PCR and analysis, and H-Ras–mediated S100A8 and S100A9 induction Reverse Transcription-PCR was further confirmed by immunoblot analysis. Importantly, The microarray data of several cluster 4 genes (up-regulated small interfering RNA (siRNA)–mediated knockdown of by H-Ras) were then validated by real-time PCR and reverse S100A8 or S100A9 expression significantly reduced H-Ras– transcription-PCR analyses. Seven genes (S100A8, S100A9, mediated MCF10A cell invasion and MMP-2/MMP-9 induc- IL1R2, H2AFO, C3, HSD17B2, and CSF3) from cluster 4 and tion, suggesting the functional significance of S100A8 and two genes (FN1 and TNFSF7) from cluster 2 were subjected to S100A9 for induction of the invasive phenotype of human real-time PCR analysis using TaqMan assay method to validate breast epithelial cells. We also provided evidence for the the microarray results. As shown in Table 2, the real-time PCR positive feedback signaling loop between Ras-activated ERK/ results of all nine genes tested were consistent with the p38 MAPK and S100A8/S100A9 expression. Taken together, microarray data. the present study provided information revealing changes in Among genes in cluster 4, S100A8 and S100A9, the gene profiles potentially associated with cell invasion, separate members of the S100 protein family within the Ca2+-binding from proliferation/transformation. EF-hand protein superfamily (13), were most prominently up- regulated by H-Ras. Because mounting evidence suggested oncogenic activities of S100 proteins in the regulation of cell Results motility and invasion during human cancer progression and Microarray Analysis metastasis (14, 15), the subsequent study focused on these two To unveil the gene expression profile associated with cell proteins. In order to ensure the real-time PCR data of S100A8 invasion apart from cell proliferation/transformation, differen- and S100A9 obtained by TaqMan assay, the SYBRGreen assay tial gene expression patterns of MCF10A, N-Ras MCF10A, and was done on these two genes and the same results were H-Ras MCF10A cells were examined by Whole Human obtained (data not shown). In addition, reverse transcription- Genome Oligo Microarrays. Using ANOVA, a total of 1,542 PCR analysis was conducted to further confirm the microarray genes were identified as differentially expressed between and real-time PCR data. As shown in Fig. 1D, the expressions control and H-Ras MCF10A cells with P V 0.001 and 3% of S100A8 and S100A9 were selectively induced in H-Ras false discovery rate, whereas 1,012 genes were differentially MCF10A cells. expressed between H-Ras MCF10A and N-Ras MCF10A cells with P V 0.001 and 4% false discovery rate. As shown in Fig. 1A, the intersection of these two sets provided 412 genes Roles of S100A8 and S100A9 in Invasive and Migratory that were differentially expressed in H-Ras MCF10A cells Phenotypes compared with MCF10A controls and also differentially In order to determine the functional significance of S100A8 expressed when comparing H-Ras MCF10A and N-Ras and S100A9 in the H-Ras–induced invasive phenotype in MCF10A cells. This approach provides a list of candidate MCF10A cells, we selectively knocked down the expression of genes involved in the H-Ras–specific induction of the invasive S100A8 or S100A9 using an siRNA approach (Fig. 2A). phenotype in MCF10A cells. It excludes, however, such genes Following S100A8 down-regulation, the invasiveness of H-Ras that are activated by both H-Ras and N-Ras and may cooperate MCF10A cells was significantly inhibited by 66% as with some other H-Ras–specific gene(s) to induce the invasive determined by in vitro invasion assay (Fig. 2B). Interestingly, behavior. a more drastic inhibition (85%) of invasion was observed In order to identify genes that were uniquely up-regulated or when S100A9 expression was knocked down. Similarly, H-Ras down-regulated in invasive/migratory H-Ras MCF10A cells MCF10A cell motility was inhibited by knockdown of S100A8 compared with noninvasive MCF10A cells or N-Ras MCF10A and S100A9 by 57% and 80%, respectively (Fig. 2B). cells, hierarchical clustering was done with the 412 genes Knockdown of S100A8/9 did not affect the proliferation of identified in the statistical analysis. Four clusters with distinct H-Ras MCF10A cells as evidenced by 3-(4,5-dimethylthiazol- expression profiles were identified as depicted in Fig. 1B, and 2-yl)-2,5-diphenyltetrazolium bromide assay (data not shown). the fluorescent spot intensity signals in each cluster are shown The results show that S100A8 and S100A9 are critical

components of H-Ras–mediated cell migration and invasion. Our previous study showed critical roles for MMP-2 and, to When both S100A8 and S100A9 were knocked down, a lesser degree, MMP-9 in the H-Ras–mediated migratory and inhibition of invasive/migratory abilities was similar to that invasive phenotype in MCF10A cells (8). To test whether exerted by siRNA for S100A9 alone, suggesting that S100A8/9 S100A8 and/or S100A9 mediate H-Ras–induced MMP-2 and/ did not act synergistically in promoting the invasion and or MMP-9 expression, we examined the levels of MMP-2 and migration of H-Ras MCF10A cells. MMP-9 by gelatin zymography and reverse transcription-PCR

FIGURE 1. Clusteringof microarray data of differentially expressed genes in MCF10A (10A), H-Ras MCF10A (H- Ras), and N-Ras MCF10A (N-Ras) cells. A. Venn dia- gram showing overlap of genes differentially expressed in 10A vs. H-Ras and H-Ras vs. N-Ras comparison; the intersection of these two sets provided 412 genes. B. Four clusters with distinct expression profiles obtained from hierarchical clustering. C. Fluorescent spot intensity signals (the ratio of intensity values of each sample compared with the reference pool) in each cluster, representing relative expression in which, for each gene, the expression ratio of each sample was divided by the mean of all samples for that gene. D. Reverse transcription-PCR analysis of S100A8 and S100A9 in MCF10A (10A), H-Ras MCF10A (H), and N- Ras MCF10A (N) cells. h-Actin was used as a control.

analysis in H-Ras MCF10A cells following S100A8 or S100A9 MAPK and ERK pathways, respectively. As shown in Fig. 3C, knockdown. As shown in Fig. 2C, inhibition of S100A8 inhibition of p38 MAPK and ERK pathways resulted in a expression resulted in the down-regulation of MMP-9 and, to a prominent inhibition of S100A8 and S100A9. The mRNA lesser degree, MMP-2 in H-Ras–activated MCF10A cells, levels of S100A8 and S100A9 were also decreased by whereas S100A9 down-regulation drastically reduced the levels SB203580 and PD98059, as evidenced by reverse transcrip- of MMP-2 and, to a lesser extent, MMP-9 expression. These tion-PCR, suggesting that the down-regulation of these proteins results suggest that H-Ras–induced S100A8 may exert a was not due to protein degradation (data not shown). The crucial role in MMP-9 up-regulation, whereas S100A9 is more results show that both p38 MAPK and ERK signaling pathways closely associated with MMP-2 expression in human breast play a critical role in the H-Ras–mediated S100A8 and S100A9 epithelial cells. expression in MCF10A cells. We next asked if expressions of Our data suggest that S100A8 and S100A9 may be S100A8 and S100A9 affected the activation of ERK and p38 responsible for H-Ras–specific induction of invasive and MAPK. As shown in Fig. 3D, the levels of phosphorylated migratory phenotypes in MCF10A cells. To validate the results ERK and p38 MAPK were decreased by knockdown of obtained by siRNA-mediated knockdown of S100A8/A9, we S100A8 or S100A9 by siRNA molecules, suggesting that did a gain-of-function experiment. We tested whether the expressions of S100A8 and S100A9 are important for the induction of S100A8 and S100A9 can confer invasive and activation of MAPK signal transduction pathways. migratory abilities in MCF10A cells which are noninvasive and nonmigratory. To address this issue, we transfected the parental Discussion MCF10A cells with either S100A8 or S100A9 DNA constructs Microarray analysis produced a list of genes selectively up- and examined their invasive and migratory properties. Invasion regulated or down-regulated by H-Ras or N-Ras. The main and migration were significantly increased by the induction of merits of this study are 2-fold. First, the comprehensive gene S100A8 and S100A9 to a similar extent, demonstrating that the expression profile of the control, N-Ras–transfected and induction of S100A8 and S100A9 could confer invasive and H-Ras–transfected MCF10A cells provide valuable informa- migratory phenotypes in the MCF10A cell line (Fig. 2D). tion which may lead to a better understanding of the molecular signature specific to many different cellular processes which are Role of Calcium in Invasion, Migration, and S100A8/A9 manifested in an H-Ras–specific or N-Ras–specific manner in Expression our experimental model system. Second, some of the signature Recent studies showed that S100 proteins form tetramers or genes may play causative roles in a particular biological even higher-order oligomers in a calcium-dependent manner, process, helping us understand a particular cellular process at and that oligomerization is an essential prerequisite for their the molecular level and thereby providing the potential biological functions (16, 17). To further investigate the molecular targets for therapy. In this study, we investigated significance of S100A8/S100A9 in H-Ras–mediated cell the functional significance of S100A8 and S100A9 in H-Ras– invasion/migration, we temporarily depleted intracellular calci- mediated cell invasion, two of the genes most abundantly um in H-Ras MCF10A cells using BAPTA/AM, an intracellular up-regulated by H-Ras in MCF10A cells. These genes were calcium chelator (18). As shown in Fig. 3A, H-Ras–induced also chosen based on recent studies which suggest the role of invasion and migration of MCF10A cells were significantly S100 proteins in mediating various cellular processes including inhibited by treatment with 50 Amol/L of BAPTA/AM, cell growth, differentiation, cell migration, and cell adhesion demonstrating the requirement of the intracellular calcium (20-22). Among S100 proteins, S100A4, S100A8, and S100A9 currents for cell movement of H-Ras MCF10A cells. Interest- have emerged as potential prognostic markers for poor patient ingly, S100A8 and S100A9 proteins were barely detectable survival and metastasis of a number of cancers (14, 23-25). when intracellular calcium was depleted (Fig. 3B, right), S100A8 and S100A9 were shown to be overexpressed in whereas the RNA levels of S100A8 and S100A9 were not gastric cancer (26, 27), prostate cancer (28), breast cancer (29), altered (Fig. 3B, left). These results suggest that calcium may lung adenocarcinomas (30), pulmonary adenocarcinoma (30), be critical for the protein stability of S100A8 and S100A9. A and hepatocellular carcinoma (31). Consistent with the present possibility that calcium may affect the translation of these study, we previously showed that S100A8 and S100A9 play proteins, however, cannot be excluded. critical roles in the invasive phenotype of a human gastric cancer cell line, SNU484 (27). In agreement with our study, Effects of ERK and p38 MARK Pathways on S100A8 S100A8 and S100A9 were shown to induce myeloid cell and S100A9 Expression recruitment and tumor cell invasion in premetastatic lung (19). S100A8 and S100A9 have been shown to induce migration Interestingly, expression profiling of signal transducers and of macrophages and tumor cells into the lung through activation activators of transcription 3–transformed MCF10A and HME of p38 MAPK signaling (19). Our previous study showed that breast epithelial cells identified 23 overexpressed genes p38 MAPK and ERK pathways were critical to the invasive including MMP-9, S100A8, and S100A9 (32). Taken together, and migratory phenotypes induced by H-Ras in MCF10A cells S100A8 and S100A9 may serve as potential markers associated (8). To investigate the functional relationship between S100A8/ with tumor cell invasion and migration. S100A9 and the previously known signaling cascades mediated At the molecular level, the S100A8/A9 complex was shown by H-Ras, we examined the levels of S100A8 and S100A9 to activate nuclear factor nB and to induce phosphorylation of following the treatment of H-Ras MCF10A cells with p38 MAPK and ERK in prostate cancer cells (33). Calcium SB203580 and PD98059, pharmacologic inhibitors of the p38 binding induces tetramer formation or oligomerization of

Table 1. List of Genes Differentially Expressed in MCF10A, H-Ras MCF10A, and N-Ras MCF10A Cells by Microarray Analysis (Clusters 1-4)

Symbol Accession No.* Sequence Description 10A H N Log2N/10A Log2H/10A Comparison

Genes down-regulated by H-Ras in MCF10A cells (Cluster 1) BC010432 BC010432 CDNA clone IMAGE:3528357 1.04 0.36 1.10 0.09 À1.54 N>10A>H CAV1 NM_001753 Caveolin 1, caveolae protein, 22 kDa 1.20 0.62 1.20 0.00 À0.95N=10A>H PLS3 NM_005032 Plastin 3 (T isoform) 1.44 0.78 1.03 À0.48 À0.87 10A>N>H CAV1 NM_001753 Caveolin 1, caveolae protein, 22 kDa 1.10 0.60 1.09 À0.01 À0.87 10A>N>H CAMSAP1 AL834528 Calmodulin regulated spectrin-associated protein 1 1.09 0.63 1.30 0.26 À0.79 N>10A>H IGHG1 L04336 Immunoglobulin heavy constant g1 1.01 0.58 1.19 0.24 À0.78 N>10A>H VPRBP BC022792 KIAA0800 gene product 1.17 0.69 0.93 À0.33 À0.76 10A>N>H NRP1 AF280547 Neuropilin 1 0.99 0.59 1.16 0.22 À0.76 N>10A>H RP9 NM_203288 Retinitis pigmentosa 9 (autosomal dominant) 1.37 0.81 1.06 À0.38 À0.76 10A>N>H MADH4 NM_005359 ElaC homologue 1 (E. coli) 1.09 0.67 1.16 0.09 À0.71 N>10A>H CAV3 NM_001234 Caveolin 3 1.09 0.67 1.11 0.02 À0.70 N>10A>H ZFP95 NM_014569 Zinc finger protein 95 homologue (mouse) 0.88 0.56 1.20 0.44 À0.66 N>10A>H HMG1 NM_002128 High-mobility group box 1 1.23 0.78 1.00 À0.30 À0.6510A>N>H MET NM_000245Met proto-oncogene (hepatocyte growth factor receptor) 1.18 0.76 1.02 À0.20 À0.62 10A>N>H EPB41L1 NM_012156 Erythrocyte membrane protein band 4.1-like 1 1.00 0.65 1.09 0.12 À0.61 N>10A>H GYPB NM_002100 Glycophorin B (MNS blood group) 1.00 0.69 1.04 0.06 À0.53 N>10A>H A_32_P235289 Unknown 1.03 0.71 1.02 À0.01 À0.53 10A>N>H A_24_P127042 Unknown 1.04 0.73 1.14 0.13 À0.52 N>10A>H A_32_P219554 Unknown 1.04 0.73 1.00 À0.06 À0.50 10A>N>H THC1856939 AK122508 mKIAA1424 protein (Mus musculus), 1.00 0.72 1.10 0.14 À0.49 N>10A>H partial (25%; THC1856939) Genes up-regulated by N-Ras in MCF10A cells (Cluster 2) PRG1 NM_002727 Proteoglycan 1, secretory granule 0.16 1.14 5.48 5.12 2.85 N>H>10A VIM NM_003380 Vimentin 0.051.00 1.46 4.83 4.28 N>H>10A VIM NM_003380 Vimentin 0.07 1.01 1.48 4.46 3.91 N>H>10A SECTM1 NM_003004 Secreted and transmembrane 1 0.10 1.03 1.94 4.28 3.37 N>H>10A FN1 NM_002026 Fibronectin 1 0.14 1.00 2.57 4.17 2.80 N>H>10A A24P596251 Unknown 0.150.99 1.29 3.07 2.69 N>H>10A TIMP1 NM_003254 TIMP metallopeptidase inhibitor 1 0.35 1.01 2.62 2.92 1.54 N>H>10A DUSP23 NM_017823 hypothetical protein FLJ20442 0.38 1.00 2.36 2.62 1.38 N>H>10A CTGF NM_001901 Connective tissue growth factor 0.59 0.96 3.38 2.53 0.71 N>H>10A TNFSF7 NM_001252 Tumor necrosis factor (ligand) superfamily 0.25 1.00 1.46 2.52 1.97 N>H>10A C14orf172 NM_152307 Chromosome 14 open reading frame 172 0.29 1.02 1.64 2.50 1.83 N>H>10A ALOX5AP NM_001629 Arachidonate 5-lipoxygenase-activating protein 0.47 0.97 2.38 2.34 1.06 N>H>10A IGFBP7 NM_001553 Insulin-like growth factor binding protein 7 0.53 0.99 2.61 2.29 0.89 N>H>10A A23P255111 Unknown 0.56 0.97 2.23 2.00 0.80 N>H>10A LTBP1 NM_000627 Latent transforming growth factor h binding protein 0.44 1.04 1.74 1.98 1.22 N>H>10A FBN1 NM_000138 Fibrillin 11 0.56 1.05 2.20 1.96 0.90 N>H>10A DUSP1 NM_004417 Dual specificity phosphatase 1 0.49 1.00 1.74 1.84 1.03 N>H>10A GPSM1 AL117478 Likely orthologue of rat activator of G-protein signaling 3 0.54 1.02 1.86 1.79 0.92 N>H>10A QSCN6 NM_002826 Quiescin Q6 0.52 0.99 1.80 1.78 0.92 N>H>10A PSCA NM_005672 Prostate stem cell antigen 0.46 1.01 1.74 1.74 1.12 N>H>10A Genes down-regulated by N-Ras in MCF10A cells (Cluster 3) EPN3 NM_017957 Epsin 3 3.73 0.99 0.26 À3.87 À1.91 10A>H>N UNC5B NM_170744 Unc-5homologue B ( C. elegans) 4.251.17 0.53 À3.01 À1.86 10A>H>N TNNI2 NM_003282 Troponin I type 2 (skeletal, fast) 3.66 1.10 0.44 À3.05 À1.74 10A>H>N CDH1 NM_004360 Cadherin 1, type 1, E-cadherin (epithelial) 3.19 1.02 0.19 À4.06 À1.6510A>H>N TACSTD1 NM_002354 Tumor-associated calcium signal transducer 1 3.08 1.00 0.08 À5.29 À1.62 10A>H>N TACSTD2 NM_002353 Tumor-associated calcium signal transducer 2 2.96 0.98 0.51 À2.54 À1.60 10A>H>N SULF2 NM_018837 Similar to glucosamine-6-sulfatases 2.77 0.99 0.40 À2.81 À1.49 10A>H>N LEPREL1 NM_018192 Hypothetical protein FLJ10718 2.88 1.09 0.70 À2.04 À1.40 10A>H>N LAD1 NM_005558 Ladinin 1 2.54 1.00 0.07 À5.12 À1.34 10A>H>N PPP1R14C NM_030949 Protein phosphatase 1, regulatory (inhibitor) subunit 14C 2.44 0.98 0.39 À2.66 À1.32 10A>H>N SFRP1 BC036503 Secreted frizzled-related protein 1 2.49 1.00 0.25 À3.29 À1.32 10A>H>N S100A14 NM_020672 S100-type calcium binding protein A14 2.52 1.01 0.11 À4.49 À1.31 10A>H>N TFCP2L1 NM_014553 LBP protein; likely orthologue of mouse CRTR-1 2.48 1.00 0.72 À1.78 À1.30 10A>H>N OVOL2 NM_021220 putative zinc finger protein from EUROIMAGE 566589 2.49 1.02 0.51 À2.30 À1.29 10A>H>N CKMT1 NM_020990 Creatine kinase, mitochondrial 1A 2.48 1.050.22 À3.47 À1.24 10A>H>N SFRP1 NM_003012 Secreted frizzled-related protein 1 2.351.00 0.31 À2.91 À1.24 10A>H>N FAM84B NM_174911 Family with sequence similarity 84, member B 2.30 1.01 0.31 À2.89 À1.19 10A>H>N MAL2 NM_052886 Mal, T-cell differentiation protein 2 2.28 1.00 0.63 À1.86 À1.19 10A>H>N TP73L NM_003722 Tumor protein 63 kDa with strong homology to p53 2.34 1.04 0.23 À3.36 À1.17 10A>H>N KRT17 NM_000422 Keratin 17 2.48 1.13 0.31 À2.98 À1.13 10A>H>N Genes up-regulated by H-Ras in MCF10A cells (Cluster 4) CSF3 NM_00759 Colony-stimulating factor 3 (granulocyte) 0.26 1.95 1.02 1.96 2.89 H>N>10A S100A8 NM_002964 S100 calcium binding protein A8 (calgranulin A) 1.03 7.48 0.70 À0.55 2.86 H>10A>N S100A9 NM_002965 S100 calcium binding protein A9 (calgranulin B) 0.87 3.69 0.54 À0.68 2.08 H>10A>N IL1R2 NM_004633 Interleukin 1 receptor, type II 0.75 3.13 0.78 0.07 2.07 H>N>10A H2AFO NM_003516 Histone 2, H2aa3 0.75 2.67 1.00 0.41 1.83 H>N>10A C3 NM_000064 Complement component 3 0.65 2.13 1.05 0.70 1.72 H>N>10A HSD17B2 NM_002153 Hydroxysteroid (17-B) dehydrogenase 2 0.99 2.70 0.87 À0.18 1.45 H>10A>N

(Continued on the following page)

Table 1. List of Genes Differentially Expressed in MCF10A, H-Ras MCF10A, and N-Ras MCF10A Cells by Microarray Analysis (Clusters 1-4) (Cont’d)

Symbol Accession No.* Sequence Description 10A H N Log2N/10A Log2H/10A Comparison

SMOX NM_175839 Hypothetical protein 0.80 1.76 0.93 0.22 1.15 H>N>10A CYP1B1 NM_000104 Cytochrome P450, family 1, subfamily B, polypeptide 1 0.90 2.00 0.91 0.01 1.15 H>N>10A COL12A1 NM_004370 Collagen, type XII, a 1 0.96 2.07 0.95 À0.01 1.11 H>10A>N TNFRSF6B NM_032957 Tumor necrosis factor receptor superfamily, 0.92 1.94 0.97 0.08 1.08 H>N>10A member 6b, Decoy ADAM8 NM_001109 ADAM metallopeptidase domain 8 0.99 2.050.60 À0.72 1.05H>10A>N LOC51333 NM_016643 Mesenchymal stem cell protein DSC43 0.98 1.78 0.84 À0.22 0.86 H>10A>N HMCN2 AK093583 Hemicentin 2 0.76 1.39 1.00 0.38 0.86 H>N>10A H2AFJ NM_177925hypothetical protein FLJ10903 0.98 1.750.62 À0.650.83 H>10A>N DPM3 NM_018973 Dolichyl-phosphate mannosyltransferase polypeptide 0.88 1.53 0.98 0.16 0.80 H>N>10A MGC10955 BC004960 Hypothetical protein MGC10955 0.93 1.60 0.90 À0.06 0.78 H>10A>N COL6A1 NM_001848 Collagen, type VI, a1 0.93 1.60 0.89 À0.36 0.77 H>10A>N ID1 NM_002165Inhibitor of DNA binding 1, dominant 0.97 1.66 0.76 À0.36 0.77 H>10A>N negative helix-loop-helix protein PFKL NM_002626 Phosphofructokinase, liver 0.951.61 0.83 À0.19 0.75H>10A>N

NOTE: Microarray analysis was done on MCF10A (10A), H-Ras MCF10A (H), and N-Ras MCF10A (N) cells. The most prominent top 20 genes in clusters 1 to 4 are listed. The ratios represent the expression level with respect to the pooled reference sample. (À) indicates decreased gene expression. *Gene bank accession number. Genes validated by real-time PCR are written in bold character.

S100A8 and S100A9, a prerequisite for their biological MAPK pathway may play a major role in the H-Ras–mediated activities (16). The conformational changes allow the phos- S100A9 expression, whereas the ERK pathway is more critical phorylation of S100A9 by p38 MAPK and polymerization of for S100A8 up-regulation. Interestingly, the siRNA-knockdown microtubules during migration of phagocytes (12). The of S100A8 resulted in a more efficient down-regulation of S100A8/S100A9 complex binds to tubulin filaments in a MMP-9 than MMP-2 whereas knockdown of S100A9 more calcium-dependent manner and in the presence of S100A8/A9 drastically reduced the level of MMP-2 (Fig. 2C). These results complex, the number of tubulin filaments is significantly support the notion that S100A8 is more closely associated with increased (12). In agreement with previous reports, the present MMP-9 expression mediated by the ERK pathway, and that study provided evidence of critical roles for S100A8 and S100A9 may play a major role in MMP-2 up-regulation, which S100A9 in H-Ras–mediated cell invasion/migration in a is dependent on p38 MAPK signaling. These results may unveil calcium-dependent manner. Interestingly, intracellular calcium subtle but potentially significant differences between S100A8 seems to be critical for the protein stability of S100A8 and and S100A9 for the regulation by intracellular signal S100A9 in MCF10A cells in addition to its previously reported transduction pathways. At present, the molecular basis for role in the induction of oligomerization and conformational these functional differences is unclear. However, taking these changes of the S100A8/A9 complex necessary for their results together with our previous study which showed a critical biological activity. role for H-Ras–specific activation of p38 MAPK signaling Our previous study showed that H-Ras activation of the p38 pathway leading to MMP-2 up-regulation and induction of the MAPK pathway is critical for MMP-2 induction, whereas the invasive phenotype of human breast epithelial cells (8, 9, 34), it ERK pathway leads to MMP-9 induction (34). Inhibition of the is plausible that S100A9 may play an important role in H-Ras– p38 MAPK signaling by SB203580 resulted in a more drastic induced invasion and migration of MCF10A cells in addition to inhibition of S100A9, whereas blocking the ERK pathway by its cooperative role with S100A8. PD98059 inhibited the expression of S100A8 more efficiently Hierarchical clustering of 412 genes that were differentially than that of S100A9 (Fig. 3C). These data suggest that the p38 expressed among MCF10A, N-Ras MCF10A, and H-Ras

Table 2. Real-time PCR Analysis for MCF10A, H-Ras MCF10A, and N-Ras MCF10A

Symbol Accession No.* Sequence Description Microarray Real-time PCR (Fold Changes)

Comparison H vs.10A H vs. N N vs. 10A Comparison

S100A8 NM002964 S100 calcium-binding protein A8 (calgranulin A) H>10A>N 23.74 165.903 À6.99 H>10A>N S100A9 NM002965S100 calcium-binding protein A9 (calgranulin B) H>10A>N 7.788 6488.21 À1000 H>10A>N CSF3 NM000759 Colony-stimulating factor 3 (granulocyte) H>N>10A 12.148 2.139 5.679 H>N>10A HSD17B2 NM002153 Hydroxysteroid (17-h) dehydrogenase 2 H>10A>N 26.154 32.142 À1.22 H>10A>N IL1R2 NM004633 Interleukin 1 receptor, type II H>N>10A 5.859 10.883 À1.85H>10A>N C3 NM000064 Complement component 3 H>N>10A 3.944 4.32 À1.09 H>10A>N H2AFO NM003516 H2A histone family, member O H>N>10A 1.08 1.027 À1.05H=10A>N FN1 NM002026 Homo sapiens fibronectin 1, transcript variant 1, RNA N>H>10A 13.93 À1.898 26.445N>H>10A TNFSF7 NM001252 Tumor necrosis factor (ligand) superfamily, member7 N>H>10A 4.999 À1.2 6.019 N>H>10A

NOTE: Real-time PCR analysis was done on MCF10A (10A), H-Ras MCF10A (H), and N-Ras MCF10A (N) cells. (À) indicates decreased gene expression. *Gene bank accession number.

MCF10A cells produced four groups of genes which are molecular actions during normal physiologic processes and uniquely up-regulated or down-regulated by H-Ras or N-Ras. pathologic conditions in the future. The list of genes presented In this proof-of-principle study, we showed that two gene here may also be useful to generate a diagnostic chip for pattern products induced by H-Ras are the downstream targets of the recognition to predict a particular cellular process. H-Ras signaling pathways. Considering the diverse roles of H-Ras and N-Ras in the regulation of central signaling Materials and Methods networks critical for many different cellular processes, the Cell Culture gene profiling data presented in this article may be useful for The development and characterization of MCF10A, H-Ras many investigators seeking to unveil Ras isoform–specific MCF10A, and N-Ras MCF10A cells were described previously

FIGURE 2. Effects of S100A8 and S100A9 knockdown on invasive/migratory phenotypes and MMP-2/ MMP-9 expressions in H-Ras MCF10A cells. A. Knockdown of S100A8 and S100A9 by siRNA was confirmed by reverse transcription-PCR analysis (left) and Western blot analysis (right) in cells transfected with siRNA. Control cells were treated with Stealth RNAi-negative control duplexes. B. Cells transfected with siRNAs for S100A8 and S100A9 were subjected to in vitro invasion assay (left) or Transwell migration assay (right). The number of invaded or migrated cells per field was counted (Â400) in 13 arbitrary visual fields. Columns, mean of results in triplicate; bars, SE. *, P < 0.01, statistically different from controls using the two-tailed Student’s t test. C. Gelatin zymogram assay (left) was done on cells transfected with siRNAs for S100A8 and S100A9 to analyze the gelatinolytic activities of secreted MMP-2 (72 kDa) and MMP-9 (92 kDa). Reverse transcription-PCR analysis (right) was done to detect mRNA levels of MMP-2 and MMP-9. Relative band intensities were quantitated and plot- ted. D. MCF10A cells were transiently transfected with S100A8 and S100A9 constructs. The transfected cells were subjected to in vitro invasion assays (left)or Transwell migration assays (right). The number of invaded or migrated cells per field was counted (Â400) in 13 arbitrary visual fields. Columns, mean of results in triplicate; bars, SE. *, P < 0.01, statistically different from controls using the two-tailed Student’s t test.

FIGURE 3. Roles of calcium and signaling pathways on invasive/migratory phenotypes of H-Ras MCF10A cells. A. H-Ras MCF10A cells were pretreated with 50 Amol/L of BAPTA/AM for 30 min and subjected to in vitro invasion assay (left) and migration assay (right)for18hinthe presence of the compound. The number of invaded cells or migrated cells per field were counted (Â400) in 13 fields. Columns, mean of results in triplicate; bars, SE. *, P < 0.01, statistically different from controls. B. H-Ras MCF10A cells were treated with various con- centrations of BAPTA/AM for 24 h and subjected to reverse transcription-PCR analysis (left) and Western blot analysis (right) to detect expressions of S100A8 and S100A9. h-Actin was used as a control. C. H-Ras MCF10A cells were treated with SB203580 or PD98059 at various con- centrations for 24 h. Expres- sion of S100A8 and S100A9 were determined by Western blot analysis. Relative band intensities were quantitated and normalized to no- treatment control. D. H-Ras MCF10A cells were transfected with siRNA for S100A8 and S100A9. Phosphorylated ERK and p38 MAPK were detected by Western blot analysis. Relative band intensities of phosphorylated ERK (pERK) or p38 MAPK (pp38) were quantitated and normalized to total ERK or p38 MAPK.

(7). Cells were cultured in DMEM/F12 supplemented with 5% each RNA sample was tested using an Agilent 2100 horse serum, 0.5 Ag/mL of hydrocortisone, 10 Ag/mL of insulin, Bioanalyzer (Agilent Technologies). Fluorescent labeling was 20 ng/mL of epidermal growth factor, 0.1 Ag/mL of cholera done on 500 ng of total RNA for each sample using the Agilent enterotoxin, 100 units/mL of penicillin-streptomycin, 2 mmol/L Low RNA Input Linear Amplification Kit according to the of L-glutamine, and 0.5 Ag/mL of amphotericin. Cells were vendor’s protocol (Agilent). A common reference pool was maintained in a humidified atmosphere with 95% air and 5% constructed by pooling equal amounts of total RNA from all j CO2 at 37 C. samples. Each sample was individually cohybridized on a microarray with the reference pool, in which the sample and Microarray reference pools were labeled with contrasting fluorescent dyes, Total RNA was prepared using Trizol reagent (Invitrogen) cyanine 5(Cy5)and cyanine 3 (Cy3). A pair of dye-swapped according to the manufacturer’s instructions. The quality of arrays was completed for each sample: Cy5-labeled sample

cohybridized with Cy3-labeled reference pool and Cy3-labeled 2 min for 1 cycle, then at 94jC for 30 s, 55jC for 45s, 72 jC sample cohybridized with Cy5-labeled reference pool. A total for 45s for 25to 30 cycles, and 72 jC for 7 min. Equal volumes of 12 microarray analyses were done in the study including two of each PCR product were analyzed by agarose gel electro- RNA samples used for each cell line, each sample having a dye- phoresis and bands of S100A8 (119 bp), S100A9 (400 bp), swapped pair of arrays. Labeled cRNA was purified using MMP-2 (119 bp), and MMP-9 (124 bp) were detected. Qiagen RNeasy Mini kit (Qiagen). Hybridization was done for 17 h at 60jC using Agilent Whole Human Genome Oligo Western Blot Analysis Microarrays, each having 41,000 probes (Agilent). Slides were Anti-S100A8 and anti-S100A9 antibodies were purchased scanned with an Agilent dual laser scanner set at 100% from Santa Cruz Biotechnology. Anti–h-actin antibody was photomultiplier tube. purchased from Sigma-Aldrich. Equal amounts of protein extracts in SDS-lysis buffer were subjected to 12% SDS-PAGE Data Analysis analysis and electrophoretically transferred to polyvinylidene Scanned (Tiff) images were analyzed using Agilent’s feature difluoride membrane. Enhanced chemiluminescence (Amer- extraction software to obtain fluorescent intensities for each spot sham-Pharmacia) system was used for detection. Relative band on the arrays. Local background subtraction and linear and intensities were determined by quantitation of each band with Lowess normalization was done on intensity values. ANOVA an Image Analyzer (Vilber Lourmat). was done with Rosetta resolver using error-weighted averaging of replicates (35). Hierarchical clustering was done with siRNA Preparation and Transfection GeneSpring (version 7.3, Agilent Technologies) using Pearson The siRNA sequences targeting S100A8 (5¶-CCAUCAU- correlation of expression ratios normalized on a per gene basis. CAACACCUUCCACCAAUA-3¶)andS100A9(5¶-CCUU- Each gene was normalized across all samples by dividing the GAACUCUAUCGACGUCUA-3¶) were purchased from expression ratio of each individual sample measurement for that Invitrogen. H-Ras MCF10A cells in the exponential phase of gene by the mean ratio obtained from all samples for the given growth were plated in six-well plates at 1.5 Â 105 cells/well, gene. The net effect is to center all gene profiles around a ratio of grown for 24 h then transfected with 25pmol of siRNA for one, thereby facilitating a comparison of the expression profiles. 6 h using Lipofectamine 2000 reagent (Invitrogen) and OPTI- MEM reduced serum medium (Invitrogen). Control cells were Real-time PCR treated with Stealth RNAi-negative control duplex (Invitrogen). The total RNA of each sample was used to generate cDNA and comparative real-time PCR reactions were carried out Transfection of MCF10A Cells with S100A8 and S100A9 following the protocol for the TaqMan Reverse Universal Constructs Transcription Master Mix (Applied Biosystems). Each sample S100A8 and S100A9 expression constructs were purchased was assayed in triplicate for each gene. Assays were completed from Invitrogen. MCF10A cells in the exponential phase of using the ABI Prism 7500 Fast Sequence Detection System growth were plated in six-well plates at 3.5 Â 105 cells/well, (Applied Biosystems). The initial setup for the thermocycler grown for 24 h, then transfected with 4 Ag of each plasmid for 6 condition was 20 s at 95jC. After the initial setup, 40 cycles h using Lipofectamine 2000 reagent (Invitrogen) and OPTI- were run with two steps, 3 s at 95jC followed by 30 s at 60jC. MEM reduced serum medium (Invitrogen). For cotransfection Data were analyzed using the comparative cycle threshold (Ct) of S100A8 and S100A9 plasmids, equal amounts of the method. Relative quantification was done comparing gene plasmids were added to the transfection mix. Control cells were expression between the two cell lines and using Polr2a as an transfected with empty vectors of the expression construct. endogenous control for normalization of genes (36). Gelatin Zymogram Assay Reverse Transcription-PCR Detection Cells were cultured in serum-free DMEM/F12 medium for RNA was reverse transcribed with RT-Superscript III reverse 48 h. The gelatinolytic activity of the conditioned medium was transcriptase (Invitrogen). Primers for S100A8 and S100A9 determined by gelatin zymogram assay as previously described were designed according to the sequences of S100A8 and (7). Areas of gelatinase activity were detected as clear bands S100A9 mRNA (GenBank no. NM_002964 and NM_002965, against the blue-stained gelatin background. respectively). The forward and reverse primers for S100A8 were 5¶-CCGAGTGTCCTCAGTATATCAGGA-3¶ and 5¶- In vitro Invasion Assay and Transwell Migration Assay GGCCATCTTTATCACCGAATGA-3¶, respectively. The for- In vitro invasion assay and transwell migration assay were ward and reverse primers for S100A9 were 5¶-GTCGCAGC- done for 18 h using a 24-well transwell unit as previously TGGAACGCAACA-3¶ and 5¶-CCTGGCCTCCTGAT- described (8). TAGTGG-3¶, respectively. The forward and reverse primers for MMP-2 were 5¶-AATGCCATCCCCGATAACC-3¶ and Densitometry Measurements 5¶-AAACTTCACGCTCTTCAGAC-3¶, respectively, and the Relative band intensities were determined by quantitation of forward and reverse primers for MMP-9 were 5¶-TCTTCCAG- each band with an Image Analyzer (Vilber Lourmat). TACCGAGAGAAAG-3¶ and 5¶-AGGATGTCATAGGTCA- CGTAG-3¶, respectively. For reverse transcription-PCR analy- Disclosure of Potential Conflicts of Interest sis, the following amplification conditions were applied: 94jC No potential conflicts of interest were disclosed.

References 19. Hiratsuka S, Watanabe A, Aburatani H, Maru Y. Tumour-mediated 1. van de Vijver MJ, He YD, van’t Veer LJ, et al. A gene-expression upregulation of chemoattractants and recruitment of myeloid cells predetermines signature as a predictor of survival in breast cancer. N Engl J Med 2002;347: lung metastasis. Nat Cell Biol 2006;8:1369 – 75. 1999 – 2009. 20. Li C, Zhang F, Lin M, Liu J. Induction of S100A9 gene expression by 2. Sotiriou C, Wirapati P, Loi S, et al. Gene expression profiling in breast cancer: cytokine oncostatin M in breast cancer cells through the STAT3 signaling cascade. understanding the molecular basis of histologic grade to improve prognosis. J Natl Breast Cancer Res Treat 2004;87:123 – 34. Cancer Inst 2006;98:262 – 72. 21. Kerkhoff C, Klempt M, Kaever V, Sorg C. The two calcium-binding proteins, 3. Dupont VN, Gentien D, Oberkampf M, De Rycke Y, Blin N. A gene S100A8 and S100A9, are involved in the metabolism of arachidonic acid in expression signature associated with metastatic cells in effusions of breast human neutrophils. J Biol Chem 1999;274:32672 – 79. carcinoma patients. Int J Cancer 2007;121:1036 – 46. 22. Li ZH, Bresnick AR. The S100A4 metastasis factor regulates cellular 4. Ma Y, Qian Y, Wei L, et al. Population-based molecular prognosis of breast motility via a direct interaction with myosin-IIA. Cancer Res 2006;66:5173 – 80. cancer by transcriptional profiling. Clin Cancer Res 2007;13:2014 – 22. 23. Arai K, Teratani T, Kuruto-Niwa R, Yamada T, Nozawa R. S100A9 5. Schuetz CS, Bonin M, Clare SE, et al. Progression-specific genes identified by expression in invasive ductal carcinoma of the breast: S100A9 expression in expression profiling of matched ductal carcinomas in situ and invasive breast adenocarcinoma is closely associated with poor tumour differentiation. Eur J tumors, combining laser capture microdissection and oligonucleotide microarray Cancer 2004;40:1179 – 87. analysis. Cancer Res 2006;66:5278 – 86. 24. Rudland PS, Platt-Higgins A, Renshaw C, et al. Prognostic significance of 6. Franks LM, Teich NM. Cellular and molecular biology of cancer. Oxford the metastasis-inducing protein S100A4 (p9Ka) in human breast cancer. Cancer University Press; 1997. Res 2000;60:1595 – 03. 7. Moon A, Kim MS, Kim TG, et al. H-ras, but not N-ras, induces an invasive 25. Ott HW, Lindner H, Sarg B, et al. Calgranulins in cystic fluid and serum from phenotype in human breast epithelial cells: a role for MMP-2 in the H-ras-induced patients with ovarian carcinomas. Cancer Res 2003;63:7507 – 14. invasive phenotype. Int J Cancer 2000;85:176 – 81. 26. El-Rifai W, Moskaluk CA, Abdrabbo MK, et al. Gastric cancers overexpress 8. Kim MS, Lee EJ, Kim HR, Moon A. p38 kinase is a key signaling molecule S100A calcium-binding proteins. Cancer Res 2002;62:6823 – 26. for H-Ras-induced cell motility and invasive phenotype in human breast epithelial 27. Yong HY, Moon A. Roles of calcium-binding proteins, S100A8 and S100A9, cells. Cancer Res 2003;63:5454 – 61. in invasive phenotype of human gastric cancer cells. Arch Pharm Res 2007;30: 9. Song H, Ki SH, Kim SG, Moon A. Activating transcription factor 2 mediates 75– 81. matrix metalloproteinase-2 transcriptional activation induced by p38 in breast 28. Hermani A, Hess J, De Servi B, et al. Calcium-binding proteins S100A8 and epithelial cells. Cancer Res 2006;66:10487 – 96. S100A9 as novel diagnostic markers in human prostate cancer. Clin Cancer Res 10. Odink K, Cerletti N, Bruggen J, et al. Two calcium-binding proteins in 2005;11:5146 – 52. infiltrate macrophages of rheumatoid arthritis. Nature 1987;330:80 – 2. 29. Cross SS, Hamdy FC, Deloulme JC, Rehman I. Expression of S100 proteins 11. Zwadlo G, Bruggen J, Gerhards G, Schlegel R, Sorg C. Two calcium-binding in normal human tissues and common cancers using tissue microarrays: S100A6, proteins associated with specific stages of myeloid cell differentiation are S100A8, S100A9 and S100A11 are all overexpressed in common cancers. expressed by subsets of macrophages in inflammatory tissues. Clin Exp Immunol Histopathology 2005;46:256 – 69. 1988;72:510 – 5. 30. Arai K, Teratani T, Nozawa R, Yamada T. Immunohistochemical 12. Vogl T, Ludwig S, Goebeler M, et al. MRP8 and MRP14 control microtubule investigation of S100A9 expression in pulmonary adenocarcinoma: S100A9 reorganization during transendothelial migration of phagocytes. Blood 2004;104: expression is associated with tumor differentiation. Oncol Rep 2001;8:591 – 6. 4260 – 8. 31. Arai K, Yamada T, Nozawa R. Immunohistochemical investigation of 13. Donato R. Perspectives in S-100 protein biology. Review article. Cell migration inhibitory factor-related protein (MRP)-14 expression in hepatocellular Calcium 1991;12:713 – 26. carcinoma. Med Oncol 2000;17:183 – 8. 14. Helfman DM, Kim EJ, Lukanidin E, Grigorian M. The metastasis associated 32. Dechow TN, Pedranzini L, Leitch A, et al. Requirement of matrix protein S100A4: role in tumour progression and metastasis. Br J Cancer 2005;92: metalloproteinase-9 for the transformation of human mammary epithelial cells 1955 – 8. by Stat3-C. Proc Natl Acad Sci U S A 2004;101:10602 – 7. 15. Zhang S, Wang G, Liu D, et al. The C-terminal region of S100A4 is 33. Hermani A, De Servi B, Medunjanin S, Tessier PA, Mayer D. S100A8 and important for its metastasis-inducing properties. Oncogene 2005;24:4401 – 11. S100A9 activate MAP kinase and NF-nB signaling pathways and trigger trans- 16. Leukert N, Vogl T, Strupat K, et al. Calcium-dependent tetramer formation of location of RAGE in human prostate cancer cells. Exp Cell Res 2006;312:184 – 97. S100A8 and S100A9 is essential for biological activity. J Mol Biol 2006;359: 34. Shin I, Kim S, Song H, Kim HR, Moon A. H-Ras-specific activation of Rac- 961 – 72. MKK3/6-p38 pathway: its critical role in invasion and migration of breast 17. Vogl T, Leukert N, Barczyk K, Strupat K, Roth J. Biophysical epithelial cells. J Biol Chem 2005;280:14675 – 83. characterization of S100A8 and S100A9 in the absence and presence of bivalent 35. Weng L, Dai H, Zhan Y, et al. Stepaniants SB, Bassett DE, Rosetta error cations. Biochim Biophys Acta 2006;1763:1298 – 06. model for gene expression analysis. Bioinformatics 2006;22:1111 – 21. 18. Muallem S, Khademazad M, Sachs G. The route of Ca2+ entry during 36. Radonic A, Thulke S, Mackay IM, et al. Guideline to reference gene reloading of the intracellular Ca2+ pool in pancreatic acini. J Biol Chem 1990; selection for quantitative real-time PCR. Biochem Biophys Res Commun 2004; 265:2011 – 16. 313:856 – 62.

Mol Cancer Res 2008;6(10). October 2008 Downloaded from mcr.aacrjournals.org on September 23, 2021. © 2008 American Association for Cancer Research. Global Gene Expression Profiling Unveils S100A8/A9 as Candidate Markers in H-Ras-Mediated Human Breast Epithelial Cell Invasion

Aree Moon, Hae-Young Yong, Jae-In Song, et al.

Mol Cancer Res 2008;6:1544-1553.

Updated version Access the most recent version of this article at: http://mcr.aacrjournals.org/content/6/10/1544

Cited articles This article cites 35 articles, 14 of which you can access for free at: http://mcr.aacrjournals.org/content/6/10/1544.full#ref-list-1

Citing articles This article has been cited by 7 HighWire-hosted articles. Access the articles at: http://mcr.aacrjournals.org/content/6/10/1544.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://mcr.aacrjournals.org/content/6/10/1544. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.